--- Log opened Wed Jul 01 00:00:20 2015 --- Day changed Wed Jul 01 2015 00:00 < gleki> alta: lo noi xebni be le prokure vau tcadu cu mo 00:00 < mensi> ([FA {lo <(¹noi [{FA ZOhE} {CU KU}]²) BEhO> vau}] KUhO¹) (¹NS tcadu NF¹)> KU}] [cu {NS mo NF} VAU]) 00:01 < gleki> mensi: doi xalbo in "alta" hyperparser i made {lo noi xabni vau tcadu} parsable again. 00:01 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.xalbo.gy. di'a cusku da 00:15 < gleki> im not sure how {i broda noi} should work. compare { i broda noi mo noi mo}. either two {lonu} are to be injected before {broda} or the second is to be injected before {mo}. 00:15 < gleki> {i fasnu fa lo nu broda noi fasnu fa lo nu mo noi mo}? 00:16 < gleki> then where {ke'a} of the first relative clause should go to? 01:20 < akmnlrse> xu su'o .admine be lo BPFK mriste cu jundi 02:01 < gleki> en: admine 02:01 < mensi> admine = x1 administers x2 |>>> glekizmiku 02:04 * nuzba @fotono: lo #lojban zantufa tricu cu milxe banro .i coi ma'oi LOhOI .e ma'oi GOhOI .e lo so'o cnino cmavo http://guskant.github.io/gerna_cipra/zantufa_farvi.svg [http://bit.ly/1Hu1rgR] 02:27 < gleki> akmnlrse: oi clani 02:29 < akmnlrse> je'u .i pe'i xagmau lo nu na jungau 02:36 < gleki> ie i la ilmen so'usai roi jungau 02:38 < gleki> mensi: doi zipcpi i replied to ur case against lo: http://mw.lojban.org/papri/Talk:zipcpi:_The_case_against_%22lo%22 02:38 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.zipcpi.gy. di'a cusku da 02:38 < gleki> fenki: tell gleki coi 02:38 < fenki> gleki: You can tell yourself that. 02:45 < gleki> fenki: doi gleki coi 02:48 < gleki> fenki: doi gleki coi 02:48 < gleki> fenki: tell gleki coi 02:48 < fenki> TypeError: not all arguments converted during string formatting (file "/home/gleki/fenki/modules/tell.py", line 114, in f_remind) 02:48 < gleki> je'e 02:49 < gleki> fenki: tell gleki coi 02:50 < gleki> fenki: doi gleki coi 02:50 < fenki> TypeError: not all arguments converted during string formatting (file "/home/gleki/fenki/modules/tell.py", line 114, in f_remind) 02:51 < gleki> fenki: doi gleki coi 02:51 < gleki> fenki: tell gleki coi 02:52 < gleki> fenki: doi gleki coi 02:52 < fenki> TypeError: not all arguments converted during string formatting (file "/home/gleki/fenki/modules/tell.py", line 114, in f_remind) 03:08 < zipcpi> Gleki: The two apples - yeah as you said that's somewhat unnatural. Any two apples - you just rendered it {lo}, which many people would also render the "the" case with 03:08 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: i replied to ur case against lo: http://mw.lojban.org/papri/ 03:08 < mensi> Talk:zipcpi:_The_case_against_%22lo%22 | 2015-07-01T09:38:11.362Z 03:10 < zipcpi> And I'm not sure I like ca'e for definitional/essentialistic. {ca'e} means "It is true because I say so; I am, by fiat, defining a word to mean something, or perhaps I am doing something like assigning responsibility or some subjective value" 03:10 < zipcpi> That's the reason I made kairpau / kai'a, even though they both use the English word "define" 03:10 < gleki> zipcpi: that's the problem of English. namely, articles are necessary. In Russian there is no a/the articles needed. 03:10 < zipcpi> It means something rather different 03:11 < zipcpi> gleki: What about "any" though? 03:11 < gleki> how to say "source code"? 03:11 < gleki> zipcpi: {lo} 03:11 < gleki> if you mean your paper. 03:11 < gleki> en: code 03:11 < zipcpi> gleki: I mean Russian 03:11 < mensi> 8151 da se tolcri: mifra, banmu'isu, iunkoda, jvinjiata, jviso, jvisuai, samselpla, se'e, tolmifygau, aski, banbu'a'a, 03:11 < mensi> banbu'abu, banbu'acu, banbu'a'e, banbu'afu, banbu'agu, banbu'aje, banbu'aju, banbu'aku, banbu'alu, banbu'amu, banbu'anu, 03:11 < mensi> banbu'a'o, banbu'apu, banbu'aru, banbu'asu, banbu'a'u, banbu'ave, banbu'avu, banbu'axe, banbu'axu, banbu'azu, banbubu'a, 03:11 < mensi> banbububu, banbubucu, banbubudu, banbubu'e, banbubufu, banbubugu, banbubu'i, banbubuje, banbubuju, banbubuke, banbubuku, 03:11 < mensi> banbubulu, banbubumu, banbubunu, banbubu'o, banbubupu, banbuburu, banbubusu, banbubutu, banbubu'u, banbubuve, banbubuvu, 03:11 < mensi> banbubuxe, banbubuxu, banbubuzu, banbucu'a, banbucubu, banbucucu, banbucudu, banbucu'e, banbucufu, banbucugu, banbucu'i, 03:11 < mensi> banbucuje, banbucuju, banbucuke, banbucuku, banbuculu, banbucumu, banbucunu, banbucu'o, banbucupu, banbucuru, banbucusu, 03:11 < mensi> banbucutu, banbucu'u, banbucuve, ... 03:11 < gleki> oi 03:11 < gleki> zipcpi: additional word 03:12 < zipcpi> samselpla? Might not be very specific though, I dunno 03:12 < gleki> Russian/Spanish/English are very different in any/some/no distinction 03:12 < zipcpi> I need to refresh my memory on how Chinese does it... 03:12 < gleki> krasi uenzi? source text? 03:13 < gleki> krasi se platu? but platu is an event 03:13 < gleki> platu2 03:13 < zipcpi> Nah, that loses the programming sense 03:13 < gleki> .dict code 03:13 < fenki> code — noun: 1. A short symbol, often with little relation to the item it represents, 2. A body of law, sanctioned by legislation, in which the rules of law to be specifically applied by the courts are set forth in systematic form; a compilation of laws by public authority; a digest — verb:[...] 03:13 < gleki> (programming, uncountable) Instructions for a computer, written in a programming language; the input of a translator, an interpreter or a browser, namely: source code, machine code, bytecode. 03:13 < zipcpi> Well I was also looking for a word for "script" as in the acting sense 03:14 < gleki> .dict source code 03:14 < fenki> source code — noun: 1. (computing, uncountable) Human-readable instructions in a programming language, to be transformed into machine instructions by a compiler, assembler or other translator, or to be carried out directly by an interpreter 03:14 < zipcpi> I remember that Chinese used demonstratives if the deictic case is necessary 03:14 < zipcpi> zhe xie; na xie 03:14 < zipcpi> ~ tinoi / tanoi 03:14 < gleki> Russian is the same in this regard 03:15 < zipcpi> But the problem with that is that it extends Lojban demonstratives 03:15 < gleki> although not only xie 03:15 < zipcpi> zhei ge 03:15 < zipcpi> na ge 03:15 < gleki> zhe ge kafe, na ge kafe - this coffee, that coffee 03:15 < zipcpi> Yeah 03:15 < gleki> "ge kafe" is kinda werd tanru or even lujvo 03:15 < gleki> *weird 03:16 < zipcpi> But they are sometimes used metaphorically in a way that Lojban demonstratives often are prescribed not to be meant to be used that way 03:16 < gleki> en: compile 03:16 < mensi> 2 da se tolcri: samrkompli, bainri 03:16 < gleki> en: samrk[3~mompli 03:16 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 03:16 < gleki> en: samrkompli 03:16 < mensi> samrkompli = x1 is a file compiled from sources / resources x2 by compiler x3. |>>> True fu'ivla, borrowing "compile". 03:16 < mensi> Contextual type-4 kompili. The x1 can be used to refer to binaries or other compiled files. |>>> 03:16 < mensi> djeikyb 03:16 < zipcpi> I mean we *could*, but I'm not sure that's the best way to go. Especially the whole {ti noi} thing has that issue with the bridi tail anyway 03:17 < gleki> yes, in Lojban to work with text another set of words is used 03:17 < zipcpi> No, I don't mean text 03:17 < zipcpi> I mean something that isn't actualy present, or demonstratable 03:17 < gleki> how it can even be? 03:18 < zipcpi> I mean... like "the play you just watched", IIRC in Chinese you can use demonstratives metaphorically to refer to it 03:18 < gleki> i translated that example 03:18 < gleki> it's co'ephora 03:19 < zipcpi> Yes... which is also what you use {le} for 03:19 < zipcpi> Like I often say, there isn't much difference between our {le}s; we only disagree on the philosophy behind the definitions 03:19 < gleki> me? maybe. but that's me, and my {le} is just a convention 03:19 < gleki> my {le} is not needed 03:20 < gleki> UD, scope and variables/constants do all the work 03:20 < zipcpi> It's convenient though, that's my point, vs having to qualify {da} all the time 03:20 < gleki> hm, so "to compile" is "to translate" 03:21 < gleki> zipcpi: tersmu version 99.0.0.1 will have to do that anyway 03:21 < ctefa`o> ...coi 03:21 < gleki> en: uenzi 03:21 < mensi> uenzi = x1 is an utterance or text, in language x2 |>>> cusku, bangu, gerna |>>> 03:21 < mensi> ahernai 03:23 < gleki> Quelltext 03:23 < zipcpi> And I'm not sure {da} always means co'ephorical either 03:23 < gleki> {da} is about UD so yes 03:23 < zipcpi> I mean I suppose it has the preference of concreteness over abstractness 03:24 < gleki> i dont understand those words 03:24 < zipcpi> I mean in terms of su'a/lo'e generalizations 03:24 < zipcpi> Meaning that da tends to not mean su'a/lo'e 03:25 < zipcpi> Though I sometimes wonder about "Alice and Bob" discussions... 03:26 < gleki> en: mi'esku 03:26 < mensi> mi'esku [< minde cusku ≈ Command express] = x1=m1=c1 expresses/states command x2=c2 to recipient(s) x3=m2=c3 via 03:26 < mensi> expressive medium x4=c4. |>>> m3 of minde (the desired result) is presumably cointained in he expression of the command. 03:26 < mensi> x1 commands x3 with expression x2. |>>> rlpowell 03:27 < zipcpi> samselpla: x1=p2 is computer program source code created by x2=p1 with intended result (state/process) x3=p3=s2 on computer x4=s1. 03:27 < zipcpi> Well I guess that lujvo is about as good as any other 03:29 < zipcpi> If I wanted to emphasize it I might make a tanru with {krasi} but I'm not sure if that's good 03:29 < gleki> luvjo: se minde cusku nu stika tutci 03:29 < zipcpi> A xavjvo? ie'i 03:29 < gleki> lujvo: se minde cusku nu stika tutci 03:29 < mensi> selmi'eskununtiktci[18143], selmi'ecusnuntiktci[18163], selmidyskununtiktci[18754], selmidycusnuntiktci[18774], selmindyskununtiktci[19784], ... 03:30 < gleki> se mi'esku nunstika tutci 03:30 < gleki> source code editor 03:30 < zipcpi> Oh 03:30 < zipcpi> se mi'esku isn't very specific though 03:30 < zipcpi> samselpla is probably better 03:31 < gleki> im not even sure it should be limited to computers 03:31 * ctefa`o cu ^^ 03:31 < gleki> likewise compliation isnt necessary 03:31 < gleki> *compilation 03:32 < zipcpi> But if it isn't, then it's just a text editor 03:32 < gleki> it should run commands using source code. interpreters also use source code 03:33 < zipcpi> What kind of interpreters? 03:33 < zipcpi> They're all skami or skami-related right? 03:34 < gleki> .dict source code 03:34 < fenki> source code — noun: 1. (computing, uncountable) Human-readable instructions in a programming language, to be transformed into machine instructions by a compiler, assembler or other translator, or to be carried out directly by an interpreter 03:34 < gleki> hm, i can imagine using an animal or a huamn instead of a computer here 03:34 < gleki> *human 03:34 < zipcpi> But then you'd probably not use programming code 03:37 < zipcpi> The thing about "any" is this... all it means is that... there is a set I care about ("apples"), but the differences within the members of the set is not important to me 03:38 < zipcpi> Which really reminds me of {nalvaidza}, but I'm not sure I want a lujvo for that 03:38 < zipcpi> Maybe just for definitional purposes 03:42 < zipcpi> nalvaimenre 03:43 < zipcpi> re xe'e plise -> re nalvaimenre be lu'i lo plise 03:44 < zipcpi> Or lu'i ro plise, depending 03:44 < zipcpi> No wait... not sure if menre works on sets that way 03:44 < zipcpi> Perhaps just {lo plise} or {ro plise} is safer 03:46 < zipcpi> The difference between "Give me any two apples" and "Give me the two apples" is that "Which two apples?" is a meaningful question to the latter, not to the first 03:47 < zipcpi> Which is also similar to that "Which apple do you want" thing 03:47 < zipcpi> "Any one" 03:48 < zipcpi> Google translates any two apples as "Gěi wǒ rènhé liǎng gè píngguǒ" 03:49 < zipcpi> rènhé 03:49 < zipcpi> Not sure how natural that is 03:49 < zipcpi> 任何 03:50 < zipcpi> http://www.zdic.net/c/b/14a/325810.htm 03:51 < zipcpi> rexe'e le mu plise -> Any of the five apples -> re nalvaimenre be le mu plise 03:52 < zipcpi> Substitute {le} for {lo} if it makes you more comfortable here 03:59 < gleki> meaningful for the second due to being limited to UD 04:01 < zipcpi> What I mean about metaphorically-extended demonstratives is that IIRC "nage" would be acceptable for the play question 04:01 < gleki> ofc. 04:01 < gleki> Russian does the same 04:01 < zipcpi> Where you want to exophoricallly refer to the play, but it's not actually there for you to demonstrate 04:01 < gleki> most deictic pronouns or articles are derived from demonstrtatives 04:01 < zipcpi> Mm 04:02 < zipcpi> Yes, but in Lojban we want to avoid using ti/ta/tu for that. Also it's often useful to separate deictic/co'ephorical from demonstrative 04:02 < zipcpi> Well the latter is a subset of the former 04:02 < zipcpi> But yeah 04:04 * nuzba @palfem_kleet: @oka_iu_tcan というかHaskellのパッケージのところにProlog with lojban……. [http://bit.ly/1KqKQLy] 04:04 < gleki> you can say " I want these apples, aples that are here" - which ones? - any 04:05 < zipcpi> Hence nalvaimenre 04:05 < zipcpi> xe'e is just a shorthand 04:08 < zipcpi> I sometimes wonder about {tu} 04:08 < gleki> it's just aquele in spanish 04:08 < zipcpi> Whether that might be extensible; since often you don't point to something that far away 04:09 < gleki> malgli 04:09 < zipcpi> "All you have to do is meet him there"? 04:09 < zipcpi> Maybe 04:10 < zipcpi> I do remember Malay has a three-level distinction too 04:10 < zipcpi> sini, situ, sana 04:10 < zipcpi> ~ bu'u ti, bu'u ta, bu'u tu 04:10 < zipcpi> Or vi va vu 04:11 < zipcpi> Hmm... could {vu} be deictic though? Cause IIRC correctly it just means "somewhere far" 04:11 < zipcpi> coi ilmen 04:12 < gleki> http://www.youtube.com/watch?v=5YloC_1yVVs 04:12 < zipcpi> The thing is in Malay sana typically is beyond pointing 04:13 < zipcpi> Crap need Spanish subtitles 04:13 < zipcpi> Can't make out the phonemes 04:14 < gleki> i only linked to show the difference in meaning 04:14 < zipcpi> Ah 04:14 < zipcpi> Yeah I got it 04:15 < zipcpi> The whole close to me, close to you, somewhere beyond either of us 04:15 < zipcpi> Malay has three levels for places, but I think only two levels of demonstratives 04:16 < zipcpi> OK cool 04:16 < zipcpi> How to translate "meet him there" though? 04:16 < zipcpi> bu'u... 04:17 < zipcpi> bu'u le co'e 04:17 < zipcpi> Bah 04:17 < zipcpi> bu'u le vu co'e 04:17 < gleki> bu'u lo vu xo'i bu'u 04:17 < zipcpi> lol 04:18 < zipcpi> But wait you made {lo} mean any :p 04:18 < gleki> anywhere over there 04:18 < zipcpi> Nope 04:19 < zipcpi> Just "anywhere far away" 04:19 < zipcpi> Which... isn't right 04:19 < zipcpi> Yeah not practical, especially under xorlo where many people use {lo} for deictic as well 04:20 < zipcpi> And I still use {lo} the most, so... 04:20 < zipcpi> Not sure I really mean {nalvaimenre} everytime I use it 04:20 < zipcpi> Probably not 04:20 < zipcpi> Hey Ilmen another magic predicate for you :p 04:20 < gleki> ok, far away 04:21 < zipcpi> Some where far away, but somewhere specific; i.e. that place and not anywhere else 04:21 < gleki> oh, {bu'u da ne vu} then 04:25 < zipcpi> I'd use {le vu co'e} myself, but whatever 04:26 < zipcpi> I mean it's not too much different from your use of {le fetsi} for she 04:27 < zipcpi> vu qualifies distance. co'e because {vu} and {bu'u} are all the semantic qualifiers we need 04:27 < zipcpi> I mean it could be {bu'u le vu stuzi} if we really wanted to avoid co'e 04:28 < zipcpi> https://tatoeba.org/eng/sentences/show/4207298 04:28 < gleki> as i said {le} is just a convenience. especially for such cases as Tatoeba where you need to emulate natlangs 04:29 < zipcpi> Well same argument for "over there" 04:30 < zipcpi> And it's not just malgli either. Chinese also uses the demonstrative 那 04:30 < zipcpi> And CKTJ wants bu'u -> zai lol 04:30 < gleki> malSAE then 04:30 < zipcpi> Well there's also bu on the page but I prefer zai; cause bu is still useful 04:31 < gleki> we need a word for SAE 04:31 < zipcpi> SAE = ? 04:31 < gleki> https://en.wikipedia.org/wiki/Standard_Average_European 04:31 < zipcpi> Chinese ain't European 04:31 < gleki> hm, then i need one more word 04:32 < gleki> :P 04:32 < gleki> malmetfo then 04:32 < zipcpi> So yeah but my point is that Chinese doesn't have definite vs indefinite articles, but it also uses constructs to arrive at the same meaning 04:33 < zipcpi> Namely demonstratives 04:33 < gleki> same for russian 04:33 < zipcpi> But in Lojban we want to avoid using demonstratives for that; and I agree that the separation is sometimes helpful 04:34 < zipcpi> Hence perhaps the malgli solution is the best solution sometimes 04:37 < zipcpi> Hah under alta we won't even need co'e in that example 04:38 < zipcpi> exp: penmi le fetpre bu'u le vu 04:38 < zipcpi> Oh... 04:39 < gleki> cant see what happened. looks like some irc problem 04:40 < zipcpi> si'au la naxle na akti 04:40 < gleki> ofc. 04:41 < gleki> vrici problems 04:41 < gleki> la naxle is waiting for mensi to start 04:43 < gleki> http://mw.lojban.org/extensions/ilmentufa/camxes-exp.html 04:43 < zipcpi> Yeah tested it there; but I wanted to test alta too 04:43 < zipcpi> alta: penmi le fetpre bu'u le vu 04:44 < mensi> ([FA ZOhE] [CU {penmi NF} { } VAU]) 04:44 < gleki> when will i incorporate all changes from ilmentufa and zantufa ... 04:44 < zipcpi> brb 04:53 < zipcpi> Back 04:54 < ghhtzt> le le le uenmo ku mutce ku mo 04:56 < zipcpi> Hm ju'ei is also useful within poi'i... 04:57 < ghhtzt> pehi naku ly. finti zo juhei 04:57 < zipcpi> ue la latro'a pu cusku lo sedu'u ly finti 05:02 < gleki> en: ju'ei 05:02 < mensi> ju'ei = [JUhEI] Tight scope bridi separator; analogous to .i without ending the abstractor or relative clause. |>>> "nu 05:02 < mensi> broda ju'ei brode" is equivalent to "nu ju'e gi broda gi brode". Similarly "ko'a noi broda ju'ei brode" is equivalent to 05:02 < mensi> "ko'a noi ju'e gi broda gi brode". |>>> latros 05:02 < zipcpi> Starts a new bridi within a sub-clause 05:02 < zipcpi> Like NOI or NU 05:03 < zipcpi> ju'e isn't an adequate solution because it only accepts two 05:03 < zipcpi> So then you'd need extreme forethought if you want more than two, and ugly nested ju'egi constructs 05:04 < zipcpi> In fact I don't know what ju'e is good for lol 05:04 < gleki> sounds like curtis'es new cmavo 05:04 < gleki> en: ju'o'e 05:04 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 05:04 < gleki> en: je'o'e 05:04 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 05:05 < gleki> en: ji'o'e 05:05 < mensi> ji'o'e = connective: elliptical/generic/vague |>>> Like zo'e or co'e, but for connectives. The connectives referred to 05:05 < mensi> may be logical or non-logical; they take the form proper for the syntax/context (probably is best implemented using 05:05 < mensi> post-original CLL connective system modification such as the MAD proposal or the system developed by selpa'i); in 05:05 < mensi> standard original CLL usage, this would be the inter-tanru after-thought connective. Synonymous with but morphologically 05:05 < mensi> and phonotactically allowed version of .y'i. |>>> krtisfranks 05:05 < zipcpi> Oh that one. That one's a connective though, and I'm pretty sure that's not what ju'e/ju'ei means 05:06 < gleki> what else can bare {i} be for 05:06 < zipcpi> Are all .i -> .i ji'o'e? 05:06 < zipcpi> Dunno 05:06 < zipcpi> Ask Curtis lol 05:07 < gleki> wait no. conenctives bind scopes 05:08 < gleki> http://jbovlaste.lojban.org/comments.html?valsi=27873;definition=63971 05:08 < gleki> and i said that myself O_0 05:08 < zipcpi> Right 05:09 < gleki> and even worked on alta after that O_0 05:10 < zipcpi> lol gleki have you been drinking? zo'oxo'osai 05:10 < gleki> did xo'osai just modified zo'o? 05:10 < zipcpi> ... dunno 05:10 < gleki> i think not 05:11 < zipcpi> Can't connect UI x.x 05:12 < zipcpi> There's toi'e but no equivalent to anti-toi'e 05:12 < zipcpi> Or UI-je 05:13 < ghhtzt> uidje 05:13 < zipcpi> u'i 05:22 < ghhtzt> mo fado doi jainu dahi spirofu 05:22 < zipcpi> Hm i'au can also be used for sumtcita. Or just for closing sumti with a lot of subclauses to continue the main bridi 05:56 < zipcpi> dacti is a physical object; how to refer to "object" in the abstract, agnostic as to physicality? 05:56 < zipcpi> Like in programming 05:56 < zipcpi> dzadza maybe lol 05:56 < gleki> you may paste a piece of code and we'll translate that 05:57 < zipcpi> No wait dzadza is a number 05:57 < zipcpi> No, it's not code; it's the concept of an object in OOP 05:58 < zipcpi> Also useful when we're talking about "objects" that sumti refer to in Lojban 05:59 < zipcpi> Was thinking of using that to define nalvaimenre, but {da} might do for now 06:02 * ctefa`o warms up a big fiery fire 06:02 * ctefa`o throws appx half of last week's ideas on it 06:03 < ctefa`o> hey guys 06:04 < ctefa`o> .i fa mi noi do'e ke'a barda cu klama 06:04 < ctefa`o> How to expandorino? 06:11 < zipcpi> do'e mi barda .ije mi klama ? 06:12 < samcinki> join #ckule 06:12 < zipcpi> ... fail? 06:16 < ctefa`o> .ije xmm 06:22 < ctefa`o> Kinda works but comletely kills more advanced expansions 06:24 < zipcpi> Well I'm not sure what else you mean by "expand". Also it can be as long as you want, with complicated logical groupings I might not be quite sure how to use 06:24 < zipcpi> But expansions need not be pretty 06:25 < ctefa`o> My hair rises from any in-sentence expansion suddenly crossing a sentence boundary 06:26 < zipcpi> Hmm... that... not sure how to fix that 06:26 < zipcpi> Cause that's also how I expand {be} 06:26 < zipcpi> I guess maybe {poi'i}? 06:27 < zipcpi> lo me mi je poi'i do'e ke'a barda cu klama... 06:27 < zipcpi> Not sure if that's any better 06:28 < ctefa`o> pe'i Not really 06:28 < zipcpi> But that's part of the problem... I think like a programmer, you see, so if you ask me to expand objects that have sub-predicates, I automatically split them into several statements 06:28 < zipcpi> Declare all my values, then use them 06:28 < ctefa`o> I want that noi do'e added to a bridi trail 06:29 < zipcpi> xoi? Not sure that's the best solution. I mean you can use {xoi do'e lo me'au ke'a cu barda} but... 06:30 < ctefa`o> I want the clause in the noi do'e added to a bridi trail/tail of mi klama 06:31 < zipcpi> If you don't mean xoi, I don't know what else you mean 06:31 < ctefa`o> 15:04 .i fa mi noi do'e ke'a barda cu klama 06:31 < zipcpi> No, I mean how you want it expanded 06:32 < zipcpi> You don't want multiple sentences, you don't want xoi, so what do you mean 06:32 < ctefa`o> 15:30 I want the clause in the noi do'e added to a bridi trail/tail of mi klama 06:32 < zipcpi> I don't know what you mean by bridi tail 06:32 < ctefa`o> sub-bridi? 06:32 < zipcpi> {mi klama jecu jai do'e barda}? 06:33 < ctefa`o> lmao 06:33 * ctefa`o gives up, goes to study connective system in-depth 06:33 < zipcpi> That's the second use of jai; to move a sumtcita place to x1 06:34 < zipcpi> The real x1 is then moved to fai 06:34 < zipcpi> mi jai gau barda fai ti -> gau mi barda fa ti 06:34 < zipcpi> Or gau mi ti barda 06:35 < ctefa`o> So where did klama go 06:35 < zipcpi> {mi klama jecu jai do'e barda} 06:36 < ctefa`o> expand jai 06:36 < zipcpi> I can't without another sentence 06:36 < zipcpi> Because {jecu} would reuse all sumti before the first bridi tail 06:36 < ctefa`o> Then it is not the solution I seek 06:37 < zipcpi> Well there's nothing else. You are making two statements about {mi} 06:37 < ctefa`o> iep 06:38 < zipcpi> So if you don't want two statements, if you don't want xoi-me'au-hacks, you don't want jecu, there's nothing else 06:38 < ctefa`o> I want jecu without a jai 06:38 < zipcpi> Impossible, because the do'e place needs to be moved to the x1 06:39 < ctefa`o> I actually have 1 solution but it doesn't work in all cases either 06:40 < ctefa`o> But the real nightmare is xoi on sub-bridi 06:40 < zipcpi> And jecu is simply shorthand for {ije} while reusing the sumti before the bridi tail 06:41 < zipcpi> Usually the x1 06:41 < ctefa`o> Maybe .ije is what I want then 06:42 * ctefa`o goes back to study connectives in-depth 06:42 < zipcpi> Yeah... perhaps you are taking {.i} too literally. Often what English uses ", and" for, gets translated into .ije in Lojban 06:42 < zipcpi> Thus Lojban counts two sentences where English only counts one 06:43 < menli> broda .ije brode == ge broda gi brode 06:43 < mensi> menli: cu'u la'o gy.Ilmen.gy.: da bu'a | 2015-06-04T21:46:26.298Z 06:43 < mensi> menli: cu'u la'o gy.Ilmen.gy.: de na bu'a | 2015-06-04T21:46:53.018Z 06:43 < zipcpi> Oh yeah that's another "solution" 06:44 < menli> {.ije} doesn't behave like {.i}, because in the former case da-variables within zo'u prenexes span over both bridi 06:44 < zipcpi> True 06:44 < menli> si both of the ije-linked bridi 06:44 < menli> which isn't the case with {da zo'u broda .i brode} 06:45 < ctefa`o> Hmm I wasn't aware of that. .ije may be what I seek then 06:45 < menli> {.ije} is like it's own word, you shouldn't pay too much attention to the {.i} part :p 06:45 < menli> zo'o ru'e 06:46 < ctefa`o> Well I need to fully understand the connectives anyway to make a formal expandalizer. Might as well learn that now. 06:47 < menli> tersmus: da zo'u da brode .i je da brodo 06:47 < menli> .ui nai la .tersmus. na jundi 06:47 < ctefa`o> Does the .ije interfere with anaphoric sentence sumti? 06:49 < menli> I don't really know. I seem to recollect that go'i usage disregards .ije, but I'm unsure. 06:49 < ctefa`o> pro-sumti* 06:50 < zipcpi> When I'm doing expansions I tend to use lerfu-sumti or goi anyway 06:50 < menli> I mean, that in the case of {broda .i je brode .i go'i}, I think in usage go'i refers only to {brode}. But then I'm unsure. 06:50 * ctefa`o cringe 06:50 < zipcpi> Much better than relying on anaphora for such situations 06:51 < zipcpi> Anaphora are only shorthand. It'd be extremely difficult or impossible to do an expansion in such a way as to not touch them at all 06:51 < zipcpi> touch -> change their values 06:52 < ctefa`o> Welcome to my job 06:52 < zipcpi> I don't see the point. Anaphora are only shorthand 06:52 < menli> Things like {ri xi N} rely on the surface sentence encoding, and not on the underlying logical proposition 06:53 < zipcpi> Anaphora themselves get expanded, and their referent relies on the source, not the expansion 06:53 < ctefa`o> Well if you insist I can make my generalized formal expander STT only 06:54 < ctefa`o> And you can have one version each for regular jbobau to suit your personal usage:) 06:54 < menli> For example, logically {lo brode lo brodo cu broda} = {lo brodo lo brode cu se viska}. Yet, {lo brode lo brodo cu broda .i ri brodu} ≠ {lo brodo lo brode cu se broda .i ri brodu} 06:55 < ctefa`o> I am not saying it is going to be easy. 06:55 < menli> So in order to extract the referent of {ri}, you have to analyze the surface sentence and the order in which its elements appear 06:55 < zipcpi> I'm saying that there is no point; your expander will have to rely on the source to expand anaphora anyway 06:57 < zipcpi> Anaphora depends on surface structure of the source, not the logical form 06:57 < menli> anaphora are kinda part of the source / surface representation; they have to be logically expanded as well if you want to fully get ride of the surface representation and only keep the underlying logical proposition 06:57 < menli> if I'm not mistaken 06:57 < menli> *get rid 06:57 < ctefa`o> I have nothing more to say. I am going to study connectices now and can't focus on chat for a while. co'o 06:58 < menli> co'o di'ai 06:58 * menli quietly paces back into the shadows 07:11 < menli> http://mw.lojban.org/index.php?title=Expansion_of_logical_connections 07:12 < menli> Also, you should play around with tersmus (it's available on github I think), as it's already a pretty complete expander to logical forms 07:12 < menli> https://gitorious.org/tersmu/ 07:13 < menli> mu'o 07:34 < gleki> maybe i should try grouping gismu into te bridi klesi 07:35 < gleki> alta: lo broda cu brode 07:35 < mensi> ([FA {lo KU}] [cu {brode NF} VAU]) 07:35 < gleki> actually it's quite easy to detect words that are to be translated as nouns now. 07:36 < gleki> those are just NF KU 07:36 < gleki> SF is gone ofc. 07:38 * nuzba @osa_k: context free grammars - Lojban http://mw.lojban.org/papri/context_free_grammars [http://bit.ly/1KsOgiC] 07:40 * nuzba @uitki: context free grammars - http://mw.lojban.org/papri/context_free_grammars by Gleki [http://bit.ly/1KsOpTh] 07:54 * nuzba @L16777216: 東工大、みんなEsperantoの話してばかりだし、lojban勉強会みたいなのしよーぜ [http://bit.ly/1KsPxWO] 07:54 * nuzba @tokoro10g: lojban講習会?? [http://bit.ly/1KsPElb] 07:54 * nuzba @tokoro10g: Lojban For Beginners — velcli befi la lojban. bei loi co'a cilre http://www.tlg.uci.edu/~opoudjis/lojbanbrochure/lessons/book1.html [http://bit.ly/1KsPDha] 07:56 * nuzba @L16777216: lojbanがEsperantoに対して優れている点として、lojbanだと合成語が一意に分解できるというのがあるらしいが、こういうの東工大の人は好きそうじゃない? [http://bit.ly/1KsQ9M9] 08:02 * nuzba @L16777216: lojbanだと括弧があるというか構文木が見えるから楽しい [http://bit.ly/1KsQJd1] 08:02 * nuzba @tokoro10g: やいやい言わずにわかりやすそうな方で学習してみるか,lojban [http://bit.ly/1KsQKgX] 08:02 * nuzba @L16777216: lojbanの紹介で行きつくところ、ここなんだよな 頭が赤い魚を食べた猫 - Togetterまとめ http://togetter.com/li/567493 [http://bit.ly/1KsQOgz] 08:04 * nuzba @tokoro10g: lojban勉強会??? [http://bit.ly/1KsQRsP] 08:04 * nuzba @L16777216: 萌え萌え(死語)なlojban教材ないものかね [http://bit.ly/1KsQYoh] 08:38 < ghhtzt> mu 08:46 < ghhtzt> lu mu~~~~ cu voi lo sofybakni ki cusku tahe. 09:04 < ghhtzt> lu ienai cu voi lo jbopre ki cusku tahe. 09:06 < so-uai-du-iu> ije lo gohi cu voi do cusku 09:10 < ghhtzt> ienai 09:12 < ghhtzt> domodoi matnysfani 09:12 < so-uai-du-iu> zo matnysfani 09:12 < so-uai-du-iu> i 09:14 < so-uai-du-iu> xrm. i lo ni lo nu diha tahe pilno lo bangu cu nandu cu bramau lo jai se kanpe. 09:17 * ctefa`o rubs hands 09:17 < ghhtzt> ii 09:17 < ctefa`o> Latest round of crazy ideas are now ready 09:20 < samsunga> doi nictyfam. lenu ko'a se bangu ko'e na se sarcu lenu ko'a mulno jimpe ko'e 09:22 * nuzba @lojytan_bot: はじめてのロジバン(http://w.livedoor.jp/hajiloji/d/)を作っています。今までにはないゆるふわな感じの講座を目指しているので、ロジバンに興味のある方はどうぞ!#lojban #はじめてのロジバン #ロジバン [http://bit.ly/1Jx3MKb] 09:22 < ghhtzt> mi senpi gajenai lo du fenki ginai lo du na cizrymau la gadganzu 09:34 * xalbo comes out of his cave, waves his fist ineffectually at the damn kids to get off his lawn and stop misusing {voi}, and goes back into his cave. 09:39 * ghhtzt cu voi ki djica lonu ki jiha ponse su kevna zdani poi lamji fa su srasu foldi 09:40 * samsunga cu barkla lemi kevna gi'e ckaji lenu mi sligau lemi bolxa'e fa'a le mabla verba noi se zukte fi lenu ri canci lemi sasfoi gi'e co'u malpilno zo lo kei gi'e za'ure'u nerkla le kevna 09:51 < ctefa`o> voi xalbo 09:51 < ctefa`o> si si xoi xalbo 09:53 * nuzba @tarik_saz: Gece gece arkadaslarla takIlmaktayIz... @ jumeirah lojban https://instagram.com/p/4mdQyZMRza/ [http://bit.ly/1FRvMQM] 09:54 < ctefa`o> sa poi xalbo 09:54 < ctefa`o> su joi xalbo 09:54 * ctefa`o cu sisti 09:55 < samsunga> ki'e le dragau .i mi na se slabu ro lei mabla cnino 09:57 < ghhtzt> mijiha pirolei mabla cnino na se slabu 10:00 < mikrysofto> zohoi h coha misno 10:01 < samsunga> Unrecognizable word 'zohoi' at line 1 column 1 10:02 < samsunga> ju'inai ju'inai 10:05 < keidji> coi ro do 10:05 < samsunga> coi la keidjis. 10:06 < mikrysofto> ju'i la samsunga 10:06 < samsunga> re'i la mikryralci 10:07 < mikrysofto> ju'i la samsunga 10:07 < samsunga> mo 10:07 < mikrysofto> do se panci lo skami 10:08 < keidji> Who can I talk to about jbovlaste? Specifically XML export 10:08 < samsunga> mi se nandu lenu jimpe lenu ta'i makau ka'e se panci le na ka'e nerkla be le nazbi 10:08 < samsunga> keidji: la mukti ka'e sidju 10:08 < mikrysofto> sa'u lo dacti cu dirce lo panci 10:09 < mikrysofto> keidji: go on, what is it bothering you 10:12 < keidji> Clicking on the "XML Export" link for jbovlaste, it takes me to a new HTML web page 10:12 < keidji> I was hoping it would export a structured XML document I could download. 10:13 < gleki> no, it's not html 10:13 < gleki> just save it, it's an xml file 10:14 < gleki> it's just your browser rendering it as a web page 10:14 < gleki> done for your convenience 10:14 < gleki> http://mw.lojban.org/extensions/ilmentufa/ircbot/dumps/ 10:14 < samsunga> 10:15 < gleki> ^ that link is the same but you can see file extensions 10:15 < keidji> http://jbovlaste.lojban.org/export/xml-export.html 10:15 < mukti_> Gleki: could you delete "Kevin Stringfellow" from the Facebook group? 10:15 < mukti_> Mobile interface won't let me 10:16 < keidji> gleki: Ah! *MUCH* better. 10:16 < keidji> ki'ecai 10:18 < gleki> mukti_: i think what keidji just told is a bug 10:19 < gleki> keidji: arent you logged in into jbovlaste? 10:19 < keidji> nope. quick and dirty just wanted the XML 10:19 < gleki> then that link 10:19 < gleki> keidji: btw pls log in and upvote 5 questions here area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 10:19 < keidji> the XML export link is available to the public 10:19 < durka42> if you aren't logged in there is a captcha 10:20 < keidji> gelki: I have already upvoted 10:20 < gleki> keidji: all 5? 10:20 < keidji> yup 10:21 < gleki> durka42: pls try doing that without logging in using incognito mode 10:22 < durka42> http://www.seas.upenn.edu/~aburka/tmp/Screen%20Shot%202015-07-01%20at%201.22.13%20PM.png 10:25 < durka42> gleki: this is why bots like vlasisku and mensi have to store their session cookie 10:25 < gleki> strange. 10:25 < gleki> and ur browser is strange too :) 10:26 < gleki> keidji: your bug report isnt confirmed. if u want u can list the steps you took to get that bug 10:28 < samsunga> la tricu staile tapla cu ge ka prami gi ka jmive 10:29 < gleki> uanai 10:30 < samsunga> http://piro.sakura.ne.jp/xul/_treestyletab.html.en (to .o'i rinka lenu lenu gunrivzu'e cu zenba leni frili toi) 10:31 < gleki> ua 10:33 < gleki> la romge na kakne 10:42 < keidji> gleki: 1. (Using Safari 8.0.7 [shut up, I don't care]) Navigate to http://jbovlaste.lojban.org 2. Click the "XML Export" link 3. Click the "English" link 4. Complete the Captcha 5. Page is loaded as xml-export.html with otherwise unparseable nodes. EXPECTED: XML document is generated and downloaded. 10:45 < samsunga> keidji: when I save that generated page it's saved as parsable XML (using Chromium and Firefox though) 10:46 < samsunga> the extension is wrong, ie 10:46 < keidji> Well, when I download the page and change it to .xml 10:46 < keidji> the nodes are still messed up (dd, dl, dt). The nodes are set up for the stylesheet that comes with it 10:47 < samsunga> ue 10:48 * samsunga assumes Safari saves the result of the transform instead of the original 10:50 < keidji> Not sure why it would do that at all..........fair enough. The bot dump is perfectly usable. Thanks. 10:51 < gleki> indeed, one could just add a link to the dumps that would be downloaded e.g. every day 10:55 < bigcentaur> i have a question about using mekso as quantifiers 10:56 < gleki> go on 10:57 < gleki> bigcentaur: btw pls log in and upvote 5 questions that have rating below 10 here area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 10:58 < bigcentaur> so the cll says i'm allowed to put mathematical expressions as quantifiers of descriptions, e.g. {mi ponse su'o ci cutci} 10:58 < gleki> yes 10:58 < bigcentaur> let's say i wanted to make the mathematical expression more complicated 10:59 < bigcentaur> if i were translating say macbeth act 1 scene 3: "Weary se'n nights nine times nine / Shall he dwindle, peak and pine" 10:59 < bigcentaur> imagine i want to quantify a week and say "9 times 9 weeks" 11:00 < bigcentaur> if i write {so pi'i so jeftu} and try to use that in a sentence, camxes goes red 11:00 < gleki> exp: so pi'i so jeftu 11:00 < mensi> ([{ } jeftu KU] VAU) 11:01 < gleki> you want {lo jeftu be li so pi'i so} btw 11:01 < latro`a> just did my SE votes 11:01 < gleki> latro`a: all 5? 11:01 < latro`a> yep 11:01 < bigcentaur> right, but now i'm just curious about why camxes is kicking up a fuss about it 11:02 < bigcentaur> http://www.lojban.org/camxes/?text=so+pi%27i+so+jeftu 11:02 < gleki> no idea. this camxes might have bugs that are unlikely to be ever fixed 11:02 < gleki> mw.lojban.org/extensions/ilmentufa/camxes-exp.html 11:02 < latro`a> as I recall you wind up needing at least one paren to put a MEX in a quantifier 11:02 < gleki> ^ better 11:02 < latro`a> vlaste: parenthesis 11:02 < vlaste> 3 results: to, toi, koi'e 11:02 < latro`a> bah 11:03 < latro`a> what was the mex parenthesis again 11:03 < gleki> vei? 11:03 < latro`a> ve'o is the right one, I think 11:03 < latro`a> vei is the left one 11:03 < latro`a> camxes: veii so pi'i so jeftu 11:03 < camxes> SyntaxError: Expected [,] but "j" found. 11:03 < latro`a> camxes: vei so pi'i so jeftu 11:03 < camxes> ([{vei <(¹so BOI¹) (¹pi'i [so BOI]¹)> VEhO} jeftu KU] VAU) 11:03 < bigcentaur> ah, too bad 11:03 < latro`a> I'm not sure whether this is a bug or a feature 11:03 < gleki> latro`a: u upvoted a question that hd 11 votes already. please unvote it and upvote something else :) 11:04 < latro`a> ki'e 11:04 < latro`a> I know that in the YACC parser, those sorts of parens were necessary 11:05 < latro`a> actually, that's why ME SUMTI MOI is a thing 11:05 < latro`a> historically 11:05 < latro`a> because MOI was unable to take MEX by itself, so instead of writing MEX MOI you would write me li MEX MOI 11:07 * nuzba @lojytan_bot: これはものすごく個人的な意見ですが、lojbanの音体系はそんな特筆すべき魅力がないように思うのですが、どうなんでしょう? [http://bit.ly/1IOiUwX] 11:09 <@xalbo> Yes, I'm fairly certain the reason is a parser limitation. But I agree, I'm not sure whether it's really a bug or a feature. You can definitely get there, and it's actually sort of nice being warned up front that something complicated is going on pe'i 11:09 < mensi> xalbo: cu'u la'o gy.gleki.gy.: in "alta" hyperparser i made {lo noi xabni vau tcadu} parsable again. | 11:09 < mensi> 2015-07-01T07:01:08.468Z 11:09 <@xalbo> ki'e la gleki 11:09 <@xalbo> .i'e 11:10 < bigcentaur> thanks for the help :) 11:11 <@xalbo> (and yes, the best answer is {lo jeftu be li so pi'i so}, but that is missing the point) 11:12 <@xalbo> .i la .leks.lutor. cu zerle'a vono titnanba no'u vei vo pi'i pano titnanba .i tcepalci 11:12 < zipcpi> Hm interesting http://mw.lojban.org/papri/punctuation 11:13 < zipcpi> Hm I like semicolon as a "natlang comma" better 11:13 < zipcpi> Well maybe it'd be useful to have two levels sometimes 11:13 < zipcpi> Their hyphen contradicts mine though 11:14 < zipcpi> And is not parseable probably 11:14 < zipcpi> Well I thought kg- -> kygy, but apparently here it's used as a selbri 11:14 <@xalbo> it's probably meant as shorthand for {ki'ogra} 11:15 < zipcpi> Yeah 11:15 < zipcpi> Might be useful to have some kind of lerfu-string equivalent for selbri though 11:15 <@xalbo> There was a really old convention of using an initial - as a makeshift me'oi. {.i mi citka lo -quiche} 11:16 <@xalbo> (Well, to fill the role that was eventually also filled by {me'oi}) 11:17 < zipcpi> lerfu-string equivalent for selbri: Could be {me'au}. {cei me'au kygy} 11:17 < gleki> 14 lojbanists have used all their upvotes by far 11:17 < zipcpi> A bit long though 11:18 < zipcpi> Another crazy idea I had was perhaps reserve all zam+[Xu/xe/ke/ve/je] zi'evla 11:18 < zipcpi> Actually that would also include some gismu 11:18 < zipcpi> Not sure if any gismu already exist in that space 11:19 < zipcpi> Essentially, zam being the rafsi of zai'e/za'e, and the syllables acting like the ISO zi'evla 11:19 <@xalbo> There are no /zam../ gismu yet. 11:20 < zipcpi> zamru would clash with zanru though :p 11:21 < zipcpi> Not sure I care too much about that though; these are only for cei assignments, not meant for permanent definition 11:21 <@xalbo> Would ?{zamcu} clash with {samcu}? I haven't internalized those rules yet. 11:21 < zipcpi> I think so 11:21 < zipcpi> I don't know those rules completely eiethr 11:22 < zipcpi> Not to menting {zamje} definitely clashes with {zamju} 11:22 < zipcpi> Oh and {zamzu} is banned >.< 11:22 < zipcpi> Whyyyyy 11:22 < gleki> nz 11:22 < zipcpi> Oh {zammu} definitely won't work either 11:23 < zipcpi> At this point I would really make the CLL-whale cry by using {zamyzu}, {zamymu}, etc lol 11:24 < akmnlrse> zagzu zagmu 11:24 < akmnlrse> rafsi: zag 11:24 < mensi> zo zargu se rafsi zo'oi zag 11:24 < zipcpi> u'i 11:25 < zipcpi> Well could save myself the trouble if I was willing to accept {zamr-} {zamn-} as a prefix 11:25 < akmnlrse> ki'u tu'a lo karsna zo'u .ei pilno zo'oi zamn- 11:26 < akmnlrse> .i lu za mrabu li'u zo zamnabu 11:27 < zipcpi> Or... 11:27 < zipcpi> exp: brodabu 11:27 < mensi> (CU [brodabu VAU]) 11:27 < zipcpi> exp: broda'abucu 11:27 < mensi> (CU [broda'abucu VAU]) 11:27 < zipcpi> Infinite brodV space 11:28 < akmnlrse> zipcpi: ta'o tu'a lo re plise zo'u do mo lo nu da'i zo ko ka'e se sko'opu na'e bo lo brarai 11:29 < zipcpi> I'm not sure how to deal with scope here, or how that solves the problem 11:29 < zipcpi> The semantic I want is in this bridi: http://jbovlaste.lojban.org/dict/nalvaimenre 11:30 < akmnlrse> da'i sai: {re plise zo'u ko ri mi dunda} -> "there are two apples such that I want you to give them to me" 11:30 < akmnlrse> {ko zo'u ko re plise mi dunda} -> "give me two apples, doesn't matter which" 11:30 < zipcpi> How about {re lo mu plise}? 11:31 < zipcpi> Or zipcpibau: {rexe'e lo mu plise} 11:31 < zipcpi> {re nalvaimenre be lo mu plise} 11:31 < zipcpi> There is a set of five apples I care about 11:31 < zipcpi> But I don't care which two within the set 11:33 < akmnlrse> quantifying nalvaimenre outside of the inner bridi would seem to contradict itself 11:33 < akmnlrse> "there are two apples out of the five, that I don't care which ones they are, but still want you to give me exactly those two" 11:33 < zipcpi> I care about the number two 11:34 < zipcpi> It's why I made a UI instead of a gadri, because I think it makes the most sense attaching to numbers 11:35 <@xalbo> What does it do attached to things other than numbers (and, I guess, gadri)? 11:35 <@xalbo> .i tu gerku xa'a 11:35 <@xalbo> ? 11:35 < zipcpi> I think gerku xe'e is nonsense right now 11:36 < zipcpi> lo xe'e gerku, in my mind, is simply an unquantified "any dog" 11:36 < gleki> xe'e was initially a PA btw 11:36 < zipcpi> Ah 11:37 < zipcpi> ko'a xe'e , however, could mean {lo nalvaimenre be ko'a} 11:39 < zipcpi> da xe'e might be useful, ju'ocu'i 11:41 * nuzba @lojytan_bot: カタカナで表せないこともないって点が魅力的かもしれません(lojban) [http://bit.ly/1IOjgUx] 11:41 * nuzba @lojytan_bot: あ、教材リストのリンクを貼っておきます。 http://mw.lojban.org/papri/%E5%AD%A6%E7%BF%92 あと、ロジバン文法のPEG定義も貼っておきます。 http://subvert-the-dominant-paradigm.net/~jbominji/code/lojban_grammar.peg [http://bit.ly/1IOjFGq] 11:42 < zipcpi> There's also another potential problem. "Choose any number of apples" 11:42 < zipcpi> Where it is the *number* itself that is the subject of nalvaimenre 11:42 < gleki> use klani or its allegedly more precise synonyms 11:43 < zipcpi> But in the end you are choosing the apples, not the number 11:43 < zipcpi> It merely means I don't care about how many you choose 11:44 <@xalbo> Thinking about that, it seems like that {xo'e xe'e plise}, and {re xe'e plise} has the {xe'e} misplaced. 11:44 < zipcpi> Yes... 11:44 <@xalbo> {re lo xe'e plise} or something 11:44 < zipcpi> Maybe 11:45 < zipcpi> What about re lo mu plise though? 11:45 < zipcpi> Will we have to resort to lo xe'e me re lo mu plise? 11:46 < zipcpi> Sounds like perhaps it's a new PA 11:46 <@xalbo> {re lo xe'e mu plise} may work, if you're being picky about there being two apples from the five, but the choice of which ones is arbitrary 11:46 <@xalbo> I guess the real question is what does {xe'e} do in general? 11:46 < zipcpi> But I'm picky about the set of five 11:47 < zipcpi> We could call it {le mu plise} if that strengthens that intent better 11:48 < zipcpi> I'm picky about the set of five, I'm picky about the number two, but I'm not picky about which two as long as it comes from the set 11:49 < zipcpi> That's basically my intention with {xe'e}, to mark what distinctions I care about and what I don't 11:50 < zipcpi> Right now as I has prescribed it, {rexe'e} doesn't say "I don't care" to the number, but rather "I don't care" to how the quantification is implemented, when choosing members of the set 11:51 < zipcpi> THat makes saying "I don't care" to the *number* itself difficult though 11:55 < zipcpi> lo plise poi zilkancu da xe'e 11:55 <@xalbo> I think the problem is how you'd get a UI to do that. The "which ones" part isn't explicitly written anywhere where it's convenient to get a UI to attach to. 11:55 < zipcpi> lo plise poi zilkancu lo nalvaimenre be lo namcu 11:56 < zipcpi> Well, if you have any better ideas... 11:56 <@xalbo> Yeah, I'm not coming up with anything helpful. 12:01 < zipcpi> Hmm... {rone'e bi'i ro} might be what is desired here 12:02 < zipcpi> For "Choose any number of apples" 12:02 < zipcpi> Obviously here {rone'e} = {no} 12:07 < zipcpi> {su'o}? 12:08 < zipcpi> If "any number" is a PA, *xe'o 12:08 < zipcpi> It could also be useful as a digit 12:08 < zipcpi> lo nanca pe li renopaxe'o 12:12 < Ilmen> I recollect that {xo'e} is "any digit" when it's not alone on its own, but is among a digit string. 12:13 < zipcpi> But could it also mean "a specific contextually indicated number"? 12:13 < zipcpi> Essentially a le'elkai be lo ka namcu 12:13 < zipcpi> E.g. xo'epamumoi nanca 12:14 < Ilmen> It would be "some unspecified/obvious/unimportant digit", I think 12:14 < zipcpi> Hm 12:14 < Ilmen> paralleling zo'e/co'e/do'e and friends 12:15 < zipcpi> How about "Choose any number of apples" though? 12:16 < zipcpi> And of course, "Choose any number of these five apples" 12:16 < Ilmen> "ko cuxna lo su'o plise" 12:17 < zipcpi> ko cuxna su'o le meti me mu plise? 12:17 < zipcpi> For the latter 12:17 < ldlework> I always thought of xo'e as the zo'e of quantifiers. It is a quantifier, the speaker is just not saying which. 12:17 < zipcpi> Or le'i mu plise, jo'au gadganzu 12:17 < ldlework> xo'e allows for the listener to pick all or none of the apples 12:17 < ldlework> (or some, of course) 12:18 < ldlework> oh you guys just said that 12:18 < ldlework> u'u 12:18 < zipcpi> The problem here is quite similar to the problem of {le}, or how any of the -o'e things work. Does it mean "I have something I care about, but you should already know which by context" 12:18 < zipcpi> Or, "I really don't care"? 12:19 < ldlework> xorlo? 12:19 < zipcpi> Not sure what xorlo has got to do with the -o'e set 12:19 < ldlework> lo'e means something new in gg, right? 12:20 < zipcpi> Well except being defined based on zo'e, which we're still not sure what it means 12:20 < ldlework> oh, non-gadri -o'e things 12:20 < zipcpi> Actually I don't think {lo'e} differs much from actual usage, among the people who use them 12:20 < zipcpi> Oh yeah, {lo'e} is not part of the -o'e set 12:21 < ldlework> but anyway I think your suggested interpretation is right 12:21 < zipcpi> {lo'e} is, roughly, CLL {lo'e} and CLL {le'e} rammed together 12:21 < zipcpi> GG-{lo'e} 12:21 < zipcpi> Which people have already sort-of been doing in practice 12:22 < ldlework> veridical and non-veridical slammed together to create indefinite generalizations... 12:22 < ldlework> if you say so :) 12:22 < zipcpi> http://jbovlaste.lojban.org/comments.html?valsi=29129;definition=66765 12:22 < zipcpi> Not *exactly* veridicality-non-veridicality, but among the same vein 12:25 < ldlework> la jenca cu pu'i jukpa lo puktce ckafi 12:27 <@xalbo> Neither CLL-lo'e nor CLL-le'e had anything to do with veridicality, really. 12:27 < zipcpi> Well it did, in its explanation, tie it to lo'i and le'i 12:27 < zipcpi> The chosen forms was, at the very least, meant to be a mnemonic 12:28 <@xalbo> Yeah, there's a mnemonic there. 12:28 <@xalbo> I really think people read way more into veridicality than CLL ever intended. Almost all speech is verdical, it's nonveridicality that's unusual. 12:28 < Ilmen> « Does it mean "I have something I care about, but you should already know which by context" Or, "I really don't care"? » −− It's an ellipsis, so the speaker doesn't see a reason to bother expliciting its value/referent, probably because it's obvious or unimportant 12:28 < zipcpi> And it does suffer from a similar intentionality problem, as when one generalizes, it might be difficult to determine if a stereotype is involved 12:29 < ldlework> Ilmen: the important thing though is that there is a referent 12:30 < zipcpi> xalbo: That might be right. Perhaps it's just that worry about calling someone {le nanmu} when it really turns out "he" is a woman 12:31 < zipcpi> Or {lo nanmu} 12:32 < zipcpi> In the end though your intention is to describe what you thought was a man... that you erred there doesn't negate everything else 12:32 < ldlework> xalbo: I agree with that. I go further and say that non-veridicality is so unusual its typically obvious and demand a central lingusitic role. 12:32 < ldlework> and doesn't demand* 12:32 < ldlework> u'u 12:33 < Ilmen> zipcpi: depends on whether the speaker commit to the claim that X nanmu 12:33 < Ilmen> or if loka nanmu is used as some identification device, but is not asserted to be the case 12:33 < ldlework> referring descriptions only mean to identify the referrent 12:34 <@xalbo> In CLL, the point of {le nanmu} is to get the listener to the right referent. It doesn't claim that that referent fills nanmu1, it just claims that describing it that way is the most convenient way to convey the appropriate referent. 12:34 < ldlework> if the listener correctly picks out the right ambiguously man/woman, it is the predication that is asserted upon the referent that we care about whether its true or not 12:34 < Ilmen> So it's like nonce-la 12:34 <@xalbo> ldlework: Exactly. 12:36 <@xalbo> So if we're looking at two statues and {le nanmu le ninmu cu zmadu lo ni melbi}, then I'm talking about the statue of a man, and the fact that it doesn't actually nanmu is irrelevant. 12:36 < zipcpi> So then... what did xorlo do exactly? In the end it doesn't look like CLL-le differs from GG-le that much 12:36 < ldlework> the illocutionary act of reference is to identify, and the intended perlocutionary effect is for the listener to correctly make that identification. predication really happens after this designation of referential description. 12:36 <@xalbo> xorlo didn't change {le} at all, except for quantifiers 12:37 < ldlework> My favorite example of this is 12:37 < Ilmen> Otherwise it would have been called {xorle} ;-) 12:37 < ldlework> "You, get out!" 12:37 < ldlework> said to more than one person 12:37 < zipcpi> Except that now we have {ju'a cu'i} for when the speaker specifically wants to divorce themself from the description 12:37 <@xalbo> The thing that changed in xorlo is {lo}. In CLL, {lo} is verdical. Now that I've gone into what non-veridicality is, veridality is the opposite: what you say is what you mean. 12:37 < ldlework> No listener can decide whether to be effected by the perlocutionary imperative until the referent of the referring description is resolved. 12:38 <@xalbo> So {lo nanmu} is "a man". Not "something that really and truly honestly is a man, I've checked the chromosomes and everything", but "a man". 12:38 < zipcpi> lol 12:39 < ldlework> xalbo: yeah imagine two men dressed up as balerina and a lumberjack 12:39 < zipcpi> So, any man, some man, could be a specific man, could be a generalization about men...? 12:39 < ldlework> "I chose the man" 12:39 < ldlework> Likely identifies the lumberjack 12:39 < ldlework> taking classic gender roles into account I suppose :( 12:39 <@xalbo> zipcpi: That's the big difference with xorlo. In CLL, {lo nanmu} really does mean something very, very much like {su'o da poi nanmu}. 12:40 < zipcpi> Ah... 12:40 <@xalbo> So it's satisfied if there's any man such that... 12:40 <@xalbo> So never a specific man (or never mandatorily one) 12:40 < zipcpi> So in CLL, lo nanmu also asserts the "man-ness", rather than merely identifying a referent 12:40 < ldlework> Just like the default illocutionary act is assertive I think the default referential act is veridical, that's all 12:41 < zipcpi> Hm 12:41 < zipcpi> Oh 12:41 < zipcpi> I get it now 12:41 < zipcpi> "satisfied if there's any man such that..." 12:41 <@xalbo> zipcpi: Yes, because it doesn't have anything else to go on. It's never specific, so it's always just "anything such that it satisfies nanmu1". If you remove the veridicality, you end up with literally anything. 12:41 * zipcpi nods 12:41 < ldlework> but that's okay 12:42 < ldlework> because truth-valueness doesn't depend on whether the referents of descriptions actually satisfy the descripitons used to identify them 12:42 < ldlework> its whether the predication of the referents is true, which is the agenda of the speaker 12:43 < zipcpi> ldlework: I think, under CLL, {lo nanmu} is not to refer to a man, it's to say "There exists a man" 12:43 < zipcpi> Not refer to a specific man 12:44 <@xalbo> Of course, once you say "There exists a man" you have a referent to him, and you're probably making a statement about him, but yes. 12:44 < ldlework> xalbo: ie 12:44 < zipcpi> Yeah 12:44 < ldlework> I want to hazaard blaming it on taking the formalization of fopl statements too seriously. 12:45 < ldlework> Like when you see "There exists X such that" 12:45 < zipcpi> I suppose then, what xorlo did, was to say "Why do we need so many ways to represent {da}?" 12:45 < zipcpi> "Why is there no difference between {pa nanmu} and {pa lo nanmu}?" 12:45 < zipcpi> I mean {lo pa nanmu} 12:47 <@xalbo> What xorlo did is object to some of the consequences of {lo nanmu} meaning {su'o da poi nanmu}. The classic examples were {mi kalte lo pavyseljirna} and {mi nitcu lo mikce}. 12:48 * zipcpi nods 12:48 <@xalbo> In the former, I can hunt for unicorns without there being any unicorns in the world such that I hunt for them. In the latter, likewise. 12:48 < ldlework> I like the idea that da merely creates room for a referent, which you then relativize sufficiently to identify what fills the reference, and then a second illocutionary act takes place where the referents, not their descriptions undergo a process of being involved in a propositional act where they are related to each other of themselves via verbs or predicates 12:48 <@xalbo> There are doctors, but it's not the case that there's a doctor out there that I need. Rather, I need that there be a doctor. 12:48 < ldlework> Then, depending on the illuctionary force indicator (xu, etc) the proposition is used to perform some speech act 12:49 < ldlework> such as asserting, or asking questions, or denying, or promising 12:49 < zipcpi> xalbo: Yeah... 12:49 <@xalbo> ldlework: That sounds pretty exactly like CLL lojban. 12:50 < ldlework> xalbo: this doesn't assert that the referent that is eventually identified by way of the description is actually true of the description though 12:51 <@xalbo> CLL lo is {su'o da poi}. CLL le is {su'o da voi}. 12:51 < ldlework> since assertion is a specific illocutionary act that takes place after the propositional act, and the dependent references 12:51 <@xalbo> Sorry, {ro da voi} 12:51 < latro`a> that sounds like you more or less want a voi-cu 12:51 < zipcpi> Anyway, back to "any number", I was also thinking about MTG, where "any number of target creatures" is a common phrase, and includes zero. I suppose {su'o no} or {su'e ro} would work? 12:51 < ldlework> Sure, then its something like, the truth value of gadri-less expansions don't depend on relative clauses 12:51 < latro`a> I've suggested that and some related things before 12:52 < ldlework> Reference is an act of language. 12:52 < zipcpi> I'm thinking of entering {su'ero} just to gloss it as "any number" 12:52 < zipcpi> Cause I don't think it means anything else 12:52 < ldlework> I think that when expanding it out to existential forms that the logical entailment of the statemetns truth-value only consists of the predecation of referents in the larger propositional act. 12:53 < ldlework> Once you reach the propositional act, where you have referents and you are predicating them, the descriptions are no longer involved since they have served there purpose of identifying the referents. 12:53 < latro`a> the catch is that there really are no actual referents "in" a fully quantified statement. they're out there in the universe of discourse, but you can't safely instantiate them to re-use them in more sentences, partly because the existentially quantified ones might be functions of previously universally quantified ones 12:54 < latro`a> people struggle with this in mathematics constantly 12:54 < ldlework> The proposition deals with the referents and the predicates involved. Then, without any special illocutionary force identifer to change the defeault act from assertion, an assertion is made about the referents in terms of the predicate. 12:54 < ldlework> The problem is mistaking logical calculus for speech, I think. 12:55 < latro`a> as an example: {ro da poi mulna'u ku'o de mulna'u gi'e zmadu da}; {de} can never be instantiated 12:55 < ldlework> The logical calculus is there, but at a very specific part of the process of the speech act. 12:55 < latro`a> (sorry to be sorta talking over you, ldlework) 12:55 < ldlework> I argue it lies right at what Searle calls the propositional act. 12:55 < ldlework> Which comes after references are designated, but before the resulting proposition undergoes illocutionary force. 12:56 <@xalbo> zipcpi: I feel slightly more comfortable with {su'o no} than with {su'e ro}, because there are contexts where {ro} is ill-defined, but I don't think there are any where {no} isn't fine. Does that make sense? 12:56 < latro`a> "some number of sets which don't contain themselves" 12:56 < latro`a> :p 12:56 < zipcpi> On the other hand though, {no} isn't necessarily the smallest number 12:56 < latro`a> (sorry) 12:56 < zipcpi> In some contexts, negative numbers make sense 12:56 <@xalbo> zipcpi: Good point. 12:57 < zipcpi> That's even why I entered {no'ai} and {no'e'u}; because when working on {xokauva'e} scales, I realized that zero could mean two very different things depending on the scale used 12:59 < latro`a> incidentally, I have thought about actually talking in lojban in a framework like you're saying, ldlework 12:59 < ldlework> When you want to consider the truth value of a statement, it is enough to know whether the surface assertion is true of the referents of descriptions. The question and interrogation of whether intended referents are true of their descriptions only has value in the resolution of unsuccessful perlocutionary designation. 13:00 < latro`a> where I have a bunch of sentences that don't actually say anything, but merely define the relevant objects and give them shorthand symbols, and then we actually talk about them 13:00 < latro`a> it's like declarative programming 13:00 < zipcpi> ua 13:00 < ldlework> latro`a: I like that metaphor 13:00 < ldlework> I used it last night on zipcpi 13:02 < ldlework> I think that this observation has a lot of effects on what we're actually trying to do when we expand lojbanic statements to more existentially logical forms 13:03 < ldlework> Sometimes I think we get distracted by our referring descriptions when we probably usually really want to understand the surface assertion or at least proposition 13:03 <@xalbo> ldlework: I think that's taking things too far. Are you saying that "I'm eating an apple" isn't a statement about whether or not you're eating an apple, but rather "an apple" is just trying to get some referent (doesn't matter what), and then you're asserting that you're eating that? 13:03 < ldlework> xalbo: very much like the default illocutionary force is assertive, the default referential one is veridical 13:03 < zipcpi> What about post-xorlo {mi citka lo plise} then, xalbo? 13:03 < ldlework> xalbo: in english say it with some inflection 13:03 < ldlework> I'm eating "an apple" 13:03 < ldlework> I'm eating an _apple_ 13:03 < ldlework> etc etc 13:04 < zipcpi> JVS isn't cooperating 13:04 < latro`a> I always came back to the same basic problem: to go from referential to existential, you have to explicitly label the domain 13:04 < zipcpi> Guess it really doesn't like {su'ero} 13:04 < ldlework> if you say "mi citka lo plise" without strange context, or any other force indication, I must understand that you are indeed really eating an apple 13:05 < latro`a> sorry, to go from referential to existential while preserving all meaning 13:05 < zipcpi> Well I'm trying to stretch existing cmavo to avoid inventing a new one; but if too many people object I might have to 13:06 < ldlework> communication is a dance of illocutionary desire and perlocutionary effects 13:06 < ldlework> words don't really have inherent meaning 13:06 < ldlework> by the fact we can perform utterance acts without actually meaning or saying anything 13:07 < ldlework> all of this gets in the way of just wanting to consider the logical entailment of some lojban statements 13:08 < ldlework> I think when you're doing logical proofs and such, you just work with the assumption that everything involved is veridical 13:09 < ldlework> Since it literally makes no sense for any logician to ever work with non-veridicality in in their work 13:09 < ldlework> But then its just an awkward thing when you try to analyze actual speech this way 13:10 < ldlework> And further to say to tell people that when they use the language they have undue logical consequences 13:11 < ldlework> Because of constraints that really only apply utility in a specific domain of analytics 13:26 < zipcpi> Ugh... having trouble accessing JVS and LMW right now 13:29 < durka42> maybe you've been rate-limited :p 13:30 < ldlework> u'i 13:32 < zipcpi> Error: You have used up all your cmavo space allocated for... forever 13:32 < zipcpi> To buy more cmavo space, send a trillion dollars to Robert LeChavelier. Thank you 13:34 < zipcpi> btw wrt-to "wrong hyphens", would you feel a bit better if it was -l-, since -l- can never be a hyphen 13:35 < zipcpi> ci'onmau could still be a problem though :p 13:35 < durka42> yeah but I love {ci'onme'a} so no big :) 13:35 < zipcpi> :p 13:43 < zipcpi> What might be nice wrt typographical conventions though, if there was some method of unambiguously shortening e.g. bypyfyky into BPFK 13:43 < zipcpi> Some people use capital letters, but that's ambiguous with the use of capital latters for syllabic stress or cmevla 13:44 < durka42> if you use capital letters for lerfu, then you'd better use accent marks for stress :) 13:44 < zipcpi> ie 13:44 < durka42> however, it should be able to disambiguate them, anyway 13:44 < durka42> you never stress a single-syllable cmavo 13:44 < durka42> so any word with all the letters uppercase is a lerfu string 13:45 < durka42> whereas a word with just one uppercase letter is a stress marker 13:45 < zipcpi> Actually I often do in actual speech... logalraiKENkaBElofraGAri 13:45 < durka42> s/just one/just some 13:45 < fenki> durka42 meant to say: whereas a word with just some uppercase letter is a stress marker 13:45 < durka42> s/letter/letters 13:45 < fenki> durka42 meant to say: whereas a word with just some uppercase letters is a stress marker 13:45 < zipcpi> Ah. 13:46 < latro`a> I'm actually not sure how to split that, zipcpi 13:46 < durka42> I mean what you said can't be an abbreviation/lerfu string 13:46 < zipcpi> exp: logalraiKENkaBElofraGAri 13:46 < mensi> ([lo {galraikenka } KU] VAU) 13:46 < durka42> why did you stress the BE 13:46 < latro`a> why stress the be 13:46 < durka42> I guess it's allowed 13:46 < latro`a> yeah 13:46 < latro`a> that confused me 13:46 < zipcpi> Just my speech style 13:46 < latro`a> because I expected the next word to be a brivla 13:46 < latro`a> but then it would have to fragment 13:47 < durka42> oh you were objecting to me saying nobody stresses single-syllable cmavo 13:47 < durka42> yeah so then, use accent marks :) 13:47 < latro`a> oh, I see 13:47 < latro`a> that's...weird 13:47 < latro`a> that's like stressing "of" in english 13:47 < zipcpi> lol 13:47 < durka42> co'o 13:47 < ctefa`o> en: ju'ei 13:47 < mensi> ju'ei = [JUhEI] Tight scope bridi separator; analogous to .i without ending the abstractor or relative clause. |>>> "nu 13:47 < mensi> broda ju'ei brode" is equivalent to "nu ju'e gi broda gi brode". Similarly "ko'a noi broda ju'ei brode" is equivalent to 13:47 < mensi> "ko'a noi ju'e gi broda gi brode". |>>> latros 13:47 < zipcpi> ctefa'o: I thought ju'ei could be replaced by jeNOI, but it can't within NU 13:48 < zipcpi> Without ju'ei I had to resort to {kei jo'u LE NU}, which is... yuck 13:51 < zipcpi> latro'a: lol maybe it's weird... I was actually thinking it more like some kind of silly TV advert 13:51 < zipcpi> logalraiKENka! BElofraGAri! 13:53 < zipcpi> I kindof-randomly stress cmavo; I do have to be careful though because I hear if you stress them before a brivla you need a mandatory pause 13:54 < zipcpi> LO.jboBANgu vs LOjbo BANgu 13:54 <@xalbo> If you stress the syllable before a brivla that has su'oci syllables, then yes, it breaks differently. 13:54 * zipcpi nods 14:04 < Donald_ET3> no lojban.org 14:04 < zipcpi> ie 14:05 < Amnadya> ür 14:05 < zipcpi> ro da pe la'o u lojban.org u cu spofu 14:09 < zipcpi> All spoffed up 14:09 < Donald_ET3> auuu :( 14:09 < Donald_ET3> I want to learn lojban. 14:10 < Donald_ET3> I only got as far as learning the pronounciation. 14:10 < zipcpi> Huh funnily this works: http://mw.lojban.org/extensions/ilmentufa/i/en/index.html 14:10 < Ilmen> Isn't it because it uses your local cache? 14:10 < zipcpi> Maybe 14:11 < zipcpi> I have no idea how sutysisku works :p 14:11 < Ilmen> Donald_ET3: You can also learn Lojban here, on IRC; there's a #ckule (school) channel for beginners 14:12 < zipcpi> This is about the point where I would link you to the Crash Course and the Wave lessons but they're both on the wiki 14:13 < zipcpi> Which is currently spoffing up 14:14 < Ilmen> Donald_ET3: Unfortunately it's pretty frequent that the main site gets down, until some admin restarts it 14:14 < Donald_ET3> Oh. Okay. 14:15 < Ilmen> Maybe you should consider saving the Crash Course or the Wave Lessons when possible on your disk, so that you can read them offline 14:16 < Ilmen> en: coi 14:16 < mensi> coi = [COI] vocative: greetings/hello. |>>> officialdata 14:16 < zipcpi> Crash course is being updated constantly though 14:16 < Ilmen> at least the dictionary bot is responding 14:16 < Donald_ET3> okay 14:16 < zipcpi> Oh yeah here's your first word for today: 14:16 < zipcpi> en: spofu 14:16 < mensi> spofu = x1 is broken/inoperable/broken down/non-utile/not usable for function x2. |>>> Agentive break, cause to become 14:16 < mensi> inoperable (= pofygau, pofyzu'e); accidentally break, as a result of an event, non-agentive (= pofyja'e, nutpo'uja'e). 14:16 < mensi> See also daspo, katna, porpi, se xrani, cikre. |>>> officialdata 14:16 < menli> lol 14:19 < zipcpi> "spoff" is an English slang term I invented based on {spofu} :p 14:26 < ldlework> people at Docker are saying "krillin it" 14:26 < ldlework> its unbearable 14:26 < ldlework> zo'o 14:30 < zipcpi> ta'o How about sumdza for "object" in the sense of object-oriented-programming, or any object that can go into a sumti? 14:31 < zipcpi> Cause dacti means physical object 14:47 < ctefaho> skami dacti 14:47 < ctefaho> dun 14:48 < ctefaho> skami zei dacti 14:48 < ctefaho> skami zau dacti 14:48 < ctefaho> skami'ydacti 14:48 < ctefaho> all beautiful 14:49 < zipcpi> No... I said that it also works for any object that goes into a sumti 14:50 < ctefaho> SKACTI 14:50 < zipcpi> That's a lujvo 14:50 < zipcpi> I think skari citka 14:50 <@xalbo> {skacti} is a lujvo, {skari zei citka}, colored-eat 14:51 < ctefaho> THAT'S RIGHT 14:51 < ctefaho> beware of the color monster 14:51 * ctefaho out 14:51 < zipcpi> Sorry, we're not implementing selpa'is free-lujvo yet 14:51 < ctefaho> I have an alternative to merging zi'evla and lujvo 14:51 < ctefaho> so you can compete with disliking sepla's or my idea 14:52 * ctefaho out2 14:52 <@xalbo> I do think something based on {sumti} is pretty good for an "object" in the OO sense. 14:57 < zipcpi> The BSFK has no objections to zi'ejvo per-se (though I can see the argument about obscuring etymology, especially for reljvo with common rafsi; after all, I recognized those rafsi without being anywhere near rolrafcre), but the BSFK doesn't decide what gets to be entered into JVS 14:58 < zipcpi> If I actually tried to enter a zi'ejvo in there I'd be downvoted right away 14:59 < zipcpi> So nope, too much too quickly 14:59 < Ilmen> co'o ro do 14:59 < zipcpi> co'o 15:08 < zipcpi> Maybe sumti zei dacti is usable, but I'm not sure how many people like an invisible -pev- or -zam- attached to the tertau of the veljvo 15:09 < em> do you capitalize the word lojban? 15:09 < em> when referencing it? 15:09 < zipcpi> In English, yes, in Lojban, no, although in Lojban we often use {lo jbobau} 15:09 < ldlework> or la lojban 15:10 < zipcpi> Yeah 15:10 < em> so if im talking about Lojban in English then I capitalize it? 15:10 < zipcpi> Yep 15:10 < zipcpi> English treats names of languages as proper nouns and we see no reason to try to break that tradition there 15:12 < zipcpi> I hear in French it is different? 15:12 < zipcpi> Not sure 15:15 < zipcpi> rafsi: dacti 15:15 < mensi> zo'oi dai .e zo'oi dact rafsi zo dacti 15:15 < zipcpi> sumdai seems nice except for that nitpick I mentioned 15:15 < zipcpi> I could just enter it and brace for downvotes 15:16 < ctef> you can capitalize it if you use other methods than capital letters to indicate stress, em 15:17 < ctef> la .Lòjban. 15:17 < ctefaho> obligatory no standard ortography reminder 15:17 * ctefaho out3 15:20 < zipcpi> k: broda'abucudu'efuguxe'ijukulumunu'opukerusutu'uvuvexujezu 15:20 < mensi> (CU [Z:broda'abucudu'efuguxe'ijukulumunu'opukerusutu'uvuvexujezu VAU]) 15:23 < zipcpi> Or make {zamne} a gismu for "jargon" lol 15:24 < zipcpi> Then all zi'evla with zamn- is assignable 15:25 < zipcpi> Another solution; {cei} to cmevla, but relies on BCM 15:28 < zipcpi> k: brody'a 15:28 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "b" found. 15:28 < zipcpi> That's too zbusufukai right now li'a 15:33 < zipcpi> k: zamy'itloku 15:33 < mensi> (CU [L:zamy'itloku VAU]) 15:36 < zipcpi> Still spoffed. Think we need to tell Robin 15:38 < zipcpi> But yeah cmevla would have the advantages of being nice and short like lerfu-sumti. ki'ogra cei .kygys. - though those pauses are a bummer 15:50 < ldlework> doi ro do ju'i .i mi nitcu lo nu jungau lo vajni 15:51 < ldlework> .i mi te benji di'e poi nuzba .i lo mi desku ladri cu mukti lo nu ro le nanmu cu zvati le bimselsru 15:51 < ldlework> mu'o 15:52 < ldlework> di'a pu co'e 15:58 * nuzba @Pontama314: 日本語話者でlojbanを話題してる人が、そこそこいるな。 so'o ponpre cu tavla la lojban ui [http://bit.ly/1T7B18u] 16:22 < ctefaho> does anyone here practice jbobau handwriting? 16:23 < dutchie> i did, when i was working through the lessons 16:23 < dutchie> did all the exercises longhand 16:25 < ctefaho> longhand? 16:25 * ctefaho has learnt a new word today 16:27 < ctefaho> also learnt that the schwa doesn't seem to ever have ´ or ` 16:28 < dutchie> why do you ask about handwriting lojban? 16:28 < ctefaho> was just curious. am starting to practice it myself 16:30 < ctefaho> just wrote down what I call "unicode lojban" (` for stress, ŭ and ị) 16:30 < ctefaho> no keyboard restrictions with handwriting:) 16:31 < ctefaho> heh it feels way different writing it by hand for some reason 16:38 < ctefaho> almost like a real language zo'o 16:40 * nuzba @bgcarlisle: Short story prompt for Lojbanists: la cizra mensi http://www.bgcarlisle.com/blog/2015/07/01/short-story-prompt-for-lojban-enthusiasts-la-cizra-mensi/ #lojban [http://bit.ly/1GOPM8D] 16:44 < bigcentaur> hey what happened to mw.lojban.org/extensions/ilmentufa/camxes-exp.html ? 16:44 < ctefaho> whole lojban.org seems to have catched fire 16:44 < bigcentaur> ahhh 16:44 < bigcentaur> blast ha 16:44 < ctefaho> cll I can access at least 16:45 < ctefaho> oh god the macro expansions get so much easier when you can draw the whole thing 16:47 < ctefaho> why have i never tried this before 17:01 < zipcpi> ldlework: Wow I don't know how we're going to translate this. The little I read from Amazon's preview talks a lot about the meanings of specific English words, as well as English metalinguistic concepts like "noun" 17:02 < zipcpi> It would... basically need a rewrite if we're going to talk about similar Lojbanic concepts 17:02 < ldlework> zipcpi: I don't know why you think lojbanists cannot talk about 'nouns' in the wider linguistic sense 17:02 < zipcpi> ldlework: We "can", but he refers to specific English words 17:03 < ldlework> Noun isn't an idea specific to english though. 17:03 < zipcpi> Yes, but it'd be very hard to choose comparable words 17:03 < ldlework> "Noun" is the english word for the idea for which the word rperesents. 17:03 < ldlework> We have a word for that too 17:03 < zipcpi> Like there's the part that talks about the word "witch" for example, which relies on the history of the usage of the word "witch" 17:04 < zipcpi> Lojban doesn't have a comparable word 17:05 < ldlework> sumti is the comparable word 17:05 < zipcpi> We'd essentially have to rewrite those parts in Searle's place, and I'm not up for that there. 17:05 < zipcpi> No, not the word "noun" per se 17:05 < zipcpi> But the specific words he considers nouns 17:05 < ldlework> Also there's nothing wrong with quoting english words when talking about them 17:05 < zipcpi> Theoretically, yes, but is this book supposed to be about English? 17:06 < ldlework> There's a difference between being written in a language and talking about specific words in a language. 17:06 < ctefaho> coho 17:06 < ldlework> When he uses the word noun, he's merely accessing the semantic which exists behind the english word noun. 17:06 < zipcpi> Yes, but that basically means that this translation can only be understood by English speakers, who understand the history behind the English word "witch" 17:07 < ldlework> zipcpi: Which is probably fine since it is a book about linguistics. 17:07 < ldlework> And language philosophy. 17:07 < ldlework> I mean there has to be a way to translate such works. 17:08 < zipcpi> If we're going to translate it in such a way that it can only be understood by English speakers, what's the point of it being in not-English? 17:08 < ldlework> We can even just subtext right in the discourse what the quoted english word means 17:08 < ldlework> In lojban 17:09 < ldlework> (to du zo termafyfe'i toi) 17:09 < zipcpi> The problem is that he doesn't just rely on denotation, but connotation 17:09 < ldlework> Which we can subtext. 17:10 < zipcpi> I don't know how to describe all the nuances of connotation in Lojban 17:10 < ldlework> Specific connotations? 17:10 < ldlework> Not the premise of connotation itself. 17:10 < ldlework> Don't exaggerate the task :) 17:11 < ldlework> zipcpi: in any case reading the book will be value on its own even if we don't translate it 17:11 < zipcpi> je'e 17:11 < ldlework> but it should standout as worrying that we can't translate an arbitrary work 17:11 < zipcpi> It's difficult to translate these sorts of texts into *any* language, not just Lojban 17:12 < zipcpi> I'm sure there are translators much more certu than me, but... 17:13 < zipcpi> But even say... French; I'm sure they have lots of words with a similar connotational history to "witch"; heck it might even be their same word 17:13 < zipcpi> Lojban, being a constructed language that isn't even a century old, doesn't have that 17:13 < ldlework> I think it comes down to convention 17:13 < ldlework> That's all 17:14 < ldlework> Got a problem? Come up with a convention to give the reader the missing context or information 17:14 < ldlework> subtext, footnotes, etc 17:14 < zipcpi> Problem is that I don't know what the missing information is myself. Connotation is a very... bellyfeel thing 17:15 < ldlework> I am considering writing an essay that frames lojban upon the thinking in this book and can therefore just write the information in the way that's most natural 17:16 < ldlework> Giving a sizable background on that thinking before delving into its consequences for lojban 17:17 < ldlework> zipcpi: so did you get it? 17:18 < zipcpi> Not yet; just reading through the preview 17:19 < ldlework> how long is it? 17:19 < zipcpi> It ends a bit past 1.4 "Why study speech acts?" 17:19 < ldlework> ah 17:20 < ldlework> zipcpi: also there's fu'ivla right? 17:20 < ldlework> and zi'evla 17:20 < ldlework> doesn't solve the problem entirely 17:24 < zipcpi> Yes, I have that idea for things like illocutionary act -> itloku nunzu'e, without bothering over exactly what itloku means, but for importing the connotational history of "witch", it doesn't quite work 17:25 < ldlework> There's gotta be a way to talk about another language's synonyms, homonyms and so on from a lanaguage other than that language. 17:25 < zipcpi> Like I said, it is possible if we want this translation to be about English 17:26 < zipcpi> Not so much if we want it to be a general reference about language philosophy 17:26 < ldlework> Well I think the examples being english isn't a problem really 17:26 < ldlework> Specially if you want to talk about linguistic effects that lojban makes impossible 17:27 < zipcpi> Hm 17:27 < ldlework> I do think it may potentially be more worth while to write our own book 17:27 < ldlework> :) 17:27 < ldlework> That summarized the content of this book and frames it with regard to the details of our langauge. 17:28 < ldlework> But we can wait to see if you see the same value in its content that I do 17:28 < zipcpi> selzu'e, not nunzu'e 17:29 < _mukti_> coi jbopre 17:29 < ldlework> coi 17:29 < zipcpi> ... who defined {selpre} even though {prenu} isn't defined with an x2? lol 17:30 < _mukti_> .u'i 17:30 < _mukti_> Not a jvajvo, apparently 17:31 < zipcpi> ue da poi jbopre cu zbusufukai semau mi iau xo'oru'e 17:31 < _mukti_> .u'i 17:32 < _mukti_> ua la tijlan cu finti 17:32 < _mukti_> .i mi se spaji 17:33 < zipcpi> ta'i ma do djunybi'o .i la jbovlaste ca spofu 17:34 < ldlework> la jbovlaste po'o nai cu jbovlaste 17:35 < zipcpi> ku'i la sutysisku na arco lo finti 17:36 < zipcpi> en: selpre 17:36 < mensi> selpre [< se prenu ≈ 2nd conversion people] = p2 is the personality/persona exhibited by p1. |>>> x2 can be other than a 17:36 < mensi> human. Cf. seltra, selzu'e. |>>> tijlan 17:40 < _mukti_> vlaste: selpre (finti) 17:40 < vlaste> selpre (finti) = tijlan 17:46 < zipcpi> rafsi: zo'e 17:46 < mensi> y no da se tolcri 17:46 < zipcpi> Why can't zo'e have a rafsi... lol 17:48 < zipcpi> What are the status of co'e-lujvo anyway? Do any exist? 17:48 < zipcpi> Well, {comco'e} lol 17:48 < zipcpi> x1 is a thingamajig / thing which the speaker can't put a predicate upon 17:50 < zipcpi> Still thinking about "object" as in "physicality unspecified; any object that can be the subject of a noun/sumti, or "object" in object-oriented-programming 17:50 < zipcpi> Candidates I have right now are {sumdza}, {sumdai}, and {sumco'e} 17:51 < zipcpi> Which would be best? 17:54 < zipcpi> I like the sound of {sumdai} the best, though that somewhat abuses the tertau 17:55 < zipcpi> And I don't know what JVS users tend to think of that 17:56 < _mukti_> xu do co'e fo do 17:56 < _mukti_> si da 17:56 < zipcpi> .y. ki'a 17:58 < _mukti_> si'au da ve co'e .i mi te preti fi do gasnu lo nu go'i 17:58 < _mukti_> (fi lo du'u) 18:16 * nuzba @464161niftg: @Pontama314 .ui do facki lo du'u go'i zo'e fi la lojban kei .uadai .u'adai [http://bit.ly/1CLhe5h] 18:16 * nuzba @464161niftg: zo'o su'o do ka'e cilre la lojban la lojban .i si'a zo'o su'o do ka'e tavla la lojban la lojban [http://bit.ly/1f38pyU] 18:20 * nuzba @464161niftg: ji'a sai su'o do ka'e cilre la lojban la lojban la lojban la lojban gi'e ka'e tavla la lojban la lojban la lojban vau vau zo'o .i baljbo sai [http://bit.ly/1CLhKjU] 18:41 < zipcpi> Anyone home? 18:42 < noncomcinse> cipra 18:42 < zipcpi> coi 18:42 < noncomcinse> .i lonu cipra cu noroi snada 18:42 < noncomcinse> coi la zipcpi 18:43 < noncomcinse> .i lodu'u go'oi cipra na jetnu 18:43 < noncomcinse> .i lonu cipra cu ro roi snada va'o lonu cilre 18:43 < noncomcinse> doi la zipcpi do mo 18:49 < zipcpi> milxe lo ka badri le nu la jbovlaste je la lobyuitki na akti 18:49 < zipcpi> mi simsa lo'e vinta iau u'ise'i 18:50 < noncomcinse> jbo: akti 18:50 < mensi> akti = x1 ca'a zilpilno x2 |>>> ko catlu fe ji'a zo pilno |>>> gleki 18:50 < zipcpi> tolspofu 18:50 < zipcpi> va'i jy je ly spofu 18:50 < noncomcinse> je'e 18:51 < noncomcinse> .y. le lojbo cu tolspofu ua 18:51 < noncomcinse> en: iau 18:51 < mensi> 7 da se tolcri: mlatrmiau, mriiaucmo, bantumu'u, bankubu'i, banpunu'a, jvisuai, jdikyplini 18:51 < zipcpi> i'au 18:52 < zipcpi> en: i'au 18:52 < mensi> i'au = [IhAU] reset bridi-level to zero |>>> Auto-inserts all necessary terminators to reset to the top-level bridi of 18:52 < mensi> the sentence (i.e. everything that i would insert except the last vau). May be used with xoi. May also be used with 18:52 < mensi> UI-cmavo; UI-cmavo placed immediately after would apply to the entire sentence; see ji'au, fu'e, fu'o. |>>> 18:52 < mensi> spheniscine 18:52 < phma> coi 18:52 < zipcpi> ki'u lo nu mi cafne co pilno zo i'au kei mi nalcatni snigau zo iau 18:52 < noncomcinse> zo .i'au cu panra lu cu'au xi da'a li'u .i ku'i zo cu'au cu spogau 18:53 < noncomcinse> sa .i zo xi spogau tu'a zo cu'au 18:53 < zipcpi> u'i 18:53 < noncomcinse> en: cu'au 18:53 < mensi> cu'au = universial famyma'o: terminates the most recently opened construct or clause. |>>> cu'au looks back for the most 18:53 < mensi> recently opened construct that has not been terminated, and emulates whatever famyma'o would terminate it. It can also 18:53 < mensi> be subscripted with xi, and will terminate that many times. Note that that means grammatical function is being put in a 18:53 < mensi> xi clause, so be careful when using it. Additionally, cu'au xi ro will terminate all the way up to the last 18:53 < mensi> sentence-starting word (.i mi klama lo zarci pe lo pendo be mi cu'au xi ro -> .i mi klama lo zarci pe lo pendo be mi 18:53 < mensi> cu'au-be'o-ku-ge'u-ku-vau). This will also terminate to sentences started in lu (will NOT emulate li'u UNLESS used 18:53 < mensi> multiple... 18:53 < mensi> [mo'u se katna] http://jbovlaste.lojban.org/dict/cu'au 18:53 < noncomcinse> coi la'oi .phma. 18:53 < zipcpi> More like cu'au xi da'a, but works like cu'au xi ro for attitudinals 18:55 < noncomcinse> xu zo .i'au zo'u lo zbasu cu pensi zo .i joi zo vau .i'au vau pe'i 18:55 < noncomcinse> exp: .i mi klama lonu zgike .i'au ui 18:55 < mensi> SyntaxError: Expected [oO] or [yY] but "i" found. 18:55 < noncomcinse> .i lu co'e .i'au .i'au .i'au .i'au .i'au .i'au li'u gendra 18:56 < zipcpi> na ca'a se sanji da poi gentufa 18:57 < noncomcinse> .i ku'i lo su'o re me zo .i'au na jai se nitcu 18:57 < zipcpi> ja'a go'oi zbasu 18:57 < phma> coi la noncomcinse .i mo 18:57 < noncomcinse> pensi lo cizra valsi gi'e ke cizra je se cizra ke'e be su'oda 18:58 < noncomcinse> gi'e troci lonu cuxna fi lo me'oi .musicplayerdaemon. samselpla 18:59 < zipcpi> .ije mi sarji sai lo du'u zo i'au cu inda lo ka binxo lo pamei slaka cmavo 19:01 < zipcpi> ki'u lo nu zifre lo nu na pensi lo gerna stura ja lo du'u lo xokau famyma'o cu jai sarcu 19:02 < noncomcinse> je'e nai la zipcpi 19:02 < zipcpi> ma na se jimpe 19:03 < noncomcinse> lodu'u go'oi slaka 19:03 < noncomcinse> cu cfipa 19:03 < zipcpi> zo iau 19:03 < _mukti_> Web site still ailing I see. 19:04 < zipcpi> ja zo go pebau la tcekitaujau 19:04 < zipcpi> mi nalcatni snigau zo iau 19:05 < noncomcinse> je'e la zipcpi 19:06 < phma> inda ki'a .i la jbovlaste cu si'au spofu 19:06 < zipcpi> en: inda 19:06 < mensi> inda = x1 deserves / merits / is worthy of x2 (property of x1) |>>> See also jerna. |>>> 19:06 < mensi> Ilmen 19:06 < zipcpi> ja'aku jyvysy spofu 19:06 < _mukti_> ie 19:06 < _mukti_> za'a 19:07 < _mukti_> lo selfu skamu cu na frati 19:07 < noncomcinse> camxes: tu'a ko'a .e tu'a ko'e 19:07 < camxes> ([tu'a {ko'a } LUhU] VAU) 19:07 < _mukti_> s/skamu/skami/ 19:07 < fenki> _mukti_ meant to say: lo selfu skami cu na frati 19:07 < noncomcinse> TIL {tu'a} has a terminator. 19:08 < noncomcinse> noroi djuno pu 19:08 < zipcpi> zo lu'u famyma'o ma'oi LAhE 19:09 < noncomcinse> .i pau da'i ganai me lo'u ma'oi COI broda le'u gi mi ko'oi sruma tu'a zo lo lu'u ba'e ji tu'a zo la lu'u 19:11 < noncomcinse> (sei lo'u ma'oi COI broda le'u lo'u ma'oi COI ba'e lo broda le'u se'u) 19:11 < zipcpi> pe'i le gadri poi se claxu cu ambigu 19:14 < zipcpi> lu li'u cu binxo lu li'u ja lu li'u 19:14 < noncomcinse> .i mi djica tu'a zo la fi'o te jinvi: lonu lakne fa lonu lo broda ka'e me la broda 19:15 < zipcpi> jdima pe'a lo nu vimcu le gadri 19:15 < noncomcinse> lu zo'e ze'ei lo li'u co'e xu 19:16 < zipcpi> je'u do'a 19:16 < noncomcinse> camxes: je'u 19:16 < camxes> je'u 19:16 < noncomcinse> .i mi fliba tcidu zo do'u 19:16 < zipcpi> zo do'a 19:16 < noncomcinse> .i mi ba co'arsanji 19:17 < noncomcinse> 21:14 < noncomcinse> .i mi djica tu'a zo la fi'o te jinvi: lonu lakne fa lonu lo broda ka'e me la broda 19:18 < noncomcinse> .i di'u srera 19:18 < noncomcinse> .i mi na djuno lodu'u makau tadji lo mi skami .i'au vau li'a 19:19 < zipcpi> u'i 19:58 < zipcpi> Thinking of new slate of zi'evla 20:00 < zipcpi> ki'orbi me'orbi gi'orbi te'orbi pe'orbi xe'orbi ze'orbi go'orbi 20:00 < zipcpi> Binary prefixes 20:01 < zipcpi> Rafsi -ki'orb(y)- etc 20:02 < zipcpi> Gleki has proposed bitsi for bit, and bivbitsi for bype 20:02 < zipcpi> So ki'orbybivbitsi = kibibyte / binary kilobyte. Hm... a bit long though 20:04 < zipcpi> vlaste: bivbi (gimka) 20:04 < vlaste> zo bivbi gimkamsmikezypro no catni gismu .i zo bivbi gimkamsmikezypro no cipra gismu 20:04 < zipcpi> vlaste: bitsi (gimka) 20:04 < vlaste> zo bitsi gimkamsmikezypro no catni gismu .i zo bitsi gimkamsmikezypro no cipra gismu 20:13 < zipcpi> exp: broda cei me'au by 20:13 < mensi> (CU [broda {cei }] VAU) 20:13 < zipcpi> broda cei citka be lo mlatu 20:14 < zipcpi> brode cei plipe be lo mumplini 20:16 < zipcpi> Wait, which way does cei/goi go? o.o 20:16 < zipcpi> durka, which way does cei/goi go? o.o 20:17 < zipcpi> I think it's ambiguous 20:17 < zipcpi> The sky is falliiiiing 20:17 < zipcpi> Call the logic police 20:18 < durka42> what do you mean? 20:18 < zipcpi> ko'a goi la .alis. ; la .alis. goi ko'a 20:19 < zipcpi> mlatu cei murmura ; murmura cei mlatu 20:19 < zipcpi> Help 20:20 < durka42> well, it's whichever side the variable is on 20:21 < durka42> the only question is what happens when a variable is on both sides 20:21 < zipcpi> What about ko'a goi ko'e then? 20:21 < durka42> right 20:21 < zipcpi> Also I dunno, I can see using nonce-names as variables 20:21 < durka42> don't know off the top of my head 20:21 < durka42> LMW is down again or I would check what the BPFK has to say 20:21 < zipcpi> ... probably will need a new UI T.T 20:22 < durka42> there might be one already 20:22 < noncomcinse> cei/goi? 20:23 < noncomcinse> I assume {X cei/goi Y} implies Y=return(x) 20:23 * zipcpi searches for cei and goi on sutysisku... couldn't find any weird cmavo that indicates direction 20:23 < durka42> noncomcinse: what? 20:23 < durka42> zipcpi: just stick a {ca'e} to whichever one you're assigning 20:24 < noncomcinse> en: bu'a 20:24 < mensi> bu'a = [GOhA] logically quantified predicate variable: some selbri 1. |>>> officialdata 20:24 < noncomcinse> bu'a cei broda 20:24 < zipcpi> I have no idea what those are 20:24 < noncomcinse> en: ca'e 20:24 < mensi> ca'e = [UI2] evidential: I define. |>>> See also mitcu'a, mitsmu. |>>> officialdata 20:24 < durka42> {bu'a} are supposed to be like {da} for brivla 20:24 < durka42> but they did not catch on at all 20:24 < zipcpi> lol 20:25 < noncomcinse> Isn't {lo} formalized as {zo'e poi ke'a}? 20:25 < durka42> usually noi, but yeah 20:25 < noncomcinse> So ultimately, lo bu'a =/= da 20:25 < noncomcinse> Kinda? 20:26 < durka42> it's more like {bu'a} is supposed to abstract over relations 20:26 < noncomcinse> kindysorta 20:26 < noncomcinse> uanai 20:26 < durka42> the same way {da} abstracts over sumti 20:26 < durka42> abstracts is the wrong word 20:26 < zipcpi> durka42: {ca'egoi} vs {goica'e}? 20:26 < zipcpi> Or {goi ko'a ca'e} 20:26 < durka42> zipcpi: ko'aca'e goi ko'e vs ko'a goi ko'eca'e 20:27 < zipcpi> Seems cromulent, but I don't know what it'd do for potential expansions of UI. It's probably already a mess :p 20:27 < durka42> I mean do we even have an expansion for {goi} 20:28 < zipcpi> Good point lol 20:28 < noncomcinse> IMO, UI should be able to be placed anywhere without affecting ANY other cmavo grammatically. 20:29 < noncomcinse> (ta'o, whenever I look at a sentence with only cmavo, it's just... mibnardjuno) 20:29 < zipcpi> Not quite so simple... there's always those weird exceptions like {zo} and {zei} :p 20:29 < noncomcinse> zo VALSI = one word 20:29 < durka42> I mean, say {ko'a goi ko'e} expands to {mi goirgau ko'a ko'e}, then if we stick a {ca'e} on one of them, we have {ko'a ca'e goi ko'e} => {mi goirgau ko'a ca'e ko'e} => {mi goirgau ko'a ko'e xoi mi ca'engau do'e ko'a} 20:29 < durka42> so it works 20:29 < noncomcinse> like with zo 'bu' 20:29 < noncomcinse> .i ca'e lu zo bu bu li'u sumti 20:29 < noncomcinse> .i ca'e lu zo bu bu bu bu bu bu bu li'u sumti 20:29 < zipcpi> goingau, you mean :p 20:30 < noncomcinse> .i me pa sumti 20:30 < durka42> yes 20:30 < durka42> though I checked, ra'oi goi rafsi no da 20:30 < zipcpi> u'i 20:30 < durka42> it's going... going... goingau! 20:30 < noncomcinse> The only problem is you can't stick UI in between *eks and {nai}. 20:30 < noncomcinse> .u'i 20:30 < durka42> can't you? 20:30 < zipcpi> Yay for magic predicates 20:31 < noncomcinse> camxes: .i mi klama je .ui nai klaku 20:31 < camxes> (i [mi CU] [klama {je } klaku] VAU) 20:31 < durka42> oh oh right 20:31 < durka42> because the nai attaches 20:32 < noncomcinse> Unrelated, is there a way to do illogical forethought connectives in ba'e classical Lojban? 20:32 < zipcpi> Oh I have a new slate of zi'evla in mind... guess what they do 20:32 < zipcpi> ki'orbi me'orbi gi'orbi te'orbi pe'orbi xe'orbi ze'orbi go'orbi 20:32 < zipcpi> ju'e 20:32 < zipcpi> ju'egi 20:32 < noncomcinse> en: ju'e 20:32 < mensi> ju'e = [JOI] vague non-logical connective: analogous to plain ".i". |>>> officialdata 20:32 < noncomcinse> Except ju'e doesn't start a new bridi. 20:33 < durka42> mi ka'enai smadi 20:33 < zipcpi> ju'e gi broda gi brode 20:33 < zipcpi> durka42: Binary prefixes. Kibi(byte), mebi(byte), etc 20:33 < durka42> what the note is saying is that where {ge broda gi brode} is like {broda .ije brode}, {ju'egi broda gi brode} is just like {broda .i brode} 20:33 < noncomcinse> off: .i mi fa'u gi klama gi citka fa'u gi lo zarci gi lo titnanba 20:33 < mensi> (i [mi CU] [fa'u gi] [klama VAU] gi [citka { gi } VAU] VAU) 20:33 < zipcpi> Of course, then we need words for bit and byte 20:34 < noncomcinse> zipcpi: pa-? 20:34 < zipcpi> pa- what? 20:34 < noncomcinse> pa'oibi? 20:34 < noncomcinse> pa'orbi* 20:34 < durka42> I would use the existing gismu for those prefixes if I were you 20:34 < durka42> some of them even have rafsi 20:35 < phma> coi rodo 20:35 < noncomcinse> vlatai: ki'oirbi 20:35 < zipcpi> durka42: But they're different from the decimal counterparts 20:35 < noncomcinse> What's the valsi for exponential? 20:35 < durka42> ah these are the binary ones 20:35 < durka42> I always forget which is which 20:35 < noncomcinse> en: exponent 20:35 < mensi> 15 da se tolcri: dektenfa, gei, ni'e'ei, ni'e'oi, te'au'u, tenfa, te'o, aigne, be'ei'oi, cpolinomi'a, 20:35 < mensi> dekykemymecnontenfa, dekyki'otenfa, nonsmitenfa, te'a, te'ai 20:35 < noncomcinse> en: gei 20:35 < mensi> gei = [VUhU2] trinary mathematical operator: order of magnitude/value/base; [b * (c to the a power)]. |>>> 20:35 < mensi> officialdata 20:35 < zipcpi> Yeah ki'orbi doesn't mean kilobyte. It means 1024 20:36 < durka42> .i'e 20:36 < zipcpi> ni'e'ei? lol 20:36 < durka42> krtis... 20:36 < durka42> vlaste: ni'e'ei 20:36 < vlaste> ni'e'ei = digit/number: Niven's greatest-exponent prime factorization constant; lim_(n->∞) (avg_n (H)) ≈ 1.705211… 20:36 < phma> ma po'u na'ebo la .camgusmis. ka'e refcfagau le selfu? 20:36 < noncomcinse> How about a valsi for 'x1 is a digital data storage unit with byte-count x3 to the power of x2'? 20:37 < zipcpi> Well gleki has proposed {bitsi} for bit 20:37 < noncomcinse> This def uses byte-count 20:37 < zipcpi> But that leaves byte 20:37 < noncomcinse> cei broda 20:38 < zipcpi> He wanted to use {bivbitsi}, but {ki'orbybivbitsi} is a bit long 20:38 < durka42> bibybitsi 20:38 < durka42> oh whoops 20:38 < noncomcinse> bit = lo broda be 2 bei -3 20:38 < noncomcinse> 2^-3 is 1/8, right? 20:38 < noncomcinse> I'm not screwing up my math? 20:38 < durka42> baitsi 20:38 < durka42> yes 20:39 < noncomcinse> er, lo broda be -3 bei 2 20:39 < noncomcinse> Because I put the power as x2 20:39 < durka42> you'd want three places if you were going to split up the number like that though 20:39 < zipcpi> I don't believe in these predicates when pretty much all computers in the world uses 8-bits per byte; not to mention bits aren't arbitrary 20:39 < durka42> in order to be able to say, e.g. 10GB 20:39 < durka42> but I would just have one place for the number 20:39 < durka42> zipcpi: that's not culturally neutral :p 20:39 < zipcpi> 8-bits per byte is, but pretty much all computers in the world use it 20:40 < durka42> even the C standard doesn't guarantee 8 bits per byte 20:40 < noncomcinse> x1 = x2 * (x4^x3) -- pei 20:40 < durka42> right that's basically how {gei} works 20:40 < zipcpi> durka42: For what, aliens? I don't give a mabla about cultural neutrality wrt imaginary cultures; if they need words for their concepts, they can make their own 20:41 < noncomcinse> 10GB bitsi 10 9 10 20:41 < noncomcinse> 5GB bitsi 4 9 10 20:41 < noncomcinse> s/4/5/ 20:41 < fenki> noncomcinse meant to say: 5GB bitsi 5 9 10 20:41 < noncomcinse> vlaste: fenki 20:41 < vlaste> fenki = x1 (action/event) is crazy/insane/mad/frantic/in a frenzy (one sense) by standard x2. 20:41 < noncomcinse> xu sampre 20:41 < phma> when I was learning computers, there were computers with 36-bit words and computers with 18-bit registers and 60-bit registers 20:42 < zipcpi> Hmm... then maybe there can be predicates for those too 20:42 < noncomcinse> 5GiB bitsi 5 9 8 20:42 < zipcpi> But 8-bit has become a standard and therefore deserves a word 20:42 < phma> lo itsi cu bitsi 20:42 < zipcpi> Rather than forcing us to fill extra places all the time 20:42 < durka42> bitsi li mu li no li pano / bitsi li pano gei mu ge'a no 20:42 < noncomcinse> en: makfu 20:42 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 20:43 < noncomcinse> en: makfa 20:43 < mensi> makfa = x1 is magic/supernatural/sorcery/witchcraft/wizardry to x2, performed by person/force/deity x3. |>>> See also 20:43 < mensi> krici, manci. |>>> officialdata 20:43 < noncomcinse> zipcpi: ko djuno lodu'u makau makfa gi'e me zo zo'e 20:43 < durka42> zipcpi: why not just something like baitsi = x1 is a byte / set of x2 (default 8) bits 20:43 < zipcpi> xo'e? 20:43 < noncomcinse> 5GB bitsi 5 9 zo'e 20:43 < noncomcinse> zo'e = lo'o xo'e 20:44 < zipcpi> durka42: That'd give a chance for zmico to be useful D: 20:44 < noncomcinse> I need a diode equals sign 20:44 < durka42> eww 20:44 < zipcpi> lol 20:44 < noncomcinse> en: zmico 20:44 < mensi> 11 da se tolcri: gafyzmico, sumyzmico, dai'o, dau'a, de'au, de'oi, di'au, di'ei, di'oi, do'ai, 20:44 < mensi> doi'a 20:44 < noncomcinse> y 20:44 < zipcpi> downvoted lol 20:44 < noncomcinse> vlaste: zmico 20:44 < vlaste> 10 results: gafyzmico, sumyzmico, dai'o, dau'a, de'au, de'oi, di'ei, di'oi, do'ai, doi'a 20:44 < noncomcinse> ue 20:44 < noncomcinse> na se nelci 20:45 < durka42> ni'udzau 20:45 < zipcpi> It's a controversial set of cmavo that Curtis put in 20:45 < noncomcinse> I read downvoted as JVS-downvote before remembering Reddit uses downvotes and questioning it. 20:45 < zipcpi> Yes, I mean JVS downvoted 20:45 < noncomcinse> ju'o 20:45 < noncomcinse> losu'u me'oi .reddit. co'e cu xamsku 20:46 < noncomcinse> si se xamsku 20:46 < zipcpi> But from what I gather, basically it messes with defaults 20:46 < noncomcinse> What does? 20:46 < noncomcinse> zmico? 20:47 < zipcpi> Yep 20:47 < phma> rlpowell: la .lojban.org. spofu .i.e'o ko refcfagau 20:47 < noncomcinse> default whats 20:47 < zipcpi> default terbri specifications; e.g. that default 8 20:47 < zipcpi> or the default 1 for lo se mitre 20:47 < noncomcinse> ma default default? 20:47 < noncomcinse> xu zo zo'e 20:48 < durka42> the zmico are a set of like 1000 cmavo that curtis created in order to obsessively micromanage the defaults for the sumti places in a given sentence, and they would be impossible for humans to use in practice 20:48 < zipcpi> lol 20:48 < phma> what's zmico? 20:48 < durka42> ^ 20:48 < noncomcinse> >what's zmica 20:48 < noncomcinse> noda djuno 20:48 < noncomcinse> sa 20:48 < noncomcinse> pa da djuno 20:48 < noncomcinse> .i lo zbasu cu djuno 20:48 < noncomcinse> .au 20:49 < noncomcinse> .a'u 20:49 < zipcpi> no drata be zy cu djuno iau li'a 20:49 < noncomcinse> si .a'o 20:49 < phma> zmoci 20:49 < noncomcinse> .i mi na se slabu lo ro cmavo be ma'oi UI 20:49 < noncomcinse> de'a 20:50 < zipcpi> durka42: Yeah you're probably right. I mean if anyone really needed {baitsi be lo pabi} they should just {cei} it to something else 20:51 < zipcpi> Er, li pa bi I mena 20:51 < zipcpi> *mean 20:51 < phma> what does the seed force to happen? 20:52 < durka42> ki'a 20:53 < noncomcinse> .idi'a 20:53 < phma> or does something force the seed to sprout? 20:53 < noncomcinse> I should honestly take flash cards for UI 20:53 < durka42> physics? 20:53 < noncomcinse> the nutrients inside cause a chemical reaction? 20:53 < phma> xu lo baitsi cu me lo tsiju poi bapli? 20:54 < durka42> ua .u'i 20:54 < durka42> forgot ts- was a valid initial 20:54 < zipcpi> Oh oops 20:54 < zipcpi> zbaiti? 20:55 < noncomcinse> tsani, tsali 20:55 < zipcpi> k: zbaiti 20:55 < mensi> (CU [Z:zbaiti VAU]) 20:55 < noncomcinse> melo tsiju? 20:55 < zipcpi> ki'orbyzbaiti 20:55 < noncomcinse> I always elide that space ,u'i 20:56 < durka42> melo yellow 20:56 < phma> jgari, jbari, jdari 20:56 < noncomcinse> dz* 20:56 < phma> melo bisli 20:56 < noncomcinse> 'melo'='' 20:56 < durka42> "among the" 20:57 < noncomcinse> yeah 20:57 < noncomcinse> fair 20:57 < noncomcinse> I always get confuse 20:57 < noncomcinse> Because someone said that 'melo broda'='broda' 20:58 < phma> red=grey 20:58 < phma> lo julne cu terlanme 20:58 < phma> julne ja seltcana 20:58 < zipcpi> How do we even define defaults in Lojban anyway? 20:59 < zipcpi> ... zmico? *ducks* 21:00 < noncomcinse> pe'i 21:00 < noncomcinse> all are zo'e 21:00 < noncomcinse> You don't get to define a default, it's just zo'e 21:00 < zipcpi> Some words already have defaults 21:01 < noncomcinse> We could just ignore them, and pretend zo'e? 21:01 < noncomcinse> I dunno. 21:01 < zipcpi> I'm not gonna do that for {mitre} 21:01 < noncomcinse> Contextually, you can generally assume someone means 1. 21:02 < noncomcinse> Actually 21:02 < noncomcinse> Having defaults is a good idea. 21:03 < durka42> jbo:mitre 21:03 < mensi> mitre = x1 klani x2 lo sornai rauci'e ni clani 21:03 < durka42> en:mitre 21:03 < mensi> mitre = x1 is x2 meter(s) [metric unit] in length (default 1) measured in direction x3 by standard x4. 21:03 < durka42> yyy 21:03 < durka42> is that supposed to be a lujvo 21:03 < durka42> 'cause it's not 21:03 < durka42> jbo: mitre 21:03 < mensi> mitre = x1 klani x2 lo sornai rauci'e ni clani |>>> klani; sornai; rauci'e; clani; minli; gutci |>>> 21:03 < mensi> xorxes 21:04 < noncomcinse> ma lujvo xudai 21:04 < zipcpi> Probably defined before we defined sornairauci'e 21:04 < durka42> lu rau ci'e li'u 21:04 < zipcpi> Oh oops 21:04 < zipcpi> OOPS 21:04 < noncomcinse> camxes: fabu klani febu lo sornai rauci'e ni clani 21:04 < camxes> ([{ BOI} CU] [klani {<(¹fe bu¹) BOI> KU} VAU]) 21:04 < zipcpi> xorxes made a mistake! GASP 21:04 < durka42> .u'i 21:04 < noncomcinse> .i gendra 21:04 < durka42> yeah there's no way that is intended 21:04 < durka42> seems like he did leave out the default though 21:04 < noncomcinse> I'm not sure what all those gismu mean 21:04 < durka42> klani clani 21:05 < noncomcinse> en: ni 21:05 < mensi> ni = [NU] abstractor: quantity/amount abstractor; x1 is quantity/amount of [bridi] measured on scale x2. |>>> 21:05 < mensi> officialdata 21:05 < noncomcinse> Right. 21:05 < noncomcinse> I know all the CV cmavo 21:05 < noncomcinse> but not all the time 21:05 < noncomcinse> If I think of the English, I can recall the Lojban pretty well. 21:06 < durka42> wait a minute why the hell did that sentence parse 21:06 < noncomcinse> IKR 21:06 < durka42> a quantified BAI-term...? 21:06 < noncomcinse> rau < PA ni 21:06 < durka42> camxes: rau bai broda 21:06 < camxes> ([{rau BOI} {bai broda} KU] VAU) 21:07 < noncomcinse> oh, rau ci'e 21:07 < noncomcinse> What. 21:07 < durka42> oh right it's just a selbritcita 21:07 < noncomcinse> What the hell...? 21:07 < noncomcinse> You can quantify tcita? 21:07 < durka42> rau ci'e ni clani => rau da poi ci'e zo'e ni clani 21:08 < noncomcinse> camxes: .i mi klama su'o mu'i 21:08 < durka42> no it's an unfilled tcita 21:08 < camxes> SyntaxError: Expected [,], [nN] or [rR] but "m" found. 21:08 < noncomcinse> camxes: +s fabu klani febu lo sornai rauci'e ni clani 21:08 < camxes> ([{ BOI} CU] [G:klani {<(¹FA:fe BU:bu¹) BOI> KU} VAU]) 21:08 < noncomcinse> Um 21:08 < zipcpi> k: kirbizbaiti 21:08 < mensi> (CU [Z:kirbizbaiti VAU]) 21:08 < durka42> no it makes sense I figured it out :) 21:08 < durka42> tersmus: jbo: fabu klani febu lo sornai rauci'e ni clani 21:08 < tersmus> cy no sornai .i rau boi da poi ci'e ku zo'u ke'a ni ce'u clani kei ku'o zo'u fa bu klani fe bu cy no da 21:08 < zipcpi> Hmm... maybe those are better "lujvo" 21:09 < zipcpi> Still keep the ki'orbi etc though 21:09 < durka42> hey, tersmus agrees with my expansion :) 21:09 < noncomcinse> zo ki'orbi na lujvo 21:09 < noncomcinse> What's the expansion? 21:09 < durka42> hence the scare quotes 21:09 < durka42> rau ci'e ni clani => rau da poi ci'eku ni clani 21:09 < durka42> it's a quantified sumti 21:09 < noncomcinse> rau ci'e ni clani = PA BAI NU GISMU 21:09 < durka42> it just looks really strange :p 21:10 < noncomcinse> = BAI PA da poi BAI ke'a ku NU GISMU -- xu 21:10 < noncomcinse> s/BAI PA/PA/ 21:10 < fenki> noncomcinse meant to say: = PA da poi BAI ke'a ku NU GISMU -- xu 21:11 < noncomcinse> I just had this thought, like, "I wonder if this bot can do extended regular expressions?" 21:11 < noncomcinse> .i mi cipra co sampre 21:11 < durka42> poi BAI ke'a NU GISMU, or poi ke'a BAI zo'e NU GISMU? 21:12 < durka42> noncomcinse: well it support s/x/y/gi at least, not sure what else? 21:12 < zipcpi> k: kirbibitsi 21:12 < mensi> (CU [Z:kirbibitsi VAU]) 21:12 < durka42> s/SUPPORT/supports/i 21:12 < fenki> durka42 meant to say: noncomcinse: well it supports s/x/y/gi at least, not sure what else? 21:12 < noncomcinse> .i mi cipra co sampre 21:12 < noncomcinse> s/cipra ?/cIpra/ 21:12 < noncomcinse> .i mi cipra co sampre 21:13 < durka42> guess not 21:13 < noncomcinse> s/[cip]/_/g 21:13 < noncomcinse> Aw. 21:13 < noncomcinse> ui nai 21:13 < noncomcinse> tolcmila bu 21:13 < noncomcinse> It least it does /g? 21:13 < noncomcinse> .u'i 21:14 < noncomcinse> inb4 bot for `tr'. 21:14 < noncomcinse> .tr "a-z'" "A-Zh" 21:14 < noncomcinse> {.tr} is the call for eobot to translate on #Esperanto. 21:15 < noncomcinse> .tr eo en novas 21:15 < noncomcinse> It didn't like translating to Lojban. 21:15 < noncomcinse> 23:15 .tr eo jbo novas 21:15 < noncomcinse> 23:15 .tr: nekonata komando. Por helpo vidu http://www.zz9pza.net/eobot/helpo 21:16 < noncomcinse> eo: novas 21:16 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 21:16 < noncomcinse> jbo: novas 21:16 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 21:16 < noncomcinse> zo'oi .novas. cu simsa zo cnino 21:16 < noncomcinse> eo: cnino 21:16 < mensi> cnino = x1 estas nova/nekonata por observanto x2 en trajto x3 (ka) laŭ normo x4; x1 estas novaĵo |>>> Vidu ankaŭ nuzba, 21:16 < mensi> slabu, citno, se djuno. |>>> xorxes 21:17 < zipcpi> I prefer something like membitsi for arbitrary bit-group, then define zbaiti as membitsi be li bi 21:17 < noncomcinse> la'oi .EoBot. pilno la'oi .GoogleTranslate. xoi mi jimpe 21:18 < zipcpi> Cause leaving it as zo'e is precisely why we have confusion right now with {tirxu} and {minli} 21:18 < durka42> at first I thought you meant "tr" the unix command, .u'i 21:20 < noncomcinse> .tr "a-z'" "A-Zh" -- was 21:20 < noncomcinse> at least it was an idea for a fenki-style bot 21:20 < zipcpi> That's why I don't think too much about cultural neutrality. I define words for what I want them to mean; other people who find other concepts useful can define words to mean what *they* want to mean 21:20 < noncomcinse> ./msg eobot .tr LANG_IN LANG_OUT [text] 21:21 < noncomcinse> ku'i pilno GoogleTranslate 21:21 < zipcpi> There's enough zi'evla space, and some would say gismu space even, for all of us 21:21 < zipcpi> Not sure about cmavo though :p 21:22 < noncomcinse> Not enough room in the, uh 21:22 < noncomcinse> vlaste: city 21:22 < vlaste> 16 results: tcadu, tcabriju, jarbu, tcakamni, banxe'osu, gugdevu'a, latmnroma, bryklyn, tapla, taksi… 21:22 < noncomcinse> tcadu 21:22 < noncomcinse> va'u re mi'o 21:22 < noncomcinse> or zu'e re mi'o 21:22 < noncomcinse> Not enough room in this tcadu for re mi'o. 21:25 < noncomcinse> nardjuno 21:27 < zipcpi> OK... that's 43 words to enter D: 21:28 < noncomcinse> what iis 21:28 < zipcpi> ki'orbi me'orbi gi'orbi te'orbi pe'orbi xe'orbi ze'orbi go'orbi 21:28 < zipcpi> bitsi bitsymei zbaiti 21:28 < zipcpi> kirbizbaiti merbizbaiti girbizbaiti terbizbaiti perbizbaiti xerbizbaiti zerbizbaiti gorbizbaiti 21:28 < zipcpi> kiltyzbaiti megyzbaiti gigyzbaiti tetyzbaiti petyzbaiti xexyzbaiti zepyzbaiti gotyzbaiti 21:28 < zipcpi> kirbibitsi merbibitsi girbibitsi terbibitsi perbibitsi xerbibitsi zerbibitsi gorbibitsi 21:28 < zipcpi> ki'orbitsi megbitsi gigbitsi tetybitsi petybitsi xexybitsi zepybitsi gotybitsi 21:28 < noncomcinse> ueua 21:28 < noncomcinse> Can we script JVS? 21:28 < noncomcinse> Or just have a script directly modify the database? 21:29 < durka42> gleki has an autoimporter script 21:29 < noncomcinse> Cool. 21:29 < zipcpi> I dunno if he'll share with me though T.T 21:29 < noncomcinse> Or, you know, just make a gismu for it. 21:29 < zipcpi> I don't think he trusts me T.T 21:29 < zipcpi> {bitsi} is a gismu 21:29 < noncomcinse> I break stuff. It's like my specialty. I should never have such a script. 21:30 < noncomcinse> The first time I got my hands on AutoHotKey... 21:30 < zipcpi> Gleki invented it, but didn't enter it into JVS 21:31 < noncomcinse> x1 is data storage with byte-size x2 * x4 ^ x3. 21:31 < noncomcinse> Alternatively, 21:31 < noncomcinse> x1 is data storage with byte-size x2 * x4 ^ ( x3 * 3 ). 21:31 < noncomcinse> So that way a 1KB = lo bitsi be fi li pa 21:32 < noncomcinse> Rather than having to work out multiples of 3 in your head. 21:32 < zipcpi> I don't want to do that mental calculation just to know what it corresponds to 21:32 < noncomcinse> Fair. 21:32 < noncomcinse> xoi stace zo'u I already do that. 21:33 < noncomcinse> KB is 1-power of a thousand in my head 21:33 < zipcpi> There's a reason those SI prefix stuff exists though 21:33 < noncomcinse> Or, alternatively even, just put the x4 as lipaki'o 21:33 < noncomcinse> Wait, are those prefixes here, in lojbobau? 21:33 < noncomcinse> Beyond ki'o, I mean. 21:33 < zipcpi> Yes, kilto, megdo, gigdo, etc 21:35 < noncomcinse> ua 21:35 < noncomcinse> How high does that go? 21:35 < noncomcinse> peta, tera? 21:35 < zipcpi> gotro 21:35 < noncomcinse> en: gotro 21:35 < zipcpi> All the way to yotta 21:35 < mensi> gotro = x1 is 10^24 of x2 in dimension/aspect x3 (default is units). |>>> See also grake, mitre, snidu, stero, delno, 21:35 < mensi> molro, kelvo, xampo, gradu. litce, merli, centi, decti, dekto, femti, gigdo, gocti, kilto, megdo, mikri, milti, nanvi, 21:35 < mensi> petso, picti, terto, xatsi, xecto, xexso, zepti, zetro |>>> officialdata 21:35 < noncomcinse> rafsi: gotro 21:35 < mensi> zo'oi got .e zo'oi gotr rafsi zo gotro 21:35 < noncomcinse> pa valsi 21:35 < noncomcinse> Damn. 21:35 < noncomcinse> And I thought terabytes were big. 21:36 < noncomcinse> Until I started using flacs. 21:36 < noncomcinse> zo'o 21:36 < durka42> .u'i 21:37 < noncomcinse> .i ku'i lo zgike datni zo'u bitsi li repa li ci li paki'o 21:37 < noncomcinse> 21GB 21:38 < noncomcinse> .i du'e vidni cu jai datni 21:38 < durka42> gigyzbaiti li repa 21:39 < durka42> or is it girbizbaiti 21:39 < durka42> I'm already confused 21:39 < zipcpi> lol 21:39 < zipcpi> Do we need a set for "I dunno which"? 21:39 < zipcpi> Kinda sucks that we're in that mess 21:40 < zipcpi> The ones who came out with KB for 1024 weren't Lojbanists, that's why 21:40 < noncomcinse> gigyzbaiti has a redundant y 21:40 < zipcpi> mumble mumble grumble 21:40 < zipcpi> It's for consistency 21:40 < noncomcinse> nelci mutce 21:41 < noncomcinse> Just don't try to start putting ybu everywhere 21:41 < zipcpi> Because tetzbaiti is not a valid zi'evla 21:41 < noncomcinse> you know the /[aeiou]'y/ reform? 21:41 < zipcpi> Yes 21:41 < noncomcinse> We can't do y'y for ybu 21:41 < noncomcinse> Because y'y is for {'} 21:41 < zipcpi> I entered {iy'y} 21:41 < noncomcinse> But we could, but where would {'} go? 21:41 < noncomcinse> .i xu cmavo gendra 21:41 < zipcpi> But yeah that inconstency sucks 21:42 < zipcpi> I think it's morphologically acceptable 21:42 < noncomcinse> Minor annoyances of Lojban: 21:42 < noncomcinse> ter 21:42 < noncomcinse> ji 21:42 < noncomcinse> .iy'y 21:42 < noncomcinse> ({gi'i} is okay, though. 21:42 < zipcpi> ji is fixed by selpa'i-s connective system 21:42 < noncomcinse> Whooo. 21:43 < noncomcinse> The thing is 21:43 < noncomcinse> thatt .iy'u is CREATED by a proposal 21:43 < noncomcinse> It wasn't a problem before. 21:43 < zipcpi> Yeah 21:44 < noncomcinse> I'll bet it was started because someone said {lu zo .abu li'u na drani} 21:44 < zipcpi> exp: zo abu 21:44 < mensi> ([{ bu} BOI] VAU) 21:44 < zipcpi> OK yeah lol 21:44 < zipcpi> exp: zo ybu 21:44 < mensi> ([zo ybu] VAU) 21:45 < zipcpi> Hahahahaha 21:45 < zipcpi> ybu is special 21:45 < noncomcinse> exp: zo .abu 21:45 < mensi> ([{ bu} BOI] VAU) 21:45 < noncomcinse> uat 21:45 < noncomcinse> sa'ei 21:45 < noncomcinse> .uat. 21:45 < noncomcinse> Why? 21:45 < noncomcinse> exp: zo .ybu 21:45 < mensi> ([zo ybu] VAU) 21:45 < zipcpi> Because {y} is invisible anywhere else 21:45 < noncomcinse> off: zo .ybu 21:45 < mensi> ([zo ybu] VAU) 21:46 < noncomcinse> What. 21:46 < zipcpi> So {ybu} gets combined in the very first level of processing 21:46 < noncomcinse> I'm done. 21:46 < zipcpi> lol 21:46 < noncomcinse> Well 21:46 < noncomcinse> See, on any other chat system, I'd type /quit, but here, that would actually quit me. 21:46 < durka42> add that special case to the minor annoyance list please 21:46 < noncomcinse> % exec sh quit 21:46 < noncomcinse> Yes. 21:46 < noncomcinse> also 21:47 < zipcpi> {xa'i mi oicli'a}? 21:47 < noncomcinse> camxes: .cmen. mi tavla ma 21:47 < camxes> (cmen [mi CU] [tavla {ma VAU}]) 21:47 < noncomcinse> This. 21:47 < zipcpi> exp: cmen mi tavla ma 21:47 < mensi> SyntaxError: Expected [yY] but "a" found. 21:47 < noncomcinse> ti'e Remnant of Loglann. 21:47 < noncomcinse> s/nn/n/ 21:47 < fenki> noncomcinse meant to say: ti'e Remnant of Loglan. 21:47 < noncomcinse> I just like seeing this bot fix things. 21:48 < noncomcinse> .i la'e di'u 21:48 < noncomcinse> mutce loka ce'u gasnu lonu mi ckaji 21:48 < noncomcinse> ce'u se zdile 21:48 < zipcpi> durka42: Do you prefer {kirbizbaiti} or {kirbibaiti}? 21:48 < noncomcinse> vau kei ku vau kei ku vau kei ku vau 21:49 < zipcpi> And I still dunno what to do about the ambiguous case 21:49 < durka42> the second one I think 21:49 < durka42> which ambiguous case? 21:49 < zipcpi> Where you don't know whether to translate "KB" as {kiltyzbaiti} or {kirbibaiti} 21:49 < noncomcinse> What's the difference in the rafsi? 21:50 < noncomcinse> Oh, is it just a y->i? 21:50 < durka42> zipcpi: that's not lojban's problem... 21:50 < noncomcinse> lty->rbi 21:50 < zipcpi> First is based off kilto, second is based off ki'orbi 21:50 < durka42> kirbibaiti is a zi'evla 21:50 < zipcpi> kilto = 1000 21:50 < zipcpi> ki'orbi = 1024 21:50 < noncomcinse> Well, KB =/=KiB 21:50 < noncomcinse> KB = 1000 21:50 < noncomcinse> KiB = 1024 21:51 < noncomcinse> .isampu 21:51 < zipcpi> Because the malpre who misused SI prefixes weren't Lojbanists, that's why 21:51 < noncomcinse> Make both for every version? 21:51 < noncomcinse> Or 21:51 < noncomcinse> What do you have so far? 21:51 < noncomcinse> Do the ones you have so far use base-2 or base-10? 21:51 < zipcpi> I'm making a set for both decimal and binary 21:52 < zipcpi> I mean two sets 21:52 < zipcpi> one for decimal, one for binary 21:52 < zipcpi> The problem is that some systems *mumble Windows grumble* still use KB for binary 21:53 < zipcpi> binary: kirbibaiti merbibaiti girbibaiti terbibaiti perbibaiti xerbibaiti zerbibaiti gorbibaiti 21:53 < durka42> yes 21:53 < zipcpi> decimal: kiltyzbaiti megyzbaiti gigyzbaiti tetyzbaiti petyzbaiti xexyzbaiti zepyzbaiti gotyzbaiti 21:53 < durka42> them and deceptive hard drive manufacturers 21:53 < zipcpi> No, the deceptive hard drive manufacturers use the decimal 21:53 < zipcpi> Then when you load them up on Windows... eh? 21:53 < durka42> right 21:54 < noncomcinse> Widows uses KiB? 21:54 < zipcpi> Yes, but calls it KB 21:54 < noncomcinse> And HDD make-- 21:54 < noncomcinse> There is a very good reason I don't use Windows, then. 21:54 < zipcpi> lol 21:55 < noncomcinse> It's official, never going back. 21:55 < noncomcinse> Ever. 21:55 < noncomcinse> Nope. 21:55 < noncomcinse> At least some people tell you how many bytes you have. 21:55 < zipcpi> I dunno. I have people tell me they just don't like the binary prefixes 21:55 < zipcpi> And I'm like... I prefer to know exactly how many bytes I'm talking about tyvm 21:55 < noncomcinse> But it's not exactly the HDD manufacturers that are lying. 21:56 < durka42> well they sound stupid 21:56 < noncomcinse> They say "1 TB". 21:56 < zipcpi> Especially since the higher up the ladder you climb, the more the discrepancy grows 21:56 < durka42> but they do make more sense :) 21:56 < noncomcinse> Not "1 TiB" 21:56 < noncomcinse> Okay. Screw the prefixes, I'm counting it by sectors. 21:56 < zipcpi> Yes, it's Windows that "lies"; but mostly it's more inertia than anything 21:56 < durka42> ok, bedtime 21:56 < durka42> co'o 21:57 < noncomcinse> co'o la durkavore 21:57 < zipcpi> That ambiguity has existed since the beginning of time, then the computer-world equivalent of xorxes came up with these new prefixes 21:58 < zipcpi> And RAM is still often given in terms of "GB" when "GiB" is meant 21:58 < zipcpi> Heck it dates all the way back to floppy disks. 1.44 MB becomes 1.38 MiB when you load it up in Windows 21:59 < zipcpi> And back then MiB didn't exist 21:59 < noncomcinse> But Windows still says it's 1.44MB 21:59 < noncomcinse> s/44/38/ 21:59 < fenki> noncomcinse meant to say: But Windows still says it's 1.38MB 21:59 < zipcpi> Exactly 21:59 < zipcpi> But back then MiB didn't exist 21:59 < zipcpi> So MB was ambiguous 21:59 < noncomcinse> And that's why it said 1.38M 21:59 < noncomcinse> B 22:00 < noncomcinse> cipra cipra 22:00 < zipcpi> Then someone came up with the new prefixes 22:00 < zipcpi> And now they're trying to make it standard 22:00 < noncomcinse> s/CIPRA/snada/gi 22:00 < fenki> noncomcinse meant to say: snada snada 22:00 < zipcpi> But it hasn't trickled down completely yet 22:00 < zipcpi> So MB is still ambiguous for now 22:00 < noncomcinse> je'e 22:00 < noncomcinse> ie 22:00 < noncomcinse> Just count it in straight sectors. 22:00 < noncomcinse> Count literal hardware. 22:00 < zipcpi> lol 22:01 < noncomcinse> cfdisk says use M/G/T for MiB/GiB/TiB 22:02 < noncomcinse> .oi 22:03 < noncomcinse> First off, I don't even know the way to say "terminal emulator", and second off, mi na djuno lodu'u makau mi se nitcu lonu lo samselpla cu tolspofu 22:05 * nuzba @lycaon_ebooks: I'M GONNA explode (it's 😉😏😘 I feel the need to get to the one hand, how can a being living such a shitty teen lojban is the "um, actually [http://bit.ly/1C0UwLC] 22:11 < Simorgh> Ìû 22:20 < zipcpi> ta'o I have an idea for how to deal with Mr./Ms. etc 22:20 < zipcpi> Namely cmevla prefixes 22:20 < zipcpi> e.g. .ionan. .ionin. 22:20 < zipcpi> e.g. {la .ionan.djonz. 22:20 < zipcpi> } 22:21 < zipcpi> Relies on BCM for brivlacme though 22:21 < zipcpi> {la .ionan. zipcpi} 22:22 < zipcpi> We could make them zi'evla but then it'd rely on BCM for the opposite case :p 22:23 < zipcpi> I have a pretty simple system for them though 22:23 < zipcpi> iolna iolni iolpe iosna iosni iospe 22:24 < zipcpi> iol- may be married or unmarried; ios- is for married 22:24 < zipcpi> Up to your culture how you want to use them 22:26 < zipcpi> But being brivla can also be useful.... {doi iolna} = "Sir," 22:26 < zipcpi> Wait that also works with cmevla 22:28 < zipcpi> Still, {lo iolna} or {le iolna} could prove useful 22:40 < zipcpi> Tentative definition: e.g. iolnan: {x1 is a Mr. (male, married or unmarried, bearer of honorific title), recognized by x2} 22:44 < zipcpi> iolna sorry 22:49 < zipcpi> I think these are a nice set though. Even gives one for married males to use if they desire, where English doesn't make the distinction 22:49 < zipcpi> And if there are other cultures that could use more distinctions, they can add them 22:57 < gleki> lo iolna be fa la djonz 22:58 < gleki> mister jones 22:58 < zipcpi> Yeah that's another possibility 23:00 < gleki> because e.g. "Miss" can be a personal name 23:00 < zipcpi> {lo iolna po'u la .djonz.} also works 23:00 < zipcpi> Right 23:03 < gleki> oh well ishould work on eng2jbo dictionary 23:03 < gleki> my Google Spreadsheet formulae exploded due to being very complex. 23:10 < gleki> jb: bitri 23:10 < mensi> bitri = bitri — x1(entity) is x2(number) bits in size 23:10 < mensi> :lo se bitri — bit (data). 23:10 < mensi> :lo vi vreji cu bitri lo kilto — This file is 1000 bit. 23:10 < mensi> :Related words: datni 23:10 < gleki> en: bitsi 23:10 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 23:13 < zipcpi> Wait you decided to make it bitri rather than bitsi? 23:13 < zipcpi> vlaste: bitri (gimka) 23:13 < vlaste> zo bitri gimkamsmikezypro no catni gismu .i zo bitri gimkamsmikezypro no cipra gismu 23:13 < gleki> because it'd sound like "pizza" 23:13 < zipcpi> u'i 23:13 < gleki> bizzi 23:15 < zipcpi> Anyway I wanted to have zbaiti for byte (yeah I know bivbitsi works but it will make lujvo that much more cumbersome 23:15 < zipcpi> Then I wanted binary prefixes 23:16 < zipcpi> Because I want to keep the SI gismu decimal 23:16 < zipcpi> ki'orbi me'orbi gi'orbi te'orbi pe'orbi xe'orbi ze'orbi go'orbi 23:16 < zipcpi> So then I ended up with a whole slate of words 23:16 < zipcpi> kirbibaiti merbibaiti girbibaiti terbibaiti perbibaiti xerbibaiti zerbibaiti gorbibaiti 23:16 < zipcpi> kiltyzbaiti megyzbaiti gigyzbaiti tetyzbaiti petyzbaiti xexyzbaiti zepyzbaiti gotyzbaiti 23:16 < zipcpi> kirbibitri merbibitri girbibitri terbibitri perbibitri xerbibitri zerbibitri gorbibitri 23:16 < zipcpi> ki'orbitri megbitri gigbitri tetybitri petybitri xexybitri zepybitri gotybitri 23:17 < gleki> i'd avoid C- prefixes 23:17 < gleki> try something like 23:17 < gleki> k:baitri 23:17 < mensi> (CU [L:baitri VAU]) 23:17 < gleki> oops 23:17 < gleki> wait 23:17 < zipcpi> Yeah that's why... 23:17 < gleki> no no 23:18 < gleki> wait 23:18 < gleki> k:baitni 23:18 < mensi> (CU [Z:baitni VAU]) 23:18 < zipcpi> Hm 23:18 < gleki> a lot of illegal front clusters 23:18 < gleki> i just want to choose something reasonable 23:18 < gleki> same for {sfaile}. too much damage to the opriginal form 23:18 < zipcpi> k: kirbibaitni 23:18 < mensi> (CU [Z:kirbibaitni VAU]) 23:19 < gleki> baitli, 23:19 < zipcpi> I like the ni 23:19 < gleki> in french it's octet 23:20 < gleki> sn- and cn- are legal initial clusters 23:21 < gleki> thus any /.+[sc]/ word can't be always suffixed with -ni 23:21 < zipcpi> Right 23:21 < gleki> there is always a danger of it being a lujvo 23:21 < zipcpi> But ml kl, cl, are legal too 23:22 < gleki> i'd rather ban cn/sn 23:22 < gleki> there are many more cluster with -l so no 23:22 < gleki> *clusters 23:22 < zipcpi> I dunno... that will kill a lot of gismu 23:22 < gleki> well, /.+n/ can occur too so a problem will be always 23:23 < gleki> vlaste: cn??? 23:23 < vlaste> 7 results: cnano, cnebo, cnemu, cnici, cnino, cnisa, cnita 23:23 < gleki> vlaste: sn??? 23:23 < vlaste> 8 results: snada, snanu, snidu, snime, snipa, snuji, snura, snuti 23:23 < zipcpi> Yeah a zi'evla-algorithm is somewhere out there, in the land of unicorns and fairies la'a 23:23 < gleki> oops i dont want that 23:23 < gleki> failyni? 23:23 < zipcpi> lol 23:24 < gleki> robin is not replying 23:24 < zipcpi> T.T 23:25 < Lawrence> Does anybody have advice on where to start in lojban? I speak Esperanto but the resources for lojban aren't as organized 23:26 < gleki> good question. our web server is down :/ 23:27 < gleki> but try opening it e.g. tomorrow mw.lojban.org/papri/Learning 23:27 < Lawrence> Oh.. I noticed that. 23:27 < gleki> just bookmark that link 23:27 < Lawrence> Alright thank you! 23:27 < gleki> we need to make a copy of most valuable pages 23:32 < zipcpi> Ugh and during the time it's been down I've came up with 50 words 23:33 < zipcpi> Funnily all of them brivla this time --- Day changed Thu Jul 02 2015 00:03 < gleki> 23 more questions with a score of 10 or more 00:31 < zipcpi> exp: le iosna je le iosni vu'o po'u la djonz 00:31 < mensi> ([{ } {vu'o }] VAU) 00:32 < zipcpi> (coi mudri. Testing out new brivla for honorific titles) 00:36 < mudri> .ue .uanai 00:36 < zipcpi> iolna iolni iolpe / iosna iosni iospe 00:36 < zipcpi> iol - unspecified marital status, ios - married 00:37 < zipcpi> na - nanmu, ni - ninmu, pe - prenu 00:37 < gleki> then {baitni} is a byte for women 00:37 < zipcpi> u'i 00:38 <@Broca> What, no prefix for unmarried? 00:39 < zipcpi> I had thought of it but I'm not sure how useful it is; because the married set is for recognizing one as part of their own family 00:39 < zipcpi> Not sure what a set specifically for unmarried would serve 00:40 < zipcpi> Or what form it should take... 01:01 < zipcpi> Hmm... how about "Doctor"? iocre? (since it doesn't just refer to medical doctors) 01:04 < zipcpi> x1 is a "Doctor/Professor" (honorific for having achieved a certain level of education or knowledge), recognized by x2 01:08 < gleki> la djonz [ku] noi profesore 01:09 <@Broca> sensei 01:09 < gleki> no idea why{.io} should be there 01:09 < gleki> en: sensei 01:09 < mensi> [< senpi sepli ≈ Doubt apart] lo nu mulno sisku zo'u: y no da se tolcri 01:09 < mensi> sensei[5867], sensep[5898], sensepli[7937], senpysei[7997], senpysep[8028], ... 01:09 < gleki> doubts apart this is a useful lujvo 01:21 < zipcpi> u'i 01:21 < zipcpi> gleki: Same reason the io is part of all the others; it's an honorific 01:21 < zipcpi> Though a brivla specifically for "professor" might be useful 01:22 < zipcpi> Because, culturally, "Doctor" as a title doesn't necessarily refer to a literal doctor 01:22 < gleki> en: professor 01:22 < mensi> 2 da se tolcri: balkulctu, ctuca 01:22 < gleki> en: balkulctu 01:22 < mensi> balkulctu [< banli ckule ctuca ≈ Great school teach] = x1 is a professor teaching audience x2 ideas/methods/lore x3 01:22 < mensi> (du'u) about subject(s) x4 by method x5 (event). |>>> Cf. balcu'e, balkultadni. |>>> 01:22 < mensi> tettaman 01:22 < gleki> .doct professor 01:22 < zipcpi> But merely as a recognition of having achieved a certain level of education, i.e. a PhD 01:23 < gleki> .dict professor 01:23 < fenki> professor — noun: 1. A teacher or faculty member at a college or university, 2. A higher ranking for a teacher or faculty member at a college or university. Abbreviated Prof 01:23 < gleki> 1 word in query. please wait until JVS is back online. 01:55 < zipcpi> Hm I'm actually starting to like the idea of elidable co'e in gadri 01:56 < zipcpi> bu'u lo vuku / bu'u le vuku 01:56 < gleki> alta: lo cu mo 01:56 < mensi> ([FA {lo COhE KU}] [cu {mo VAU}]) 01:57 < zipcpi> {leku} could be a useful idiom for anything that you might use a demonstrative for but which actually isn't there to demonstrate 01:58 < zipcpi> It's like "You should know what I'm talking about by context already; there is no need for description. "This" 01:59 < gleki> alta: bu'u le vu cu mo 01:59 < mensi> ([{bu'u } {FA ZOhE}] [cu {mo VAU}]) 01:59 < gleki> alta: bu'u le vu mo 01:59 < mensi> ([{bu'u } {FA ZOhE}] [CU {COhE VAU}]) 01:59 < gleki> alta: mi djica lo nu lo plise 01:59 < mensi> ([FA mi] [CU {djica } KEI] KU¹)> VAU}]) 02:00 <@Broca> How does the PEG rule look like? I don't see how you can make an elidable something that isn't a terminator. 02:00 < gleki> are you asking how this alta works? 02:00 <@Broca> I mean, if you make a description *without* co'e, the grammar doesn't insert it, so it doesn't work like an elidable something at all. 02:00 <@Broca> Yes. Yes, I am. 02:01 < gleki> alta: mi djica lo nu lo plise 02:01 < mensi> ([FA mi] [CU {djica } KEI] KU¹)> VAU}]) 02:01 < zipcpi> Alta is very forgiving 02:01 <@Broca> Because it's just an alternative PEG ruleset, right? 02:01 < zipcpi> I dunno if it runs on PEG 02:01 <@Broca> Oh my. Lojban, DWIM-style. 02:02 < gleki> here, Alta is waiting for a selbri but if there isn't then after"/" operator there is another option, namely,selbri_elidible which requires mepty string and outputs selbri_elidible node 02:02 < gleki> yes, Alta is fully in PEG 02:02 < gleki> i didnt want any javascript on top of PEG 02:02 < gleki> cirko is working on an alternative 02:02 < gleki> using javascript, it seems 02:03 < zipcpi> broca: Specifically though, {co'e} is too semantically tied to "thingy" in people's minds, so I'd like to be able to avoid it in the usage scenarios I described 02:03 < gleki> i didnt want anything on top of PEG because the current PEG itself is very strange and not elucidating 02:03 < zipcpi> "All you have to do is meet here there" 02:03 < zipcpi> *meet her there" 02:04 < zipcpi> do bilga fe po'o lo nu penmi le ninmu bu'u... 02:04 < zipcpi> {tu} might not be right if you can't actually point or otherwise signal to it 02:04 <@Broca> I imagine if you feed Alta the empty string it auto-completes to "mi klama le zarci" or something :-) 02:04 < gleki> indeed 02:05 < gleki> i just havent come to autocorrecting everything 02:05 <@Broca> alta: 02:05 < mensi> ([FA ZOhE] [CU {COhE VAU}]) 02:05 <@Broca> Ahhahaha! 02:05 < zipcpi> So what I really mean is "there is a place that I think you already know about by context, merely by the description of it being far away 02:05 < gleki> ideally it shouldnt output any errors 02:05 < gleki> a more funny example is 02:05 < gleki> alta: lo nu noi ku 02:05 < mensi> ([FA {lo <(¹nu [{FA ZOhE} {CU }] KEI¹) (¹noi [{FA ZOhE} {CU }] KUhO¹)> ku}] [CU {COhE VAU}]) 02:05 <@Broca> You mean, the purpose of the project is that it should accept everything? 02:06 < gleki> yes, autocorrect everything 02:06 < gleki> it's one of the purposes 02:06 <@Broca> alta: abracadabra 02:06 < mensi> SyntaxError: Expected ['h] or ['hiI] but "b" found. 02:06 < gleki> another one is te sumti detection using PEG only 02:06 < gleki> Broca: oops, probably not in morphology 02:06 <@Broca> :-) 02:06 < gleki> but that could be autocorrected by prefixing with {la'oi} 02:07 < zipcpi> u'i 02:11 < zipcpi> Specifically though I think elidable selbri in gadri can work without going all the way to alta 02:11 < zipcpi> Just instead of saying "It accepts exactly one selbri", now it's "accepts zero or one" 02:12 < zipcpi> If {ku} or some incompatible thing is encountered, just close it 02:12 < zipcpi> Kinda like how COI can be left empty 02:13 < zipcpi> exp: coi vu 02:13 < mensi> ([coi DOhU] [vu KU] VAU) 02:13 < zipcpi> Right COI doesn't accept sumtcita lol 02:13 < gleki> in alta injecting {co'e} is needed to avoid ruining the concept of selbri 02:13 < zipcpi> Hm 02:14 < gleki> it should be {coi zo'e}, yes. a known issue 02:14 < zipcpi> lol 02:14 < zipcpi> Gotta go. co'o 02:18 * nuzba @464161niftg: lojban › A chart of audio of Lojban phonemes https://groups.google.com/d/topic/lojban/QbFMY_690J0/discussion / la cirko がWikiMediaの音声を集めて表を作つたので、或はそれも有用かも。http://mw.lojban.org/extensions/ilmentufa/cirkotci.html [http://bit.ly/1CcxC49] 02:20 * nuzba @464161niftg: .uenai .uinai la http://lojban.org nau spofu cando .i ba ko'oi vitke [http://bit.ly/1Ccy3LY] 03:03 < _mukti_> .oi la jukni cu ca 03:03 < _mukti_> s/ca/ca'o bilma/ 03:03 < fenki> _mukti_ meant to say: .oi la jukni cu ca'o bilma 03:10 < gleki> mumble: kmi pu benji SMS boi CGM i ku'i na spusku 03:10 < gleki> *mi 03:10 < gleki> i oise'i 03:10 < gleki> _mukti_: mi pu benji SMS boi CGM i ku'i na spusku 03:11 < _mukti_> ki'e la gleki 03:11 < gleki> i xu la jukni na spuda 03:12 < _mukti_> ie la jukni na frati 03:13 < gleki> mi ca zbasu lo skripto pe lo nu zbasu lo vlaste poi traduki lo glico lo lojbo 03:15 < _mukti_> ue la cirko cu te xatra byfy boi bau lo jbobau se pi'o zo'oi h 03:15 < gleki> sepiho zohoi h ie 03:15 < _mukti_> .u'i 03:15 < gleki> _mukti_: ei ma'a zukte ma va'o lonu la jukni soi'roi co'a spofu 03:16 < gleki> i pe'i lo nu zbasu lo drata kibystuzi cu sarcu 03:16 < _mukti_> lo nu do zbasu lo skripto kei ua .i'e 03:17 < gleki> sa'u mi pu trci lo ka pilno lo formula pe la gugle gredile i ji'asai GQL i ku'i le gredile co'a du'eva'ei masno 03:17 < _mukti_> Switching to English for a moment, because my geekspeak in lojban is ironically impaired 03:17 < gleki> *troci 03:17 < gleki> i never use any lujvo 03:17 < _mukti_> Once jukni is accessible again, I'm going to check the logs for indications of what caused it to become wedged 03:17 < gleki> well, with most people 03:17 < gleki> oh 03:18 < gleki> i just think a github website might be a good alternative 03:19 < _mukti_> I wonder if the internet backbone attack is somehow related to Jukni's outage... 03:19 < gleki> uanai 03:19 < gleki> en: backbone 03:19 < _mukti_> I'm seeing a lot of sites inaccessible, and the server is located in the bay area ia 03:19 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 03:21 < ctefaho> does za'e have a UI version? 03:22 < gleki> en: zi'a 03:22 < mensi> zi'a = [UI3a] nonce-word indicator; indicates previous word is nonce-creation and may be nonstandard |>>> Equivalent to 03:22 < mensi> za'e but in selma'o UI. See also za'e, pe'a, xo'o |>>> gleki 03:22 < gleki> and ba'e=> bi'a 03:22 < keegi> I am reading https://lojban.github.io/cll/18/10/ and I have some questions. 1) xu pano ju'u paxa du paxa 2) xu pano ju'u reno du pano 03:22 < ctefaho> ki'e 03:22 < ctefaho> bi'a I know, just wondered if za'e had an equivalent 03:23 < gleki> xu li pano ju'u pa xa du li paxa 03:23 < ctefaho> zi'a is experimential, correct? 03:23 < gleki> yes 03:23 < gleki> and bi'a 03:24 < ctefaho> bi'a too? 03:24 < _mukti_> ua zo ju'u na nitcu zo li 03:24 < ctefaho> hmm I had gotten the impression that was standard 03:24 < gleki> keegi: im think you are right although just add {li} as i showed 03:24 < gleki> *i think 03:24 < keegi> gleki: thanks, indeed missed {li} 03:24 < _mukti_> A postfix {za'e} ? 03:25 < gleki> en: bi'a 03:25 < mensi> bi'a = [UI3a] emphasis indicator; indicates the previous word is especially emphasized |>>> Equivalent to ba'e but in UI. 03:25 < mensi> See also ba'e |>>> gleki 03:25 < gleki> just was shutting some selmaho 03:25 < gleki> *shutting down 03:26 < keegi> what about {li pano ju'u papa} is it like base 16 (since we have numbers for it) or is it like base 20 ? 03:26 < ctefaho> so ba'e and bi'a are just forethought vs afterthought then? 03:27 < gleki> oh 03:27 < gleki> i think i got your question 03:27 < gleki> it should be {li pa no ju'u re no cu du li re no} 03:28 < _mukti_> li pa no ju'u re no cu du li re no ju'u pa no 03:28 < gleki> and no lojban still doesnt have words for more than 15 03:28 < keegi> gleki: that's contradicting example 10.8) in that article 03:29 < gleki> keegi: ok. then you are right 03:30 < _mukti_> li pano ju'u reno ju'u pano cu du li reno ju'u pano ju'u pano 03:31 < keegi> _mukti_: li pa pi'e no ju'u re no cu du li re no ju'u pa no 03:33 < _mukti_> pei li no ju'u ma kau du li no 03:33 < _mukti_> .i je li pa ju'u ma kau du li pa 03:34 < _mukti_> li pa ju'u pa cu li pa 03:35 < _mukti_> li papa ju'u li pa du li re ju'u li pano 03:35 < _mukti_> s/ju'u li pa/ju'u pa/ 03:35 < fenki> _mukti_ meant to say: li papa ju'u pa du li re ju'u li pano 03:35 < _mukti_> la fenki cu masno milxe 03:37 < _mukti_> ua nai zo pi'e na slabu mi 03:39 * nuzba @lojban_org: @464161niftg http://vrici.lojban.org/~gleki/mediawiki-1.19.2/extensions/ilmentufa/cirkotci.html [http://bit.ly/1NyOK4A] 03:39 * nuzba @lojban_org: @464161niftg http://vrici.lojban.org/~gleki/mediawiki-1.19.2/extensions/ilmentufa/cirkotci.html [http://bit.ly/1NyOMcJ] 03:40 < _mukti_> keegi: Thank you, I see what you mean 03:41 < _mukti_> {pano} being a digit in base 20, {pa pi'e no} being one in the twenties place and zero in the ones place 03:41 < _mukti_> .u'a ua 03:42 < keegi> yes, with exception to base 16 (and possibly base 11..15) where there are explicit digits for 10..15 03:42 < keegi> if i understand it correctly 03:43 < _mukti_> It sounds like you do. So were you saying earlier you found an inconsistency? 03:43 < keegi> no, just unsure about bases 11..15 03:43 < _mukti_> How so? 03:47 < keegi> is { li pano ju'u papa } equal 10 or 11 ? is the correct sequence {li so ju'u papa} {li pano ju'u papa} {li pa pi'e no ju'u papa} or {li so ju'u papa} {li pau ju'u papa} {li pano ju'u papa} ? 03:47 < keegi> s/pau/dau 03:47 < fenki> keegi meant to say: is { li pano ju'u papa } equal 10 or 11 ? is the correct sequence {li so ju'u papa} {li pano ju'u papa} {li pa pi'e no ju'u papa} or {li so ju'u papa} {li dau ju'u papa} {li pano ju'u papa} ? 03:48 < phma> en:fiction 03:48 < mensi> 14 da se tolcri: cfika, cumymu'efi'a, fikcku, frafi'a, mi'irlarfi'a, brano, clinu, daskemxarpresmitra, fikrfantasi, 03:48 < mensi> kunrkripto, lisri, ridrdverga, skefi'a, xarpresmitra 03:49 < _mukti_> If {li pano ju'u reno} is equal to decimal 10, as 10.8 suggests then I would assume {li pano ju'u papa} is also equal to decimal 10, with {pano} in each case representing the digit ten 03:49 < _mukti_> i.e. equivalent to {dau} 03:52 < keegi> _mukti_: .u'a ua 03:55 < _mukti_> The rule seems to be, for bases larger than decimal, any number larger than 9 but smaller than the base is assumed to be a digit, and {pi'e} (thank you, new trick for me) is needed if you want multiple digits 03:56 < keegi> except base 16, where you can say {li dau fei gai ju'u pa xa} 03:57 < _mukti_> Right, I mean only when using digits no through so 03:58 < gleki> _mukti_: robin is on vacation 03:58 < _mukti_> ua 03:58 < gleki> you can ask Tene and banseljaj in #jbopre 03:59 < keegi> hm, is then {li pa no ju'u pa xa} same as {li dau ju'u pa xa} ? I thought it was {li pa no ju'u pa xa} same as {li pa pi'e no ju'u pa xa} 04:01 < _mukti_> gleki: Ok, I msg'd them. But in the case they aren't tuned in, and so you know, I'm going to be away at work in the next hour or so. 04:01 < gleki> okay, 04:01 < gleki> im not an expert here anyway 04:02 < _mukti_> keegi: It would seem from 10.8 that hex is a special case where pano is not assumed to be a digit 04:03 < keegi> .ua u'a 04:03 < _mukti_> Or actually, I'm not sure. Perhaps it's consistent. Is there an example that uses pano ju'u paxa 04:03 < _mukti_> ? 04:03 < _mukti_> pa pi'e no ju'u paxa is unambiguous 04:04 < phma> mensi:spatrxorki 04:04 < gleki> en:spatrxorki 04:04 < mensi> spatrxorki = s1 is an orchid of genus Orchis of species/strain/cultivar s2. 04:04 < gleki> mensi: help 04:04 < mensi> Parsers: type "exp:" (experimental), "off:" (camxes), "gerna:" (jbofi'e), or "yacc:" (official yacc) followed by the text to show the structure of sentences. 04:04 < mensi> Lojban dictionary: type "language-code: word", where language code is one of jbo,en,ru,es,fr,f@,ja,de,eo,zh,hu,sv. This searches in both directions. 04:04 < mensi> Type "language-code:word" (i.e. without a space after ":") to get a shorter definition. 04:04 < mensi> "selmaho: ca'a" gives "CAhA", "selmaho: CAhA" gives "bi'ai, ca'a, ..." 04:04 < mensi> "rafsi: kulnu" gives "klu", "rafsi: klu" gives "kulnu" 04:04 < mensi> Other conlang dictionaries: "toki:", "laadan:", "loglan:" 04:04 < mensi> Lojban <-> Loglan conversion (incomplete): "coi:", "loi:" 04:04 < mensi> "Tatoeba: klama" gets a random example sentence using "klama" 04:04 < mensi> Delayed messaging: type "mensi: doi user message" to send "message" to "user" when they return 04:05 < _mukti_> ue ua 04:05 < gleki> one more function 04:05 < gleki> mensi: gadri 04:05 < mensi> lo [PA] broda = zo'e noi ke'a broda [gi'e zilkancu li PA lo broda] 04:05 < mensi> la [PA] broda = zo'e noi lu [PA] broda li'u cmene ke'a mi 04:05 < mensi> lo PA sumti = lo PA me sumti 04:05 < mensi> la PA sumti = zo'e noi lu PA sumti li'u cmene ke'a mi 04:05 < mensi> loi [PA] broda = lo gunma be lo [PA] broda 04:05 < mensi> lai [PA] broda = lo gunma be la [PA] broda 04:05 < mensi> loi PA sumti = lo gunma be lo PA sumti 04:05 < mensi> lai PA sumti = lo gunma be la PA sumti 04:05 < mensi> lo'i [PA] broda = lo selcmi be lo [PA] broda 04:05 < mensi> la'i [PA] broda = lo selcmi be la [PA] broda 04:05 < mensi> lo'i PA sumti = lo selcmi be lo PA sumti 04:05 < mensi> la'i PA sumti = lo selcmi be la PA sumti 04:05 < mensi> PA sumti = PA da poi ke'a me sumti 04:05 < mensi> PA broda = PA da poi broda 04:05 < mensi> piPA sumti = lo piPA si'e be pa me sumti 04:05 < gleki> a secret periodic table of gadri :) 04:06 < keegi> agreed that using pi'e is unambiguous, no, i haven't found examples that would clarify it 04:06 < b_jonas> coi 04:07 < _mukti_> coi la .ionac. 04:07 < phma> en:curve 04:07 < mensi> curve = x1 is pure/unadulterated/unmitigated/simple in property x2 (ka). 04:07 < phma> en:kruvi 04:07 < mensi> kruvi = x1 is a curve/turn/bend in x2, at locus x3, and defined by set of points/properties x4. 04:08 < phma> en:centaureia 04:08 < mensi> centaureia = x1 is a cornflower of type x2 04:10 < _mukti_> Seriously bad Internet weather today 04:10 < _mukti_> I'm having trouble reaching google news 04:11 < f8l> I can’t reach lojban.org. :-( 04:12 < _mukti_> Yes, lojban.org is down ui nai sai 04:13 < gleki> i think a simpe, static website is necessary 04:13 < gleki> *simple 04:13 < gleki> _mukti_: any suggestions? 04:14 < _mukti_> gleki: Well, one suggestion that occurs to me is that it would be good if access to lojban.org's registration were better distributed. I don't know if Tene or banseljaj has it. 04:14 < _mukti_> This way, we could, in a case like this, at least point lojban.org at a static site. 04:15 < banseljaj> You mean DNS, mukti? 04:15 < _mukti_> Yes, although depending on the registration the DNS may be self-hosted, etc. 04:16 < banseljaj> Yeah I don't know where the DNS is 04:16 < banseljaj> Probably on gandi.net 04:16 < _mukti_> Yep 04:17 < _mukti_> So the registrar does appear to being doing the DNS, too 04:18 < banseljaj> Tene would be the next person to ask 04:18 < gleki> _mukti_: i mean constructing the actual website. any sugegstions of a supersimple CMS? 04:18 < _mukti_> Ok, I msg'd him as well 04:18 < gleki> i only know that MW cant be converted into a static website 04:18 < gleki> automatically 04:18 < phma> DNS doesn't look like the problem. 04:19 < banseljaj> Failover is a problem. 04:19 < _mukti_> No, DNS is not the problem, but we're brainstorming how to better handle outages like this in the future, and if we had the ability to alter the DNS to point away from jukni, we'd be able to at least get a site on line 04:20 < phma> What I'm worried about is, the form for reserving the rooms for jbonunsla have to be signed by today, the names and addresses are mixed up, 04:20 < phma> and with lojban.org down, la lojbab can't get his email. 04:20 < phma> I sent an email to the hotel gal explaining the problem. 04:20 < _mukti_> phma: Maybe I can help with this? 04:21 < b_jonas> i lo tcadu sorprekarce cu se nenri lo goi ko'a su'o seljai .i lo prenu poi sanli cu jgari pa .a re ko'a te zu'e lo nu ra na snuti farlu ka'e lo nu lo karce cu spaji jabre .i ma lojbo valsi lo ka me ko'a 04:21 < phma> maybe. are you an officer? 04:21 < _mukti_> Yes, I'm officially Sec/Treas now. 04:22 < phma> are you Shunpiker? 04:22 < _mukti_> That's my email, yes. 04:23 < phma> Then you should have been getting the emails. Send Lorit an answer and let's see if we can fix this problem. 04:23 < _mukti_> Ok, I'm on it. 04:23 < phma> ki'esai do 04:23 < phma> also cc me and lojbab 04:24 < _mukti_> No problem. Is the contract from June 26 the latest? 04:24 < etgdfgfd> coi 04:25 < _mukti_> coi 04:27 < gleki> i'd like to copy most important pages into that backup website and our html/js apps 04:27 < gleki> like camxes.js, sutysisku 04:27 < _mukti_> I think poor Lorit is likely to be bewildered by all of that detail. I'm going to try to make things simpler for her. 04:29 < b_jonas> i ko'a me lo nalmuvdu jinme tubnu gi'a dasri gi'e bo dandu ri 04:32 < etgdfgfd> jgari grana ja garna 04:34 < b_jonas> je'e 04:51 < phma> mi puzi zvati le jupku'a 04:52 < phma> all what detail? 04:52 < _mukti_> The last email about the domain being down, etc. 04:54 < _mukti_> I will just serve as the main contact henceforth. Thank you for doing the legwork of setting everything up! 04:55 < phma> The latest version has my phone number and Lojbab's address. She may not answer email till evening, so you may want to call her. 04:56 < _mukti_> I'm not going to wait. I'm going to print the contract, amend and sign it in my name, and then send it to her. If she wants to print out another version, I'll return that too, but this way at least she will have something in hand today. 04:57 < phma> ok 04:58 < _mukti_> I'll continue to CC you and Bob on all communications 05:00 < gleki> _mukti_: can you finally oblige organizers of logfests report on how they went? iotherwise it's a waste of time 05:01 < _mukti_> gleki: I agree that reports are useful, but they're not unprecedented ... banseljaj reported on the gathering he organized a couple of years ago 05:01 < banseljaj> Speaking of, I have the pictures and a bunch of video too. 05:02 < gleki> _mukti_: what in your opinion logfests are for? where are textbooks, dictionaries, fluent speakers that they produced? there are none 05:02 < banseljaj> I lack the skills to turn them into something presentable and would be willing to hand it over 05:02 < _mukti_> But yes, I agree that it would be valuable ... in organizing future events, it would be good to know what worked well, what didn't work, etc. 05:02 < gleki> it'd be much more helpful to send NooLearning to Germany of if he doesnt want then to me. 05:02 < _mukti_> gleki: I think there are intangibles. We're not meeting to draft a dictionary. 05:03 < banseljaj> gleki: Seeing as we are a very niche community (Someone once described lojbanists as the skinny nerd in the back that even the nerds pick on), Just the existence of logfests is great as it allows people to meet. 05:03 < gleki> banseljaj: why it's not a waste of money and time in your opinion? 05:04 < banseljaj> Because otherwise you don't have an annual gathering and more people lose interest. 05:04 < gleki> banseljaj: have you tried once? 05:04 < gleki> prove your claim 05:04 < banseljaj> Theoretically, you can say "we organized a dozen logfests" and people would not dig further. 05:04 < gleki> this is how i see it. 05:04 < banseljaj> Tried not having a logfest? 05:05 < gleki> yes 05:05 < gleki> people dont speak much lojban there. so no fluent speakers? 05:05 < gleki> then what are logfests for? to impress someone that lojban is alive? 05:05 < banseljaj> We have three fluent speakers, relatively 05:05 < gleki> well, not lojban but lojbanistan 05:05 < gleki> banseljaj: okay, how many of them taught new people speak lojban? 05:05 < banseljaj> gleki: Pretty much the primary purpose of any sorts of convention. 05:06 < gleki> banseljaj: whom have you impressed? 05:06 < banseljaj> I know two people who have taught extensively, both of them taught me. There's robin, who's teaching his daughters. bob taught a lot of people back in the day 05:06 < gleki> banseljaj: during logfests? 05:07 < banseljaj> Not during logfests 05:07 < gleki> banseljaj: who or what benefited during the logfest you were present at? 05:07 < _mukti_> I think part of the value of a gathering like this is that you don't know what is going to emerge from it. A single person may be inspired to do something as a result of it. Or people may form a bond which enables them to work on something else together. Etc. 05:07 < gleki> banseljaj: whom did you impress during or as the result of that logfest? 05:07 < banseljaj> We are *not* a structured or hierarchical community 05:07 < gleki> _mukti_: what did emerge as the result of past logfests? 05:08 < banseljaj> it'd be foolish to accept tangible goals like that. 05:08 < banseljaj> gleki: the CLL v1 05:08 < _mukti_> My understanding is that a lot of the language was designed at the early gatherings. 05:08 < gleki> I state that logfests, primarily last 3 logfests were a waste of time and money. Now prove me wrong. 05:08 < gleki> documentally 05:09 < gleki> My counterproposal: instead of a logfest spend some money on making selpa'i and NooLearning meet. 05:10 < gleki> he may not agree ofc, then his name can be replaced. 05:10 < banseljaj> gleki: Well I organized one of them. I can assure you that not much money was spent and it was generally a celebration of the 25 years we have had lojban for. We reflected on how we got there and what we should do moving forward 05:10 < gleki> banseljaj: why a skype conference couldnt be done forthat? 05:10 < banseljaj> gleki: Nope 05:11 < gleki> So my counterproposal goal: record the whole meeting, teach representatives of NooLearning basic Lojban, upload the recordings to the LLG 05:11 < gleki> audio would be enough 05:11 < _mukti_> gleki: I think you also have to factor in what people want to do. 05:11 < banseljaj> What is NooLearning? 05:12 < _mukti_> People want to meet. There's no indication, for example, that people want to teach representatives of NooLearning. If they do, great, let's talk about supporting that. 05:12 < gleki> banseljaj: just a random org. well, they want to create an e-learning tool 05:12 < _mukti_> There is, however, an indiction that people want to meet. 05:12 < gleki> _mukti_: if people want to meet then why? 05:12 < gleki> why should we support their desire? 05:13 < banseljaj> Uh. Because humans are wrired to meet others? 05:13 < _mukti_> Lojbanists meeting is consonant with the mission of LLG. 05:17 < _mukti_> Your question, gleki, provokes another question: Why should we support things that people don't desire? 05:17 < _mukti_> If you, for example, desire to teach representatives of NooLearning -- great! Put togeter a proposal and bring it to the Board! 05:18 * PrincessBecca tried organizing a washington toki pona meetup. ended up being her and one other person the first time. and the other person never came back 05:18 < samcinki> .oi dai 05:18 < _mukti_> ike 05:18 < PrincessBecca> je'e 05:20 < banseljaj> The thing is persistence, PrincessBecca 05:20 < banseljaj> I've had a lot of meetings where I was the only one or one of two peopl 05:21 < PrincessBecca> yeah probably. i 05:21 < PrincessBecca> may have been i 05:21 < PrincessBecca> n over my head tryi 05:21 < PrincessBecca> ng to organi 05:21 < PrincessBecca> ze esperanto and toki pona at the same time 05:21 < PrincessBecca> ugh stupid keyboard 05:21 < PrincessBecca> i ended up just sticking with eo and neglecting toki pona 05:23 * nuzba @xeno_256: @kuina_ch 人工言語だけど、ロジバン(lojban)も主語を省略できて、空気を読む形での推測です。面白そうなので勉強中… [http://bit.ly/1Cd0EAL] 06:04 < noncomcinse> .i cipra 06:04 < noncomcinse> s/cipra/snada/ 06:04 < fenki> noncomcinse meant to say: .i snada 06:04 < noncomcinse> s/snada/fliba/ 06:04 < fenki> noncomcinse meant to say: .i fliba 06:05 < noncomcinse> s/fliba/mi kakne loka pilno za'ure'u ku pa notci/ 06:05 < fenki> noncomcinse meant to say: .i mi kakne loka pilno za'ure'u ku pa notci 06:05 < noncomcinse> .i cipra cipra 06:06 < noncomcinse> s/cipra/snada/ 06:06 < fenki> noncomcinse meant to say: .i snada cipra 06:06 < noncomcinse> .i cipra cipra 06:06 < noncomcinse> s/cipra/snada/g 06:06 < fenki> noncomcinse meant to say: .i snada snada 06:06 < noncomcinse> s/SNADA/cipra/ 06:07 < noncomcinse> .i cipra cipra 06:07 < noncomcinse> s/cipra/snada/g 06:07 < fenki> noncomcinse meant to say: .i snada snada 06:07 < noncomcinse> s/SNADA/cipra/i 06:07 < fenki> noncomcinse meant to say: .i cipra snada 06:30 < gleki> phma: is noolearning coming? 06:31 < phma> afaik yes. 06:38 < phma> en: nabmi 06:38 < mensi> nabmi = x1 (event/state) is a problem to/encountered by x2 in situation/task/inquiry x3. |>>> Also: x1 requires 06:38 < mensi> consideration by x2. See also preti, danfu, ciksi, jijnu, jinvi, nandu, pensi, sidbo, spuda, raktu. |>>> 06:38 < mensi> officialdata 06:39 < phma> en: danfu 06:39 < mensi> danfu = x1 is the answer/response/solution/[reply] to question/problem x2. |>>> (cf. ciksi, frati, preti, nabmi, spuda 06:39 < mensi> for agentive response/reply, cpedu) |>>> officialdata 06:40 < Zearen> .i coi ro jundi 06:41 < niftg> coi mi'e na'e banzu jundi 06:44 < gleki> Zearen: coi. have u upvoted at least 5 questions with rating below 10 here area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 06:47 < Zearen> .i na go'i .i mi na me la'e zoi gy. have an account gy. 06:47 < gleki> do se jaspu no da 06:47 < gleki> i e'u do zbasu lo jaspu 06:47 < Zearen> Beginnners: Don't do what I just did ^^' 06:48 < gleki> gi'e zanru su'o mu lo preti 06:48 < gleki> for tatoeba.org interface i translated "account" as {jaspu} 06:48 < gleki> {lo jaspu} to be precise 06:51 < Zearen> vi'o 06:52 < Zearen> .i ta'o la vlasisku cu fliba do'e mi 06:52 < gleki> http://vrici.lojban.org/~gleki/mediawiki-1.19.2/extensions/ilmentufa/ircbot/naxle.html 06:52 < gleki> use this 06:53 < gleki> vrici.lojban.org/~gleki/mediawiki-1.19.2/extensions/ilmentufa/i/en/index.html 06:53 < gleki> or this 06:56 < Zearen> .i ki'e .i mi ka'e tavla bau la .lojban. pi'o no valsi sorcu .iku'i lote kancu lo lojbo valsi poi jai se djuno mi cu cmalu 06:57 < gleki> sepi'o 06:57 < Zearen> .ie .u'u 06:57 < gleki> i lo te kancu be lo lojbo valsi li'o 06:59 < Zearen> za'a mi duske loni sutra 06:59 < gleki> dukse 06:59 < Zearen> si dukse 07:00 < gleki> jb: zildukse 07:00 < mensi> zildukse [< zi'o dukse ≈ Zi'o* dukse*] = zildukse — x1(entity) is in excess, is too much for x2(property of x1) to 07:00 < mensi> take place(clause)'' 07:00 < mensi> :lo zildukse — excess. 07:00 < mensi> :lo nu mi tatpi cu zildukse lo nu mi sidju do — I'm too tired to help you. 07:00 < mensi> :Related words: bancu, banzu, ricfu, zmadu 07:00 < Zearen> .oi puzu mi se lojbo 07:00 < gleki> do du'e va'e lo ka sutra 07:00 < gleki> i do ze'u na lojbo 07:00 < gleki> i puzuku do lojbo i babo do na lojbo 07:01 < Zearen> .i .oiru'e mi bazi na tavla 07:02 < gleki> do di'a lojbo vau i'e 07:02 < gleki> i ta'o xu do na co'u morji tu'a zoi gy. area15.gy. 07:03 < Zearen> .i na go'i .i la'e zoi gy. area15 gy. mo 07:03 < gleki> http://area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 07:04 < Zearen> .ua zoi gy. area15 gy. nadu zoi gy. area51 gy. 07:05 < gleki> i oi se'i mi pu srera 07:37 * clakre wakes up this morning with "balnema" playing in head 07:38 * clakre realizes ze'a isn't terminated and is acting as a sumtcita. Song suddenly makes more sense 07:38 < gleki> oh, seems like eng2jbo dictionary will be generated in one hour using livla.js. much better than infinite time using Google Spreadsheet :D 07:38 < clakre> u'i 07:38 < Zearen> Is there a «Lojban the Hard Way» equivalent that uses metaphors like "Imaginary journeys are best understood as modifications of inner selbri like other modals, hence «da ga'uvu zu'avi muvdu» can be understood as moving that's the a little to the left, and then the a-little-to-the-left-moving is far above. I.e. each outer modal adds a segment creating a new selbri to the exiting imaginary journey of the inner selbri". 07:39 < gleki> clakre: have you upvoted 5 questions with rating below 10 here? pls do if you havent yet area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 07:39 < durka42> Zearen: heh, sounds like you should write one :) 07:39 < clakre> sorry, not a stackexchange member and haven't taken the time to become one yet... 07:39 < gleki> Zearen: sounds like Waves or CLL or Jonh Clofford 07:39 < gleki> *Clifford 07:39 < gleki> clakre: you may login 07:39 < Zearen> .i mi na banzu loni certu 07:40 < gleki> jb: banzu 07:40 < gleki> oh i see 07:41 < Zearen> Well, I actually got frustrated this way trying to understand something in the CLL. It kept making metaphors to English which made it harder to understand. 07:42 < gleki> i know 07:42 <@xalbo> The intention of the Wave Lessons was to teach Lojban qua Lojban, although it does not infrequently try to relate things back to English. 07:43 * clakre is using the Wave Lessons 07:43 * clakre needs to get back to making them into an epub with the answers on their own pages at the back 07:43 < gleki> xalbo: that could be justified if it used real life examples. but it doesnt 07:43 < Zearen> Which is fair, the CLL is targeted to beginning speakers who haven't had their brains warped by the language yet. I mean more of a strictly intermediate text. 07:44 < durka42> ie sarcu 07:44 < Zearen> Er, not targeted, but written as not to alienate beginners. 07:44 <@xalbo> I really should reread the Wave Lessons and work on improving them. 07:45 < durka42> mi'u 07:45 < gleki> ящ кфэщ уэг 07:45 < gleki> zo ra'o e'u 07:45 < PrincessBecca> omg @ "nasa" off the new lojban rap album... when he starts spitting in toki pona 07:45 < durka42> actually I don't think I've read them the first time .o'anai 07:45 < gleki> i zo mi'u pe do malgli 07:45 < durka42> PrincessBecca: .i'i 07:46 < Zearen> .i ki'u ma doi gleki 07:47 < gleki> mo ki'u makau 07:48 <@xalbo> ma ve jinvi lo du'u malgli vau zo'ei zo mi'u do doi la gleki 07:48 < Zearen> zo mi'u pe la durka malgli 07:48 < gleki> *cu 07:49 < durka42> da'i ca'iku .ei mi cusku lu mi ji'a li'u 07:49 < gleki> BPFK cu xusra lo du'u lu mi'unai li'u srana lo si'o anfora 07:49 < gleki> i mi tugni 07:49 < gleki> i plixau 07:49 < gleki> jb: mi'u 07:49 < gleki> y 07:50 < gleki> jb: mi'u 07:50 < gleki> za'a gau mi spofu 07:52 <@xalbo> I can't find any reference to {mi'unai} in CLL, although maybe I'm looking in the wrong place. All I find is ““mi'u”, adds a parallel case to the previous argument, and can also be used in tables or the like to show that something is being repeated from the previous column. It is distinct from “go'i” (of selma'o GOhA, discussed in Chapter 7), which is a non-discursive version of “ditto” that explicitly repeats the claim of the previous 07:52 <@xalbo> bridi.” 07:52 <@xalbo> http://lojban.github.io/cll/13/12/ 07:52 < gleki> but it doesnt act like {ra'o} 07:52 < gleki> it doesnt update references to the current speaker 07:53 < gleki> at least this is what all should agree i suppose 07:53 <@xalbo> Huh, http://lojban.org seems to be down. 07:55 < gleki> you are the last person to notice that 07:55 < gleki> http://webcache.googleusercontent.com/search?q=cache:PNH-ay9A7GoJ:mw.lojban.org/papri/BPFK_Section:_Highlight_Discursives+&cd=1&hl=en&ct=clnk&gl=ru 07:55 < gleki> mi'u ^ 07:57 * nuzba @fotono: lo #lojban zantufa tricu cu milxe banro .i .uo.uo co'osai ma'oi SA http://guskant.github.io/gerna_cipra/zantufa_farvi.svg [http://bit.ly/1LKjjqt] 08:00 < gleki> i think it's not thr right time to test updates tomensi 08:00 < Zearen> zo anfora ki'a 08:00 < durka42> anaphora 08:02 < gleki> jb: mi'u 08:02 < mensi> mi'u = mi'u [interjection] — ditto (the same thing) — another one 08:02 < mensi> :mi pu zukte ra mi'u — I did the same thing. 08:02 < mensi> :mi pu lebna lo cukta gi'e dunda lo cukta mi'u nai la Alis — I took a book and gave another one to Alice. 08:02 < mensi> :Related words: mintu, bi'u nai, le, ri, ra 08:02 < durka42> mi'u bi'u 08:03 < gleki> bi'u is different 08:03 < gleki> jb: bi'u 08:03 < mensi> bi'u = bi'u [interjection] — newly introduced information, bi'u nai — previously introduced 08:03 < mensi> information 08:03 < mensi> :lo bi'u nai nixli pu klama lo panka i bi'u ny ca tavla lo nanla bu'u py — That girl (mentioned earlier), she went to a 08:03 < mensi> park. What is new is that she is talking to a boy there. 08:03 < mensi> :Related words: cnino 08:03 < Zearen> I forget, is there a NOI that's like an editor's correction, e.g. "The governer was quoted saying, '[The police] have recently been criticized for...'" 08:04 < ctefaho> you mean parenthesis? to-toi if so 08:04 < durka42> there is a TO with an automatic sa'a 08:04 < durka42> I forget what it is 08:04 < durka42> to'i? 08:04 < durka42> vlaste: to'i 08:04 < vlaste> to'i = open editorial unquote (within a quote); contains grammatical text; mark with editorial insert. 08:05 < durka42> to'i = to sa'a 08:05 < Zearen> Oh ! That's it. 08:05 < gleki> le turni pu se sitna fi lo smuni be lu ... li'u 08:06 < Zearen> zo «sa'a» cu pu jai sisku mi 08:06 < durka42> sesisku* 08:06 < ctefaho> yeah to'i 08:06 < Zearen> .y. .ie 08:08 < Zearen> .i za'a mi nitcu lonu lonu mi lojbo cu zmadu fi loni cafne 08:09 < Zearen> .iku'i lonu gunka cu co'e 08:13 < Zearen> .i mi pu djica lonu cusku «lu zo mi'u pe la durka sa'a cu malgli li'u» 08:17 * nuzba @garydale: Fiction: At the GruntWerx Board of Directors Meeting: http://bit.ly/1IRyUON #lojban [http://bit.ly/1H3yoPA] 08:18 < Ilmen> za'a la lojbo kibystu za'o spofu ju pofpo'u 08:21 < Ilmen> .i mi kucli lo ka ki'u ce na jmina su voi cnino je cu zifre zmadu lo ka'e jai gau .akti be di'a bei fai la lojbo kibystu 08:21 < Ilmen> *kei 08:21 < Ilmen> s/zmadu/zmadu kei 08:22 < fenki> Ilmen meant to say: .i mi kucli lo ka ki'u ce na jmina su voi cnino je cu zifre zmadu kei lo ka'e jai gau .akti be di'a bei fai la lojbo kibystu 08:22 < Zearen> .uo pofpo'u .i za'a mutce sai loka spofu 08:23 < zipcpi> pofpofpofpofpofpofpo'u 08:23 < Ilmen> to .a'o lo mi banklecekitauvoisu na jminandu dukse toi 08:23 < zipcpi> spo'o'o'o'o'o'ofu 08:24 < Ilmen> spirofu spofu 08:24 < zipcpi> coi menli 08:24 < Ilmen> ta'o nai zo'o nai ki'u ma na jmina lo cnino lo ka'e gasnu be lo nu di'a selfu 08:25 < zipcpi> doi ma 08:25 < Ilmen> doi kaidji be loka spusku .u'i 08:26 < zipcpi> mi na sanji lo drata selfu 08:26 < rlpowell> site's back up 08:26 < zipcpi> ui 08:26 < Ilmen> .i'o 08:26 < zipcpi> mupa cnino valsi ku oi 08:28 < Ilmen> doi ro da poi pu denpa .i lo kibystu ba'o xruti 08:30 < zipcpi> "DBD::Pg::db do failed: ERROR: null value in column "natlangwordid" violates not-null constraint 08:30 < zipcpi> DETAIL: Failing row contains (null, 66867, 0). at /srv/jbovlaste/current/dict/editdef.html line 171." 08:32 < zipcpi> ... having difficulty adding gloss words to my word cause of this error 08:32 < Ilmen> la jbovlaste su mikce cu nitcu 08:32 < zipcpi> u'i 08:33 < zipcpi> http://jbovlaste.lojban.org/dict/sumdai 08:34 < samcinki> basically {bu'a} 08:34 < zipcpi> Maybe... but I don't know what {bu'a} means when used metalinguistically :p 08:35 < Ilmen> bu'a = me'au da 08:35 < Ilmen> so basically "satisfy at least one predicate 08:35 < Ilmen> " 08:36 < Ilmen> which seems somewhat tautological 08:36 * zipcpi nods 08:37 < Ilmen> Reminds me of the Toaq Dzu predicate "x1 is anything" 08:37 < zipcpi> I guess that's why almost no one uses it lol 08:38 < zipcpi> http://jbovlaste.lojban.org/dict/iolpe Slate coming 08:39 < Ilmen> .u'i 08:39 < zipcpi> iol - unspecified marital status, ios - married (thought of a set for unmarried, but I don't know how useful it'd be [married set is for recognizing one as part of their own family], and I don't know what they should look like) 08:39 < zipcpi> na - nanmu, ni - ninmu, pe - prenu 08:40 < zipcpi> iocre is Dr/Prof 08:41 < Ilmen> je'e 08:41 < Ilmen> ni'o mi de'a jundi 08:53 < cliva> ki'e doi gasnu be lo nu la jbovlaste cu se pagbu lo xunterjo'e 08:54 < zipcpi> ie zabna 08:57 < cliva> (to ba'a ja'ai pei lo nu setca zo xunterjo'e la jbovlaste cu rinka lo vilti'a toi) 08:57 < zipcpi> vilti'a ma u'i 08:57 < durka42> lo xunterjo'e cu mo 08:58 < zipcpi> xunre samterjo'e 08:58 < cliva> redli'inki 08:58 < zipcpi> .y. kibyterjo'e 08:58 < durka42> ua 08:58 < durka42> mi gasnu .o'a 08:59 < cliva> .ei pu ba'e zu se pagbu .i ku'i zabna ja'a ku 09:00 < cliva> .i ta'o ru'e ga'a do ma ni lo ka fanva tu'a lo jbovlaste sazycimde lo jbobau cu nandu 09:02 < zipcpi> uo http://jbovlaste.lojban.org/dict/iocre 09:02 < durka42> ua 09:02 < durka42> .ai mi bazu .iocre 09:03 < ldlework> ua 09:04 < ldlework> yoshreh durka 09:05 < durka42> zipcpi: pei da terbri jecu srana lo se tadni 09:06 < zipcpi> xy'y .ie pilxau ka'e 09:06 < gleki> where is toaqzdu's dict. and how to use this "anything" brivla? 09:07 < zipcpi> $x_1$ is a Dr./Prof. (honorific for achieving a certain level of education or knowledge in field $x_2$), recognized by $x_3$ 09:07 < zipcpi> ? 09:08 < zipcpi> Not sure if it's better to put the more useful place in x2, or keep consistency with the rest of the slate 09:08 < cliva> gleki: https://docs.google.com/spreadsheets/d/1pyR0DkYVnjTWtmdibSHPKr-k7MQRMBerBzL7WfpKjkM/edit?pli=1#gid=0 (to laldo sai ku'i toi) 09:08 < gleki> "recognized by" - i dont think it's a popular approach these days due to incosistency of place structures. compare melbi2 09:09 < zipcpi> Yeah... 09:09 < gleki> we are moving into guaspi 09:09 < zipcpi> u'i 09:09 < gleki> ti itku'ile si melbi ga'a mi 09:10 < ldlework> gleki: I think recognized by is in an inherent aspect to doctorial status though 09:10 < gleki> cliva: ba'anairu'e i ki'e 09:10 < gleki> ldlework: it's inherent to many many concepts 09:10 < ldlework> Not really 09:10 < gleki> that lack this place 09:10 < ldlework> Certification is intrinsically defined by some external authority 09:11 < gleki> i just perceive this place as "by standard", only a bit narrower in semantic range 09:11 < ldlework> Mere association with some approving entity isn't the same 09:11 < gleki> ca'i 09:12 < ldlework> better remove cinmo3 09:12 < ldlework> since you have sera'a 09:12 < ldlework> I imagine many things will be deleted by sera'a 09:12 < ldlework> let's get started 09:12 < gleki> jb: cinmo 09:12 < mensi> cinmo = cinmo — x1(entity) feels emotion x2(property of x1) about x3(entity) 09:12 < mensi> :lo cinmo — emotional (live being). lo se cinmo — emotion, feeling (impression). 09:12 < mensi> :mi cinmo lo ka gleki vau lo nu mi'o simxu tavla bau la Lojban — I feel happy that we talk in Lojban. 09:12 < mensi> :Related words: cumla, jilra, nelci, xendo, ckeji, cortu, jgira, kecti, kufra, marvele, prami, steba, zungi, badri, 09:12 < mensi> burna, gleki 09:12 < gleki> no idea how to remove it 09:12 < ldlework> sera'a 09:13 < gleki> neither sera'a nor ne sera'a seem be working here 09:13 < ldlework> right. of course not. 09:13 < gleki> what should i attach sera'a to? 09:13 < ldlework> but ca'i is a justified replacement for the authority in the act of certification 09:13 < gleki> to cinmo2? 09:13 < ldlework> gleki: whatever you'd fill x3? 09:13 < ldlework> what? I explicitly said cinmo3 09:13 < ldlework> nevermind, gleki 09:13 < gleki> i mean how to use this sera'a instead of cinmo3 09:14 < gleki> 09:12 < ldlework> better remove cinmo3 09:14 < gleki> how? 09:14 < Zearen> I think having a place that is typically unfilled is the same is assuming it's zo'e and not zi'o. I.e. iocre typically has a governing or approving body assumed, and fi zi'o would seem odd. 09:14 < ldlework> in the same exact sense that you would remove iocre2 09:14 < ldlework> Zearen: exactly 09:14 < Zearen> Same thing with cinmo3, actually 09:14 < zipcpi> It's iocre3 now 09:14 < gleki> i ithink for iocre ma'i or ca'i is fine 09:15 < gleki> en: nolraitru 09:15 < mensi> nolraitru [< nobli traji turni ≈ Noble superlative govern] = t1=n1 is a regent/monarch of t2 by standard n2. |>>> Cf. 09:15 < ldlework> Zearen: sure, I only suggested cinmo3 to point out silly gleki's position on iocre is 09:15 < mensi> nobli, traji, turni, balnoltru, noltruni'u, noltrunau, noltroni'u, noltronau, nolse'u, nolkansa, gugja'a, truralju. |>>> 09:15 < mensi> arj 09:15 < ldlework> I don't think we should actually remove cinmo3 09:15 < gleki> en: nobli 09:15 < mensi> nobli = x1 is noble/aristocratic/elite/high-born/titled in/under culture/society/standard x2. |>>> Also upper-class; 09:15 < mensi> high/low, upper/lower are poor Lojban metaphors; note x2 standard applies when the title/nobility is not recognized 09:15 < mensi> culture/society wide; this would include self-assumed titles. See also banli. |>>> 09:15 < mensi> officialdata 09:15 < Zearen> I mean, you still can use those if you want. I use tavla bau ... instead of tavla fo all the time because it's more explicit. 09:15 < ldlework> Zearen: also it seems like its ben a looooooooong time since I've seen you 09:16 < Zearen> It has :p I've been working, burnt out, and now I'm feeling like a human again. 09:16 < ldlework> fi'i lo tolcliva 09:16 < ldlework> Zearen: I'm not sure how its more explicit though 09:17 < ldlework> It seems perfectly arbitrary and dependent on style rather than semantic 09:17 < ldlework> that said, I prefer a heavy tag style too 09:17 < zipcpi> Just easier for a reader to pick out without having to remember what {lo ve tavla} is 09:17 < ldlework> ie 09:17 < Zearen> Er, "explicit" is not the best word. It gives more hints to a human reader who may not have memorized that many places of tavla. 09:17 < Zearen> .ie 09:18 < ldlework> :) 09:18 < ldlework> Zearen: I'm glad to hear you're on an upswing. 09:19 < zipcpi> All the honorific slate has a recognition place; my reasoning is to be able to render things like "A gentleman always keeps his word" 09:19 < Zearen> .i doi cadgu'a xu do ca ka'e tavla bau la .lojban. po'o 09:19 < durka42> zipcpi: how's that use the x2 09:20 < zipcpi> Implicit zo'e 09:20 <@xalbo> Yeah, to me a degree always has a granting institution, that's an intrinsic part of what a degree *is*. If it's not granted by a particular institution, it isn't a degree, it's a piece of paper (and often, not even that). 09:20 < zipcpi> I mean I imagine liars and cheats get called "mister/sir" all the time, but in that sentence the speaker is expressing an opinion on what *he* considers to be worthy of that title 09:20 < durka42> je'e 09:21 < durka42> hmm so then we need "an elephant (status recognized by x2) never forgets" 09:21 < durka42> zo'o 09:21 < zipcpi> u'i 09:21 < ldlework> .i doi zy. mi ka'e go'i du'i tu'a lo'e jbopre 09:22 < zipcpi> u'i 09:22 < durka42> I agree with xalbo but if these words are just honorifics we don't need to bloat them with places. you can always use another selbri to ask "doctor of what according to whom?" 09:22 < ldlework> Why have multi-place predicates at all then? 09:22 < ldlework> Where does the valuation stand? 09:23 * durka42 cu binxo lo taksi 09:23 < Zearen> Well, the above could easily be «mu'i lodu'u da .iople kei, da ze'e nupre lo zasti» 09:23 < zipcpi> iolpe 09:23 < Zearen> Actually, to "keep ones word" is a little trickier to translate 09:24 < zipcpi> Hmm... yeah 09:24 < zipcpi> gasnu ro lo nupre be ri? 09:24 < zipcpi> Er se nupre 09:25 < Zearen> na go'i 09:25 < Zearen> upre = x1 (agent) promises/commits/assures/threatens x2 (event/state) to x3 [beneficiary/victim]. 09:25 < zipcpi> Yeah... it's an event, so usable for gasnu2 09:26 < Zearen> Oh, NVM, I missed your translation, I thought you were correcting mine. 09:27 < zipcpi> ua 09:27 < Zearen> If I could replace any place with a modal, I'd replace all these agent places with gau -_- 09:27 < Zearen> *especially* for kancu 09:28 < zipcpi> kancu... dunno though. Does gau mean *make* it that number, or does it just mean "count"? 09:28 < Zearen> Though I suppose a promiser is intrinsic in the relationship of a promise. 09:28 <@xalbo> You think? 09:29 < zipcpi> I think zilkancu might be better rendered as {meirlai} or something 09:29 < zipcpi> {memlai} 09:30 <@xalbo> meirmei 09:30 < zipcpi> That's usable too, just thought that {klani} is useful here 09:30 < Zearen> Hmm, that's a good question. a model like kancu fi'o facki da would be better. I just found the x1 place of kancu more of a hinderance than a help. 09:33 < ldlework> xalbo: lol 09:33 < ldlework> zipcpi: there's a whole chapter on promises in Speech Acts 09:33 < ldlework> Its the first illocutionary force he analyzes in full 09:33 < zipcpi> Right 09:34 < Zearen> Oh right, mei. I ta'e forget about «mei»... 09:34 < ldlework> Zearen: have you heard selpahi's full album Za'o? 09:35 < Zearen> na go'i 09:35 < ldlework> ko'oi sai 09:36 < ldlework> Zearen: also we have a 3D virtual lojbanistan now 09:36 < ldlework> If you ever want to hop on I can help get you setup 09:36 < Zearen> .ua.uo 09:36 < Zearen> .i go'i sai 09:37 < ldlework> camxes: la zearen ku 09:37 < camxes> SyntaxError: Expected [,] but "z" found. 09:37 < ldlework> Zearen: send me a lojban name and a password and I'll set up an account 09:37 < ldlework> Has anyone logged onto jbogu'e lately? Is it still running? 09:37 < ldlework> I was getting pretty bad lag the last couple times I logged in 09:37 < Zearen> la'o xy. zearen xy. 09:38 < Zearen> camxes: la'o xy. zearen xy. 09:38 < camxes> ([la'o xy xy] VAU) 09:38 < ldlework> Zearen: can't have such spaces and such, easier if you have a preferred lojbanification of your name 09:39 < ldlework> not even dots work :( 09:39 < Zearen> Oh, my lojbanified name is la .ziren. 09:39 < ldlework> okay pm me a password 09:40 < ldlework> Zearen: you'll need a client, everyone seems to like Singularity Viewer 09:40 < zipcpi> ... why did ilmen write an alternative definition for {zilkancu}? Looks like he misplaced something meant for another word 09:41 < Zearen> ge'e la'o gy. second life gy. co'e 09:41 < ldlework> doi ziren zo'ei do ca se zbasu 09:41 < ldlework> Zearen: yeah but its private, its not on actual secondlife 09:41 < ldlework> It runs on the opensource OpenSim 09:42 < ldlework> So it runs a lot nicer and its all about lojban 09:42 < zipcpi> Made {memlai} a synonym of {zilkancu} 09:42 < Zearen> .i lo mi pamoi nu jmive cu banzu loni nandu 09:42 < ldlework> but yeah same tech essentially 09:43 < ldlework> Zearen: that's fine. its for enjoying lojban in a spatial place so you can stretch your vocabulary legs 09:43 < ldlework> We have some neat lojban 'toys' in there too 09:44 < ldlework> I'm hoping to eventually encourage some structured roleplay too 09:44 < ldlework> And get some people to play lojban primtionary 09:45 < Zearen> .i je'e .i mi terpa tu'a la'o gy. Second Life gy. noi drata 09:45 < ldlework> Zearen: you can ask some of the others who have vistied whether or not it is terpa inducing 09:46 < zipcpi> It's not part of the Second Life thingy itself; just runs off the same engine 09:46 < ldlework> Its a small island, some lojban toys, a sky, and not much else. 09:46 < ldlework> Some hidden lojbo pendos 09:47 < ldlework> Zearen: in anycase your account is setup so if you get the client you can connect to http://sim.ldlework.com:9000 09:48 * ldlework cu sisku su'o ckafi 09:51 < Ilmen> "... why did ilmen write an alternative definition for {zilkancu}? Looks like he misplaced something meant for another word" -- What do you mean, Zipcpi? 09:52 < zipcpi> "x1 (plural) is such that x2 (number) of its members satisfy the property x3" is quite different to "x1 numbers/adds up to/consists of x2 (number) counting by units x3; x1 are x2 in number." 09:52 < ldlework> coi Ilmen 09:54 < Ilmen> zipcpi: Please elaborate. 09:54 < Ilmen> .e'o sa'ulnaisku 09:54 < zipcpi> units x3 is rather different from "satisfy the property x3", pe'i 09:55 < Ilmen> People used to say things like "lo xanto cu zilkancu li ci lo re xanto", which is weird to me 09:55 < zipcpi> I don't even know oll 09:56 < Ilmen> How would you use {zilkancu}, as you understand it from its definition? 09:56 < zipcpi> I dunno; I tend to ignore the x3 09:57 < zipcpi> 2 apples -> lo re plise -> lo plise poi zilkancu li re 09:57 < zipcpi> I have no idea what should go into x3 09:57 < zipcpi> But to me your definition could be {lo plise poi zilkancu li re lo ka xunre} 09:58 < Ilmen> How would you say "I'm counting the red apples in this pile of apples"? 09:58 < zipcpi> But you are actually pointing to a bunch of apples, 2 of which are red... 09:58 < zipcpi> kancu isn't defined by property either 09:59 < Ilmen> en: kanji 09:59 < mensi> kanji = x1 calculates/reckons/computes x2 [value (ni)/state] from data x3 by process x4. |>>> See also kancu, jdice, 09:59 < mensi> skami. |>>> officialdata 09:59 < zipcpi> mi kancu lo xunre poi cmima lei po'uti plise 10:00 < zipcpi> Or lei'i, jo'au gadganzu 10:01 < zipcpi> Yeah no idea what goes into x3, but I'd imagine it's important for those cases where a "unit" might be less well defined 10:02 < Ilmen> en: memkai 10:02 < mensi> memkai [< mei ckaji ≈ Cardinal selbri feature] = x1 is/are x2 (li) in number, and are among x3 |>>> This definition is 10:02 < mensi> based on the BPFK definition of mei. See also mei. |>>> Ilmen 10:02 < zipcpi> oi... and I just defined memlai 10:03 < zipcpi> Too many synonyms 10:05 < durka42> see? search before you reinvent words :) 10:05 < zipcpi> I will never learn my lesson 10:05 < ldlework> someone should kansa Zearen who is vitke la jbogu'e for the pamoi 10:05 < zipcpi> Never 10:06 < ldlework> akmnlrse: ki'e 10:08 < zipcpi> mi kancu lo xunre me le'i plise 10:10 < zipcpi> ldlework: ta'o {re'u} means "for the nth time" 10:13 < akmnlrse> ldlework: je'u simlu lo ka snomau lo jai purci 10:13 < gleki> en:snomau 10:13 < mensi> [< masno zmadu ≈ Slow more] lo nu mulno sisku zo'u: y no da se tolcri 10:13 < mensi> snomau[5847], snozma[5858], snozmadu[7897], masnymau[7997], masnyzma[8008], ... 10:16 < zipcpi> mi pu pensi lo snime remtai 10:17 < zipcpi> mi se malglixlu 10:17 < akmnlrse> xe'e si .u'i 10:17 < gleki> mi ca'o snada favgau lo zmiku be loka zbasu lo vlaste poi traduki fe lo glico 10:18 < akmnlrse> .i si'a si'au la vlaji'u ku poi ctuca fo lo canlu valsi cu spofu (to pa lo jvinu no roi canci toi) 10:19 < gleki> en: vlaji'u 10:19 < mensi> [< valsi jvinu ≈ Word view] lo nu mulno sisku zo'u: y no da se tolcri 10:19 < mensi> vlavin[5878], valvin[5898], vlaji'u[6367], valji'u[6387], vlajvinu[7897], ... 10:20 < gleki> akmnlrse: xu itku'ile 10:20 < akmnlrse> cmene lo pa dacti klesi ne la jbogu'e 10:20 < gleki> je'e 10:20 < zipcpi> u'i ma finti zo kresi'e 10:20 < akmnlrse> .i skicu lo lojbo jufra se pi'o lo virtuuale dacti 10:21 < akmnlrse> ... ma smuni .i ma srana lo kerfa gi'e snime 10:21 < zipcpi> dandruff 10:21 < akmnlrse> ua 10:21 < zipcpi> za'a la .totus. cu finti 10:23 < xunlohu> coi la gleki 10:23 < gleki> xunlohu: coi i xu do co'i zanru su'e mu lo preti pe la'o StackExchange 10:23 < gleki> sisi 10:23 < gleki> la'oi StackExchange 10:24 < xunlohu> go'i na 10:24 < xunlohu> .u'u 10:25 < gleki> xunlohu: xu do ba zanru 10:25 < xunlohu> Um 10:25 < xunlohu> I misunderstand 10:25 < xunlohu> Understood 10:25 < xunlohu> Sorry 10:25 < gleki> area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 10:26 < xunlohu> I'm not on stack exchange 10:26 < xunlohu> I did see it, I'll sign up when I have Wi-Fi 10:26 < gleki> just login and upvote five questions with rating below 10 10:26 < ldlework> akmnlrse: I can restart the whole server if you'd like 10:26 < gleki> Zearen: will you register there and upvote questions too? 10:27 < Zearen> I have. .i ji'a mi pu finti lo preti 10:27 < xunlohu> I'll do so, gleki. Like I said, I'll do it when I get Wi-Fi to sign up for it on 10:28 < gleki> Zearen: so im marking you as upvoted 5 questions, right? 10:28 < zipcpi> Hmm... looks like someone already defined {samki'o} for 1024 10:28 < Zearen> Teah. 10:28 < Zearen> *Yeah. 10:29 < gleki> ok 10:29 < Zearen> relki'o might be better. 1024 isn't only for computers :) 10:29 < zipcpi> Not sure about extensibility though 10:29 < akmnlrse> ldlework: na sarcu (to ba'a ba ze'a ku no da pilno (to ge'e toi) toi) 10:29 < zipcpi> True... though I came up with my own slate of zi'evla too 10:30 < ldlework> akmnlrse: yeah 10:30 < zipcpi> ki'orbi me'orbi gi'orbi te'orbi pe'orbi xe'orbi ze'orbi go'orbi 10:30 < zipcpi> And then using them to make informal "rafsi", kirbi- merbi- etc 10:30 < zipcpi> e.g. kirbibaitni, kirbibitri 10:31 < zipcpi> bitri will be a new gismu for "bit" 10:31 < gleki> like if there was an old one 10:32 < gleki> loglan: bithi 10:32 < mensi> bithi = B measures C in bit/binary digit (s) default 1. 10:32 < akmnlrse> 10.8) lo ratcu cu cmalu .iku'i lu'i ri barda 10:32 < akmnlrse> cu'u CLL .i xu ja'ai ca da drani fa di'u 10:33 < gleki> uanai 10:33 < durka42> zipcpi: what happened to bitsi? 10:33 < gleki> jbo: lu'i 10:33 < mensi> lu'i = [LAhE] sumgadri le se li'erla'i sumti le du'u sinxa le selcmi be ro ri .e no drata |>>> selcmi sumgadri; ce; la'i, 10:33 < mensi> le'i, lo'i |>>> xorxes 10:33 < durka42> vlaste: bitsi (fapro) 10:33 < vlaste> zo bitsi gimkamsmikezypro no catni gismu .i zo bitsi gimkamsmikezypro no cipra gismu 10:33 < durka42> vlaste: bitri (fapro) 10:33 < vlaste> zo bitri gimkamsmikezypro no catni gismu .i zo bitri gimkamsmikezypro no cipra gismu 10:34 < gleki> i suppose bitsi was used by me at some point 10:34 < zipcpi> durka42: gleki wanted it to be bitri, because he thought bitsi sounded too much like pizza lol 10:34 < durka42> fair 10:34 < durka42> though {bitri} is dyslexically close to {birti} 10:34 < zipcpi> u'i 10:34 < akmnlrse> da'i nibli lo du'u pu tu'a la xorlo lu'u ji'a sai la iolpe broda cu zasti 10:35 < gleki> and {bitni} to "beatnik" 10:36 < zipcpi> Erm, better decide before I enter 40+ words lol 10:36 * akmnlrse cu zmanei ru'e zo bitsi 10:36 < zipcpi> mi ji'a go'i 10:37 < gleki> the earliest mentioning of bitsi is from february 2015 by cirko. 10:37 < gleki> but by hat time it was already {bitri} in la bangu 10:37 < gleki> *tht 10:37 < akmnlrse> ue 10:37 < zipcpi> When I discovered it in la bangu it was bitsi 10:37 < akmnlrse> .i ba'a nai ru'e su'o lojbo proga cu se pagbu zo bitsi 10:38 < gleki> i cant see it in logs earlier! 10:38 < gleki> en: bitsi 10:38 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 10:38 < gleki> oh tatoeba probably 10:38 < akmnlrse> cumki 10:38 < akmnlrse> Tatoeba: bitsi 10:38 < mensi> no da se tolcri 10:39 < gleki> no 10:39 < gleki> okay someone erased our memory 10:39 < akmnlrse> da 10:40 < akmnlrse> se 10:40 < akmnlrse> tolcri 10:40 * zipcpi hides his MIB co'e 10:40 * akmnlrse to'o bajra soi se fanta lo nu Z di'a tolmipri le co'e 10:41 < zipcpi> u'i 10:41 <@xalbo> lo .itsi bitsi jukni cu cpare lo djacu tolrinci 10:42 < akmnlrse> na nabmi .i lo bitsi cu cmacma .i ja'o mapti sai 10:54 < Zearen> What's with this recent trend of using reduplication for 'tce' ? 10:55 < zipcpi> I've only seen it with brabra, cmacma 10:55 < zipcpi> And perhaps tcetce, but that's justifiable 10:55 < zipcpi> And I think it's as old as dirt 10:56 < Zearen> Yeah, tcetce makes sense. But also, pofpo'u earlier. 10:56 < zipcpi> Oh right lol 10:56 < zipcpi> I think that wasn't serious 10:56 <@Broca> It's not recent; I'm pretty sure it was common when I discovered Lojban 18 years ago :-) 10:56 < gleki> it's strange. 10:56 < gleki> en: cmacma 10:56 < mensi> cmacma [< cmalu cmalu ≈ Small small] = c1 is tiny/miniature/diminutive/very small in property c2 with criterion c3. |>>> 10:56 < mensi> Cf. cmalu, cmalu, cmacai, ruble, toltce; antonyms: brabra, tcebra. |>>> brtais 10:56 < akmnlrse> s/cmacma/cma'alu 10:56 < fenki> akmnlrse meant to say: na nabmi .i lo bitsi cu cma'alu .i ja'o mapti sai 10:56 < gleki> how is it gonna expand? 10:56 < Zearen> It's just interesting to see Lojban evolve lije a natlang would 10:56 < akmnlrse> xu mabla mleca 10:57 < zipcpi> I dunno... I guess if you stretch it... ko'a barda lo ka barda lo ka co'e 10:57 < gleki> cmalu lo ka cmalu? then it's not tiny at all 10:57 < zipcpi> lol 10:58 < zipcpi> barda je barda 10:58 < Zearen> I mean, I suppose small-type-of-small makes sense. 10:58 < zipcpi> cmalu je cmalu 10:58 < gleki> although probably cmalu is about physical size only 10:58 < Zearen> I.e. the tanru 10:59 < zipcpi> Or maybe barda xoi zi'o karbi ke'a lo barda 10:59 < gleki> jb: tcemi 10:59 < akmnlrse> barda ma'i lo du'u lo barda cu cmalu 11:00 < zipcpi> Oh wait barda has a fi place 11:00 < zipcpi> barda fi lo barda 11:00 < zipcpi> barda fi lo'e barda 11:01 < gleki> jb: tcemi 11:01 < mensi> tcemi = tcemi — x1(entity) is compared with x2(entity) 11:01 < mensi> :mi poi tcemi do barda — I am big compared to you. 11:01 < mensi> :Related words: zmadu 11:01 < gleki> i stole that from ceqli 11:01 < zipcpi> ku'i se ni'udzau 11:01 < gleki> a truncated karbi 11:02 < gleki> im reluctant to stealing anything unless people need words 11:02 < gleki> so im not going to import the whole toaqdzu into this dict. 11:03 < zipcpi> zo tcemi ca dzaunlai li ni'upa 11:06 < gleki> oh i almost finished creating a script of eng2jbo dictionary 11:07 < gleki> zipcpi: ua i ja'o so'i da mutce xebni le valsi 11:08 < zipcpi> je'u 11:08 < durka42> bu'u la jbovlaste lo remei cu so'imei iau uinai 11:09 < durka42> sei mi stace mi ji'a na nelci 11:09 < mensi> ji'a mi mutce nelci i ie 11:09 < durka42> ko mrobi'o bu'u lo fagri doi zmiku 11:10 < zipcpi> u'i 11:10 < gleki> btw Japanese FrameNet added new frames that ... English FrameNet needs too! 11:10 < gleki> but some are interesting 11:11 < gleki> e.g. "to be distributed",i.e. non agentive "to distribute" 11:11 < durka42> .au lo do besna lo terdi cu me'au zoi gy to be distributed gy 11:12 < durka42> zo'o 11:15 < gleki> http://www.lrec-conf.org/proceedings/lrec2012/pdf/1053_Paper.pdf 11:15 < gleki> The intransitive 11:15 < gleki> verb tiru ‘get scattered’ in (1a) is an inchoative verb used 11:15 < gleki> to describe particles or small objects falling. It is difficult 11:15 < gleki> to find a semantic frame in the current FN database with 11:15 < gleki> the meaning. As for the morphologically-related transitive 11:15 < gleki> counterpart tirasu ‘scatter’ in (1b), on the other hand, we 11:15 < gleki> assume that the Dispersal frame is involved, since the 11:15 < gleki> corresponding English verb scatter evokes the 11:15 < gleki> Dispersal frame (defined in FN as “an AGENT or a 11:15 < gleki> CAUSE disperses or scatters INDIVIDUALS from the 11:15 < gleki> SOURCE, a relatively confined space, to a the GOAL_AREA, 11:15 < gleki> a broader space”). 11:15 < gleki> en: fatri 11:15 < mensi> fatri = x1 is distributed/allotted/allocated/shared among x2 with shares/portions x3; (x2/x3 fa'u). |>>> Also spread, 11:15 < mensi> shared out, apportioned; agentive distribution (= fairgau, fairzu'e). See also cmavo list fa'u, fendi, preja, katna, 11:15 < mensi> tcana. |>>> officialdata 11:22 < zipcpi> xu lau'e zbusufukai zo'u: bau la tcekitau iau ca'e zo cei simbasti zo cu'i 11:22 < zipcpi> exp: xu ti zbusufukai zo'u broda 11:22 < mensi> SyntaxError: Expected [,] but "b" found. 11:23 < zipcpi> oi zo zo'u naka'e co'e lo bridi 11:23 < zipcpi> xu lau'e zbusufukai tu'e 11:23 < durka42> mi stidi zo .i 11:24 < zipcpi> ie ka'e ji'a 11:24 < niftg> sumsmi porsi sepli bo cmavo vau ju'o cu'i 11:25 < niftg> xu da za'e brismi .a'u cu'i vau ta'o ju'i cu'i 11:27 * nuzba @464161niftg: za'e kairji'i lo jbobau lo ka sruri pandi bangu [http://bit.ly/1gd6GqY] 11:28 < durka42> lo jbobau cu me lai .lisp. xu 11:28 < zipcpi> simlu lo ka go'i iau u'i 11:29 < zipcpi> .y. me'ei go'i 11:30 < zipcpi> lu zi'o simlu li'u ju'ocu'i oi 11:31 < akmnlrse> zo jetmlu ka'e basti 11:31 < zipcpi> ie 11:32 < zipcpi> doi cirko do'u pei go'oi cei 11:32 < niftg> .ua zo jbobau ji'a jai se sisku la nuzba 11:33 < zipcpi> mi ja'a za'o pilno zo cei .iku'i la'a na pavyslakycmavy'inda 11:33 < akmnlrse> zo cei zo'u mi na pante 11:33 < akmnlrse> doi lo zipcpi 11:33 < zipcpi> zo lo .i mi ja'a zipcpi xoi se ca'engau 11:34 < akmnlrse> verdikala ja'a~~~~~i 11:34 < zipcpi> u'i 11:34 * akmnlrse bregause'i lo ka se darxi lo jai crori'a 11:34 < ldlework> oi mi cortu lo mi birka 11:35 < akmnlrse> coi zi'o 11:35 < ldlework> heh I like how the first result for searching for "stress" is "krtis" 11:35 < akmnlrse> .u'i 11:37 < Zearen> .i .u'i ta'o lo mi danlu bambola cu se cmene o zipcpi .imu'ibo mi se cfipu lonu zo'e cusku le cmene 11:37 < durka42> vlaste: krtis (n) 11:37 < vlaste> krtis (notes) = Perhaps unfortunately, the stress of the Lojbanized name falls on the final syllable ("tis"); this version enunciates the /t/, rather than reducing it to a tap. 11:37 < durka42> hmm 11:37 < durka42> is that true? 11:37 < zipcpi> mi ji'a ponse lo zipcpi bambola 11:37 < durka42> the rules are in conflict 11:38 < akmnlrse> je'u mi ta'e bacru do'e zo'oi kr,tìs 11:38 < durka42> you should stress the second-to-last syllable, but you can't stress a syllabic r 11:38 < durka42> je'e 11:38 < durka42> ai mi ningau lo mi slaka samtutci 11:38 < durka42> ji'a zo .kyrtis. da'i mo 11:38 < zipcpi> mu'i ma paunai na pilno zo .kyrtis. 11:39 < akmnlrse> pe'i dunli (to ku'i .ei lo drata cu spuda toi) 11:39 < zipcpi> pe'i ra'a lo'e cmevla le se basna slaka na rinju 11:40 < durka42> ie 11:41 < durka42> $ ./camxes.py -t syllable krtis 11:41 < durka42> "kr,TIS" 11:42 < akmnlrse> ba zu ma lo do tutci co'a gubni 11:44 * nuzba @uitki: ce ki tau jau - http://mw.lojban.org/papri/ce_ki_tau_jau by Spheniscine [http://bit.ly/1GRocaI] 11:45 < durka42> la lujvo cu jimca lo se .urli be la'oi github.com/durka/camxes-py 11:45 < zipcpi> exp: la'oi github.com/durka/camxes-py 11:45 < mensi> ([{la'oi ~} {py BOI}] VAU) 11:45 < zipcpi> .y. oi 11:45 < durka42> .... 11:45 < zipcpi> za'a mi pofygau da 11:45 < durka42> exp: mi-do-xebni 11:45 < mensi> ([mi do] [CU {xebni VAU}]) 11:45 < durka42> ua 11:45 < durka42> did you spoff the preprocessor 11:46 < ldlework> u'i 11:46 < zipcpi> I just brought up to Gleki of an idea to make the hyphen an allograph for the space 11:46 < zipcpi> So that I can connect strongly related brivla with cmavo, e.g. tsali-sai 11:46 < durka42> well, it's not like la'oi is formalizable anyway 11:47 < durka42> exp: la'oi .mi-do. 11:47 < mensi> ([{la'oi ~} do] VAU) 11:47 < zipcpi> I think la'oi should ignore my zbusufukai reform though 11:47 < zipcpi> I was actually testing for the period, not the hyphen 11:48 < durka42> well, AFAIK the la'oi parsing is done by the parser 11:48 < durka42> but the special character replacements are done by regexes before the parser is even run 11:48 < durka42> exp: mi?do?xebni 11:48 < mensi> ([mi do] [CU {xebni VAU}]) 11:48 < durka42> exp: la'oi brok?en 11:48 < mensi> ([la'oi ~] VAU) 11:48 < zipcpi> Hm 11:48 < durka42> huh, maybe not 11:48 < durka42> exp: la'oi brok-en 11:48 < mensi> ([la'oi ~] [CU {en VAU}]) 11:48 < zipcpi> exp: mi.do.xebni 11:48 < mensi> ([mi do] [CU {xebni VAU}]) 11:48 < durka42> why are those different 11:49 < zipcpi> exp: la'oi mi.do.xebni 11:49 < mensi> ([la'oi ~] VAU) 11:51 < durka42> dots are definitely handled by the grammar itself 11:51 < zipcpi> Yeah so "all we have to do" is make the hyphen be treated like the dot or the question mark 11:52 < zipcpi> Nothing could go wrong, right? 11:53 < zipcpi> to mi sefsi dapsku toi 11:53 < gleki> i dont even know whether la'oi should eat dots or not. i consider it an edge case 11:54 < durka42> vlaste: la'oi 11:54 < vlaste> la'oi = single-word non-Lojban name; quotes a single non-Lojban word delimited by pauses and treats it as a name 11:54 < durka42> xm 11:54 < durka42> I thought it said spaces in writing 11:54 < durka42> vlaste: zo'oi 11:54 < vlaste> zo'oi = quote next non-Lojban word only; quotes a single non-Lojban word delimited by pauses (in speech) or whitespace (in writing) 11:54 < durka42> yeah 11:54 < durka42> they are the same selma'o 11:54 < durka42> so... 11:55 < gleki> so what is the solution to detri system? 11:55 < gleki> li cy xi pa ce li my xi reno? 11:55 < durka42> zipcpi's lerfu :) 11:55 < zipcpi> Erm, right now my solution is li'ei 11:55 < gleki> you mean mixed number/lerfu? 11:55 < zipcpi> But, experimental selma'o and all that 11:55 < gleki> li'ei is superstrange. one cmavo only to handle detri1? 11:56 < durka42> not just for detri1 11:56 < zipcpi> No. it's also useful for addresses, passwords, etc 11:56 < zipcpi> license plates 11:56 < zipcpi> Anything where there are arbitrary mixed letters and numbers 11:56 < gleki> i dont even understand what this xorxe's change would solve. 11:56 < durka42> exactly 11:56 < durka42> well actually it's not hard to understand 11:57 < gleki> still {re boi remoi} 11:57 < durka42> it would solve {abu za'u re'u co'e} and that's it 11:57 < zipcpi> Anyway li'ei can actually be easily implemented, at least if we want to do it quickly. Just make it shorthand for {li pe'o joi'i} 11:57 < gleki> {abu za'u re'u co'e} is a imstake similar to {lo mlatu pinxe} 11:58 < zipcpi> Well, from my point of view, the problem is that it's hard to remember both that za'u is grammatically identical to pa, and that it attaches to a letter before it 11:58 < zipcpi> You see I actually support his proposal, except where it messes with LI...LOhO T.T 11:59 < gleki> looks like i need to rework my brivla system then 11:59 < zipcpi> Huh? Don't see how it affects brivla 11:59 < gleki> because it's impossible to finish CC then 11:59 < gleki> your endless rants 11:59 < gleki> they damage the language 11:59 < ldlework> What does attaching a letteral infront of za'u do? 12:00 < gleki> and make CC unfinishable 12:00 < durka42> camxes: abu za'u re'u co'e 12:00 < camxes> SyntaxError: Expected [oO] but "e" found. 12:00 < durka42> er 12:00 < zipcpi> My rants? I don't think I said anything about brivla 12:00 < durka42> ldlework: see, {abu za'u} combines into one lerfu string 12:00 < durka42> so it breaks 12:00 < durka42> camxes: abu boi za'u re'u co'e 12:00 < camxes> ([{ boi} CU] [{za'u re'u} co'e] VAU) 12:00 < gleki> yours and others' rants about how lerfu should work 12:00 < durka42> exp: abu za'u re'u co'e 12:00 < mensi> ([{a bu} BOI] [CU { co'e} VAU]) 12:00 < zipcpi> gleki: I didn't start the fire 12:01 < gleki> i dont care who started 12:01 < gleki> wat if tomorrow i propose a brivla system and xorxes says "i think brivla should be killed" 12:01 < ldlework> what 12:02 < zipcpi> ... I guess you're just not in the mood then... 12:02 < gleki> i cant finish CC because the language is changing 12:02 < durka42> ldlework: remember that lerfu strings can contain numbers, i.e. {li ry re dy re} 12:02 < durka42> ldlework: the proposed reform is to change that, so they wouldn't combine 12:03 < durka42> gleki: stop being a drama queen .u'i 12:03 < durka42> all languages change 12:03 < dutchie> some faster than others 12:04 < zipcpi> I was against the part that made {li'ei}, {xi}, or other solutions necessary 12:04 < akmnlrse> camxes: za'u re'u co'e 12:04 < camxes> ([{za'u re'u} co'e] VAU) 12:04 < gleki> what was the vote then? 12:04 < zipcpi> I'm just sticking with {li'ei} because it's the "safest" solution as I see it 12:04 < ldlework> Usually the constructed ones that are incomplete who have communities full of linguists, logicians and language philosophers. 12:04 < durka42> I didn't see any vote yet 12:05 < zipcpi> And I'm not sure that {boi} thing, i.e. the actual useful part of xornunsep, would affect your materials that much 12:06 * ldlework ta'o jinga lo na srana ui sai 12:06 < ldlework> uo o'u 12:06 < zipcpi> Also I favor {li'ei} as a compromise because it only adds one syllable, and can be used for any arbitrary charstring 12:07 < zipcpi> No need to patch things together with xi, ce, or joi'i 12:08 < ldlework> I have just solved an insanely difficult systems engineering problem that has been plaguing my team for months 12:08 < zipcpi> So yeah I don't see how I'm to blame here 12:08 < ldlework> I can't wait until they get back from lunch 12:15 < zipcpi> Bah... the only reason that xornunsep wants to extend to LI... LOhO too is because of that darn mekso grammar, la'a 12:15 < ldlework> Complain to all the mathematicians using lojban 12:16 < zipcpi> ldlework: Oh, but even the mathematicians disagree; there's a page with a long rant saying that mekso should be ditched 12:16 < zipcpi> In favor of predicates 12:17 < zipcpi> http://teddyb.org/robin/tiki-index.php?page=Lojban,+Math,+mekso,+and+bridi+cmaci 12:21 < zipcpi> I mean, the mekso grammar is about as hard to learn as the rest of the language combined, it uses a bunch of unfamiliar cmavo, which represents acres of cmavo space... 12:21 < ldlework> No complaints here 12:21 < ldlework> Lojbanists should use the language of the field when working in the field. 12:22 < ldlework> No other language requries an "in-language" grammar for specialized fields. 12:22 < zipcpi> Maybe Ithkuil lol 12:22 < ldlework> Ithkuil has a mesko grammar too? 12:22 < zipcpi> Last I heard 12:22 < zipcpi> I don't know the details 12:23 < ldlework> I have to wonder how mesko falls under the purview of the language's goals. 12:23 < ldlework> Lojban that is. 12:23 < ldlework> Since mathematical forms are neither ambiguous or having anything to do with language in the first place. 12:23 < zipcpi> "about as hard to learn as the rest of the language combined" Oh and with less documentation too 12:24 < ldlework> And usage. 12:24 < zipcpi> Mhm 12:24 < ldlework> And active certus. 12:33 < Zearen> .i mi citka lo keisdiia noi kukte 12:34 < zipcpi> "mathematical forms are neither ambiguous" Erm, technically they are. "f(x)" can mean either the function f over x, or f*x 12:34 < zipcpi> Not to mention many, many letters are overloaded 12:35 < Zearen> Well, it actually fills a very important need: 12:35 < zipcpi> zo keisdiia ki'a 12:36 < Zearen> Turning written math into speech. There are no languages that do this unambiguously 12:36 < zipcpi> Predicates can do that too 12:38 < zipcpi> li pano sumji li ny le ni'ai ke'a pilji li re li pai li xy :: "10 = n + (2 * pi * x)" 12:38 < zipcpi> ni'ai is just mathematical poi'i 12:38 < zipcpi> Makes it clear you expect a number 12:39 < Zearen> fu'ivla fo «zoi sy. quesadilla sy. » lo spano bangu 12:39 < zipcpi> ua 12:41 < Ilmen> ckaji la .kesadiĭas. 12:41 < Ilmen> -sd- isn't a permitted cluster 12:42 < Zearen> I don't think mekso is the best system for math, granted. I also think it would be better served as a series of predicates, but it definitely fits in line with the languages goals. 12:42 < Zearen> Ah, right. Is kesadiĭas vlatai ? 12:43 < Zearen> camxes lo kesadiias cu co'e 12:43 < durka42> ke sa di ias 12:44 < Zearen> ^ that's what I thought, and why I dropped the a. (I forgot about voiced/unvoiced) 12:45 < Zearen> camxes: lo cidjrkeisadiia 12:45 < camxes> ([lo cidjrkeisadiia KU] VAU) 12:45 < Zearen> Oh, nvm, cmevla, I see. 12:47 < Zearen> la jbobau cu pluja tcetce 12:48 < ldlework> Why is being able to speak math part of the language's goals? 12:48 < ldlework> It just seems like a non-sequitur 12:48 < ldlework> Why not a grammar for speaking written music unambiguously? 12:50 < Zearen> Because you could just sign it, or some such. Mathematics in commonly spoken in most modern languages of advanced cultures 12:51 < Zearen> s/sign/sing/ 12:51 < fenki> Zearen meant to say: Because you could just sing it, or some such. Mathematics in commonly spoken in most modern languages of advanced cultures 12:51 < Zearen> .ie 12:52 < ldlework> Zearen: you can sing a tune so accurately as to unambiguously transmit to me the information required to perform it or transcribe it? 12:52 < durka42> I mean it's one of the goals because it was one of the goals :p 12:52 < ldlework> Mathematics is commonly -described- in most modern languages 12:53 < ldlework> durka42: I meant why does it fall under the purview of the more general and fundamental goals 12:53 < ldlework> but u'i 12:53 < durka42> ie 12:53 < Zearen> Me ? No. Someone who needed to ? Maybe 12:53 < zipcpi> in b4 bunch of cmavo to represent this unambiguously: https://en.wikipedia.org/wiki/Juggling_notation 12:53 < ldlework> zipcpi: indeed 12:53 < ldlework> zipcpi: you just _juggle_ the pattern 12:53 < ldlework> that's how you communicate the pattern to me 12:55 < zipcpi> Though I have somewhat wondered about "private use" cmavo space... or perhaps specialized lo'u ... le'u 12:55 < durka42> zipcpi: that notation is just digit strings :p 12:55 < durka42> ldlework: then you have to be able to juggle to talk about juggling 12:55 < ldlework> durka42: but its ambiguous that thos digit strings represent a juggling pattern 12:55 < ldlework> we need specific cmavo for it 12:55 < zipcpi> durka42: Yay lerfu tcita detri 12:56 < durka42> oh I misinterpreted 12:56 < durka42> that's only one cmavo, not a bunch though 12:56 < durka42> vlaste: juggle 12:56 < vlaste> no results. http://vlasisku.alexburka.com/juggle 12:56 < Zearen> Well, note above I think that mekso is overly complex and would be better represented lojbanically as a collection of... GOhA ? or whatever the list for cmavo shaped brivla is. 12:56 < durka42> boo 12:56 < zipcpi> No just make a new magic brivla in the vein of detri 12:56 < durka42> za'e malbarizmo 12:57 < zipcpi> x1 (li'ei string) describes juggling pattern x2 under string interpretation system x3 12:57 < durka42> fu'ivla fu lo spano 12:57 < zipcpi> Done 12:57 < ldlework> u'i 12:57 < ldlework> Same thing can be used for any encoding system 12:58 < Zearen> But we should certainly be able to speak/describe/communicate math in Lojban. Since it's Lojban, it should also be able to be done unambiguously. 12:58 < durka42> better to just say "li 414414414 mifra fi lo malbarizmo ciste" 12:58 < ldlework> Zearen: I do agree that if we went to a predicate system it shouldn't be trivialized, math is important of course. 12:58 < zipcpi> Zearen: That page already describes how to do it with predicates 12:58 < zipcpi> "How do Lojbanists do it?" <.< 12:59 < Zearen> Oh, I just saw "How can mekso be one of Lojban's goals" and I misunderstood. 13:00 < zipcpi> http://teddyb.org/robin/tiki-index.php?page=Lojban,+Math,+mekso,+and+bridi+cmaci 13:00 < durka42> lojbanists do it with a lot of unfilled places 13:00 < Zearen> I am in strong agreement that a whole new system of cmavo is less than useful. 13:00 < zipcpi> Lojbanists do it with broken light. 13:00 < Zearen> .u'i 13:01 < Zearen> The times I've felt like I wanted to arbitrarily use prefix notation, postfix notation, convert a predicate into a counter, etc. is basically never. 13:02 < zipcpi> There's no whole new system of cmavo on that page... I saw two. There were a few more but I'm not sure how they're supposed to be used yet 13:02 < zipcpi> And one of those two have some general usage (although right now most people use xoi for that kind of function) 13:03 < Zearen> Especially since you can just PA SELBRI. 13:03 < zipcpi> And that replaces all the bunch of cmavo for mekso 13:03 < Zearen> I was talking about the existing mekso cmavo :p 13:03 < zipcpi> Right 13:04 < ghhtzt> lojbanists do it if and only if they do it, under conditions x4 13:04 < zipcpi> with epistemology x5 13:05 < dutchie> u'isai 13:05 * ctefaho cringe cringe 13:05 < Zearen> The only things I might add as a mekso system is some stuff to make dealing with lambdas (via «ka»), e.g. a way to do folds and zips. 13:06 < dutchie> that sounds more like a computer science system than a maths one 13:06 < dutchie> the line is blurred of course 13:06 < Zearen> Then you could define su'i as x1 is the sum of x2 and x3 13:08 < zipcpi> We already have sumji 13:08 < Zearen> And define foldrai = x1 is the result of folding the binary property x2 over list/sequence x3 13:09 < zipcpi> I have no idea what that does, but that's exactly how great this is; just define predicates for what you need them to mean 13:09 < zipcpi> It's like OOP 13:09 < Zearen> dutchie: It probably is, but I think that's more useful here. 13:10 < Zearen> Or lambda calculus/functional programming. 13:10 < zipcpi> Mhm 13:10 < durka42> ie, foldrai would be nice :) 13:10 < Zearen> Oh. And a map 13:10 < durka42> then we could have a consistent definition for {sujysi'u} 13:11 < zipcpi> And hey you can even use my new rafsi -zam-... though cmacyzam is a bit long... 13:11 < durka42> when I complained that {sujysi'u} is wrong, xorxes said that {simxu} should be vague :/ 13:11 < durka42> ie nai sai 13:11 < Zearen> This has an equal expressive power of mekso, if not more, and doesn't introduce new concepts. You can use the Lojban you already know. 13:12 < durka42> ie 13:13 < Zearen> 13:14 < ghhtzt> hi, can someone teach me ithkuil 13:14 < dutchie> i'm not even sure if the annoyingly pedantic logical answer is yes zo'oru'e 13:14 < Zearen> .u'i 13:15 < durka42> ko cpedu fi la selpa'i 13:15 < durka42> la selpa'i cu .samyuan. 13:16 < zipcpi> So I still don't know whether I should use {bitri} or {bitsi} 13:16 < durka42> wait, in the future we might need a gismu for B-Trees 13:16 < durka42> zo'o 13:16 < zipcpi> lol 13:17 < Zearen> .i za'a lonu se .irci la'oi #lojban cu krinu lonu ctuca fi la .itku'il bau la .lojban. 13:17 < zipcpi> vlaste: bitni (gimka) 13:17 < vlaste> zo bitni gimkamsmikezypro no catni gismu .i zo bitni gimkamsmikezypro no cipra gismu 13:18 < zipcpi> gleki convinced me to change zbaiti to baitni 13:18 < zipcpi> But... 13:18 < ghhtzt> eho xu do ctuca mi la itkuhile 13:19 < dutchie> pe'i do nitcu zo tu'a 13:19 < ghhtzt> dojiha nitcu bahe tuha zo tuha 13:20 < durka42> .u'i 13:20 < zipcpi> Actually I think nitcu-2 doesn't need to be an abstraction 13:20 < dutchie> mi jinvi lodu'u zo nitcu na nitcu zo tu'a .iku'ibo mi kakne lonu srera 13:20 < Zearen> skamrbytri = x1 is a node of a B-tree with child list/sequence x2 with minimum number of children x3, maximum x4 and comparison function x5 13:21 < zipcpi> That doesn't work 13:21 < zipcpi> At least under current doctrine 13:21 < dutchie> my sentence or skamrbytri? 13:21 < durka42> no -y- in fu'ivla 13:21 < zipcpi> skamrbytri 13:21 < ghhtzt> zo itkuhile cmene loduhu lo itkuhilu cu mokau ijaho na sarcu fatuha zo tuha 13:21 < zipcpi> Also I'm allergic to type-3-fu'ivla lol 13:22 < durka42> sambiti 13:22 < zipcpi> k: sambitricu 13:22 < mensi> (CU [Z:sambitricu VAU]) 13:22 < Zearen> .i lujvo 13:22 < zipcpi> k: sambiti 13:22 < mensi> (CU [Z:sambiti VAU]) 13:22 < zipcpi> na lujvo 13:23 < dutchie> perhaps the word for byte should refer to eight in some way 13:23 < Zearen> camxes: lo samby.zeitricu 13:23 < camxes> SyntaxError: Expected [,] but "s" found. 13:23 < dutchie> bivka'u or something? 13:23 < ghhtzt> xahozasti 13:23 < zipcpi> dutchie: Well the first candidate was bivbitsi, but that ended up long when I wanted to attach prefixes to it 13:23 < Zearen> .i'e 13:24 < zipcpi> I mean {bivbi} is open, but xu gimy'inda 13:25 < dutchie> {bivjboka'u} even perhaps 13:25 < zipcpi> <.< 13:25 < dutchie> er, not jbo 13:25 < dutchie> whatever the rafsi for lojbo is 13:25 < dutchie> er, not that either 13:25 < zipcpi> It's already defined as bivjetka'u, but I'm not using that for my slate 13:25 < zipcpi> What's more it only has one place 13:26 < dutchie> what other places do you want? 13:26 < Zearen> logji = loj 13:26 < zipcpi> Well, match all the other unit brivla 13:26 < dutchie> oh right 13:26 < zipcpi> x1 is measured as x2 units 13:27 < ghhtzt> tihe zahu 5000 da itkuhile rafsi 13:27 < ctefaho> lojbo'y rafsi is best rafsi 13:27 < zipcpi> <.< 13:27 < ghhtzt> ijaho kona pante doilojbo 13:27 * ctefaho ducks 13:27 < zipcpi> I don't see why you want that when we already have four-letter rafsi 13:27 < ctefaho> I speak STT jbobau 13:27 < Zearen> In the great traditions of language, Lojban is the English name, and jbobau is the Lojban name 13:28 < ctefaho> I kinda nuked the 4-letter gismu rafsi from it 13:28 < Zearen> STT jbobau ? 13:28 < zipcpi> Yeah if STT jbobau is gonna be even more syllabic than official one, then leave me out of it 13:28 < ctefaho> SumTciTa'a 13:28 < ctefaho> it won't 13:28 < zipcpi> It will with all your broda'y nonsense 13:28 < ctefaho> I didn't say I nuked all rafsi 13:29 < ctefaho> that's Logbanu 13:29 < ctefaho> ...where rafsi kinda still exist as affixes 13:29 < zipcpi> Yes, just the ones that are the last resort of those poor rafsi-less gismu 13:29 < Zearen> argument tool writing ? 13:30 < ctefaho> Zearen: it is a dialect/fork/whatever you want to call it I am doing 13:30 < zipcpi> I don't see why you want to get rid of four-letter rafsi 13:30 < ghhtzt> i cannot find "rafsi" in the ithkuil ductionary. whats it mean 13:30 < ctefaho> zipcpi: think about it for more than 10 seconds 13:30 < zipcpi> ... still not seeing it 13:30 < ctefaho> keep thinking 13:31 < zipcpi> No, I've had years to think it through 13:32 < ctefaho> think harder 13:32 < ctefaho> harder better faster stronger 13:32 < zipcpi> This isn't going to work 13:32 < Zearen> Because when the join using an -y- hyphen, they're the same number of sounds anyway ? 13:32 < Zearen> *thry 13:32 * Zearen is not a very good typist 13:32 < zipcpi> No, gismy is two syllables, gismu'y is three 13:33 < Zearen> Wait, what ? 13:33 < Zearen> Why would you do that ? 13:33 < zipcpi> ctefaho: Is this really how you win an argument? Try to get me to read your mind? 13:33 < ctefaho> Zearen: thank the isomorphic morphology for that 13:33 < ghhtzt> doesnt it fall apart as gi smuhy 13:33 < ctefaho> zipcpi: you seem like a lost cause and I am not trying to win anything:) 13:34 < zipcpi> Then stop playing these stupid games 13:34 < ctefaho> I love playing games 13:34 < Zearen> That's... not a very nice thing to say. 13:34 < zipcpi> I'll mute you 13:34 < ctefaho> Zearen: he kinda endlessly complains about any ideas I have 13:35 < Zearen> Well, games are only fun if you have people to play them with. Hence, it's counter productive to irritate your fellow players 13:35 < ctefaho> I am just pulling zipcpi's leg:) 13:35 < zipcpi> Zearen: You see how he tried to lure me into some mind game; saying that I'd see how wonderful his idea is if only I "thought about it" 13:36 < Zearen> I feel like I'm getting stuck in the middle of something stupid, so I'm going to ignore it. 13:36 < ghhtzt> bring forth arguments or dont waste each others time. now pls teach me. i dont understand ithkuils verbs 13:36 < zipcpi> I can't teach you Ithkuil 13:37 < ctefaho> why are you asking about ithkuil in a lojban channel? 13:37 < ghhtzt> i thought this was the right place. 13:37 < ctefaho> well if selpa'i was here maybe he could help, but I know of no other jbòpre doing ithkuil 13:39 < ghhtzt> ok sry 13:39 < ctefaho> What made you think Ithkuil and Lojban were similar? they are quite different to my knowledge 13:40 < ctefaho> no problem. someone might know 13:40 < ghhtzt> they are both made up and weird 13:40 < ctefaho> .u'i ja'a 13:40 < zipcpi> ctefaho: Thanks for letting me know you are adding nothing productive to the conversation. To the mute you go 13:41 < ctefaho> calm down man 13:41 < ghhtzt> and lojban is too easy. i need a challenge. someone said i should try ithkuil 13:41 < zipcpi> I don't suffer trolls 13:41 < Zearen> .i drata fi lo xokau mei poi tavla 13:41 < zipcpi> So shut up forever 13:42 < ghhtzt> frica filoka xokaumei fa lo se bangu becehu 13:42 * ctefaho sighs 13:42 < zipcpi> Muted 13:42 < Ilmen> ŭi nai nu to'e panpi 13:42 < ctefaho> coi Ilmen 13:42 < Zearen> I don't know, Lojban has a lot of depth and complexity. itkuil is only more difficult because it's more obscure, so has less resources. 13:42 < dutchie> it's also younger 13:43 < ctefaho> to me Ithkuil seems complex in the wrong way 13:43 < zipcpi> I hate trolls and mind games 13:43 < ctefaho> the way it builds everything into the words 13:43 < Zearen> Er, yeah, frica 13:43 < ctefaho> the sumtcita are at least tweakable 13:44 < ctefaho> not saying Lojban is better than Ithkuil it was just the impression I got from studying its case system 13:44 < Zearen> Itkuil is an impressive piece of work. It's difficult due to its level of abstraction. 13:44 < Zearen> What's left to be proven is whether it's abstract in useful ways/ 13:45 < ctefaho> indeed 13:45 < Zearen> I'm not sure comparing languages on a scale of goodness is ever productive. 13:45 < zipcpi> They're only good for being rendered into soap 13:45 < ctefaho> on a scale of usefulness then:) 13:46 < Zearen> I'd argue that on a scale of usefuness, **both** itkuil and Lojban rank pretty low :p 13:46 < dutchie> .ie.u'i 13:47 < ghhtzt> ithkuil cases are not much different from lojban sumtcita 13:47 < ctefaho> *cough* STT is trying to improve on that;o 13:48 < ctefaho> I think one of the biggest problems with jbobau is the unique rafsi system 13:48 < Zearen> Not the lack of vocab ? Because that's what generally trips me up. 13:48 < ghhtzt> biggest problem is vocab 13:48 < ctefaho> because it makes an overwhelming majority of possible lujvo pretty much useless 13:48 < ctefaho> well 13:49 < ctefaho> the locked rafsi system is part of that 13:49 < samcinki> I wonder how content clauses (lodu'u...) work in Iþkuîl, provided something similar exists 13:49 < ctefaho> the idea I have is to ban 4-letter gismu rafsi so 'y always works 13:49 < ctefaho> and any word can be joined with 'y 13:49 < ctefaho> also 13:50 < Zearen> I remember latro'a making the point that lujvo are essentially a falled experiment, and that tacking an agglutinative system onto a system built around predicates doesn't make much sense 13:50 < dutchie> what does 'y offer over 4-letter rafsi? 13:50 < ctefaho> My solution: y can never be used to join words by removing a letter, 'y can always join any two lojban words, like zei 13:51 < ctefaho> any two lojban words ending in vowel 13:51 < Zearen> But for Ithkuil, which essentially is a fusional language builf around agglutination, it makes sense. 13:51 < zipcpi> Zearen: They're a useful mnemonic at least. Though the whole problem with -gau lujvo etc has caused some problems 13:51 < ctefaho> this also makes joining cmavo and brivla together a piece of cake 13:51 < zipcpi> But people still use them because it's easier than {jai gau broda fai} 13:51 < ctefaho> ko'yklama 13:52 < ctefaho> fa'ylo'ymlatu'yklama 13:52 < ctefaho> for 2 random examples 13:52 < durka42> that would not work under the current morphology 13:52 < ctefaho> correct 13:52 < ghhtzt> content clauses are made by using framed relation on a formative 13:52 < ctefaho> but in STT morphology it does 13:52 < durka42> ok 13:52 < Zearen> Well, I'd rather get rid of -gau lujvo and replace them with gau ... BRIVLA ... 13:53 < ctefaho> so for lujvo, any whole word is the only predictable rafsi of that word 13:53 < durka42> I see STT's jurisdiction has broadened beyond sumtcita and become ctefa'obau :) 13:53 < zipcpi> zearen: You're basically asking for something that can modify a predicate's place structure 13:53 < zipcpi> And... yeah right now we don't have any good solution for that 13:53 < ctefaho> eh perhaps 13:54 < ctefaho> I mostly desire to learn jbobau to think and cogitate in it 13:54 < durka42> zipcpi: what do you mean? gau ... BRIVLA ... is now 13:54 < Zearen> You mean like modals ? Don't modals do that already ? Isn't a "gau" a modal ? 13:54 < Zearen> ^ 13:54 < ctefaho> Zearen: all modals are macros expanding into xoi-clauses 13:54 < ctefaho> in my world at least 13:54 < zipcpi> Zearen... yes... but like I said, people prefer brodygau over jai gau broda fa 13:54 < Zearen> o_O 13:54 < zipcpi> *fai 13:55 < durka42> there's three solutions 13:55 < ctefaho> .i gau mi do sipna -> .i do sipna xoi mi gasnu ke'a 13:55 < durka42> gau ko'a ko'e broda 13:55 < durka42> ko'a brodygau ko'e 13:55 < ctefaho> ke'a being the outer bridi, that is "do sipna" 13:55 < durka42> ko'a jai gau broda fai ko'e 13:55 < durka42> ctefaho: ie 13:55 < durka42> sa'enai 13:55 < ghhtzt> ikal tu = mi klama. íkala tu = losuhu dahinai mi klama 13:56 < ctefaho> but as for STT, I am basically putting a whole bunch of ideas for improvement into it, mine and a few others 13:56 < Zearen> Well, there is something to be said for brevity... 13:56 < zipcpi> First one is basically "ergative case", or so gleki tells me, but it's difficult to form the forethought required for it, and it doesn't work with {jecu / gi'e} etc 13:56 < ctefaho> I also desire a logical language that is 100% formally expandable 13:57 < zipcpi> Third one is long and the {fai} often forgotten 13:57 < zipcpi> And so people resort to the lujvo option 13:57 < durka42> ah, I had my suspicions on how ghhtzt was, but now they are confirmed :) 13:57 < durka42> zipcpi: yeah and I don't see any problem with that to be honest 13:57 < ctefaho> as for STT's rafsi system, it essentially allows an exponential amount of semantics to be crammed into lujvo of 2 or 3 short rafsi 13:57 < zipcpi> The lujvo option? 13:58 < durka42> yeah 13:58 < zipcpi> The problem is that right now -gau lujvo are somewhat inconsistent 13:58 < durka42> I don't use the inconsistent ones 13:58 < ctefaho> at the expense of making lujvo word creation a bit less automatic process 13:58 < zipcpi> Some are actually more like {jai broda} 13:58 < Zearen> If a term appears before the after thought connective, then it is a term of every bridi. So yes, it does work with gi'e 13:58 < ctefaho> the alternative is to do like sepla and fuse zi'evla and lujvo 13:58 < zipcpi> Zearen: What I mean is that the gau would stick to both predicates 13:59 < zipcpi> When often that is undesirable 13:59 < ctefaho> but I still believe there is something worth having in the gismu-rafsi-lujvo-zi'evla/fu'ivla system 13:59 < Zearen> Then put it in the tail. 13:59 < zipcpi> You usually put something in the x1 for a reason 13:59 < zipcpi> Because that is the "subject"; the sumti with the most emphasis placed on it 14:00 < Zearen> vau exists. 14:00 < ctefaho> I just don't think permanently allocated short rafsi is a viable long-term solution 14:00 < ghhtzt> wedding vau 14:00 < Zearen> Also, you can put more than one term in the bridi head. 14:00 < zipcpi> I'm not sure how vou works here... say you want to say: gau mi broda ko'a .i mi brode ko'e 14:00 < durka42> I think I'm the only lojbanist who likes the lujvo system 14:00 < durka42> zipcpi: ...then say that :p 14:01 < ctefaho> durka42: including the short rafsi allocation? 14:01 < zipcpi> Imagine that instead of mi 14:01 < zipcpi> You have a long thing... or you just want to connect the two sentences using gi'e 14:01 < Zearen> gau zo'e mi broda gi'e brode / mi broda gi'e brode vau gau zo 14:01 < Zearen> 'e 14:01 < durka42> ctefaho: they're not perfect but if the alternative is losing the lujvo-veljvo associations, I'll keep 'em 14:01 < zipcpi> Where did zo'e come from? 14:02 < ctefaho> durka42: well in my system they still sort of remain 14:02 < ghhtzt> broda gaumi koha gihe brode kohe fami 14:02 < durka42> zipcpi: anyway just move the gau-term to after the selbri 14:02 < Zearen> It 14:02 < ctefaho> short rafsi can't come from any word 14:02 < zipcpi> durka42: The problem here is that you want to raise the gau term to the x1, so that it can be the x1 of both bridi 14:03 < ctefaho> and in lujvo the rafsi will in its defs have the original gismu's traced 14:03 < zipcpi> jai is supposed to do that 14:03 < Zearen> It's just a nonce word used to show structure. Here, I'll think up an actual example. 14:03 < durka42> zipcpi: eh? terms don't fill places 14:03 < zipcpi> ... 14:03 < zipcpi> I'm using the word term wrongly 14:03 < zipcpi> The gau sumti to the x2 14:03 < zipcpi> *x1 14:03 < durka42> zipcpi: {gau mi broda gi'e brode} it sticks to both selbri. {broda gau mi gi'e brode} only the first one 14:03 < ctefaho> and all existing brivla base luvjo you love and cherish can easily be brought along 14:03 < ctefaho> based* 14:03 < zipcpi> Then who brodes? 14:04 < zipcpi> Actually it's the x1 of broda 14:04 < durka42> the brodaer and brodeer are both unspecified 14:04 < ctefaho> zohepowah 14:04 < zipcpi> No, {broda gi'e brode} -> {zo'e broda gi'e brode} implies that they're the same zo'e 14:04 < durka42> okay, still unspecified though 14:05 < durka42> in both examples 14:05 < zipcpi> Yeah but it doesn't say what I wanted to say 14:05 < durka42> what do you want to say, .u'i 14:05 < durka42> welp 14:05 < dutchie> u'isai 14:05 < zipcpi> Oops 14:05 < zipcpi> gau mi broda ko'a .ije mi brode ko'e 14:06 < zipcpi> I want to make mi the subject of both sentences 14:06 < zipcpi> Both bridi 14:06 < Zearen> «gau lo te bilma mi kafke gi'e vamtu» / «mi kafke gi'e vamtu vau gau lo te bilma» 14:06 < durka42> ohh 14:06 < durka42> don't think you can do that without repeating {mi} or using {jai} 14:06 < zipcpi> Exactly my point 14:06 < ghhtzt> mijaigaubroda gihe brode 14:06 < zipcpi> Or... -gau lujvo 14:06 < durka42> or that 14:06 < ghhtzt> poihi 14:06 < Zearen> "The sickness caused me to cough and vomit." 14:07 < zipcpi> The most promising fix I've heard thus far is the two letter rafsi... they play in a morphological space that is currently unassigned 14:07 < Zearen> Though gau isn't right here unless the sickness is the descolada 14:07 < ghhtzt> le siknese purinka lenu mi kafke gihe vamtu 14:08 < zipcpi> And being unusual and nothing like current lujvo, signals that they serve a very special purpose, modifying the brivla in a very specific way 14:08 < Zearen> I.e. gau implies volition. 14:09 < zipcpi> So instead of relying on exactly how {zangau} play with each other 14:09 < zipcpi> We just define -yga to always have a single expansion 14:09 < durka42> that could work 14:10 < Zearen> Er, how about «gau mi lo selkei muvdu gi'e se sance» / «lo selkei muvdu gi'e se sance vau gau mi» 14:10 < durka42> mi nelci i ie mi nelci 14:10 < ctefaho> sounds like sort of what sepla is up to 14:10 < ctefaho> with his "affixes" 14:10 < ghhtzt> sei stace mi na mutce nelci 14:10 < Zearen> " I moved the toy and made it make noises" 14:10 < durka42> I think those are the same yes 14:10 < durka42> assuming you add the missing cu 14:11 < Zearen> Er, right. Sorry, it's been awhile ^^' 14:11 < ghhtzt> uhasai nicte toldi co barda 14:12 < ctefaho> ...and that and some other things is what STT is about. leaving you to your talkings now:d 14:13 < Zearen> Note that if you just want "I moved the toy and it made noises", you could « lo selkei cu muvdu gau mi gi'e se sance» 14:14 < Zearen> I used to be really into the far out there proposals and tweaks, but at some point people start speaking some other language with Lojbanic vocab. 14:15 < Zearen> There's nothing wrong with that, it's just that maybe all Lojban proper needs is a little polish, not an entire revamp. 14:15 < durka42> ie 14:18 < ctefaho> agreed Zearen 14:18 < Ilmen> .oi glare fa lo nicte vacri 14:19 < Zearen> .i .iesai .i mi to'e nelci lo sance be lo rutni brife 14:20 < Ilmen> .i mi ŭunmo lo xabju be lo glatce bratca be'o noi la'a ca'o febvi 14:21 < dutchie> pe'i vi dukse glare ca lo donri 14:22 < Zearen> .uanaisaidai 14:22 < dutchie> .i ku'i la'a ze'i ranji 14:23 < Ilmen> la'a nai dai cimni jai ranji 14:23 < dutchie> u'i vi lo gligu'e bazi carvi 14:24 < zipcpi> bu'u 14:24 < Ilmen> .e'u capyju'i zo vi noi jitfa pendo 14:24 < Ilmen> .i ĭe zo bu'u mapti 14:24 < zipcpi> For some reason old materials really liked vi over bu'u 14:25 < dutchie> ah yes 14:25 < zipcpi> I hear it's Loglan influence 14:25 < Ilmen> ki'u su da lo laldo ve ctuca cu mutce co zmanei zo vi zo bu'u 14:25 < Ilmen> .i ti'e dai loglo malxlu 14:26 < zipcpi> ki'e 14:27 < Zearen> «su'u» ? 14:27 < Ilmen> su'o 14:27 < zipcpi> Oh... yeah that's a CKTJ reassignment... sometimes we use it informally 14:28 < zipcpi> Also made somewhat famous by ZA'O 14:28 < Ilmen> za'o seju xa'o misno 14:28 < Zearen> .i ji'a za'a zo laldo valsi cnino vau .i'e 14:29 < zipcpi> na mutce lo ka cnino .iku'i na'e krasi gismu 14:29 < Ilmen> seznafmlu ĭe dai fa lodu zo laldo cnino 14:29 < Zearen> .i mi puze'u na'e jundi la .lojban. 14:29 < zipcpi> je'e 14:30 < durka42> zipcpi: well people also used to think {vi} tagged a location. at some point it was clarified to tag a spatial distance 14:30 < durka42> like {iz} 14:30 * durka42 {zi} * 14:31 < Ilmen> sei jinto lo pe'ople si prenu pu ta'e jinvi lo du zo vi tcika lo stuzi sumti 14:31 < Ilmen> *tcita 14:32 < Zearen> The reason it used to be enouraged was because often people would translate English to Lojban word to word, and use bu'u as "at" which was very often wrong and they meant "vi" 14:32 < durka42> why is bu'u wrong for that 14:33 < Zearen> I.e. you're typically either "vi" or "ne'i" it 14:33 < durka42> ...using CLL-vi 14:34 < Zearen> There are other versions of "vi" ? 14:34 < zipcpi> Maybe they are being superpedantic and saying you can only "bu'u" something if you are actually overlapping it, or something like that 14:34 < durka42> well the CLL-vi is polysemous, look: 14:34 < durka42> vlaste: vi 14:34 < vlaste> vi = location tense distance: here at ... ; at or a very short/tiny distance from ... 14:34 < Zearen> I always thought bu'u was discouraged for the same reason "du" is. You're rarely in the exact same place. 14:34 < durka42> ca'i BPFK, {vi} always tags a spatial distance so it's exactly like {zi} 14:34 < zipcpi> Zearen: Then what's the point of bu'u then? rofl 14:35 < Zearen> Iunno. I just learned to roll with it. 14:35 < Zearen> I like this version much better, I was just giving historical context. 14:35 < zipcpi> Yeah 14:35 < durka42> je'e 14:36 < Ilmen> I think bu'u is about overlap (otherwise, we'd use ne'i). 14:36 < Zearen> ^ 14:36 < zipcpi> OK then... am I overlapping Malaysia, the country, or my house? 14:36 < Zearen> I like bu'u ~= ca much better though. 14:36 < zipcpi> OK those can be both ne'i technically 14:37 < Ilmen> I think ne'i implies bu'u 14:37 < Ilmen> ie nai pei 14:37 < mensi> ba'e mi na tugni 14:37 < Ilmen> re'i smuske cretce no'u la mensi 14:37 < Ilmen> mi denpa tu'a lo do prije valsi 14:38 < zipcpi> But yeah it does make sense to reserve vi/va/vu for spatial proximity 14:38 < zipcpi> da xo'ibu'u da 14:39 < dutchie> mi ca'o cilre ui 14:39 < Ilmen> ta'o co'o ro do 14:39 < zipcpi> Heck even zvati was never this pedantic 14:39 < Zearen> .i ca'o 14:40 < Ilmen> Well maybe zvati allows x1 and x2 to not overlap but be close to each other, I dunno. But then we'd lack a word for spatial overlap 14:41 < Zearen> zvati was never the pendantic one. stuzi is the pendantic one. 14:42 < zipcpi> Ilmen: Well, whether I am overlapping Malaysia or my house does kinda depends on interpretation 14:44 < Zearen> Well, even if we take bu'u = spatial ca, then we're talking about points if we're strict about it. This seems useless though. 14:44 < zipcpi> Yeah... 14:45 < ghhtzt> ca doesnt imply points 14:45 < Ilmen> I tend to think of "ca/cabna" as time overlap 14:45 < zipcpi> Not to mention the common semi-metaphorical usage... due to... the Internet 14:45 < zipcpi> e.g. bu'u la jbovlaste or bu'u la uitkipediia 14:46 < Ilmen> .ai sipna 14:46 < Ilmen> co'o 14:46 < zipcpi> co'o 14:46 < dutchie> co'o 14:47 < Zearen> Hahaha, I still am afraid to use «fi'o se .irci ti» or to use «ti» talk about anything on IRC or especially about anything on the internet. 14:47 < Zearen> I... 14:48 < Ilmen> .a'o lo glare na'e dukse .e nai lo mabla toknu cu ckana fi do 14:48 < Zearen> «lo'u bu'u la jbovlaste le'u» zanru nakusai mi 14:48 < zipcpi> Zearen: Hm the demonstratives... to me the act of posting something somewhere is about equivalent as signaling it. It's a bit different from "These are the reasons I quit my job:" (la'edi'e / lau'e) in my opinion 14:49 < Zearen> .i co'o .ibu 14:49 < zipcpi> Zearen: Yeah maybe that's sketchy; we don't quite have an alternative for it though 14:49 < Zearen> fi'o 14:50 < zipcpi> The metaphor of websites being... sites... is pretty entrenched, not just in English 14:50 < Zearen> Lojban is pretty adamant at refusing antipatterns just because other natlangs do it. It's part of its appeal. 14:50 < zipcpi> fi'o... part of the problem is always having to import the brivla. fi'o... kibystu? se irci? 14:50 < zipcpi> Yeah I see your point 14:51 < zipcpi> Just don't know what's the best soulition 14:52 < Zearen> I'd use fi'o kibystu personally. Also, sometimes trying to find a modal that doesn't exist means you should probably just be using the predicate. 14:53 < zipcpi> Problem is wrapping the bridi as an afterthought 14:53 < ctefaho> xoi'oi'oi 14:53 < ctefaho> just put a predicate in xoi and you got your modal 14:53 < Zearen> Just because you'd express that with a preposition in English doesn't mean you use a modal in Lojban. 14:53 < zipcpi> Problem is that often you're already deep into a bridi before you realize you want to put it in something 14:53 < zipcpi> So there have been various solutions came up with over the years 14:54 < zipcpi> xoi 14:54 < zipcpi> gi'ei (proposed in an old blog somewhere, though I just added it{ 14:54 < Zearen> There's also «.i» :p 14:55 < zipcpi> {.i lo nu go'i} seems a bit weak though; like you're making two statements 14:56 < Zearen> Well, you are. If that really bugs you, then «.ijebo lonu go'i» 14:56 * zipcpi shrugs 14:56 < zipcpi> The thing is the point of sumtcita is to encapsulate the most common "bridi-wrapping" predicates 14:57 < zipcpi> So either we come up with some new cmavo to express kibystu, due to commonness, or we stretch the meaning of bu'u 14:57 < zipcpi> The problem is that the line blurs even more when talking about virtual 'places" 14:57 < Zearen> The original purpose is to add places to bridi. Though what you said is probably how it's used today. 14:58 < zipcpi> Because at some level of the UD, those places "exist" 14:58 < zipcpi> And they have to, unless we want to add {virtu'aly-} to everything 14:59 < Zearen> (This is why you can do weird stuff like lo mlatu be bai zo'e) 14:59 < zipcpi> mi virtu'alycitka lo virtu'alyplise 15:02 < zipcpi> Yeah... that's just not gonna work 15:04 < zipcpi> Another problem with .i is that it doesn't work within a NU 15:04 < zipcpi> It will close off the NU 15:05 < zipcpi> There is another crazy cmavo for that... {ju'ei} 15:05 < dutchie> there is a crazy cmavo for everything 15:05 < zipcpi> Yeah I took a long hiatus too, I had to pick all this stuff up when I came back 15:05 < Zearen> I mean, at some point you should stop nesting and start a new thought any way. 15:07 < zipcpi> It doesn't even require that much nesting... only any NU where you might want to express more than one bridi 15:07 < Zearen> Or just use fi'o :p 15:07 < zipcpi> fi'o has similar problems to tanru; it imports one specific place, but doesn't say how the bridi is related to it 15:08 < zipcpi> It's usable, for sure, 15:08 < zipcpi> But sometimes vague 15:08 < ctefaho> fi'o is trash use xoi or soi 15:09 < zipcpi> And cumbersome if you actually want to fill more than one place of the imported predicate 15:09 < Zearen> All modals are like that. Also, your solution to fi'o is too vague is let's make bu'u more vague. 15:09 < ctefaho> or even kai'a/kai'i depending on your needs 15:09 < zipcpi> I know what you mean about bu'u 15:09 < Zearen> So I'm confused. 15:09 < zipcpi> But I'm not sure what the best solution is 15:10 < zipcpi> But even then... is confusion really possible? 15:10 < Zearen> ctefaho: I'll speak Lojban how I please as long as it's gendra and smudra 15:10 < ctefaho> are you serious zipcpi did you make kai'a one of your UIs 15:10 < ctefaho> how many words has zipcpi added? 15:10 < Zearen> Oh ! I thought of the best, obvious solution: «do'e» :p 15:10 < zipcpi> Hah 15:10 < ctefaho> OH MY GOD 15:10 < Zearen> I'm actually pretty serious about that though. 15:11 < ctefaho> is that a hundred words 15:11 < ctefaho> I think that's a hundred words 15:11 < zipcpi> Er, no, if we have to resort to that it might as well be a new sumtcita 15:12 < ctefaho> Zearen: http://selpahi.weebly.com/lojban/proprietor-fronting 15:12 < zipcpi> I mean I do use do'e, but if you're suggesting we use it for all kibystu... yeah don't think that's gonna work out well 15:12 < ctefaho> ^if xoi/soi wasn't what you were looking after 15:12 < ctefaho> maybe you even want new-voi 15:12 < ctefaho> http://mw.lojban.org/papri/new_voi 15:13 < ctefaho> Zearen: What is it you want to do exactly? 15:13 < Zearen> Well, if the context is sufficiently obvious to figure out bu'u la'oi http://, then context should be sufficient for do'e. If it's not, fi'o kibystu still works. 15:14 < Zearen> ctefaho: "at/on ", e.g. "I found this on http://awesome.com/" 15:15 < dutchie> clicking that link is probably a bad idea 15:15 < ctefaho> bu'u zoi url http://awesome.com/ 15:15 < dutchie> but i'm rather curious 15:15 < ctefaho> bu'u zoi url http://awesome.com/ url 15:15 < zipcpi> I mean if we really needed we could come up with something that works like pe'a, but specifically for {kibro}, but... I'm not sure that works out for things like MMORPGs 15:15 < dutchie> ok that was very disappointing 15:15 < ctefaho> if you want make that just zvati directly 15:15 < ctefaho> zoi url http://awesome.com/ url zvati 15:16 < ctefaho> throw in a pe'a if you must 15:17 < ctefaho> but I think zvati is fine 15:17 < zipcpi> "All modals are like that." Well there are actually attempts to formalize the current sumtcita 15:17 < ctefaho> it is a location in "cyberspace" 15:18 < zipcpi> And I think some of them might work 15:18 < zipcpi> Not sure about {ra'a} or some others though 15:18 < zipcpi> But the point is that they *could* be more specifically defined than just by their fi'o 15:19 < Zearen> I disagree that you can be «bu'u by.» unless you can also be «ga'u by» 15:19 < ctefaho> Zearen: think like this, what is it that you want to say? 15:19 < ctefaho> dutchie: what did you expect huh 15:19 < Zearen> Websites don't exist in space. 15:19 < zipcpi> What about MMORPGs then? 15:19 < dutchie> ctefaho: mostly malware 15:20 < ctefaho> Zearen: but they exist on servers physically located away from you 15:20 < ctefaho> the website name is just a way to get to them 15:20 < Zearen> I'd be very willing to use bu'u to talk about a character being bu'u a place in the game world, but the player (human) isn't in the game. They're jorne it, or something.. 15:21 < zipcpi> Hm 15:21 < Zearen> Except that a single webpage may have data from multiple physical servers. 15:21 < ctefaho> Zearen: bu'u is just zvati, zvati isn't ok with you? 15:21 < Zearen> **A website is not a place** 15:22 < Zearen> It's not. 15:22 < ctefaho> I kinda disagree having studied IT management 15:22 < ctefaho> they are very much physical locations 15:22 < Zearen> ... 15:23 < zipcpi> I dunno, I don't really have any more arguments to move forward with; I'd make a new modal myself, but I don't know how others would feel about it 15:23 < ctefaho> a website is data located on a physical server away from you 15:23 < ctefaho> there is no pe'a in that spatiality 15:23 < Zearen> I repeat my previous point: a single webpage may have data from multiple physical servers. 15:23 < zipcpi> I'd like to get some feedback from some other peolpe 15:23 < ctefaho> Zearen: how is that a problem when the url is the location? 15:24 < durka42> lo mutpapri na zvati pe'i 15:24 < durka42> s/zvati/selzva 15:24 < fenki> durka42 meant to say: lo mutpapri na selzva pe'i 15:24 < Zearen> zipcpi: Well, ki'o is taken. What about xi'o ? :p 15:24 < Zearen> I still think do'e is sufficient though. 15:24 < zipcpi> There's always ki'oi or something 15:24 < ctefaho> hmm good question 15:25 < Zearen> The URL is not a location in space.. You cannot be left of it. 15:25 < ctefaho> maybe stuzi is the better modal then 15:25 < ctefaho> or verb 15:26 < ctefaho> durka42: Good question. There is the URL and there is what is at the URL, right? 15:26 < ctefaho> the url and the content of the webpage 15:26 < ctefaho> iepei? 15:26 < mensi> ba'e mi na tugni 15:26 < durka42> right 15:26 < ctefaho> So I believe the url would be where something is located at 15:27 < VegBerg> It's not a physical place, although it may be an abstract place. 15:27 < zipcpi> Or even something like zai'o, but that relies on the CKTJ reassignment of bu'u to zai to make mnemonic sense 15:27 < durka42> coi la .vegberg. 15:27 < ctefaho> So then the url is zvati-1, and the webpage zvati-2? 15:27 < ctefaho> err 15:27 < durka42> ctefaho: I'd rather say the URL is a handle/reference to something 15:27 < ctefaho> other way 15:28 < ctefaho> Url is zvati-2, webage zvati-1 15:28 < Zearen> What is CKTJ ? 15:28 < VegBerg> coi la .durka42 15:28 < durka42> Zearen: one of those tweaks you used to be a fan of :) http://mw.lojban.org/papri/ce_ki_tau_jau 15:28 < ctefaho> Zearen: some cmavo reshuffling 15:28 < ctefaho> to save syllables 15:28 < Zearen> Oh, u'i 15:28 < ctefaho> more used cmavo gets shorter forms 15:28 < Zearen> I never thought anyone would seriously do that ! 15:29 < ctefaho> durka42: well an url is just a name for an ip 15:29 < dutchie> (and a bit more) 15:29 < ctefaho> and the ip is like a road address 15:29 < durka42> sure 15:29 < Zearen> urli = x1 is the URL/address of web page/resource/file x2 15:29 < ctefaho> where you stop and pick up your data 15:29 < ctefaho> except the only thing moving is the data 15:29 < durka42> a road address isn't a place either 15:29 < durka42> I mean when expressed as a number or text 15:29 < ctefaho> right 15:30 < Zearen> So a URL is .urli and the content is se .urli 15:30 < durka42> the URL is an identifier 15:30 < ctefaho> but what is at the address, is a place 15:30 < ctefaho> yeah 15:30 < durka42> yeah I guess so 15:30 < durka42> not a useful place though pe'i 15:30 < durka42> since it might be in Robin's basement or a cloud server in Korea, who knows 15:30 < dutchie> or cares? 15:30 < ctefaho> "The image" zvati "www.imgur.com" 15:30 < durka42> and it may move to a different cloud server tomorrow 15:31 < durka42> it seems like a huge misuse of {zvati} 15:31 < dutchie> or maybe you get it from a different server from me 15:31 < ctefaho> I think zvati works fine 15:31 < Zearen> And "I logged in at http://example.com" is most exactly « mi jorne fi'o .urli zo'oi http://example.com» 15:33 < durka42> .u'u .ei mi cliva iau co'o 15:33 < ctefaho> co'o 15:33 < Zearen> I also highly disagree with the use of zvati. It seems tcetce mutce lo ni maltce mutce malgi 15:33 < Zearen> .i co'o 15:33 < ctefaho> your choice Zearen 15:34 < Zearen> *malgli 15:34 < ctefaho> use a simple brivla or a heavily complicated construct, you choose:) 15:35 < ctefaho> if you do believe I won't understand you if you use zvati 15:36 < zipcpi> Oh wait, forgot, zai'o is taken by some... well, there's this guy who's somewhat infamous for defining weird cmavo, but some might put me in that category as well 15:37 < Zearen> Haha, the one swap I wanted the most ce <-> fi'o isn't on there. These choices seem better though. 15:37 * ctefaho has defined 4 UI cmavo in total 15:38 < Zearen> Well, the first 5 anyway. The others seem like your typical proposal craziness. 15:38 < zipcpi> fi'o might be able to get foi, though I dunno if someone wanted that for something else 15:39 < ctefaho> Zearen: Do you truly believe I won't understand you if you use bu'u/zvati? 15:41 < Zearen> No. I also think the reason you'll understand me is because you speak English, hence why its malgli. 15:41 < Zearen> *it's 15:42 < Zearen> Especially since I think you'll understand me also if I use do'e/.urli 15:42 < ctefaho> if you accused me of having English as a first language you can enjoy, cause I am not a native speaker of it 15:42 < Zearen> Clearly you speak it, and that's all I meant. 15:43 < ctefaho> then you mean mabla natlang 15:43 < Zearen> Whatever, it's not the point. 15:43 * ctefaho concedes 15:44 < Zearen> ? 15:44 < ctefaho> over-complicate lojban as much as you want, it won't do any good 15:44 < ctefaho> I am going to go practice my jbo-handwriting before I go to sleep, co'o 15:44 < dutchie> co'o 15:44 < Zearen> .y. 15:45 < Zearen> co'o .i pluka sevna 15:45 < demize> s/vn/nv/? 15:45 < Zearen> .ua .ie .u'u 15:50 * nuzba @earthtopus: @kukkurovaca ...and that the sentence made a lot more sense in the original Lojban [http://bit.ly/1T8RGss] 15:52 < demize> mi mutce tatpi .iku'i dukse glare 15:55 < demize> lo ti kumfa cu jacke'o du'eda 15:56 < Zearen> .uinaidai 16:07 < Zearen> Is there a may to attach a mo'i to a tense tag, e.g. mo'ivi la .djak. lo mlatu cu smaji klama 16:14 < zipcpi> exp: mo'ivi la .djak. lo mlatu cu smaji klama 16:14 < mensi> SyntaxError: Expected [,] but "k" found. 16:14 < zipcpi> exp: mo'iga'o la .djak. lo mlatu cu smaji klama 16:14 < mensi> SyntaxError: Expected [,] but "k" found. 16:14 < zipcpi> mo'iga'u la .djak. lo mlatu cu smaji klama 16:14 < zipcpi> Ugh 16:14 < zipcpi> exp: mo'iga'u la .djak. lo mlatu cu smaji klama 16:14 < mensi> ([{ } {lo mlatu KU}] [cu {smaji klama} VAU]) 16:15 < zipcpi> Hmm... the tags are supposed to all have the same grammar in exp, but apparently it wasn't completely merged 16:15 < zipcpi> But I'm not exactly sure what means 16:16 < zipcpi> Do you mean "towards"? 16:16 < zipcpi> That would be fa'a 16:22 * nuzba @uitki: Logical Language Group - http://mw.lojban.org/papri/Logical_Language_Group by LeoBeltran [http://bit.ly/1gdzFeg] 16:42 < Zearen> Some thing that move tangentially is mo'i vi, but not fa'a 16:43 < zipcpi> "Tangentially"? Think there is supposed to be something else for that 16:43 < Zearen> E.g. a comet headed towards the sun. 16:43 < Zearen> Maybe 16:43 < zipcpi> zo'a? 16:44 < Zearen> Maybe ? 16:44 < Zearen> I can't tell if it implies motion. I think it does... 16:44 < zipcpi> There is mo'izo'a 16:44 < zipcpi> Dunno exactly what mo'i does though 16:45 < zipcpi> The ones working on formalization tell me it's "deep magic" lol 16:45 < Zearen> It implies a change. It's kind of like a spatial aspect WRT time. 16:46 < zipcpi> I gotta go; see ya 16:46 < Zearen> co'o 19:20 < pimlu> coi 19:21 < Zearen> .i coi .i do ca mo 23:12 < gleki> i'd say that -gau doesnt imply volition 23:12 < gleki> -zu'e does. 23:14 < gleki> pimlu: pls register and upvote questions with rating below 10 http://area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 23:20 < gleki> i ua la selkik ca'o y zukte 23:20 < gleki> gi'u lojbo zukte 23:26 < zipcpi> No one, NO ONE "pulls my leg" without getting kicked in the nuts 23:27 < zipcpi> No one plays with my emotions without getting more than they bargained for 23:27 < gleki> wrong channel xu 23:27 < gleki> u'i 23:28 < zipcpi> Ugh... he's not here 23:28 < zipcpi> I got into a fight with ctefaho 23:28 < zipcpi> I'm sick of his mind games 23:29 < gleki> this channel requires either speaking in lojban or about lojban 23:29 < zipcpi> Sigh... sorry 23:29 < zipcpi> I just dunno what to do about this crap 23:31 < gleki> any non lojba related stuff should be avoided 23:31 < gleki> sorry i didnt read the backlog 23:31 < zipcpi> It *was* about Lojban 23:32 < zipcpi> ba'e pu srana le jbobau 23:33 < gleki> then it was okay 23:37 < gleki> im very close to creating an eng2jbo dictionary 23:38 < zipcpi> Cool 23:38 < zipcpi> Do you need any more iolpe-type slates of words? :p 23:47 < b_jonas> Is {la .trabant. noi karce cu marji ma} the question I should ask if I want to ask what material something is made of, and expect an answer like {.i lo pelji} or {.i lo slasi}? 23:51 < zipcpi> http://jbovlaste.lojban.org/dict/ki'oi --- Day changed Fri Jul 03 2015 00:00 < zipcpi> b_jonas: I believe so 00:00 < b_jonas> je'e 00:04 < gleki> en: ki'oi 00:04 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 00:05 < zipcpi> lo'u fi'o kibystu le'u ja lo'u fi'o se kibzva le'u 00:07 < zipcpi> zmadu zo bu'u lo ka smudra ja filseljmi .i mu'a lu ki'oi ma li'u 00:13 < zipcpi> gleki: Not sure how grammatical your Chinese for xanto is. I have a feeling it's missing a few particles, but I'm not sure how much I trust myself 00:13 < zipcpi> shi4 ge4 da4 xiang4 00:14 < gleki> xanto? 00:14 < zipcpi> Well, that's the funny thing about Chinese though. At least in writing, it is quite understandable without those particles 00:14 < gleki> srana ma 00:14 < zipcpi> je'u zo xanto 00:15 < zipcpi> But Modern Mandarin Chinese has homophones by the reams, and thus all these particles evolved to make up for it 00:17 < zipcpi> 是个大象 00:17 < zipcpi> Also it's gone through its own version of CKTJ :p 00:18 < gleki> what is this all related to 00:18 < zipcpi> this =? 00:19 < zipcpi> Just saying by *my* intuition, 是象 should be replaced by 是个大象 00:19 < gleki> deilctic thing? 00:20 < zipcpi> I meant what you mean by "this" 00:22 * nuzba @uitki: L17-04 - http://mw.lojban.org/papri/L17-04 by Gleki [http://bit.ly/1JBEIlc] 00:24 < zipcpi> Hm... 10-15: you might want to include both {pano} etc. and the hex digits. Hex digits are a nice shorthand and have those experimental rafsi for month names, but {pano} may be easier to remember 00:25 < gleki> well, a new script-generated dictionary http://mw.lojban.org/papri/L17-04 00:25 < gleki> still some sorting problems 00:25 < zipcpi> K 00:26 < gleki> it puts × after "z" 01:00 * nuzba @uitki: punctuation - http://mw.lojban.org/papri/punctuation by Spheniscine [http://bit.ly/1RVt1VK] 01:11 < gleki> zipcpi: i got what you meant about xanto. but it wasnt me who translate gimste to chinese 01:11 < zipcpi> Ah 01:12 < gleki> it was panabsent, a native chinese speaker 01:12 < zipcpi> je'e 01:13 < zipcpi> Well Chinese grammar is pretty flexible. It just felt somewhat empty to me without those particles, but perhaps they're just for disambiguation, which is less needed in writing 01:14 < gleki> looks like the lack of good search in jvs and the lack of eng2jbo dictionary is a serious problem 01:15 < zipcpi> I've been getting by with vlasisku but that hasn't been updated in a while 01:15 < gleki> we only need to finish tagging tatoeba.org sentences and i need to implement glossword search in sutysisku to cover them 01:15 < gleki> vlasisku doesnt search in glosswords 01:15 < gleki> try searching for "and" 01:15 < zipcpi> Right 01:15 < gleki> if it does it doesnt prioritize words with "and" as a glosssword 01:16 < gleki> only my laziness prevents me from adding glossword search 01:16 < zipcpi> Well it is kinda important; I have been avoiding sutysisku for eng2jbo because of it T.T 01:17 < gleki> vlasisku is not better 01:18 < zipcpi> Maybe not. So then I resort to going through the natlang list on JVS 01:18 < zipcpi> Which... yeah, sucks, as you know 01:18 < gleki> *mo 01:20 < gleki> *no 01:21 < zipcpi> no? 01:21 < gleki> vlasisku is no better 01:21 < zipcpi> Ah 01:22 < gleki> and i was against any server solutions exactly because i saw problems with the server 01:22 < gleki> with sutysisku we only need a simple hosting for static files 01:26 < gleki> i will have to create another sutysisku based on the two La Bangu dictionaries 01:26 < gleki> as for words not listed there it's a very complex problem 01:27 < gleki> many kurtyvla are ... idk 01:27 < zipcpi> u'i 01:27 < gleki> some maybe useful 01:28 < gleki> like {dutso} 01:28 < zipcpi> And {corci} 01:32 < gleki> so i can only manually select words 01:32 < gleki> and ofc. 17 000 words are too hard to check and add immediately. 01:33 < zipcpi> Not to mention lujvo and zi'evla 01:33 < gleki> en: 01:34 < mensi> 19319 da se tolcri: a, a'a, a'acu'i, a'anai, abata'adj, abgad, abniena, abu, acaman, aclotlu, admine, advarka, adverbi, 01:34 < mensi> adverbivla, adverbu, adzau, a'e, a'enai, aftobuso, aftostopi, agbakate, agmasto, ai, a'i, a'icu'i, aicu'i, aidji, aigne, 01:34 < mensi> ainai, a'inai, aiste, aitco, aitxero, aizdo, akcoloti, akmela, akrobat, aksiptrina, akti, alba'aka, albatro, albuma, 01:34 < mensi> alcamacrike, aldeia, alfabeta, alga, algoritma, alkanse, alminiu, alna, alpaka, altamru, altare, altfor, altrici, 01:34 < mensi> alxataini, alzaitu, ambergri, ambigu'o, amble, ambuci, ambulance, amfite'atro, amfora, amgidala, amnio, amniota, 01:34 < mensi> ampigravle, ampute, amsake, amtisti, amxari, anci, andiroba, angeli, angila, anji, ankabuta, ankla, anli, ... 01:34 < gleki> okay, 19 319 words now in JVS 01:34 < gleki> most are ISO-fuhivla, okay. 01:34 < zipcpi> lol 01:34 < gleki> but what to do with them? 01:35 < zipcpi> ambigu'o? I've seen ambigu more often 01:35 < gleki> also this ocean=ve'u xamsi problem 01:36 < gleki> i think i will have to further develop La Bangu dict. i will just tag iso-fuhivla as non core words. 01:36 < zipcpi> I think braxamsi might be able to have a different sense than ve'u xamsi, but I dunno, probably not inda lo ka pagbu la bangu 01:37 < gleki> that difference needs to be explained 01:37 < gleki> but take "ex-president" 01:37 < gleki> which is {lo ba'o jatna} 01:38 < zipcpi> Well difference is that braxamsi could mean "ocean", while ve'u-xamsi could be just any "sea" that's subjectively big 01:38 < gleki> what is ocean then 01:38 < zipcpi> But yeah ex-president probably doesn't need another word 01:39 < zipcpi> Well, there is a somewhat clear separation between the bodies of water that are called "seas", and the bodies that are called "ocean"s 01:39 < zipcpi> Mostly a naming thing 01:44 < gleki> what is that separation about? 01:45 < zipcpi> Like I said, mostly a naming thing 01:45 < gleki> im not sure lojbna should immediately relex english 01:45 < gleki> *lojban 01:45 * zipcpi shrugs... perhaps 01:46 < zipcpi> I mean, yeah, "Atlantic Ocean" probably can just be rammed into xamsi 01:46 < gleki> why not 01:48 < zipcpi> May be just an expression thing. {xamsi ja braxamsi} might sound better than {xamsi ja ve'u xamsi} (is the latter even grammatical?) 01:49 < zipcpi> exp: xamsi ja ve'u xamsi 01:49 < mensi> ([CU {xamsi VAU}] [ja {CU VAU} VAU]) 01:49 < gleki> not in a sense you wanted 01:49 < zipcpi> It is under selpa'i, but it breaks in a way that might or might not be desired lol 01:49 < gleki> off: xamsi je ve'u xamsi 01:49 < mensi> SyntaxError: Expected [,] but "x" found. 01:50 < gleki> it's the same as {xamsi gi'e ve'u xamsi} 01:50 < gleki> under exp. 01:50 * zipcpi nods 01:50 < gleki> off: xamsi ja xamsi be ve'u 01:50 < mensi> ([xamsi {ja }] VAU) 01:51 < Amnadya> the problem is we need a good way to say "the young ex-president". baho jatna isnt a ziltau and using "be" changes the word order 01:52 < gleki> {lo ba'o jatna noi citno} 01:52 < zipcpi> (I tend to prefer poi) 01:52 < gleki> adjectives correspond to NOI clauses in Lojban 01:54 < gleki> oh im so happy to delete that ugly spreadsheet that eventually failed to generate eng2jbo dictionary 01:55 < zipcpi> I've been using ze'oi or xo'i where I really can't be bothered to expand a tanru, but it's a bit cumbersome 01:57 < gleki> if tanru is really needed then it's {lo citno jatna be ba'o} 01:58 < zipcpi> Yep, works, but changes the word order making it somewhat cumbersome for glosses 02:00 < gleki> then {jatna ba'o} 02:01 < gleki> you'll have to add {be} anyway when converting {klama do} into a noun 02:01 < zipcpi> Hm 02:02 < Amnadya> baho zei jatna 02:02 < zipcpi> I do prefer ba'o jatna in the absence of tanru though; it's more natural for sumtcita in general. Dunno what the best solution is here *shrugs* 02:02 < gleki> ba'o zei jatna is a lujvo 02:03 < gleki> which means it can mean anything 02:04 < Amnadya> sure. baho zei jatna: x1 is a wooden table 02:05 < zipcpi> It's kind of why I invented {ze'oi} 02:05 < zipcpi> To avoid having to use zei when wanting to use random cmavo in a tanru 02:07 < zipcpi> As a "bonus", it's also a convenient way to convert cmevla in the absence of brivla-cmevla-merge grammar *shrugs* 02:22 < gleki> exp: li re xi dy cu mo 02:22 < mensi> ([li {re BOI } LOhO] [cu {mo VAU}]) 02:23 < gleki> it's usually 01h55m 02:23 < gleki> de'i li re poi se cacra is long 02:23 < zipcpi> Is this about detri system? 02:23 < gleki> yeah 02:23 < gleki> {za lo masti be li re za lo djedi be li mu} in two months and 5 days 02:24 < zipcpi> I dunno, I just chose prefix over postfix somewhat arbitrarily, but mainly because it parallels FA/sumtcita 02:24 < gleki> even if some de'i'i can take the place of {za} it will be long 02:24 < zipcpi> There's also moi'i 02:25 < gleki> re masti could be short but how to make them splice 02:25 < zipcpi> No wait moi'o 02:26 < gleki> it's just that for each coordinate we need to specify dimension 02:28 < ctefa`o> coi 02:28 < gleki> coi 02:30 < ctefa`o> .i coi la gleki .i xu lo gleki do mintu 02:30 < zipcpi> Go away. No one "pulls my leg" without being kicked in the nuts. No one plays with my emotions without getting more than they bargained for 02:30 < gleki> su'oroi 02:30 < ctefa`o> someone is upset alright 02:31 < ueiudhfi> pls 02:35 < zipcpi> You bet I'm upset. I hate liars, flatterers, and trolls 02:35 < zipcpi> You're not the first idiot who thinks he can just poke at me like some dog in a cage 02:36 < gleki> stop imeediately 02:36 < gleki> *immediately 02:36 < gleki> change to another channel 02:39 < ctefa`o> I thought he had put me on ignore but I will just stop mentioning anything related to him then 02:39 < zipcpi> I did but then you changed your name 02:39 < gleki> change to discussing lojban 02:39 < gleki> or discussing in lojban 02:41 < ctefa`o> I am on my phone! 02:41 < gleki> u'i do cpana lo fonxa 02:42 < ctefa`o> I am on my phone! 02:43 < ctefa`o> ...which lagged a bit and said the message didn't get through? 02:43 < ctefa`o> >_> 02:44 < ctefa`o> If I am bothering you so much just put this nick on ignore too 02:49 < ctefa`o> gleki: su'uroi happens 02:49 < ctefa`o> ... 02:49 < gleki> *su'oroi 02:49 < ctefa`o> su'oooo 02:50 < ctefa`o> si su'o'o'o'o'o'o'oroi 02:51 < ctefa`o> I suppose "fasnu" would work best for that "happens"? 02:52 < gleki> yes 02:53 < ctefa`o> {.i su'oroi fasnu}? not that used to roi 02:54 < gleki> yes 02:54 < ctefa`o> je'e 02:56 * nuzba @Zelts_57AlDBr83: ロジバン [http://bit.ly/1GUbCXk] 02:59 < ctefa`o> .i lo nu mi cpana lo mi fonxa cu su'oroi fasnu 03:01 < b_jonas> ctefa`o: I suggest you try {rapli li su'o} instead of {su'o roi fasnu} 03:01 < ctefa`o> xmm 03:02 < ctefa`o> Something wrong with making that the tense? 03:04 < ctefa`o> {.i lo nu mi cpana lo mi fonxa cu rapli li su'o} is better how? 03:05 < b_jonas> ctefa`o: nothing wrong, I just think it's unnecessaraily complicated for this 03:05 < ctefa`o> je'e 03:06 < b_jonas> And I really don't like {roi} and {re'u} 03:06 < b_jonas> so I don't use them much (but I still understand them) 03:07 < ctefa`o> I suppose roi would expand to something like rapli in this case 03:07 < ghhtzt> stop fighting. sort out your problems or fight elsewhere but please keep this place nice and civil thanks. feho 03:07 < ctefa`o> I am not fighting! 03:08 < ghhtzt> end of discussion. 03:08 < gleki> ctefa`o: e'uru'e do ciksi to bau lo jbobau li'a toi lo nu pu fasnu fa makau 03:09 < gleki> i za'a lo jbopre cu so'ujonai no roi damba casnu 03:10 < ctefa`o> (And someone making up random nicks to say things behind an ignore-wall is kinda pathetic) 03:10 < ctefa`o> fa'o 03:11 < zipcpi> ghhtzt isn't me 03:14 < ctefa`o> gleki: Explain what the past event was? 03:18 < ctefa`o> Explain the {lo nu mi cpana lo mi fonxa}? 03:30 < gleki> "that i am atop of my phone" 03:30 < gleki> literally 03:30 < gleki> ctefa`o: @ 03:33 < ctefa`o> mi zutse ra ? 03:34 < gleki> not necessarily sitting, just on top 03:34 < ghhtzt> ra solri cevni 03:35 < ctefa`o> ja'a 03:37 < ctefa`o> Well, if I sit on it I am on top of it? 03:38 < gleki> yes, but not vice versa 03:40 < ctefa`o> Do you mean explaining the concept of being on top? 03:40 < gleki> if you are on top you are not necessarily sitting on it 03:40 < ctefa`o> I think that's kinda hard to do when someone is not physically present 03:40 < ctefa`o> Iep 03:42 < gleki> you may be lying on top of it 03:43 < ctefa`o> je'u 03:44 < ctefa`o> Made me think of trying to explain colors 03:45 < ctefa`o> It only reallly works when you have other colors to refer to 03:46 < zipcpi> Wuzzy made {nulpa'a} a synonym of {kanpe}... heh... 10 years 03:47 < gleki> утЖ тгдзфэф 03:47 < gleki> en: nulpa'a 03:47 < mensi> nulpa'a [< nutli pacna ≈ Neutral wish] = p1 expects/looks for the occurence of p2 (event), expected likelihood p3 (0-1); 03:47 < mensi> p1 subjectively evaluates the likelihood of p2 (event) to be p3. |>>> The value of x3 is a subjective estimate of 03:47 < mensi> likeliness according to x1, and is the basic determinant of whether kanpe means something like "hope" or "wish" or " 03:47 < mensi> expect", although kanpe never carries the connotation of desire; for that connotation see pacna. nulpa'a with x3 not 03:47 < mensi> very close to 1 has no simple equivalent in English, but for objects/states with negligible expectation it is something 03:47 < mensi> like "wishing"; if the state is plausibly likely, it is something like "hoping". In both cases, though, the English 03:47 < mensi> implication of emotional desire is no... 03:47 < mensi> [mo'u se katna] http://jbovlaste.lojban.org/dict/nulpa'a 03:48 < ghhtzt> pehi loka kanpe na srana loka pacna 03:49 < gleki> he seems to not have removed kanpe 03:49 < gleki> from notes 03:49 < zipcpi> Yeah... 03:55 < gleki> what was selpahi suggesting about mixed lerfu/number strings? somethingto do with ana phora but what exactly 03:55 < zipcpi> Er I don't think he suggested numbers? 03:55 < gleki> he suggested fixing anaphora instead 03:55 < zipcpi> I think he wanted to redefine a set of lerfu sumka'i to mean particular classes of nouns, by default 03:56 < gleki> i.e. {*abu* cu za'ure'u cusku} 03:56 < gleki> maybe that 03:56 < gleki> well, no since that is not relevant 03:56 < zipcpi> Don't think he did anything about the {boi} problem that xornunsep is trying to fix 03:56 < gleki> once he suggested fixing the meaning of letterals 03:57 < gleki> but that doesnt fix mixed strings 03:57 < zipcpi> Exactly 03:57 < gleki> fy = le fetsi etc. 03:57 < zipcpi> Yep 03:57 < zipcpi> Although he himself avoids le lol 03:57 < gleki> and i dont udnerstand wy xorxes wants that half-solution. he never replied to my {re boi remoi} counterexample 03:58 < zipcpi> I find that less problematic because I find I'm able to detect two numbers running together better 03:59 < gleki> x is a variable representing a number 03:59 < zipcpi> But while using Lojban I've often caught myself making that abu za'u mistake 03:59 < zipcpi> And then I say 03:59 < zipcpi> la xornunsep za'ure'u jinga 03:59 < gleki> it sounds like fixing system based on current misusage 04:00 < gleki> then let's fix {lo mlatu pinxe} killing tanru 04:00 < zipcpi> You could say about the same thing about dotside though 04:01 < gleki> i will just never use letterals anaphorically then 04:01 < gleki> if that bothers anyone 04:02 < zipcpi> le nu ku ; le abu ku ; *shot* 04:02 < gleki> le abu mo 04:03 < gleki> i sefsi la'a 04:05 < gleki> http://cdg.pleolight.net/Rgjh 04:05 < gleki> Blog : Puku naku wallpapers 04:06 < zipcpi> Maybe that would actually convince selpa'i of {le}; it's not too different from his {le naku} idea... hah 04:06 < ghhtzt> off: ny moi 04:06 < mensi> ([ny moi] VAU) 04:07 < gleki> exp: ny moi 04:07 < mensi> (CU [{ny BOI} moi] VAU) 04:07 < gleki> exp: ny su'i re moi 04:07 < mensi> (CU [{ } moi] VAU) 04:07 < ghhtzt> there is a proposal to merge PA and BY into PA. why can we say pamoi and pa lo broda but nymoi and not ny lo broda 04:07 < zipcpi> What kind of website is that? Looks like just a bunch of stuff glued together for some nefarious SEO purpose 04:07 < gleki> je'usai 04:08 < gleki> "n broda", why not 04:08 < zipcpi> exp: ny broda 04:08 < mensi> ([ny BOI] [CU {broda VAU}]) 04:08 < gleki> n of broda* 04:08 < zipcpi> exp: mo'e ny broda 04:08 < mensi> ([{ BOI} broda KU] VAU) 04:09 < zipcpi> Who's trying to break grammar now? lol 04:09 < gleki> li mo'e dy re mo'e ly re 04:09 < gleki> exp: li mo'e dy re mo'e ly re 04:09 < mensi> ([li {<(¹mo'e [dy BOI] TEhU¹) (¹re [mo'e {ly BOI} TEhU] re¹)> BOI} LOhO] VAU) 04:10 < gleki> exp: li re mo'e dy re mo'e ly 04:10 < mensi> ([li { BOI} LOhO] VAU) 04:10 < zipcpi> That works, but xu smudra 04:10 < zipcpi> And a little long 04:11 < ghhtzt> xornunsep and selpamnunpav 04:12 < zipcpi> Heh 04:14 < zipcpi> And my own crazy idea is to relax the gadri rules, so that {le ku} can be a nonce deixis 04:14 < zipcpi> Not sure how {le nu ku} would work without going full alta though 04:17 < zipcpi> alta: le nu ku 04:17 < mensi> ([FA {le ku}] [CU {COhE VAU}]) 04:18 < zipcpi> alta: le na ku 04:18 < mensi> SyntaxError: Expected ['h] but end of input found. 04:18 < gleki> oh 04:18 < zipcpi> Oops 04:18 < gleki> file a bug 04:18 < gleki> alta: le ku 04:18 < mensi> ([FA {le COhE ku}] [CU {COhE VAU}]) 04:20 < gleki> btw if someone doesnt know there is a tool to test your PEG grammars online http://mw.lojban.org/extensions/ilmentufa/mahantufa/ircitufa.html 04:22 * nuzba @uitki: jbonunsla - http://mw.lojban.org/papri/jbonunsla by Pier.abat - /* Logfest 2015/jbonunsla 2015 */ date and place [http://bit.ly/1JGxeum] 04:22 < gleki> that bug is limited to NA_clause 04:23 < gleki> wasnt NA_clause supposed to be merged with tense_modal? 04:24 < ghhtzt> zona na sumtcita 04:25 < gleki> ma prali 04:25 < gleki> alta: lo na ku 04:25 < mensi> SyntaxError: Expected ['h] but end of input found. 04:27 < gleki> alta: lo na ku 04:27 < mensi> ([FA {lo ku}] [CU {COhE VAU}]) 04:27 < gleki> what i didnt understand is that why tags were at selbri level but NA_clause at selbri_1 level 04:28 < gleki> now they are both at selbri level in alta 04:28 < gleki> ni'o since {lo trene} means "carvan" how to say "railroad train" then? 04:29 < gleki> trene lo karce? 04:29 < gleki> trene lo carce? 04:31 < Amnadya> laha 04:33 * nuzba @oka_iu_tcan: http://jbovlaste.lojban.org/dict/selbrisle おおおおおおお! selbrisle !(個人的に革命的単語) [http://bit.ly/1JGz8ev] 04:39 < Amnadya> mi pu facki fi soho nabmi pe la itkuhile i je mi xanka lonu la kiiadas ba spusku loseduhu na nabmi kei jeba cikre noda 04:40 < gleki> do du sy ibo xu 04:40 < gleki> i jaho na gohi 04:41 < Amnadya> gohi i zvati lo xamsi jibni 04:42 < Amnadya> i vitci ke fonxa kibro 04:42 < gleki> uasai 04:42 < gleki> i do e la cirko zoha pilno zohoi h 04:42 < gleki> sa 04:43 < gleki> i do e la cirko coha pilno zohoi h 04:43 < gleki> i ei mi stidi da pe la itkuhile 04:43 < gleki> github.com => repo => Fork 04:43 < gleki> pe'a 04:43 < gleki> u'i 04:44 < gleki> i ji'a ko jmina la to'anzu lo do se forki sorcu i taiku do finti lo cnino je kalsa glekitufa si bangu 04:45 < gleki> i xu do zvati la xunre xamsi 04:47 < Amnadya> la stuna xamsi kune la dotyguhe 04:48 < gleki> uanai 04:48 < gleki> i xu zasti 04:48 < Amnadya> berti ikuhi zo stuna cmene 04:48 < gleki> i de'i ma lo gugdepulu co'a xamsi 04:48 < gleki> i ua 04:50 < gleki> in interstellar communications even from/to the Earth/Moon 3 second delay between messages will make {go'i} useless since you wont be able to understand what you were referring to unless everyone waits for three seconds 04:50 < zipcpi> u'i 04:50 < zipcpi> Use go'oi for a little more safety 04:50 < zipcpi> Not completely safe though 04:51 < gleki> en: go'oi 04:51 < mensi> go'oi = [GOhOI] pro-bridi: quotes the next word and repeats the most recent bridi containing that word |>>> The quoting 04:51 < mensi> rules are the same as zo. |>>> durka42 04:51 < gleki> li'ei isnt in peg 04:52 < gleki> go'oi isnt in peg 04:52 < zipcpi> Yeah... 04:52 < zipcpi> GOhOI needs to be though, to canonize {ze'oi} 04:52 < zipcpi> li'ei is just short for {li pe'o joi'i} 04:53 < gleki> and i cant see how li'ei is different from li. why not making li as powerful as li'ei? how does it affect abu za'ure'u? 04:53 < zipcpi> Ask xorxes lol 04:53 < gleki> i think i will restore the original system back 04:54 < zipcpi> Like I said I support xornunsep *except* li...lo'o 04:54 < zipcpi> So to me {li} works just like {li'ei}... but obviously xorxes disagrees 04:54 < gleki> xornunsep isnt in LMW 04:55 < zipcpi> Yeah but it is in exp 04:55 < zipcpi> Maybe we need to split exp for the "common but not completely canonized", and the "actual experimental" parser 04:56 < zipcpi> So that stuff like xornunsep can go into exp without affecting the kampu usage 04:58 * nuzba @uitki: L17-01 - http://mw.lojban.org/papri/L17-01 by Gleki - /* Basic notions */ [http://bit.ly/1HAMNEt] 05:01 < zipcpi> {ze'oi} is just indispensible if we want to stem the tide of {zei}-abuse :p 05:16 < gleki> en: ze'oi 05:16 < mensi> ze'oi = [GOhOI] converts following word to selbri: "x1 is related to the meaning of this word in aspect x2" |>>> Quoting 05:16 < mensi> rules are similar to zo. Roughly, (ze'oi *coi*) ≈ (me la'e zo *coi* moi). Meant primarily to quote cmavo, to supersede 05:16 < mensi> use of zei as a way of making "tanru" out of cmavo. |>>> spheniscine 05:16 < zipcpi> cmavo/cmevla -> selbrisle 05:17 < zipcpi> Just a really rough conversion, with no very-precise meaning 05:35 < amnadya1> attawîl! 05:39 < gleki> ithkuil: attawîl 05:39 < ithkuil> attawîl: P1S1/STA-tt-OBL-DEC/FAC-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 05:41 < amnadya1> melbi ikuhi laha sidju jihinoda 05:42 < gleki> zo'oi ifl mo 05:43 < gleki> amnadya1: @ 05:45 < amnadya1> informal designation 05:46 < zipcpi> lu melbi ikuhi laha sidju jihinoda li'u u'i 05:46 < gleki> dec = declarative? 05:47 < gleki> i go'i 05:47 < gleki> ithkuil: a'ctattal 05:47 < ithkuil> a’ctattal: P1S1/STA-CTX/PPS/DEC-MNO-tt-OBL-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 05:47 < gleki> oh i almost got there 05:47 < ctefa`o> We have a ithkuil bot? o_O 05:49 < gleki> hm, but is there a default phase? 05:49 < zipcpi> ithkuil: akt’asalb 05:50 < gleki> very strange. i combined something but with phase and sanction 05:50 < zipcpi> http://www.ithkuil.net/sound_files/akarenina.mp3 Erm, do you need me to call the doctor? :p 05:50 < zipcpi> doi bacru .i xu do nitcu lo mikce 05:50 < gleki> wth is this attawil 05:51 < zipcpi> What's attawil? 05:51 < zipcpi> http://www.ithkuil.net/texts.html 05:51 < gleki> see backlog 05:52 < zipcpi> I don't know what it means... oh... is it "hello"? 05:52 < gleki> ctx/pps/dec/mno - how to remove that and how to add dec/fac after OBL? 05:53 < amnadya1> wë 05:54 < gleki> sanction, valence and phase are to be removed and illocution is to be moved 05:54 < gleki> ithkuil: a'ctattal 05:54 < ithkuil> a’ctattal: P1S1/STA-CTX/PPS/DEC-MNO-tt-OBL-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 05:54 < gleki> ithkuil: wë 05:54 < ithkuil> wë: Personal adjunct 05:55 < gleki> http://mw.lojban.org/extensions/ilmentufa/ithkuil/index1.html <-- it's just broken 05:56 < gleki> ithkuil: attawal 05:56 < ithkuil> attawal: P1S1/STA-tt-OBL-ASR/SUB-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 05:56 < gleki> ithkuil: attawîl 05:56 < ithkuil> attawîl: P1S1/STA-tt-OBL-DEC/FAC-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 05:57 < niftg> ta'o lo banzu ke itku'ile gentufa nu'o xu se finti 05:57 < gleki> ithkuil: attal 05:57 < ithkuil> attal: P1S1/STA-tt-OBL-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 05:57 < amnadya1> attál = attàwël 05:58 < gleki> ithkuil: a'ctattal 05:58 < ithkuil> a’ctattal: P1S1/STA-CTX/PPS/DEC-MNO-tt-OBL-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 05:58 < gleki> attal changes to a'ctattal when changing illocution from assertive to DEClarative 05:59 < gleki> i suppose this might be a bug since table 9 disagrees http://ithkuil.net/05_verbs_1.html#Sec5o6 06:00 < amnadya1> ie naldra 06:42 < gleki> mensi: doi amnad da djica lo ka se ctuca fi la lojban vau gi'e bredi lo ka pleji 06:42 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.amnad.gy. di'a cusku da 06:52 < gleki> en: so'iroi 06:52 < gleki> en: piso'iroi 06:53 < mensi> so'iroi = [ROI*] tense interval modifier: many times; objective tense; defaults as time tense. |>>> 06:53 < mensi> officialdata 06:53 < mensi> piso'iroi = [ROI*] tense interval modifier: often; objectively quantified tense; defaults as time tense. |>>> 06:53 < mensi> officialdata 06:54 < gleki> and those are official data 06:54 < gleki> another question. shall we finally have a BAI for {fi'o se cpana}? 07:18 < gleki> You ought to have been there; it was great fun. 07:18 < gleki> ^ translate that 07:18 < gleki> ought to = Used to indicate desirability 07:24 < gleki> "have to do with". looks like the shortest word for that is ... {gleua} 07:30 -!- Irssi: #lojban: Total of 154 nicks [2 ops, 0 halfops, 0 voices, 152 normal] 09:09 < Zearen> .i coi 09:10 < dutchie> coi 09:13 < Zearen> .i ma nuzba 09:20 < durka42> mi pinxe lo ckafi 09:20 < dutchie> ei mi zdani gunka ca bavlamjetfu .i mi na nelci ki'u lo nu mi na'e kansa lo prenu 09:20 < mensi> i mi i mi i mi mo i mi na nelci 09:21 < dutchie> ca lo* 09:21 < gleki> alta: coi do'u ma nuzba 09:21 < mensi> ([coi DO do'u] [FA ma] [CU {nuzba VAU}]) 09:21 < durka42> za'a la .cogas. cu finti lo so'o cizra cmavo 09:21 < gleki> {do} autorestored in voactives now 09:22 < gleki> in empty vocatives sa'e 09:22 < gleki> alta: coi i ma nuzba 09:22 < mensi> ([coi DO DOhU] [i {FA ma} {CU }]) 09:23 < Latro__> coi 09:24 < Zearen> .i mi na nelci lo ckafi .imu'ibo cy. dukse loni kurki 09:24 < mensi> ji'a mi mutce nelci i ie 09:24 < Zearen> .y. doi mensi do cizra 09:26 < latro`a> .i coi zy 09:26 < Zearen> .i doi la'oi durka42 cizra fi ma 09:26 < Zearen> .i coi ly. 09:26 < durka42> cizra fi lo ka cnino genstura 09:27 < latro`a> .i ma cizra 09:27 < gleki> i also dont like this "vocative relative_clauses? selbri" 09:27 < gleki> i think LE has to be restored there 09:27 < Zearen> .i lonu lonu do se kulnu lo lojbo cu ranji cu pluka mi doi ly. 09:29 < Zearen> .i dy. pu cusku lo se du lo cizra cmavo poi se finti la .cogas. 09:30 < durka42> latro`a: xu do zgana zo co'ei je zo be'oi je zo bei'oi 09:30 < durka42> srana lo sumtcita valsi 09:39 < gleki> alta: coi broda 09:39 < mensi> ([coi {LE broda KU} DOhU] [FA ZOhE] [CU {COhE VAU}]) 09:39 < gleki> LE autorestored 09:39 < gleki> alta: coi noi du vau broda noi mo 09:39 < mensi> ([coi {LE <(¹noi [{FA ZOhE} {CU }] KUhO¹) (¹broda [noi {FA ZOhE} {CU }] KUhO¹)> KU} DOhU] [FA ZOhE] [CU {COhE VAU}]) 09:40 < gleki> let it be LO 09:40 < gleki> alta: coi broda 09:40 < mensi> ([coi {LO broda KU} DOhU] [FA ZOhE] [CU {COhE VAU}]) 09:41 < gleki> and of course technically {coi selbri} is now banned 09:42 < gleki> it's only by virtue of LO autorestoration that it still works 10:15 < Ilmen> mensi: da poi bu'a 10:15 < mensi> mi na birti lo nu mi mulno jimpe lo se smusku be do 10:15 < Ilmen> exp: da poi bu'a 10:15 < mensi> ([da {poi KUhO}] VAU) 10:16 < samcinki> alta: da poi bu'a 10:16 < mensi> ([FA {da }] [CU {COhE VAU}]) 10:16 < Ilmen> Huh. Why FA ZOhE here? 10:17 < Ilmen> Does ZOhE covers ke'a? 10:18 < Ilmen> alta: mi ckaji lo ka broda 10:18 < mensi> ([FA mi] [CU {ckaji } KEI] KU¹)> VAU}]) 10:18 < gleki> ke'a isnt supported yet 10:18 < gleki> neither ce'u 10:19 < Ilmen> Also, nullary predicates may exist, so maybe autoinserting FA ZOhE is unnecessary 10:19 < gleki> im not sure they can exist 10:20 < gleki> broda1 is assumed by PEG 10:39 < gleki> okay bugs into ithkuil word generator are coming 10:40 < Zearen> .i lo .alta cu mo 10:40 < Zearen> sa .i la .alta. cu mo 10:46 < ctefaho> It seems bridi-tail tail-terms are not supported in the simplified connective system? 10:47 < ctefaho> after gi'e was cut out 10:47 < durka42> really? 10:47 < durka42> that should be a bug in the parser if so 10:47 < Zearen> gije = gi'e, so what ? 10:48 < Zearen> Is there a new parser being written for fancylojban ? 10:49 < ctefaho> ah I was mistaken 10:49 < ctefaho> nvm 10:49 < ctefaho> je-cu behaves exactly like gi'e then 10:49 < ctefaho> ki'e durka 10:51 < ctefaho> Zearen: http://selpahi.weebly.com/lojban/how-to-substantially-simplify-the-lojban-connective-system-my-connective-system 10:52 < durka42> as far as I know the experimental camxes supports GIhE, GIJE, and JECU 10:52 < Zearen> I was actually around for xalbo's balni ke jorne valsi galfi or some such 10:52 < durka42> (I was surprised that they are all compatible) 10:53 < Zearen> Er, not xalbo, xorxes 10:54 < Zearen> Except it was very new a controversial. I'd be surprised if it's still not controversial. 10:54 < durka42> well that's always going to depend on who you ask :) 10:55 * durka42 is happy using {je cu} 10:55 < durka42> iepei doi la mensi 10:55 < mensi> mi tugni i ie mi tugni 10:56 < Zearen> .i mu'i lonu la mensi cu tugni drani li'a gije banli 11:01 < latro`a> alta: .i jetnu fa pa me lo du'u broda jo'u lo du'u brode 11:01 < mensi> (i [FA ZOhE] [CU {jetnu KU} {jo'u }] KEI²) KU>}] MEhU¹) KU> VAU}]) 11:01 < latro`a> cool, that does what I wanted 11:04 < durka42> that first {fa zo'e} seems wrong 11:08 < latro`a> I'm not sure why it does that when you fill the fa place afterward 11:09 < latro`a> (that wasn't what I wanted to get out of it, though; I wanted to be sure the jo'u worked the way I expected) 11:09 < latro`a> (this was for zipcpi's mriste post from a few days ago) 11:09 < latro`a> incidentally, did we ever decide on a word for the connective "fold" operator? 11:11 < latro`a> that is, something like NEW_WORD CONNECTIVE ARG1 ARG2 ... ARGn NEW_TERMINATOR -> ARG1 CONNECTIVE ARG2 ... CONNECTIVE ARGn? 11:14 < durka42> liberating {pe'o} from its mekso confines, eh 11:16 < latro`a> I pretty much pretend the MEX cmavo except maybe su'i/vu'u/pi'i are unallocated, these days 11:17 < durka42> :) 11:18 < durka42> well maybe {pe'o} could be this NEW_WORD? 11:18 < durka42> is it possible, though? it seems like there are going to be a bunch of edge cases with termsets 11:18 < durka42> not that it would be terrible if termsets died in a fire 11:30 * nuzba @s_i_d_a: lo sicpi sance cu pagre lo kertinytci .ije menli cilmo #lojban #haiku [http://bit.ly/1NCinC5] 11:49 < gleki> that fa zo'e isnt wrong. 11:50 < gleki> simplu la alta asserts that bridi haed cant be empty of FA 11:50 < gleki> simply la alta asserts that bridi haed cant be empty of FA 11:50 < gleki> following how linkargs work (starting from x2 since x1 is always filled) 11:50 < gleki> oh, ilmen left 11:52 < gleki> mensi: doi lmen i m not sure how ke'a autorestoration can be implemented in peg parsers. consider {mi noi klama do noi ke'a ke'a du}. what each of ke'a refer to? to different {noi} or to the same? another question is {mi noi klama do noi du ke'a}. should {noi du} be restored into {noi ke'a du} or the only ke'a refers to the second {noi} already? 11:52 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.lmen.gy. di'a cusku da 11:54 < gleki> latro`a: im sure selpahi added that word into jvs 11:54 < gleki> selmaho:me 11:54 < mensi> .i lu me li'u cmavo zo'oi ME 11:54 < mensi> http://lojban.github.io/cll/5/10/ 11:54 < mensi> http://lojban.github.io/cll/18/1/ 11:54 < mensi> cmavo: me, me'au, mei'o, me'oi, me'u 11:54 < durka42> I agree, I don't think "find the first open sumti slot" can be expressed in a PEG 11:55 < gleki> i just dont know what is the algorithm when ke'a can be embedded deep inside embedded clauses 11:57 < gleki> it's not about autorestoration of ke'a but about which ke'a refers to which clause 11:58 < durka42> oh 11:58 < durka42> pe'i it's always the innermost 11:58 < durka42> you can subscript to get outwards 12:00 < gleki> the please answer my two questions 12:01 < cliva> en: mei'e 12:01 < mensi> mei'e = Accepts any number of sumti and turns them into a selbri-unit that means "x1 is among the referents of these 12:01 < mensi> sumti". |>>> See me, menre, cmima. Elidable terminator: mei'o |>>> spheniscine 12:01 < durka42> mi noi klama do noi ke'a ke'a du -- both ke'a refer to du 12:01 < durka42> sorry 12:01 < durka42> mi noi klama do noi ke'a ke'a du -- both ke'a refer to do 12:01 < durka42> mi noi klama do noi du ke'a -- du1 is unspecified, du2 is do 12:02 < gleki> but the first {noi} should search for its closest ke'a 12:02 < durka42> I think of it as ke'a search for noi, heh :) 12:03 < gleki> it should be formalized and described somewhere 12:03 * durka42 glances over the BPFK sections 12:04 < durka42> vlaste: ke'a (selma'o) 12:04 < vlaste> ke'a (class) = KOhA7 12:04 < durka42> In the case of nested relative clauses, ke'a or ke'a xi pa refers to the inner-most sumti to which the relative clause containing it is attached. ke'a xi re refers to the next outer-most relative clause attached sumti, and so on for higher numbers. 12:06 < gleki> nandu 12:07 < gleki> even more nandu when ke'a is omitted 12:07 < durka42> tersmus: mi noi klama do noi ke'a ke'a du cu co'e 12:07 < tersmus> klama(mi,do); (do == do); co'e(mi) 12:08 < durka42> tersmus: mi noi klama do noi du ke'a cu co'e 12:08 < tersmus> klama(mi,do); ( == do); co'e(mi) 12:08 < durka42> in this case I agree with la tersmus 12:08 < durka42> tersmus: mi noi klama do noi du ke'axire cu co'e 12:08 < tersmus> klama( ,do); FA x1. (do == x1); co'e(mi) 12:08 < durka42> wat 12:08 < durka42> .u'i 12:24 < gleki> so is the rule that in the absense of ke'a the first unfilled slot of the noi clause is to be filled with ke'a? 12:25 < durka42> in my view yeah 12:25 < durka42> I think that's the rule 12:25 < durka42> well, _unless_ a ke'a is explicitly specified 12:25 < durka42> then that takes precedence over the "search for the first slot" rule 12:37 < gleki> alta: lo brodi broda cu mo 12:37 < mensi> ([FA {lo KU}] [cu {mo VAU}]) 12:37 < gleki> alta: lo brodi broda ui cu mo 12:37 < mensi> ([FA {lo KU}] [cu {mo VAU}]) 12:38 < gleki> indeed NF is like a UI 12:40 < gleki> now {dansu NF} will be glossed as "dancer" and all other cases as "dances" 12:59 * mumble jorne co'a fa la'o fa menli fa 13:36 * nuzba @Rodericus: @Eleder_ Solo se me ocurre en la lengua construida lojban, donde hay varias formas decir «no» como negación… total (hay otras negaciones) y [http://bit.ly/1Cg4KIm] 13:48 < cliva> doi lo djuno be lo du'u mo kau .i lu ko'a tersumei li su'o li'u se smuni ge'i lo du'u su'o dubmau be li pa cu se pormei ro terbri be ko'a gi lo du'u ro terbri co'e su'o dubmau 13:48 < cliva> sisisi cu co'e su'o dubmau 13:50 < durka42> en:tersumei 13:50 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 13:50 < durka42> en:tersu'imei 13:50 < mensi> tersu'imei [< te sumti mei ≈ 3rd conversion argument cardinal selbri] = x1 (property/relation) is x2-ary/ 13:50 < mensi> -adic 13:51 < durka42> mi na jimpe 13:51 < zipcpi> {te'i'o}?? What have I done 13:51 < Ilmen> li'a .i glibau velski 13:51 < mensi> Ilmen: cu'u la'o gy.gleki.gy.: i m not sure how ke'a autorestoration can be implemented in peg parsers. consider {mi noi 13:51 < mensi> klama do noi ke'a ke'a du}. what each of ke'a refer to? to different {noi} or to the same? another question is {mi noi 13:51 < mensi> klama do noi du ke'a}. should {noi du} be restored into {noi ke'a du} or the only ke'a refers to the second {noi} 13:51 < mensi> already? | 2015-07-03T18:52:54.625Z 13:51 < durka42> cliva: se smuni lo du'u bridi fi su'o pa da 13:52 < zipcpi> I think ke'a always refer to the same referent, unlike ce'u 13:52 < cliva> .. ia nai 13:52 < zipcpi> At least, within one NOI clause 13:52 < durka42> zipcpi: ie 13:52 < cliva> .i fi lo pormei be li su'o pa 13:52 < durka42> iebu'o 13:52 < cliva> .i ku'i ma sko'opu stura .i xu .ambigu sa'u 13:53 < durka42> lojbo .i na .ambigu .u'i 13:53 < zipcpi> ca'e na ambigu fe'o fe'o fe'o 13:53 < cliva> le zei le 13:54 < durka42> lu le zei le li'u du zo le 13:54 < Ilmen> mensi: doi la gleki In {mi noi do viska fo'u noi ke'a ke'a}, both ke'a refers to the same referent. Any new NOI reassign ke'a's referent, till the end of the NOI clause. To access the previous ke'a-referent, {ke'a xi re} must be used. 13:54 < mensi> Ilmen: mi ba benji di'u ba lo nu la'o gy.la.gy. di'a cusku da 13:54 < zipcpi> ue mi jinvi lo du lu le zei le li'u selbrisle 13:54 < cliva> ma'i lo fadni gerna cu go'i 13:55 < Ilmen> doi la cliva lu ko'a tersu'imei li su'o li'u zo'u li'a co'e lo du'u naku su'o kunti porsi cu se zilbri ko'a 13:55 < Ilmen> .i va'i sa'unai su'o da sumti 13:56 < Ilmen> .i lo da'i sumti memkai cenba selbri zo'u lu ko'a tersu'imei za'u da li'u .e'u 13:57 < cliva> je'e 13:57 < Ilmen> .i xu lo se cusku be mi cu jai rinka lo nu do co'a djuno lo voi do pu kucli 13:57 < cliva> ja'a go'i 13:57 < Ilmen> je'e 13:57 < cliva> .i'o 13:57 < Ilmen> je'e ŭi 13:57 < bigcentaur> .i coi ro do 13:57 < zipcpi> oi za'umoi ke mekso cmavo 13:58 < cliva> kentauru co la co coi 13:58 < zipcpi> xaksu piro le cmavo canlu 13:58 < bigcentaur> lu co la co li'u ki'a 13:59 < zipcpi> lo zasni gerna pe la cliva iau u'i 13:59 < cliva> va'o su'o gerna zo co ka'e se pilno fi lo cizra 13:59 < cliva> alta: lo djedi co pu zi 13:59 < mensi> ([{ co } {FA ZOhE}] [CU {COhE VAU}]) 13:59 < zipcpi> exp: kentauru co la co coi 13:59 < mensi> SyntaxError: Expected ['h] but "i" found. 13:59 < zipcpi> alta: kentauru co la co coi 13:59 < mensi> SyntaxError: Expected ['h] but "i" found. 14:00 < cliva> (to ma'oi le zo'u nandu fa lo ka jmina .i se sarcu lo ka jdice so'o da pe la sumti rebla toi) 14:00 < zipcpi> http://jbovlaste.lojban.org/dict/ambigu ma du fobu ki'a 14:02 < zipcpi> "x1 is ambiguous, being possibly x2 or x3, according to observer x4? 14:02 < cliva> "me lo ka ganse jo'u lo ka te smuni jo'u lo simsa" 14:03 < cliva> .i mi na djuno lo du'u ma kau krinu lo nu pagbu lo terbri stura 14:03 < zipcpi> u'i 14:05 * nuzba @Eleder_: @Rodericus Muy bueno! Pensaba más bien en lenguas naturales, pero es perfecto :) Un día tengo que aprender más de lojban... [http://bit.ly/1GVA4aM] 14:21 < cliva> lo ze'irsaivi cu ci'i si'au mei 14:23 < cliva> ta'o 14:23 < cliva> http://jbovlaste.lojban.org/dict/fau'e 14:24 < cliva> .i xu ja'a ka'e ku lo cmavo na se tcita fi lo ka ma kau selma'o 14:40 * mumble jorne co'u fa la'o fa menli fa 14:43 < zipcpi> doi cliva .i ja'a .i la'a se krinu lo nu gau frili lo na'e certu be tau lo selma'o 14:44 < dutchie> .ui mi jinga fo lo caxmati .i ze'upu na jinga 14:50 < dutchie> .i mi bazi sipna .i co'o rodo 14:52 < Amnadya> coho 14:52 < mensi> Amnadya: cu'u la'o gy.gleki.gy.: da djica lo ka se ctuca fi la lojban vau gi'e bredi lo ka pleji | 2015-07-03T13:42:31. 14:52 < mensi> 411Z 14:52 < zipcpi> co'o doi dutci 14:52 < zipcpi> vlaste: dutci (gimka) 14:52 < vlaste> zo dutci gimkamsmikezypro zo tutci noi catni gismu .i zo dutci gimkamsmikezypro no cipra gismu 14:52 < mensi> vlaste: cu'u la'o gy.Ilmen.gy.: gleki In {mi noi do viska fo'u noi ke'a ke'a}, both ke'a refers to the same referent. 14:52 < mensi> Any new NOI reassign ke'a's referent, till the end of the NOI clause. To access the previous ke'a-referent, {ke'a xi re} 14:52 < mensi> must be used. | 2015-07-03T20:54:17.948Z 14:52 < zipcpi> u'i 14:52 < vlaste> 2362 results: smugau, cuntu, gleki, kevna, latmo, notci, viska, bitmu, bongu, boxfo… 14:53 < zipcpi> .y. 14:53 < zipcpi> vlaste: dutci (gimka) 14:53 < vlaste> zo dutci gimkamsmikezypro zo tutci noi catni gismu .i zo dutci gimkamsmikezypro no cipra gismu 14:53 < zipcpi> simlu lo ka la mensi cu ji'isre lo du la gleki cu cusku kei oi 14:54 < zipcpi> .y. srera lo ka sko'opygau le cnima'o 14:55 < Amnadya> da du ma doi la gleki ixu la rutytar 14:55 < zipcpi> la dutci na du la rutytar 14:56 < Amnadya> liha 14:56 < zipcpi> jo'au tcekitau na du'u u'i 15:04 < durka42> .u'i la .krtis. cu spusku 15:04 < durka42> http://jbovlaste.lojban.org/comments.html?valsi=27179;natlangword=0;commentid=1464;definition=0 15:06 < zipcpi> ku'i mi na certu lo mekso xoi banzuni lo ka jimpe 15:06 < zipcpi> ci'iba'umei cmavo porsi 15:07 < durka42> ie 15:09 < zipcpi> ki'aki'aki'aki'a 15:10 < zipcpi> ki'a'o ki'a'e ki'a'i ki'a'au 15:10 < zipcpi> ki'a'a'o'e'i'u'a'e 15:11 < durka42> clearly we need a series of ki'a'V for explaining just what it is you're confused about 15:11 < zipcpi> u'i 15:12 < durka42> {ki'a'y} is reserved for forgetting which ki'a'V applies 15:12 < zipcpi> ki'a'o'e - elliptical confusion 15:16 < zipcpi> ei lo ma'ormekso cu se daspo .i zabna fa lo brimekso 15:16 < zipcpi> a'o ze'u jmive 15:22 < zipcpi> zo te'i'o ki'a .i si'au zo xe'e co'a patfu pe'a lo cmavo poi za'uva'ei se ki'anmo 15:26 < zipcpi> ... mo le cnino jefydei ciste poi se finti la vonxlu 15:26 < zipcpi> skari srana 15:26 < durka42> mo ra'o 15:26 < durka42> ei mi tcidu lo mriste 15:26 < zipcpi> http://jbovlaste.lojban.org/dict/xundei?bg=1;langidarg=2 15:26 < durka42> za'a la .bab. cu nerkla lo nuncasnu 15:26 < durka42> xusra lo du'u lo skari cu kulnutli 15:27 < durka42> si klunu'i 15:27 < zipcpi> ua xu srana lo tanbargu 15:27 < durka42> mi na djuno 15:27 < durka42> ti'e casnu vecu'u lo mriste 15:32 < zipcpi> mi na tolcri le nu casnu 15:34 < durka42> it's buried on the "jbonunsla" thread for some reason, .u'i 15:35 < durka42> co'o 15:41 < zipcpi> I can't remember these weird day systems though. I'm very culturally biased toward the number system because of what I learned from Chinese 15:41 < zipcpi> I'm fine with calling Sunday nondei instead of zeldei but that might not work for NLDCMS so well 15:43 < zipcpi> It is based on the rainbow 15:44 < Amnadya> im very used to monday = 1, so it annoys me that ithkuil has sunday as 1. thats just not how i count 15:45 < zipcpi> Yeah 15:45 < zipcpi> Portuguese has sunday = 1 too though 15:47 < zipcpi> NLDCMS is basically a relex of ISO 8601 to be more usable for Lojban 15:49 < zipcpi> So the days of the week is based on their week-year calendar, which uses Monday = 1 16:09 < zipcpi> zo ro'au'o oi 16:17 < zipcpi> Maybe {xe'e} can be thought of as the antithesis of {mo'oi} 16:18 < zipcpi> It specifically says, "hey, {mo'oi} is not a meaningful question here!" 16:18 < zipcpi> And when used as an answer to {mo'oi}, it's like a {na'i}, but with a more specific meaning 16:19 < zipcpi> "Yes, I want apples, but I'm not picky about which ones" 16:26 < Zearen> .i xu ranji fa lonu lojbo tavla do'e la .mymbl. 16:26 < zipcpi> su'oroi 16:29 < zipcpi> {mo'oi}: Please give me more information so I can identify the referent. {xe'e}: I don't care about the specific referent. 16:36 < durka42> I don't understand krtis' la'e'au at all 16:37 < zipcpi> I *think* he means to refer to specific names as grammar systems 16:37 < durka42> but... it's in LAhE 16:37 < zipcpi> So la'e'au la xorlo, la'e'au la tcekitaujau, etc 16:38 < durka42> but 16:38 < durka42> how is {la xorlo} different from {la'e'au la xorlo} 16:38 < zipcpi> It's to make it a sumti, while {jo'au} actually code switches, I think 16:38 < durka42> it differentiates "whatever I am calling Xorlo" from ... something else that I am calling Xorlo? 16:39 < zipcpi> Specifically refer to it as a code I suppose... I dunno. I'd just prefer defining brivla for the codes themselves rather than relying on names, if that is important 16:39 < durka42> he says the referent is (somehow) specified by the grammar 16:39 < durka42> not that the referent is a grammar type? 16:39 < durka42> or maybe he is just spectacularly bad at writing definitions (which he is) 16:40 < zipcpi> Well yeah... he misplaced xu'u'i and xai'a is COI when it should be in UI after all... 16:40 < zipcpi> Or maybe it's something got to do with *shudder* mekso 16:41 < durka42> ohh 16:41 < durka42> JVS only showed me his giant explanatory comment after I posted my own 16:41 < durka42> let's see if it answered any of the questions 16:45 < zipcpi> NLDCMS needs a name... lertideci'e? 16:50 < durka42> what's NLDCMS? 16:50 < zipcpi> lerfu tcita detri 16:50 < durka42> xu zo li'ei pagbu la'e'au la lertideci'e 16:50 < zipcpi> mi na djuno u'i 16:50 < zipcpi> la'a pagbu la xornunsep 16:51 < zipcpi> si'au le du go'i cu zmadu lo ka mapti 17:05 < durka42> zipcpi: btw, in usage youtube is {la dontu'u}, not {la dontivni} :) 17:05 < durka42> and yes that's a malgli cimjvo but it's a cmevla so shut up :) 17:11 < zipcpi> u'i 17:12 < zipcpi> Go ahead and contradict me in the BPFK mriste then 17:14 < zipcpi> See this is exactly why brivlacme need some way to be added 17:14 < zipcpi> Cause I don't know what "usage" was :p 17:15 < durka42> ie 18:35 < zipcpi> lol somehow after reading Curtis' explanation I'm even more confused than before I read it :p 18:39 < zipcpi> And now I have another crazy system for days of the week. soldei and lundei are kept, but then martidei merkudei iovdei vernedei saptudei 18:41 < durka42> everyone should have their own system for days of the week 18:42 < durka42> it's the best way to promote communication 18:42 < zipcpi> u'i 18:42 < zipcpi> I wasn't exactly serious. Still though, the planets may deserve similar names 18:59 < zipcpi> On {ki'o}: I'm not sure how useful it is beyond a short hand anyway, and even then {nonono} is only one syllable longer. I think in both writing and in speech it can be customary to separate the number into easily digestible groups, but it need not be phonemic 19:00 < zipcpi> Another advantage is "solving" the problem as to whether it should be thousand, ten-thousand, or hundred-thousand :p 19:01 < durka42> I thought it was defined to be 1000 19:01 < zipcpi> Yeah it makes sense since it's based on {kilto} 19:02 < zipcpi> It's just, you know, mumblings as to how culturally neutral it is blablabla I don't really have a dog in that fight :p 19:03 < zipcpi> It's just I think {paki'o zemuciki'o pareci} is just wasted syllable 19:03 < durka42> however I don't know how to render "1,200" vs "1,000,200" 19:03 < zipcpi> *s 19:03 < zipcpi> Right 19:03 < durka42> I mean I'm sure the BPFK has specified that somewhere, but it's not in my memory banks 19:03 < zipcpi> xm 19:04 < durka42> I really need to write a bpfk looker-upper for vlaste 19:04 < durka42> vlaste: ki'o (selma'o) 19:04 < vlaste> ki'o (class) = PA3 19:04 < zipcpi> Just override the ma'oste definitions already fe'o :p 19:06 < durka42> iesai 19:06 < durka42> well 19:06 < durka42> that depends on implementing multilingual support in vlasisku 19:06 < durka42> then we just import the BPFK defs as another "language" 19:06 < zipcpi> lol 19:06 < durka42> BPFK English 19:07 < zipcpi> Then tcekitau and all the others lol 19:07 < durka42> pa ki'o recivo ki'o muxaze 19:07 < durka42> ((1*1000) + 234) * 1000 + 567 = 1234567 19:07 < durka42> ok I can work with that 19:08 < zipcpi> So the "answer" is that 1,000,200 = pa ki'o ki'o renono? 19:08 < zipcpi> Eh, I'd just go pa nonono renono 19:08 < durka42> yes I believe that's correct 19:14 < zipcpi> Or, crazy idea: pa megdo-si nonono renono (optional separator just to let the listener know what order of magnitude you're dealing with :p) 19:16 < durka42> heh 19:36 * nuzba @youxkei: 日本語版sutsisみたいなものってないんですか? #lojban [http://bit.ly/1CgBDok] 20:05 < xajmi> Hello 20:05 < xajmi> I'm new to the lojban language 20:05 < durka42> coi 20:05 < durka42> Hi! 20:05 < xajmi> but I want to say something in lojban in a video i will be making 20:06 < xajmi> the phrase I want to say is "i am a unicorn" 20:06 < durka42> mi pavyseljirna 20:07 < xajmi> cool... thank you so much!!! 20:08 < xajmi> and how do you say "damn you" 20:08 < durka42> di'ai nai 20:08 < xajmi> *sorry if it is insulting 20:08 < durka42> this sounds like an interesting video... :p 20:08 < durka42> from unicorns to curses 20:09 < vultrue> sounds like pretty typical internet reply 20:09 < vultrue> " @$%& you I'm a dragon!" 20:09 < durka42> hmm 20:09 < durka42> vlaste: dragon 20:09 < vlaste> 7 results: drakono, ranmrdrakono, blurdrakono, finprfikodurusi, finprfilopteriksi, finprsinxnatfinai, ranmi 20:09 < xajmi> hahahaha... yea... it's something funny, I hope it will be popular... 20:09 < durka42> drakono? 20:10 < durka42> di'ai nai .i mi drakono co'o 20:10 < durka42> :) 20:10 < xajmi> hahaha.... 20:11 < xajmi> Popo is about to teach you the pecking order... 20:11 < xajmi> :) 20:13 < xajmi> I have a question... Why did letter "h" was omitted from the Lojban language, if it could have been used instead of the " ' " 20:13 < zipcpi> Mostly as a reminder that the morphology doesn't treat it as a true consonant in many ways 20:13 < durka42> it's shorter and it emphasizes the fact that ' isn't really a consonant, just a vowel separator 20:14 < durka42> there is another answer involving Loglan and history, but... who wants to listen to that :) 20:14 < zipcpi> u'i 20:14 < xajmi> OK got it... 20:15 < durka42> also, it enforces cultural neutrality by breaking every command line tool and search engine 20:15 < durka42> zo'o :) 20:15 < zipcpi> u'i 20:17 < xajmi> that is a problem actually... because it is supposed to evolve eventually into acomputer-human language (supposedly), but the charater " ' " is reserved in most programming languages 20:18 < durka42> yeah, but lojban _is_ the language 20:18 < durka42> so no problem 20:20 < xajmi> mmm... I'm not quite convinced... but doesn't matter... it's still a pretty cool language 20:26 < xajmi> Well I bid farewell my friends... I hope the video be good enough, and thanks again 20:26 < durka42> co'o di'ai 20:26 < zipcpi> co'o 20:27 < xajmi> co'o 20:28 < durka42> hopefully he looked up {nai} and didn't think I was saying "bye go to hell" :) 20:28 < zipcpi> u'i 20:29 < zipcpi> lol looking back I can't believe how many iterations {i'au} has gone through 20:29 < niftg> coʰoi 20:30 < zipcpi> VAUhU -> UI(subselma'o uncertain and kept changing) -> TOIhE -> IhAU 20:30 < zipcpi> With IhAU *almost* like VAUhU 20:33 < niftg> .uinai me'oi aspirated me y'ybu na drani ke vidni tolcanci va'o tu'a tinoi ircitci 20:34 < durka42> ja'a tolcanci .i .iunkoda tezu'e lo nu jinga 20:34 < durka42> niftg: http://www.seas.upenn.edu/~aburka/tmp/Screen%20Shot%202015-07-03%20at%2011.34.03%20PM.png 20:34 < niftg> ja'a .ua 20:35 < durka42> .u ʰ 20:35 < fenki> U+02B0 MODIFIER LETTER SMALL H (ʰ) 20:37 < niftg> .uenai ti noi tutci zo'u gau mi la'oi Courier-New ju'ocu'i cu ralju fo'onto gi'e na vasru lo lerfu tarmi vreji pe le fange lerfu 20:37 < zipcpi> .u 象 20:37 < durka42> uinaidai 20:37 < fenki> U+8C61 CJK Ideograph (象) 20:37 < zipcpi> exp: la'oi one-two 20:37 < mensi> SyntaxError: Expected [,] but "t" found. 20:37 < durka42> za'a mi pilno la'ai Lucida Grande le'u 20:37 < zipcpi> za'o gau spofu 20:38 < niftg> exp: la'oi - 20:38 < mensi> SyntaxError: Expected [,] but end of input found. 20:38 < niftg> exp: zo'oi - 20:38 < mensi> SyntaxError: Expected [,] but end of input found. 20:39 < niftg> xy'y 20:39 < zipcpi> ca dubji'isre me'o vu'ubu me'o kunti bu 20:40 < niftg> .u'i je'u 20:40 < zipcpi> mi pu djisku lo se du'u lu gleki-sai li'u mu'a cu ei gendra ma'i le gentufa 20:41 < zipcpi> ku'i na drani lo ka .y. se implemente 20:41 < niftg> me'oi Lucida klesi lertai vau ji'a gi'e melbi mi 20:44 < zipcpi> ja'o lo nu mi kargau le moklu cu selja'e lo nu le gentufa cu spofu ru'e 20:48 < niftg> le ki'a .a'u ru'e moklu 20:49 < zipcpi> .y. fu'o pe'a 20:50 < niftg> sei za'u re'u krici mi se mau se pluka tu'a la me'oi Courier co cnino 20:50 < zipcpi> exp: sei za'u re'u krici mi se mau se pluka tu'a la me'oi Courier co cnino 20:50 < mensi> ([sei { krici} SEhU] [mi {CU <(¹se mau¹) (¹se pluka¹)> VAU}]) 20:51 < zipcpi> mi na se slabu le gerna pe zo sei 20:52 < durka42> lo sei zei bridi cu se fanmo lo selbri 20:53 < durka42> no da brirebla 20:53 < zipcpi> xy'y 20:53 < zipcpi> exp: la .frank. cu prami sei la .frank. cu gleki la .djein. 20:53 < mensi> ([la frank KU] [cu {prami } {la djein KU} VAU]) 20:54 < niftg> su'o da djica lo nu lo brirebla ji'a cu nenri lo sei zei stura 20:54 < niftg> mi no'e go'i 20:54 < zipcpi> ka'e pilno zo be 20:54 < durka42> ro da se djica su'o de 20:55 < zipcpi> mi djica lo ka baze'e daspo lo jbobau iau bu'a'a'a'a'a'azo'o 20:55 < durka42> ro da se djica su'o de .i su'o da djica ro de .i lo pa go'i cu se cmene zo zipcpi 20:56 < zipcpi> :p 20:56 < niftg> ju'o ru'e pu'i gentufa lo gerna poi va'o ke'a curmi lo nu lo sei zei stura cu se nenri la brirebla 20:57 < durka42> zo soi poi cnino cu vasru piro lo'e bridi jecu se famyma'o zo se'u 20:57 < durka42> ja'o ja'a cumki 20:59 < niftg> .ua zo soi noi cnino cu banli zmadu 21:00 < zipcpi> ie .i ku'i mi na djuno lo du'u makau te frica zo xoi 21:00 < durka42> camxes: +exp lo frank cu prami soi la frank cu gleki la djein 21:00 < camxes> ([lo frank KU] [cu {prami VAU}]¹) SEhU> VAU}]) 21:00 < durka42> zipcpi: sko'opu po'o frica 21:00 < durka42> camxes: +exp lo frank cu prami xoi la frank cu gleki la djein 21:00 < camxes> ([lo frank KU] [cu {prami VAU}]¹) SEhU> VAU}]) 21:01 < zipcpi> je'u .iku'i mi za'o na jimpe 21:01 < durka42> xu do jimpe lo banxa je carvi mupli 21:02 < zipcpi> ei da uitki mutpapri ko'au lo nu ciksi ju'eije mupli jufra 21:03 < durka42> xunai su'o da mupli ki'oi lo se srana be zo soi poi cnino 21:03 < zipcpi> ki'oi ma 21:04 < zipcpi> mi nitcu lo mupli poi karbi zo xoi zo soi 21:04 < durka42> ki'oi zoi urli http://mw.lojban.org/papri/new_soi urli 21:04 < zipcpi> je'e mi pu viska 21:05 < durka42> zo soi simsa zo xoi va'o lo nu muvdu lo tolfa'o be lo jufra 21:06 < zipcpi> xy'y xu zo xoi .y. kansele le sko'opu 21:06 < durka42> na go'i 21:06 < durka42> zo xoi zbasu lo cnino sko'opu 21:06 < zipcpi> uaru'e 21:07 < durka42> zo xoi dunli lo sumtcita lo ka sko'opu zbasu 21:07 < zipcpi> ma me le mupli pe le'i mutpapri ge'u poi ka'e arco le te frica 21:09 * nuzba @bripre: @youxkei .e'u pilno la'e di'e i zoizoi http://mw.lojban.org/extensions/ilmentufa/i/ja/ zoi #lojban [http://bit.ly/1CgHdqE] 21:10 < durka42> pe'i le romoi mupli cu .arco 21:10 < zipcpi> lu li'u zo'u: xu zo xoi jai gau frica 21:11 < durka42> cmalu te frica 21:12 < durka42> nu na pairnu'i kei fa lo nu su'o prenu ... 21:18 < zipcpi> le lobyuitki mutpapri pe zo xoi za'o arco lo laldo smuni 21:19 < durka42> oi 21:19 < zipcpi> smudu'i zo va'ei u'i 21:33 < zipcpi> exp: su'o prenu soi nu na pairnu'i cu xusra lo du'u ro jbopre cu lazni 21:33 < mensi> ([{ prenu KU} {soi SEhU}] [cu {xusra jbopre KU} {cu }] KEI¹) KU> VAU}]) 21:33 < zipcpi> exp: su'o prenu xoi nu na pairnu'i cu xusra lo du'u ro jbopre cu lazni 21:33 < mensi> ([{ prenu KU} {xoi SEhU}] [cu {xusra jbopre KU} {cu }] KEI¹) KU> VAU}]) 21:34 < zipcpi> tersmus: su'o prenu soi nu na pairnu'i cu xusra lo du'u ro jbopre cu lazni 21:34 < tersmus> Parse error at 15 21:34 < zipcpi> tersmus: su'o prenu xoi nu na pairnu'i cu xusra lo du'u ro jbopre cu lazni 21:34 < tersmus> Parse error at 11 21:34 < zipcpi> oi 21:35 < zipcpi> lu li'u zo'u: .y. mi pu ji ca pacna 21:36 < zipcpi> tersmus: do pu ga'orgau lo canko soi mi pacna 21:36 < tersmus> (c0); (pu)(). ga'orgau(do,c0) 21:36 < zipcpi> tersums: do pu ga'orgau lo canko xoi mi pacna 21:36 < zipcpi> tersmus: do pu ga'orgau lo canko xoi mi pacna 21:36 < tersmus> Parse error at 24 21:36 < zipcpi> na nelci zo xoi li'a 21:36 < mensi> e'u do vrude pajni gi'e nai ze'i co'a cinmo lo ka nelci 21:38 < zipcpi> tersmus: mi pacna lo nu do pu ga'orgau lo canko 21:38 < tersmus> nu[(pu)(). ga'orgau(do,c0)](c1); canko(c0); pacna(mi,c1) 21:38 < niftg> la mensi ca'o za'a farvi 21:38 < zipcpi> ie 21:40 < niftg> ma poi gentufa cu se pilno la tersmu vau .a'u 21:41 < zipcpi> ju'oinai .i mi na djuno 21:41 < durka42> mi puzu retsku .i ju'ocu'i sepli jecu simsa la camxes 21:43 < niftg> la'a lo nu sepli cu sarcu 21:44 < niftg> ta'o .ua zo ju'oi pu co'a pagbu la jbovlaste 21:45 < durka42> puzuku zo ju'oi se smuni lu ju'o cu'i li'u 21:45 < zipcpi> u'i 21:45 < durka42> ca lo drata pu se smuni lo simsa be zo zo'oi 21:45 < durka42> nau smuni ... .i ju'oinai .u'i 21:45 < niftg> kibro ra'u judri me zo'oi 21:46 < niftg> sa'e me zo zo'oi 21:48 < niftg> ju'oi cu'i 21:51 < zipcpi> zo ju'oi sinxa lo du'u le cusku cu djuno lo du'u le bridi cu jetnu kei a le danfu be le preti poi ga cy spuda ke'a gi gau cy zo kau tcita ke'a 21:51 < zipcpi> ju'oi .i mi djuno 21:51 < zipcpi> ju'oinai .i mi na djuno 21:52 < zipcpi> ju'oicu'i .i .y. mi no'e djuno iau ju'ocu'i 21:53 < niftg> .io zasni finti lo jbovelckijufra 22:12 < zipcpi> mi na broda xoi brode = I don't (broda brode-ly), mi na broda soi brode = I (don't broda) brode-ly ? 22:15 < gleki> ei mi bregau lo cnino vlaste pe xau lo gloso tutci 23:18 < gleki> gloss: lo melbi dansu cu dansu 23:18 < mensi> A(n) handsome dancer(s): dance(s) 23:19 < gleki> the first primitive noun detection! yay! 23:30 < noncomcinse> coi --- Day changed Sat Jul 04 2015 00:06 < vultrue> coi teleko 00:06 < noncomcinse> [alureon]: coi 00:06 < noncomcinse> s/[alureon]: // 00:06 < fenki> noncomcinse meant to say: coi 00:12 < gleki> coi 00:13 < gleki> gloss: lo melbi klama cu klama 00:13 < mensi> A(n) handsome comer: come(s) 01:44 < gleki> exp: lo mlatu pe klama 01:44 < mensi> SyntaxError: Expected [,] but "k" found. 01:44 < gleki> what it could even mean 01:45 < gleki> pe => poi co'e 01:45 < gleki> lo mlatu poi co'e vau cu klama? 01:46 < gleki> co'e is co'ephorics, right? 01:58 < gleki> alta: lo mlatu pe klama 01:58 < mensi> ([FA {lo <(¹mlatu NF¹) (¹pe [FIhA KU] GEhU¹)> KU}] [CU {klama NF} VAU]) 02:04 < gleki> mensi: doi mand mi fircku spusku fi do 02:04 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.mand.gy. di'a cusku da 03:19 * mumble jorne co'a fa la'o fa menli fa 03:27 * mumble jorne co'u fa la'o fa menli fa 03:28 * mumble jorne co'a fa la'o fa menli fa 03:38 * mumble jorne co'a fa la'o fa akti fa 03:38 * mumble jorne co'u fa la'o fa akti fa 03:49 * mumble jorne co'u fa la'o fa menli fa 04:24 < Amnadya2> coi 04:37 < Amnadya> veni vidi vitci falo kibro 04:37 < mensi> Amnadya: cu'u la'o gy.nwofdlfn.gy.: 11:04 mensi: doi mand mi fircku spusku fi do | 2015-07-04T11:26:12. 04:37 < mensi> 422Z 04:42 * mumble jorne co'a fa la'o fa menli fa 04:45 * nuzba @okau_junpe: sutysisku日本語のこの文章の翻訳に関わった記憶があるけど、これはまずい翻訳ではないだろうか? #ロジバン mi pu fanva lo jufra soi morji .i fanva fu lo mabla ju'oru'e http://mw.lojban.org/extensions/ilmentufa/i/ja/ [http://bit.ly/1H5WEk4] 04:46 < ctefaho> coi 04:54 < Zearen> .i coi 05:20 * nuzba @sneubalhon: ct, 電腦版。 http://mw.lojban.org/index.php?title=Presenting_Lojban&mobileaction=toggle_view_desktop [http://bit.ly/1TbRNTT] 05:27 < Ilmen> coi 05:29 < ctefaho> xoi Ilmen 05:34 < zipcpi> Yeah, I think maybe the lerfu system is the best for shorthand systems with their own defined grammar. Heck it's so flexible. 24 letters not enough? Use any old word + bu 05:35 < zipcpi> They can be brivla, or they can be any old cmavo that reminds one of that brivla. Who cares what the cmavo actually mean? For example an alternative to NLDCMS could use na'abu, ma'ibu, de'ibu, etc 05:36 < zipcpi> deibu 05:37 < zipcpi> Maybe all that is needed is something that turn any old word into a PA. And one that semantically turns one into a VUhU... but I don't really care about enforcing mekso grammar anymore 05:39 < zipcpi> exp: fa'obu 05:40 < mensi> (fa'o bu) 05:40 < zipcpi> ... hm not good. We'd need to compile a list of "bad words" :p 05:40 < zipcpi> Probably zo and zei too 05:41 < zipcpi> exp: ki'a'aibu 05:41 < mensi> ([{ki'a'ai bu} BOI] VAU) 05:42 < zipcpi> Ah... small problem... it means for every "magic word" we coin, like go'oi or ze'oi, there is a small chance of breaking some nonce {bu} somewhere 05:44 < zipcpi> Maybe a special {lo'u... le'u} is better 05:51 * mumble jorne co'a fa la'o fa niftyg fa 06:02 < samcinki> I'm jundi-ing Mumble, if anyone is interested. 06:04 * mumble jorne co'a fa la'o fa zipcpi fa 06:09 < gleki> yes, for Mayan system the prefixes would be {kin bu} etc. 06:09 < zipcpi> Yeah maybe cmevla is safer 06:09 < gleki> btw in voice chats {co'o pei}, {fe'o pei}, {je'e pei} would be more used than today 06:10 < zipcpi> ua 06:10 < zipcpi> I do use {je'epei} 06:10 < gleki> co'o/fe'o pei ~= Anything else you want to add or we finish the conversation? - co'o ~= No, goodbye, i have nothing to add 06:12 * mumble jorne co'a fa la'o fa akti fa 06:13 < zipcpi> I use {je'epei} as shorthand for {xu do jimpe fi le se cusku be mi} 06:14 * mumble jorne co'a fa la'o fa niek fa 06:14 * mumble jorne co'u fa la'o fa niek fa 06:22 * mumble jorne co'a fa la'o fa gunro fa 06:32 * mumble jorne co'a fa la'o fa Zearen fa 06:34 * mumble jorne co'u fa la'o fa Zearen fa 06:35 * mumble jorne co'a fa la'o fa ziren fa 06:48 * mumble jorne co'u fa la'o fa akti fa 07:05 * mumble jorne co'a fa la'o fa akti fa 07:08 * mumble jorne co'u fa la'o fa menli fa 07:15 * mumble jorne co'a fa la'o fa ctefaho fa 07:19 * mumble jorne co'u fa la'o fa akti fa 07:19 * mumble jorne co'u fa la'o fa ziren fa 07:30 < zipcpi> exp: pe'a plipe jebabo cu plipe jebabo cu farlu lo kevna 07:30 < mensi> (pe'a [CU {plipe VAU} {je ba} bo {cu bo VAU} VAU]) 07:31 < gleki> wow wow wow 07:32 < zipcpi> vau vau vau? 07:32 < gleki> yacc: 07:32 < mensi> /FA'O/ 07:32 < gleki> yacc: pe'a plipe jebabo cu plipe jebabo cu farlu lo kevna 07:32 < gleki> yacc: pe'a plipe jebabo plipe jebabo farlu lo kevna 07:32 < mensi> (pe'a { <[lo kevna KU] VAU>}) 07:32 < cliva> yacc: pe'a plipe gi'e ba bo plipe gi'e ba bo farlu lo kevna 07:32 < mensi> (pe'a { <[plipe VAU] [gi'e ba bo] [farlu ({lo kevna KU} VAU)] VAU> VAU}) 07:33 < gleki> la'a lo vlaste pe la bangu ba prane zmadu fukpi la jbovlaste 07:33 < gleki> i sa'u mi ba jmina lo ralju be na valsi 07:38 < gleki> en: ambigu 07:38 < latro`a> coi 07:38 < cliva> màsnò 07:38 < cliva> sei na gendra 07:38 < cliva> coi 07:38 < mensi> ambigu = x1 is ambiguous, possibly being x2 (ka) or x3 (ka), this being hard to x4 (2-place ka) |>>> Near-synonyms: 07:38 < mensi> ambigu'o, smuvrici |>>> spheniscine 07:38 < gleki> zipcpi: why x2 and x3 are such? what if >2 meanings 07:39 < gleki> x1 is ambiguous, possibly being x2 (plural of ka)... 07:39 < zipcpi> Not me; I only translated selpa'i-s def 07:39 < gleki> en: ambigu'o 07:39 < mensi> ambigu'o = x1 is ambiguous, open to interpretations x2 (plural of property of x1) |>>> See smuni, sinxa |>>> 07:39 < mensi> gleki 07:39 < gleki> i'e doi la gleki 07:40 < zipcpi> u'i 07:40 < gleki> what are possible examples of ambigu4? 07:41 < gleki> dont offer {la samtin}and {zo'e} 07:41 < cliva> latro`a: lo terjoma nu me'oi fold zo'u xu zo mei'e jo'u lo da'i cnino brivla cu jai banzu 07:41 < cliva> jbo: ambigu 07:41 < gleki> en: terjoma 07:41 < mensi> ambigu = x1 ka'e ckaji x2 ja x3 .i lo du'u x1 ckaji x2 ji kau x3 na jai frili fai lo ka ce'u ce'u me'au x4 me'u noi me 07:41 < mensi> lo ka ganse jo'u lo ka te smuni jo'u lo simsa |>>> ko catlu zo smunambigu je zo smuvrici je zo sorselsmu |>>> 07:41 < mensi> selpahi 07:41 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 07:41 < cliva> jbo:terjoma 07:41 < mensi> terjoma = x1 terjonma'o x2 boi x3 boi x4 boi x5 07:41 < zipcpi> I don't know either lol I even mistranslated it as ma'i 07:41 < zipcpi> Someone then corrected it 07:42 < cliva> .i mu'a lo di'u velcki cu jai mukti lo nu mi pilno zo ke'a ne ka'ai zo jai 07:42 < cliva> vau vau ta'o 07:44 < zipcpi> Like I said, didn't write that def, only translated it 07:44 < zipcpi> I do think I like your terbri better 07:44 < zipcpi> But {ambigu} sees more use, probably because of length 07:45 < cliva> .oi la korpora za'o na se cikre fi lo ka se pi'o ce'u sisku lo ka mupli lo valsi poi se tolfa'o lo karsna 07:47 * la_kristan may be disconnected without notice at any time. 07:48 < la_kristan> .i coi rodo 07:49 < cliva> coi 07:49 * mumble jorne co'a fa la'o fa menli fa 07:50 * la_kristan puzi mo'u tcidu la'o zoi gy. Lojban for Beginners .gy 07:51 < cliva> .u'a dai (to .e'u vimcu zo zoi toi) 07:52 * mumble jorne co'u fa la'o fa niftyg fa 07:52 * la_kristan puzi mo'u tcidu la'o gy. Lojban for Beginners .gy 07:53 < la_kristan> like that? 07:53 < cliva> go'i 07:55 < la_kristan> I just started a Memrise course to learn all the gismu, 07:56 < la_kristan> but it's kinda big to download for offline access. 07:58 < zipcpi> I don't bother memorizing, I just use, but to each their own 08:00 * mumble jorne co'u fa la'o fa zipcpi fa 08:01 < cliva> I did need flashcards at the beginning, but once I had enough to guess unknown words from context I abandoned them 08:03 < zipcpi> They say vocabulary is like Pokemon. You encounter a word in the wild, then catch it :p 08:03 < cliva> ta'o nai there's also Anki, with the gismu deck being 4 MB: https://ankiweb.net/shared/info/2744730295 08:04 < cliva> (the speech synthesizer there is horrible though: x and ' pronounced the same for example) 08:04 < zipcpi> u'i 08:04 < zipcpi> pa'a paxa pa'a paxa pa'a paxa 08:04 * clakre wrote a randomized-position cloze generator to make an anki deck 08:05 < cliva> no wait, they're different, but only enough to tell them apart when it says "pa'a paxa pa'a paxa" 08:08 < gleki> there is a set of manually uttered gismu 08:09 < gleki> there was a project 08:09 < gleki> no idea where those recordings are now 08:09 < la_kristan> I've never used anki 08:10 < la_kristan> I've always used Memrise, as evidenced by my Membrain rank. 08:10 < gleki> what is the best language learning course on memrise in ur opinion? 08:11 < la_kristan> I don't know; I haven't tried all language learning courses on Memrise. 08:12 < gleki> la_kristan: what about those that u tried. what was the best? 08:13 < la_kristan> for any language, or one in particular? 08:13 < gleki> for any language 08:14 < la_kristan> apart from one Esperanto vocab course that had some spelling errors, they've all been pretty good. 08:15 < gleki> la_kristan: have u tried any courses not only with individual words but with examples where the complexity of sentences grows from level to level? 08:16 < la_kristan> although the problem with Memrise, is that it can't get it through its head that the presence and location of . and ' actually matters! 08:17 < la_kristan> I can write "ee" instead of .e'e and it will accept it! 08:18 < la_kristan> or 'e.e or '''...ee or whatever! 08:18 < zipcpi> u'i 08:18 < gleki> la_kristan: have u registered and followed this? area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 08:18 < zipcpi> . actually doesn't matter for parsing 08:19 < la_kristan> oh, yeah, I did the Basic Russian course, it had sentences. 08:19 < zipcpi> Only matters in speech 08:20 < Zearen> Well, It matters little. E.g. «cy.co'e» ≠ «cyco'e», but typical style would use a space. 08:20 < la_kristan> gleki , No, I haven't seen that. 08:21 < gleki> la_kristan: also please upvote any 5 questions with rating below 10 08:21 < Zearen> I personally try to be denpa bu drani just to assure I remember where they go. 08:22 < la_kristan> zipcpi: it matters when I'm already having trouble remembering .ue and .u'e 08:22 < zipcpi> No, not the apostrophe, that one definitely matters 08:22 < la_kristan> and it tells me I'm right when I get the wrong one. 08:22 < gleki> english is also affected then since "its" and "it's" have different meanings 08:23 < gleki> maybe we need a course where ' is replaced with h 08:24 < la_kristan> maybe Memrise should pay attention to apostrophes. 08:24 < gleki> you may open a new discussion 08:25 < la_kristan> I don't know much about that; I just use the Android app. 08:26 < la_kristan> all the discussions are just "behind the scenes " to me 08:30 * la_kristan may be disconnected at any time without notice. 08:31 < la_kristan> In case you didn't get the first notice. 08:32 < la_kristan> .i ca lo cabdei lo merko cu salci 08:33 < la_kristan> is it lo cabdei or le cabdei 08:33 < la_kristan> ? 08:33 < cliva> ba'a nai ru'e there is an option for authors on Memrise to control which characters are whitespace 08:34 < gleki> la_kristan: {lo cabdei} is definitely correct here 08:35 < la_kristan> it was le cabdei in L4B, but that seemed weird. 08:36 < la_kristan> since it /is/ today, not just what I call today. 08:36 < zipcpi> Ah... that veridicality thing doesn't really apply anymore 08:37 < zipcpi> I'd tend to use le, because it's the cabdei in reference to something, but lo won't be wrong either 08:38 < la_kristan> I've been making mental note of rhyming gismu... 08:39 < la_kristan> but the fact that salci rhymes with kalci doesn't strike me as useful. 08:39 < zipcpi> u'i 08:39 < zipcpi> I found Lojban hard to rhyme with; I just settle for rhyming the last syllable 08:40 < la_kristan> palci might usefully rhyme with either... 08:40 < zipcpi> Part of it though is that my vocab isn't that great 08:40 < la_kristan> .i mi zanru lo tanru 08:41 < zipcpi> Well, there's a little thing you should know called "xorlo", post that, {lo} is just a generic "convert selbri into sumti" 08:41 < la_kristan> .i mi prami le skami 08:42 < zipcpi> Instead of making any existential claim like su'o da poi broda 08:42 < la_kristan> well, I read about xorlo; but I thought things were still supposed to actually be the thing in question, in some way 08:43 < la_kristan> even if it's just bear goo... 08:43 < zipcpi> lol 08:43 < la_kristan> it's still a bear, somehow 08:44 < zipcpi> There are many ways to disassociate one from a description... pe'a, ba'u, even ju'acu'i (rather recently invented compound{ 08:45 < zipcpi> Well, my current opinion is that it may not be {ca'a cribe}, but it's {pu/ba'o cribe}, and tenses can be elided given sufficient context 08:45 < clakre> near-rhymes and other assonances seem more useful than last-syllable rhymes 08:45 < zipcpi> Well I did try near rhymes 08:45 < clakre> maybe we should be making alliterative poetry zo'o 08:45 < zipcpi> But then I failed to keep it up 08:46 < la_kristan> zbasu and smacu don't rhyme, but they're close enough. 08:46 < zipcpi> Anyway the meter of the song I translated stressed the last syllable, even though that's "wrong" strictly speaking 08:47 < cliva> ZA'O has lots of words broken up by pauses, just to keep the rhyme 08:47 < zipcpi> But what can be done about that, save rewriting the entire song 08:48 < la_kristan> so, nowadays, do people still usually say "ro lo [insert brivla here]"? 08:49 < zipcpi> Yes... it means "all of the" or "all of some" 08:49 < la_kristan> cliva: yeah, I noticed that. 08:49 < cliva> "ro lo broda" still exists, but doesn't mean the same as "ro broda" anymore 08:49 < zipcpi> Yep 08:50 < la_kristan> cliva: about the album, that is. 08:50 < gleki> indeed we need a tool to recreate rhyming dictionary automatically 08:51 < la_kristan> and "lo ro broda" means... ? 08:52 < zipcpi> The brodas / some brodas which number all 08:52 < cliva> everything that brodas, but not necessarily distributively 08:52 < la_kristan> I actually purchased ZA'O 08:52 < zipcpi> But what's nice is tha lo mu broda no longer means "all the brodas in the universe, which number five", which is... rather difficult to use :p 08:53 < cliva> "lo ro prenu poi zvati ti cu bevri lo pipno": everyone who is here carries a piano (it could be each carrying one individually, or all of them together) 08:53 < la_kristan> distributively ki'a 08:53 < zipcpi> Ah... poi zvati... clever. I would've had a hard time coming up with a use case on the spot for lo ro broda 08:55 < la_kristan> and "ro lo broda poi zvati ti"? 08:55 < la_kristan> what's the difference then? 08:56 < zipcpi> Not much, the elided place is zo'e 08:56 < zipcpi> Oh wait 08:56 < zipcpi> Now it's an outer quantifier 08:56 < zipcpi> So it will be definitely distributive 08:56 < la_kristan> I purchased the whole album, just so I could read the lyrics to "nasa" 08:57 < zipcpi> I'll probably need to purchase it to analyze the lyrics at all 08:57 < zipcpi> I'm horrible at interpreting lyrics. That XKCD comic is so me 08:57 < la_kristan> I still don't know what you mean by "distributive" 08:57 < zipcpi> In English even, there's no need to mention Lojban 08:58 < zipcpi> Meaning individually 08:58 < la_kristan> what XKCD comic? 08:58 < dutchie> https://xkcd.com/1538/ i assume 08:58 < zipcpi> ro lo prenu poi zvati cu bevri lo pipno, implies that each person carried a/some pianos 08:58 < la_kristan> so, with ro first, they each carry a piano? 08:58 < zipcpi> Yep 08:59 < la_kristan> or more than one, lol 08:59 < zipcpi> Yeah 09:00 < la_kristan> about the comic; I can totally relate to that :-D 09:01 < la_kristan> no matter what language the lyrics are in. 09:02 < la_kristan> but sometimes someone does a remake of a song and I'm like "so that's what they were saying!" 09:04 * nuzba @uitki: L17-01 - http://mw.lojban.org/papri/L17-01 by Gleki - /* Vehicles */ [http://bit.ly/1KASW60] 09:04 * la_kristan is very annoyed when recent apps shows apps used a week ago, but not ones used two minutes ago. 09:06 * la_kristan is afk (and could still be suddenly disconnected at any time) 09:12 < la_kristan> if you need me, mention my name and my phone will buzz. 09:18 * mumble jorne co'u fa la'o fa ctefaho fa 09:18 * mumble jorne co'a fa la'o fa ctefaho fa 09:18 * mumble jorne co'u fa la'o fa ctefaho fa 09:20 * mumble jorne co'a fa la'o fa niek fa 09:21 * mumble jorne co'u fa la'o fa niek fa 09:21 < niek> Boop. 09:39 * mumble jorne co'u fa la'o fa gunro fa 09:43 < gleki> coi 09:44 < gleki> i fi'a ma cu'e mo 09:56 < zipcpi> fi'a ma cu'e pai'e mo pei'a pei 09:56 < zipcpi> xo 10:01 * la_kristan is still trying to figure out how to say "fireworks" in toki pona 10:01 < la_kristan> ijo pi suno seli musi 10:01 < la_kristan> maybe? 10:13 * mumble jorne co'u fa la'o fa menli fa 10:24 < gleki> i guess John Clifford on Facebook can answer your question 10:25 < la_kristan> gleki: so what is this stackexchange thing? 10:26 < gleki> a new place for lojban. if we get 230 more upvotes StackExchange will create a new domain for us 10:27 < gleki> we already got 70 followers but few of them upvoted 10:27 < la_kristan> oh. I see questions and answers. 10:27 < gleki> yes, we need to upvote those that are below 10 in rating 10:27 < gleki> first register then you can upvote 10:27 < la_kristan> but not anything for answering questions. 10:27 < gleki> you need to login to answer them 10:28 < la_kristan> okay, I'm logged in, but I still don't see it... 10:29 < la_kristan> and I confirmed the email 10:29 < gleki> "add comment" 10:29 < gleki> to create a new question there is "Propose an example question: 10:29 < gleki> " 10:29 < gleki> in the footer 10:30 < gleki> you might need to follow this section 10:30 < la_kristan> oh, I see. and what is the significance of this new domain? 10:31 < Luluf> coi 10:31 < gleki> well, some people prefer such type of websites. 10:32 < gleki> there are no universal methods of learning languages so the more options we provide the better 10:37 < astroendjisai> .i coi .i mi ctuca mi fo lo lojbo ku 10:38 < astroendjisai> If what I just said was wrong please correct me 10:38 < gleki> fine. 10:39 < gleki> although {mi tadni la .lojban.} = "I study Lojban" might be an easier option :) 10:40 < astroendjisai> Ah, cool. .ui la'edi'u pluka mi 10:40 < gleki> ui dai 10:42 < la_kristan> gleki: so I click the up arrow? 10:43 < la_kristan> if the number isn't 10? 10:45 < la_kristan> be'e la gleki 10:47 * nuzba @s_i_d_a: noda djuno lo du'u drudi je loldi .i ma'a dandu #lojban #haiku [http://bit.ly/1LOW66t] 10:48 < la_kristan> that reminds me... 10:48 < la_kristan> there's a line in the song "nasa" 10:48 < la_kristan> that goes ".i noda djuno lo du da klama ma kau" 10:49 < la_kristan> shouldn't it be "du'u", not "du"? 10:52 < la_kristan> gleki, you there? I asked you about the upvoting. 10:56 < samcinki> "du" replaces "du'u" in cekitau 10:57 < la_kristan> haiku: 10:57 < la_kristan> doi selbeiskami/ 10:57 < la_kristan> do mutce melbi blanu/ 10:57 < la_kristan> .i mi nelci do 10:57 < la_kristan> samcinki: huh? what's that? 10:59 < la_kristan> and WHERE is gleki? he's not responding! 10:59 < la_kristan> cekitau ki'a 11:00 < latro`a> la_kristan, there's a "dialect" of late, commonly called ce ki tau jau, which switches out some common CV'V with similar-sounding uncommon CV 11:00 < la_kristan> weird. 11:01 < latro`a> the main four are ce'u<->ce, ke'a<->ki, tu'a<->tau, jo'u<->jau 11:01 < latro`a> and sometimes du'u<->du 11:02 < Amnadya> coi 11:02 < la_kristan> why? they don't mean the same thing. 11:02 < latro`a> not making them interchangeable, just reversing their roles 11:03 < latro`a> because the longer ones are much more frequently used 11:03 < latro`a> (in the case of ce'u and jo'u, because of recent changes) 11:03 < latro`a> *recentish 11:03 < la_kristan> but that just sounds confusing. 11:04 < latro`a> who wrote the song? 11:04 < la_kristan> changes? aren't there enough of those already? 11:04 < la_kristan> djemynai 11:04 < latro`a> these aren't changes to the language itself but more to usage 11:05 < latro`a> and yeah, that's what I figured 11:05 < latro`a> for instance ce'u was not really expected to be a common word, especially not a common word used multiple times in a row, when the CLL was written 11:05 < latro`a> which is presumably why it got relegated to CV'V space 11:06 * la_kristan looks up ce'u 11:06 < Amnadya> It seems jau hasnt caught on, but some now use joi instead of johu. 11:06 < latro`a> ce'u isn't very clear in the ma'oste 11:06 < latro`a> ce'u is a word that's hard to put in the ma'oste because you can't use it in isolation 11:07 < latro`a> it needs {ka} (or sometimes {ni}) to even make sense, and you need an appropriate brivla to understand why you would want that anyway 11:07 * la_kristan doesn't feel very enlightened by the definition. 11:07 < latro`a> let me try to give an example to help 11:07 < latro`a> mi do dunli lo ka ce'u tadni la lojban 11:07 < Amnadya> The cmavo definitions arent always that great. There was actually a character limit 11:07 < latro`a> this literally reads "you and I are equivalent in the property "x1 studies Lojban"" 11:08 < la_kristan> oh, character limit. that does explain a lot. 11:08 < latro`a> the {ce'u} gives the place that the sumti substituted into the ka gets substituted into 11:08 < latro`a> in {dunli}, what's going on is that the truth value when you substitute in dunli1 is the same as the truth value when you substitute in dunli2 11:09 < latro`a> so "both of us study lojban or neither of us study lojban" 11:09 < latro`a> is a more idiomatic translation of my example 11:09 < la_kristan> oh, yeah, I think I remember reading about ka and ce'u 11:09 < latro`a> there are quite a few gismu with places like this, and more recently we've discovered that quite a few gismu that weren't *defined* this way should have been, such as {kakne} 11:10 < latro`a> tsani's example is that {mi kakne lo nu do citka lo plise} makes no sense 11:10 < latro`a> kakne2 as nu only makes sense when kakne1 is in the nu 11:10 < la_kristan> hmm, I see 11:10 < latro`a> which is a constraint that can be made *explicit* by using ka/ce'u instead 11:11 < la_kristan> ah. 11:12 < la_kristan> mi kakne lo ka lo gerku citka ce'u 11:12 < latro`a> you missed a ku/cu, but yes; "I can be eaten by a dog" 11:13 < la_kristan> oh man, I keep doing that! 11:14 < la_kristan> I'm having a harder time remembering to put cu where I need it, than I ever had with the Esperanto accusative case! 11:15 < durka42> coi'oi'oi'oi 11:16 < mjhat> Hi there ,I like to learn my first conlang lojban.I already know tamil and english.so far neither i can find a consistent tutorial nor a clear path or procedure to learn .there is laot going on in the community like the grammer changing. can anyone clearly tell me what to learn 11:17 < la_kristan> mjhat : quite a few people recommend the Wave Lessons 11:17 < la_kristan> I totally understand the frustration! 11:17 < mjhat> thank you does that follow cll1.1 11:18 < la_kristan> it's supposed to be more up-to-date, with dotside and whatever. 11:19 < la_kristan> but I never used it, so I don't know a whole lot about it. 11:21 < la_kristan> I just did Lojban for Beginners, and then got told how shakespearean my speech was. 11:22 < mjhat> http://vrici.lojban.org/~rlpowell/media/public/cll_prince_for_lsi_2015-03-15.pdf. Is this the new grammer when it will be published.Should i follow it. 11:22 < la_kristan> actually, I still have the excercises for the last lesson left to do... 11:23 < gleki> la_kristan: you click up if the number is below 10 11:23 < la_kristan> yay! he's alive! 11:24 < gleki> mjhat: mw.lojban.org/papri/Learning/en 11:24 < gleki> mjhat: choose some course there and then ask questions here anytime 11:24 < latro`a> la_kristan, it's not very weird, only the le usage sounds funny these days 11:25 < latro`a> the recent stuff is more minor than we make it out to be 11:25 < la_kristan> gleki: okay. I'm still not totally sure what that's supposed to accomplish, but I like to help people however I can... 11:25 < gleki> as for {ce'u} if you understand {ke'a} then you will understand {ce'u} 11:25 < gleki> la_kristan: yes, we just need upvotes to show admins that people are interested in lojban 11:26 < gleki> a protection from spammers i guess 11:26 < la_kristan> yeah, I just forgot what ce'u means; I understand it okay. 11:27 < mjhat> Yes,thank you .But those lessons are based on the old grammer right 11:28 < la_kristan> now I've apparently run out of upvotes. 11:28 < gleki> mjhat: The Crash Course although incomplete is more about the current grammar 11:29 < gleki> the two grammars are still very similar. 11:29 < gleki> la_kristan: yeah, you have only 5 11:29 < mjhat> thank you very much 11:29 < gleki> la_kristan: and if i stop replying it mean im afk. my account is connected 24/7 11:29 < gleki> to this channel 11:29 < la_kristan> well, I've given them, hope it helps... 11:30 < la_kristan> gleki: oh I see. 11:30 < gleki> e.g. im gonna sleep in ~1-2 hours 11:31 < la_kristan> oh, what time is it? it's 12:31 here. 11:31 < gleki> "A True Lojban Moment" story by selrun is now top3 of all time on reddit 11:32 < la_kristan> oh, I'm not on reddit, but that sounds interesting. 11:32 < gleki> I work in retail, and at night not much happens. I was going about my business, doing the little jobs that need to be done when there are no customers, and I was humming a lojban tune to myself. After a while, I tried to compose my own verses, which had me thinking in lojban. Then a customer walked through the door, and I almost said, "coi la ...". The only thing that stopped me was that I did not know their 11:32 < gleki> name, and then I realized what had almost happened. Without any conscious effort, my mind was acting in lojban. It was an awesome moment! 11:33 < gleki> i sampu lisri 11:34 < la_kristan> oh, that happend to me quite a bit with Esperanto. 11:34 < la_kristan> */happend/happens 11:37 < gleki> klare 11:37 < _mukti_> coi la gleki 11:38 < la_kristan> I tend to think in Esperanto, and if someone suddenly addresses me, I usually just look at them blankly for a moment, because the response comes to mind in Esperanto, and it takes a bit for me to get back to English, 11:38 < la_kristan> and by then the person's moved on. 11:39 < durka42> coi la mukti noi ca nau kansa mi lo ka viska makau poi jbovlaste kibzva 11:39 < _mukti_> coi .u'i la durkavore 11:39 < mjhat> hey glekii came across 11:39 < mjhat> The La Gleki's 11:39 < mjhat> Crash Course in Lojban 11:39 < mjhat> The guide to the naturalistic logical language 11:39 < mjhat> Published 2016" Is it okay to learn from that. 11:39 < _mukti_> .i ie mi djica lo ka kansa do 11:40 < dutchie> coi rodo 11:40 < Amnadya> coi 11:40 < _mukti_> coi la nederland 11:41 < _mukti_> coi la'o zoi amnadya zoi 11:41 < durka42> mjhat: that looks like the Crash Course yep 11:44 < _mukti_> Gleki: I'm looking at this bug: https://github.com/lojban/jbovlaste/issues/182 11:44 < _mukti_> I was able to download the XML export in Safari, and I'm not seeing the issue. Do you have Safari? 11:45 < _mukti_> Which is to say, I see the XML as rendered via the XSLT stylesheet. 11:45 < la_kristan> what's the difference between the 2nd edition cll and the one I've got? 11:46 < durka42> the one you have exists 11:46 < _mukti_> la_kristan: Still being worked out, but the 2nd edition will include some corrections 11:46 < durka42> CLL v1.1 is coming soon with just corrections but no real content changes 11:46 < la_kristan> oh, okay. 11:47 < la_kristan> when I look at the original cll pdf on my phone, the formatting is garbled. 11:47 < la_kristan> mostly the examples 11:47 < durka42> hmm 11:48 < la_kristan> many are illegible. 11:48 < durka42> I'm generally suspicious of phone PDF viewers... 11:48 < mjhat> I've read that xorlo are the changes when are they going to be included. 11:49 < durka42> xorlo is a very old change by now, it's included in the crash course 11:49 < durka42> but not yet in the CLL 11:49 < durka42> la_kristan: here is an up-to-date PDF of just chapter 1, see how the examples look http://alexburka.com/lojban/cll_1.pdf 11:49 < _mukti_> mjhat: Great question, mjhat. I think we're collectively waiting on someone to take the bull by the horns the way that John Cowan did. 11:50 < durka42> la_kristan: actually, chapter 1 doesn't have any examples. look at chapter 2 http://alexburka.com/lojban/cll_2.pdf 11:51 < durka42> I hope that as soon as CLL 1.1 is out the door, we can get to work on the new gadri chapter 11:51 < mjhat> Does cll1.1 include xorlo 11:52 < durka42> no 11:52 < durka42> just formatting and corrections 11:52 < durka42> _mukti_: re bug 182, if you save the page as a "web archive" it is indeed corrupted. but if you save as "page source" it is fine 11:53 < _mukti_> Hmm. I wonder what it means to save an XML/XSLT page as a "web archive". 11:53 < durka42> not a whole lot, apparently :) 11:53 < _mukti_> Does it try to save it as HTML ? 11:53 < durka42> I think it tries to save the HTML that it rendered, yeah 11:53 < _mukti_> Hmm. Far be it from me to suggest that Jbovlaste is not riddled by bugs, but I don't think that one is ours. 11:54 < _mukti_> Which is to say, we're not serving HTML, so if the browser tries to save HTML, it's doing something more than we contracted. 11:54 < durka42> I agree, but is there a way to hint to the client that it should download this document and not try to display it? 11:54 < _mukti_> I'll research that. 11:54 < durka42> you can't right click save target as, because there's a captcha in the way 11:55 < _mukti_> I'm not seeing a captcha, but then, I'm logged in. 11:55 < durka42> yes 11:55 < durka42> captcha iff not logged in 11:56 < _mukti_> Ok, saving as web archive (though for me, page source was offered first) and then reopening I get a rendering error. Maybe this is what Gleki is talking about? 11:57 < durka42> I think so 11:57 < durka42> when I look at the .webarchive in a text editor, I see a bunch of dt and dd 11:57 < _mukti_> Yeah, webarchive looks kind of like a tarfile. Let me see if tar can expand it. 11:58 < _mukti_> No, not a tarfile 11:58 < _mukti_> Aha? 11:59 < _mukti_>  file en.webarchive 11:59 < _mukti_> en.webarchive: Apple binary property lis 11:59 < _mukti_> (list) 12:00 < durka42> of course 12:00 < durka42> well known format 12:00 < durka42> zo'o 12:00 < la_kristan> all the examples in that chapter are legible. 12:00 < durka42> yay! 12:00 < _mukti_> plutil -convert xml1 en.webarchive 12:00 < _mukti_> (makes it at least readable) 12:01 < la_kristan> but I don't remember if those were legible in the one I have or not... 12:01 < _mukti_> Well, kind of. Opening the result, all of the data seems to be stuffed in a "WebResourceData" element, base-64 encoded 12:02 < durka42> yeah, "readable" but not really heh 12:02 < durka42> you need to save it as "page source" if you're going to import it to anything that expects our XML format 12:03 < la_kristan> mostly it was ones with long-winded Lojban sentences, the words would overlap each other. 12:03 < _mukti_> So are we looking for a way to hint to Safari that it should not offer "web archive" as an option? 12:04 < durka42> it should just download the file, as if it were a ZIP or whatever 12:04 < durka42> I dunno if you can tell it to do that 12:04 < _mukti_> Yeah, I'm skeptical. ia nai 12:05 < gleki> _mukti_: i dont have safari but note the version of it mentioned 12:06 < la_kristan> also, in the chapter on vocatives, the list was messed up in such a way that most of it was pushed off the page and I couldn't read it, which was rather disappointing 12:06 < durka42> I wonder what MIME type we serve it as 12:06 < durka42> vlaste: coi (cll) 12:06 < vlaste> coi (cll) = http://dag.github.io/cll/6/11/ http://dag.github.io/cll/13/14/ 12:06 < gleki> la_kristan: in what book? 12:06 < gleki> la_kristan: in CLL? 12:06 < la_kristan> yeah, the cll 12:06 < durka42> do you know which of those chapters? 12:07 < gleki> la_kristan: link to what you were looking at? a pdf version? 12:07 < _mukti_> durka42: application/xml 12:07 < durka42> _mukti_: http://stackoverflow.com/a/4832418/1114328 12:07 < durka42> no idea if that would make a difference 12:08 < _mukti_> Maybe the best solution would be offer an explicit link for those who wish to download the file rather than view it. The XSLT was added to make the XML viewable as a result of a different feature request. 12:08 < gleki> _mukti_: our hypothesis was that that old safari tried to render xml and to save not the source but the result of rendering 12:08 < durka42> I think that hypothesis is correct 12:08 < gleki> _mukti_: the quick fix is to detect an old version of safari and to download xml, not render it 12:08 < _mukti_> Let me see if I can extract the goodies out of the webarchive so I can at least determine what Safari is saving. 12:08 < la_kristan> gleki; it was a pdf; I don't have the link at the moment. 12:08 < _mukti_> gleki: We're not rendering anything... the browser does that 12:09 < _mukti_> But if you mean disabling XSLT if Safari is detected, we could do that. 12:09 < la_kristan> and I haven't seen if it's the same way when it's not on my phone. 12:09 < _mukti_> Alternately, we could provide an explicit download link. 12:10 < _mukti_> durka42: Interesting 12:10 < _mukti_> (text/xml) 12:11 < _mukti_> I'm not sure though if we'd consider it humanly readable without the XSLT transforms -- given that they were added because of complaints that the bare XML was not readable. 12:12 < gleki> _mukti_: i think that for safari the only link should be "download" 12:12 < gleki> for an old safari 12:12 < gleki> la_kristan: maybe this http://vrici.lojban.org/~rlpowell/media/public/cll_build-2015-06-20/cll_prince.pdf 12:13 < durka42> what's old? is safari 8 old? I have safari 7, .u'i 12:13 < _mukti_> Well, I have a newish (if not the newest) safari, and I'm seeing an issue with webarchive, so it seems to affect Safari in general -- not just the old one 12:13 < _mukti_> I'm using 8.0.6 12:13 < la_kristan> gleki: it's got a red cover, that's all I know. 12:15 < gleki> _mukti_: im sure for safari you can replace the direct link to .xml file to a link downloading that .xml file. 12:15 < gleki> _mukti_: im surprised you want to fix that issue. 12:16 < gleki> _mukti_: this bug is much more serious: i cant add glosswords anymore https://github.com/lojban/jbovlaste/issues/165 12:16 < _mukti_> gleki: Whenever something new arrives that seems easy, I try to close it quickly. But I agree there are more worthwhile things to do. 12:17 < gleki> you can just link everybody to mw.lojban.org/extensions/ilmentufa/ircbot/dumps/ 12:45 < gfsh> hello, would the lojban community be interested in Mozilla localization for lojban? 12:45 < gfsh> just throwing some feelers out there, I am not authorized to offer anything :) 12:46 < gfsh> say, something like mozilla.com, websites or even the browser (though the latter is much more complex an operation) 12:49 < durka42> more localizations are always awesome :) 12:49 < gfsh> there would have to be say... at least 1K active users and 4-5 people willing to keep the locale updated 12:50 < durka42> that first one is quite a tall order 12:50 < gfsh> I was also looking for klingon speakers, but they don't hang out on freenode :( 12:50 < durka42> heh 12:50 < _mukti_> Yes, sadly we're an order of magnitude at least away from that. 12:51 < _mukti_> The locale updating, we could probably handle. 12:51 < gfsh> in any event, if people are willing to do the work the Mozilla community will likely say 'yes' 12:51 < _mukti_> That's good to know. Where do we start? 12:53 < gfsh> _mukti_: I would begin with something maintainable by a few people, like mozilla.com 12:53 < gfsh> it's a few thousand strings to keep updated and they don't change very much over a few months 12:53 < _mukti_> I see the Mozilla wiki is MediaWiki. We've got quite a bit of localization in our installation. 12:54 < gfsh> Support Mozilla? (sumo) 12:54 < gfsh> ah, the wikipedia is something separate 12:55 < gfsh> no, sumo is too large for a small team to do, especially if there's no demand for it :) 12:55 < _mukti_> I was looking at this: https://wiki.mozilla.org/L10n:Localization_Process 12:56 < gfsh> _mukti_: https://svn.mozilla.org/projects/mozilla.com/trunk/locales/ 12:57 < gfsh> this is where localization files for mozilla.com websites are stored ^^ 12:57 < gfsh> _mukti_: if you can use subversion, the command "svn checkout https://svn.mozilla.org/projects/mozilla.com/trunk/locales/" is what you can use 12:57 < gfsh> that will grab the files from all locales 12:57 < gfsh> the en-US locale is the default one which everyone uses as a template 12:58 < _mukti_> gfsh: Thanks, I'll check it out. 12:58 < _mukti_> (Literally.) 12:58 < gfsh> now, the .lang file format is relatively simple 12:58 < gfsh> the source string is the lines that begin with ; 12:58 < gfsh> and the target string (localized string) is the line immediately after 12:58 < la_kristan> #klingon has existed since 2009 12:59 < la_kristan> but last I checked nobody was there. 12:59 < gfsh> keep the file format as UTF-8 without BOM, keep all html/xml or identifiers 13:00 < gfsh> if you wanna give it a shot, you can edit some of those files and I can give you instructions on how to set up your own private server for testing the website 13:00 < _mukti_> It sounds like you know your way around the project. 13:00 < gfsh> I am a new localizer myself too, only been doing it for about 2 months :) 13:01 < _mukti_> What's your target language(s), if you don't mind me asking? 13:01 * durka42 svn co 13:01 < gfsh> Romanian 13:01 < _mukti_> I assume Esperanto is already well represented? 13:02 < durka42> how big is this repository o_O 13:02 < gfsh> a few megabytes 13:02 < durka42> oh so it i 13:02 < durka42> s 13:02 < durka42> just a lot of small files 13:03 < gfsh> the end of line format is linux: LF 13:03 < durka42> by file size, the esperanto localization is about 50% 13:03 < durka42> dunno if that means anything 13:06 < gfsh> actually, the esperanto localization is complete for mozilla.com, just checked :) 13:06 < gfsh> just missing 4 strings 13:07 < gfsh> https://l10n.mozilla-community.org/webdashboard/?locale=eo 13:07 < gfsh> the section named "State of your lang files" 13:07 < gfsh> External Web Projects are other mozilla-related websites 13:10 * durka42 wonders what the ISO code is for klingon 13:10 < gfsh> tlh 13:11 < durka42> makes sense 13:12 < gfsh> the website in esperanto https://www.mozilla.org/eo/ 13:20 < la_kristan> I'm a FireFox user, but I can't take on another project just at present. 13:23 * la_kristan finishes slicing a seemingly endless pile of strawberries. 13:23 < gleki> gfsh: there is a group localizing Firefox to Lojban but it's not active. cirko wanted to localize Firefox 13:23 < gleki> but i havent heard any news from him about that 13:24 < gleki> translatewiki.net is for localizing MediaWiki 13:24 < gfsh> gleki: it's a somwewhat complex project that takes a full team to do 13:24 < gfsh> the browser itself 13:24 < gleki> ofc. i do translations on it once in a while since the main lojban wiki is also mediawiki 13:24 < gleki> gfsh: i also fully localized Vivaldi browser to lojban 13:24 < gleki> and tatoeba.org 13:25 < gleki> but the latter is probably not relevant 13:25 < gfsh> ah, opera 13:25 < gleki> opera doesnt accept any localizations 13:25 < gleki> they hire a company who does everything for them 13:25 < gfsh> still, vivaldi is proprietary technology 13:25 < gleki> they dont accept any volunteers 13:26 < gleki> Vivaldi was nice in that it allowed to see what we can get in the end 13:26 < gfsh> I suppose it makes sense for a proprietary company 13:26 < gleki> Chromium also has Lojban localization, even more strings than for Esperanto are translated but still not enough even for compiling a Lojban Chromium 13:27 < gfsh> Chromium is a pain in the ass to compile on your own 13:27 < gleki> i mean to translate surface GUI 13:27 < gleki> i have no clue what are the most important strings there 13:27 < gleki> it's on luanchpad.net 13:27 < gleki> launchpad.net 13:28 < gleki> Also I wish someone got developer acccess on gerrit to manually add some localizations files for Mediawiki 13:28 < gfsh> gleki: hm... if I were to do it, I would manually explore the interface, write down the strings in some file with descriptions, then grep -riI the localization file for the source strings 13:28 < gfsh> recursive grep in the source code 13:28 < gleki> e.g. "Special:" prefix isn't on translatewiki.net. it's somewhere in /languages/ subfolder of mediawiki 13:29 < gfsh> that way at least you get to see where things are stored 13:29 < gleki> gfsh: i just thought to be sincere to myself and translate everything! but i have too many tasks in Lojbanistan 13:29 < gfsh> gleki: of course, a web browser is a huge project even for just localization 13:30 < gleki> launchpad.net unlike Transifex doesnt provide good hints on what each term for Chromium may mean, what its context 13:30 < gleki> i just dont understand what some strings mean, where they are used 13:30 < gleki> not even in English 13:31 < gleki> i have nowhere to ask 13:31 < gleki> unlike with Vivaldi and Tatoeba 13:31 < gleki> and MediaWiki 13:31 < gfsh> ah, it's why I would go about it in reverse, from the interface and then locate the strings in the localization source files 13:32 < gfsh> that way at least you can do the interface even if some stuff like error messages remain untranslated 13:34 < gleki> if i knew how to compile Chromium strings to a file and immediately test what i get ... 13:34 < gleki> e.g. for Tatoeba you have dev.tatoeba.org website. Whatever you translate on Transifex in half an hour goes to that website so that you can test what you did. 13:35 < gleki> It goes to production only once a week 13:35 < gfsh> that's cool 13:35 < gleki> so you have a lot of time to polish everything 13:35 < gleki> Vivaldi does this de facto once a fortnight id say 13:35 < gleki> still fine 13:35 < gfsh> I would be apprehensive at contributing to a proprietary project, however 13:35 < gleki> with Chromium and Firefox i just have no motivation 13:36 < gleki> Vivaldi has only 400 strings, that's easy 13:36 < gfsh> not bad :) 13:36 < gfsh> firefox has oh, tens of thousands 13:36 < gleki> at least we now know how to translate GUI widgets names etc. 13:36 < ctefaho> Firefox has tens of thousands of strings for one language? 13:37 < gleki> Vivaldi localisation file can be downloaded and analyzed as a dictionary 13:37 < gfsh> well, a few thousands are the key ones, the rest are features buried deep like the debugger 13:37 < gleki> debugger should be probably only in English 13:37 < gleki> anyway we first need to ask cirko 13:37 < gfsh> well, French insist on doing every little string 13:37 < gleki> cirko! where are you?? 13:37 < gleki> do zvati ma doi la cirko 13:37 < gleki> i ma nuzba la faglo'u 13:38 < la_kristan> my phone's battery is low... 13:38 < gleki> cliva: akmnlrse: ma nuzba la faglo'u i xu do fanva 13:38 < la_kristan> and I'm too busy to stand by the charger. 13:39 < gleki> gfsh: if you have anything particular in mind you can start translating, sure. if you can make your lifer easier by providing translation or testing interface for s then ... we'll gladly accept it and help you 13:39 < gleki> *for us 13:40 < la_kristan> so I guess I have to go: I just got the low-battery warning from the phone... 13:40 < gleki> co'o la kristan 13:41 < la_kristan> co'o 13:41 < gleki> i am afk now too. 13:53 < astroendjisai> That's a long pingtime 13:58 < Ilmen> Well the grammar changes since the CLL are few, the two major official changes are Xorlo and Dotsides; Xorlo doesn't change Lojban's syntax, but only the exact meaning of a few Lojban words; Dotsides is a very minor grammar change only affecting lojbanized foreign personal names 14:00 < Ilmen> The CLL (Complete Lojban Language, the reference grammar) book is currently in the slow process of being updated 14:00 < Ilmen> but as it's a volunteering undertaking, it's slow 14:00 < Ilmen> (The CLL is a big book!) 14:02 < Ilmen> You can still very well learn Lojban using the current CLL release (which doesn't contain yet the Xorlo and Dotside reforms), because Lojban hasn't changed much since the one described in the CLL 14:04 < Ilmen> However the CLL isn't intended as a tutorial or learning guide, it's a reference grammar; starting with one of the various Lojban courses available is probably better; the Wave Lessons and the Crash Course are famous ones 14:04 < Ilmen> mu'o su'o no jundi be di'u 14:05 < ctefaho> coi Ilmen 14:07 < Ilmen> coi do 14:07 < ctefaho> do mo ma 14:11 < Ilmen> mi zo'u zilcte gi'e vacri glare 14:16 < Ilmen> .i cu'uxna fami ca'o feda foloka bazukte fami fece 14:52 < Ilmen> exp: ze'a lo nu ro tarmi lo se pensi be lo remna cu nitcu fa ma'a lo ka sanga lo jai se jalge be lo nu ri pensi 14:52 < mensi> ([ze'a {lo tarmi KU} {lo <(²se pensi²) (²be [lo remna KU] BEhO²)> KU}] [cu {nitcu I> KU} BEhO]³) KU> VAU}] KEI²) KU>} VAU]¹) KEI> KU}] VAU) 14:53 < Ilmen> ŭa nai 14:55 < Ilmen> .a'o tincpa fliba 17:48 < zipcpi> vlaste: vujnu (gimka) 17:48 < vlaste> zo vujnu gimkamsmikezypro no catni gismu .i zo vujnu gimkamsmikezypro no cipra gismu 17:56 < phma> vlaste: gimka 17:56 < vlaste> no results. http://vlasisku.alexburka.com/gimka 17:56 < phma> vlaste: ginka (gimka) 17:56 < vlaste> zo ginka gimkamsmikezypro no catni gismu .i zo ginka gimkamsmikezypro no cipra gismu 17:57 < phma> vlaste: gimka (ginka) 17:57 < vlaste> no results. http://vlasisku.alexburka.com/gimka 17:57 < phma> vlaste: ginka 17:57 < vlaste> ginka = x1 is a camp/encampment/temporary residence of x2 at location x3. 17:57 < zipcpi> lol the gimka thing is just a shortening, but too bad it's a clashing gismu 17:57 < zipcpi> I had suggested {gimpro} 17:59 < phma> I've been translating the middle third of Ki Tetze, which will be the parashat ha shavua shel jbonunsla, and which talks about camping for war, so the word is on my mind. 17:59 < zipcpi> Cool 17:59 < zipcpi> I wanted {vujnu} as the gismu for subtraction, so that it's more semantically separated from {selsumji}, which might be "addend" 18:03 < zipcpi> Gotta go. co'o 19:55 < klaku> coi lojbo prenu .i do mo 20:07 < klaku> za'a na'e jundi vau ui 21:23 < astroendjisai> Hello. I have a question. Can 'n' ever sound like 'ng' like in 'sing' (such as before G or K) or is it always 'n' 21:23 < durka42> yeah, you can pronounce it that way 21:23 < durka42> it's an allophone 21:24 < astroendjisai> Sorry my browser glitched 21:24 < durka42> but if there's a 'g' or 'k' after the 'n', you have to pronounce that too :) 21:24 < durka42> oh whoops 21:24 < durka42> yes you can pronounce it that way 21:24 < durka42> it's an allophone 21:24 < astroendjisai> Okay cool. 21:24 < astroendjisai> Thanks 22:01 < zipcpi> Why does {pilji} get a rafsi but not anything else? I suppose they ran out T.T 22:03 < zipcpi> Hmm... I'm begining to see the utility of xe'e being in PA 22:03 < zipcpi> xe'e le drata 22:03 < zipcpi> le xe'e drata doesn't quite work the same way 23:01 < zipcpi> Hm well if {xe'e} is the anti-{mo'oi} then maybe it should be in LE... but that would break using bare {xe'e} as an answer to a {mo'oi} question 23:02 < gleki> i need to also explain UD operators like {dei} somewhere 23:02 < zipcpi> Hm 23:03 < gleki> because {lo mlatu pe dei cu pinxe} - The UD cat drinks 23:03 < gleki> and {do'e nai dei mlatu cu pinxe} - Cats in general drink 23:03 < gleki> and {do'e nai dei lo mlatu cu pinxe} - Cats in general drink 23:04 < zipcpi> Interesting 23:04 < gleki> altough va'o and na'o is what is usually meant so all my previous comments on that Talk page still stand 23:05 < zipcpi> The thing is even {na'o}... it could mean {na'oku la .mitynz. cu pinxe}, or {na'oku lo su'a mlatu cu pinxe} 23:06 < zipcpi> So I'm not sure {na'o} actually solves the problem; merely throws more things to King Context 23:06 < gleki> of course it adds more things otherwise how could you get the distinction 23:07 < gleki> also note that {na'o} explicitly adds things to context, not denies them from context 23:07 < zipcpi> Yes, but there is still the problem of whether {lo mlatu} refers to a specific cat like one that might be named Mittens, or a generalization of all cats 23:08 < gleki> lo mlatu pe dei as i said 23:09 < zipcpi> Well if that's what it takes to formalize {le}, *shrugs* 23:09 < gleki> no the same as {i da poi mlatu zo'u do broda ije da brodi} 23:09 < gleki> not the same as {i da poi mlatu zo'u do broda ije da brodi} 23:09 < gleki> so not about {le}, no 23:09 < gleki> nor about {da} 23:10 < gleki> i just dont know what are your problems and you are saying that you know them intuitively. so i just added one more answer to a question that im trying to guess 23:12 < zipcpi> And I don't know what is it that you are having trouble understanding... 23:12 < gleki> why do you need {le} and what it is 23:12 < gleki> and what is "apples" in general 23:13 < zipcpi> It's the difference between the specific cat that might be named Mittens, and just a generalization of all cats without referring to any one cat 23:14 < zipcpi> You told me that Russian uses demonstratives instead of articles to refer to deictic/specific cats, but in Lojban we don't use ti/ta/tu that way 23:15 < zipcpi> Thus I think the "malgli" solution is the best solution right now 23:15 < gleki> both {da} and {dei} solutions can work here. they mean different things, though. 23:15 < gleki> i solved all your questions and added one more with {dei}. so im struggling to understand why you still need this {xe'e} 23:16 < zipcpi> Oh it's xe'e now? That is for "Choose any two apples" 23:16 < zipcpi> {ro} doesn't work here 23:17 < zipcpi> Like I said, it's the anti-mo'oi, "which" 23:18 < zipcpi> It makes no sense to ask "which two apples" after I told you choose "any two apples" 23:19 < zipcpi> {lo} here doesn't disambiguate, because I can still ask "which" 23:20 < gleki> the two apples here can be either expressed using {da} if you preserve your variable across UD or with {dei} which is just about UD 23:21 < zipcpi> ko cuxna re da poi me lo plise ? 23:22 < gleki> {re plise} 23:22 < gleki> {re da poi plise} 23:22 < zipcpi> What about the reverse case then... "Which apple do you want"? 23:22 < gleki> but for that {da} should be declared somewhere 23:22 < gleki> zipcpi, we already discussed that 23:22 < gleki> if you need logs 23:23 < zipcpi> Hm... lo se cuxna be do... 23:23 < zipcpi> Not sure if that breaks down somewhere; will have to think about it 23:24 < zipcpi> "Which of the members should complete the project" 23:25 < gleki> i have all the logs. by request i email them 23:25 < zipcpi> I do too 23:25 < zipcpi> At least, logs of when I'm present 23:27 < zipcpi> But yeah... ""Which of the members should complete the project?"... that might be where {lo se cuxna be do} doesn't quite work 23:28 < zipcpi> mo'oi cmima be le fuzykamni cu ei mulgau le gungunma 23:31 < zipcpi> Perhaps the "correct" answer is {da}, but that might not be understood 23:36 < zipcpi> Maybe I'll wait to ask a few others about {da}. Anyway I gotta nap. See ya --- Day changed Sun Jul 05 2015 01:03 < fghjebj> ithkuil: amnadya 01:03 < ithkuil> amnadya: P1S1/STA-mn-OBL-UNI/CSL/A/PRX/RPV-EXS-IFL/UNFRAMED 01:07 < fghjebj> ithkuil: sorpeka 01:07 < ithkuil> sorpeka: P1S1/STA-s-ERG-AGG/VAR/U/DEL/NRM-SSD_1/3-EXS-IFL/UNFRAMED 01:09 < fghjebj> ithkuil: arxokuna 01:09 < ithkuil> arxokuna: P1S1/STA-rx-ERG-UNI/ASO/U/DEL/NRM-AGC_1/9-EXS-IFL/UNFRAMED 01:11 < gleki> ma smuni zo'oi rx 01:12 < gleki> -RX- ‘OBLIQUELY LINEAR MOTION (relative to XYZ grid) The stems of this root are patterned after the root -K- 01:12 < gleki> wth 01:12 < gleki> i mi muvdu ma 01:12 < gleki> i ju'ocu'i mi muvdu to'o la itku'ile 01:30 < gleki> ithkuil: gleki 01:30 < ithkuil> gleki: P1S1/STA-gl-ABS-UNI/ASO/U/DEL/NRM-FNC-IFL/UNFRAMED 01:31 < gleki> ithkuil: lo mlatu cu pinxe lo ladru 01:31 < ithkuil> lo: Personal adjunct 01:31 < ithkuil> mlatu: P1S1/STA-ml-OBL-UNI/CSL/U/DEL/NRM-AMG-IFL/UNFRAMED 01:31 < ithkuil> cu: Personal adjunct 01:31 < ithkuil> pinxe: P1S1/STA-p-AFF-nx-RPS-IFL/UNFRAMED 01:31 < ithkuil> lo: Personal adjunct 01:31 < ithkuil> ladru: P1S1/STA-l-OBL-UNI/CSL/U/TRM/RPV-AMG-IFL/UNFRAMED 01:39 < gleki> alta: lo mlatu goi pinxe 01:39 < mensi> ([FA {lo <(¹mlatu NF¹) (¹goi [FIhA {ZOhE KU}] GEhU¹)> KU}] [CU {pinxe NF} VAU]) 01:39 < gleki> now tag can't be {TAG ku} but only {TAG sumti ku} 01:39 < gleki> hence ZOhE is inserted 03:51 < ctefaho> coi 03:54 < gleki> coi 04:22 < ctefaho> gleki, when you added "smaka" what did you base it on? 04:36 < Amnadya2> . 04:40 < Amnadya2> It seems lojban is way better at scope than ithkuil 04:43 < ctefaho> Does Ithkuil not have well-defined scope then? 04:45 < zipcpi> Hm... perhaps the problem with {ko cuxna re plise} actually lies in {ko} 04:45 < zipcpi> Compare 04:45 < zipcpi> re da poi plise zo'u: mi djica lo nu do cuxna da 04:46 < zipcpi> vs 04:46 < zipcpi> mi djica lo nu, re da poi plise zo'u: do cuxna da 04:47 < zipcpi> First: There are two apples such that I want you to choose them; which means I want you to choose those two apples, and nothing else 04:47 < zipcpi> Second, I want there to be two apples such that you choose them; which means I just want you to choose two apples 04:49 < zipcpi> Compare perhaps, {mi viska re prenu}; "I see two people"; even here, {mo'oi re prenu} "Which two people" might still apply, asking for more information 04:50 < zipcpi> On the other hand I don't think that {mi viska xe'ere prenu} would be meaningful 04:52 < zipcpi> Perhaps that's the similar issue with {lo}/{le}... whether the prenex starts after {mi xusra lo du'u}, or before 04:52 < gleki> ctefaho: in meaning or in etymology? 04:52 < zipcpi> (Issues with {da} vs {zo'e} and {mi nitcu lo mikce} aside) 04:53 < gleki> i ta'o ba'a sy ba tinke la itku'ile 04:53 < demize> gleki: etymology. ;) 04:53 < gleki> demize: on English, Scandinavian, Russian 04:53 < demize> Mmm.. Figures. 04:54 < demize> {smaka} is 'smaka' in Swedish. 04:54 < demize> Which was why he asked. 04:54 < gleki> smack 04:54 < gleki> [smak] in Russian 04:54 < gleki> i just open wiktionary and search there 04:54 < gleki> for lojban-looking words 04:55 < gleki> vlaste: bacru (gimka) 04:55 < vlaste> zo bacru gimkamsmikezypro no catni gismu .i zo bacru gimkamsmikezypro no cipra gismu 04:55 < gleki> vlaste: datru (gimka) 04:55 < vlaste> zo datru gimkamsmikezypro zo tatru noi catni gismu .i zo datru gimkamsmikezypro no cipra gismu 04:55 < gleki> vlaste: smaka (gimka) 04:55 < vlaste> zo smaka gimkamsmikezypro no catni gismu .i zo smaka gimkamsmikezypro no cipra gismu 04:56 < gleki> vlaste: gosta (gimka) 04:56 < vlaste> zo gosta gimkamsmikezypro zo kosta noi catni gismu .i zo gosta gimkamsmikezypro no cipra gismu 04:56 < gleki> vlaste: tasta (gimka) 04:56 < vlaste> zo tasta gimkamsmikezypro no catni gismu .i zo tasta gimkamsmikezypro no cipra gismu 04:56 < gleki> mi pu djica tu'a zo'oi gosta i ku'i za'adai 04:56 < ctefaho> je'e 04:58 < zipcpi> What did you want {gosta} to mean? 04:59 < ctefaho> me gusta 04:59 < gleki> probably what tasta means now 05:00 < zipcpi> Ah 05:05 < _mukti_> ue 05:06 < _mukti_> en: seldri 05:08 < gleki> heh https://en.wiktionary.org/wiki/Category:Lojban_phrases 05:09 < _mukti_> valste: seldri 05:09 < _mukti_> vlaste: seldri 05:09 < vlaste> seldri = x2=b1 is sad/depressed/dejected/[unhappy/feels sorrow/grief] about x1=b2 (abstraction). 05:09 < _mukti_> How can {seldri} have the same place structure as {badri} ?! 05:11 < gleki> it doesnt 05:12 < gleki> as for mensi its again vrici's problems 05:13 < _mukti_> Is it a particular bad definition? I see that {selte'a} is accurate. 05:15 < gleki> just read it carefully 05:15 < gleki> en: coi 05:15 < mensi> coi = [COI] vocative: greetings/hello. |>>> officialdata 05:15 < _mukti_> Oh, I see. It's giving x2 first. 05:16 < gleki> no idea what's the use of such definitions 05:16 < gleki> some understanding that newer generations of lojbanists lost. 05:16 < _mukti_> Yes, given that they confuse someone like myself who has had some exposure. 05:17 < gleki> neither i have that understanding 05:26 < zipcpi> Yeah I think if the SE-rafsi lujvo are to be useful, they must be tied to a gloss word somehow and reworded, rather than just changing the numbering 05:27 < zipcpi> It's like how {terve'u} might be used to convey the meaning of "buy", vs "was sold to" 05:29 < gleki> or maybe they should be ignored and never used 05:29 < zipcpi> Perhaps the problem with CLL-{le} was that it focused too much on the non-veridicality aspect, when (I think) that fact is incidental rather than definitional. The point is to describe an object to make a reference to it 05:29 < gleki> zei: mi terve'u ti 05:29 < mensi> mi te vecnu ti 05:30 < zipcpi> zei: mi selzauka'e ti 05:30 < mensi> mi se zei zanru zei kakne ti 05:31 < zipcpi> Hmm... 05:31 < _mukti_> I see the non-veridical aspect of CLL {le} as being strongly related to the notion of reference: The reference is primary, and the description is explicitly weakened in favor of it. 05:31 < zipcpi> Yeah 05:32 < gleki> Loglan and partially CLL shows the desire to embed some logician's theories into the language. 05:33 < gleki> Loglan would have turned into another Laadan or Esperanto until Jeff Prothero did his work putting the end to this disease. 05:35 < gleki> there shouldnt be any philosophies embedded into the language except the instrument of operating the text itself. 05:36 < zipcpi> gleki: Well I have observed there to be a trend for some distaste for using affixed opposites... {tolpo'u} eventually became {akti}, {tolci'o} eventually became {laldo}, {tolcri} seems to have stuck though 05:36 < gleki> jbo: akti 05:37 < mensi> akti = x1 ca'a zilpilno x2 |>>> ko catlu fe ji'a zo pilno |>>> gleki 05:37 < gleki> jbo: tolpo'u 05:37 < mensi> tolpo'u [< to'e spofu ≈ To'e spofu*] = x1 to'e spofu x2 |>>> srana fa zo spofu |>>> 05:37 < mensi> Ilmen 05:37 < gleki> jbo: spofu 05:37 < mensi> spofu = x1 na ka'e akti x2 |>>> gleki 05:37 < gleki> yes 05:37 < zipcpi> lol 05:37 < gleki> so tolpo'u != akti 05:38 < zipcpi> Erm, I think the meaning is the same, we've just refactored the definitions, changing the primitives 05:39 < gleki> the key difference is in ka'e/ca'a 05:39 < _mukti_> Well, there's a well-established tradition for the relation of reference to description. 05:40 < zipcpi> gleki: So a computer that's turned off isn't {akti}? 05:40 < _mukti_> On the one hand, there's the idea of descriptions which "rigidly designate" to the extent that they are functionally equivalent with reference. 05:40 < zipcpi> I have no idea what {zilpilno} is 05:40 < _mukti_> And on the other hand, descriptions which are shorthand for a reference which is primarily subjective. 05:41 < _mukti_> CLL {le} seems to be a nod in the direction of the latter tradition. 05:41 < gleki> akti implies ca'a 05:41 < zipcpi> zilpilno isn't even defined 05:41 < gleki> tolpo'u doesnt 05:42 < gleki> zilpilno = zi'o pilno 05:42 < gleki> that's another jvajvo rule 05:42 < _mukti_> The canonical example of {le nanmu cu ninmo} is an example of that. 05:42 < zipcpi> Yeah 05:42 < gleki> i wish la samyuan si das could create a wiki page with all jvajvo rules 05:43 < zipcpi> Is a computer that is turned off {akti}? 05:43 < gleki> ofc. only reference in the meaning of parsing text is the only justifiable concept. 05:43 < gleki> zipcpi: no since it's not ca'a se pilno gi'a ca'a zilpilno 05:44 < zipcpi> Oh OK 05:45 < zipcpi> Hm where does {cando} fit into this... 05:45 < gleki> nowhere 05:45 < zipcpi> How do you say "turn off/turn on" then? 05:46 < gleki> gasnu lo nu co'u ca'a zilpilno 05:47 < zipcpi> Hm... longish... 05:47 < gleki> gau ko'a na akti then 05:47 < gleki> we should teach people ergative case style more 05:47 < gleki> ko'a jai gau akti be na fai ... 05:48 < gleki> even if fix grammar then 05:48 < gleki> ko'a jai gau na akti fai ... 05:48 < gleki> still long 05:48 < gleki> gau ko'a na akti <-- but this isnt translated into English word by word 05:49 < zipcpi> The thing is that the "ergative case style" isn't just problematic by being difficult for English speakers; it also breaks {gi'e} / {jecu} 05:49 < gleki> then dont use gi'e 05:50 < gleki> although i dont see how it breaks 05:50 < gleki> mi klama ije gau mi lo canko cu kalri? 05:50 < zipcpi> Then what? {.i lo jai gau go'i}? 05:50 < zipcpi> The sumti might be much more complicated than {mi} 05:50 < gleki> {mi} is already a sumti 05:51 < gleki> alta: mi 05:51 < zipcpi> Even something like {da}, you want to keep the reference somehow 05:51 < gleki> vrici! 05:51 < gleki> iunai 05:51 < mensi> ([FA mi] [CU {COhE VAU}]) 05:51 < gleki> _mukti_: what's going on with vrici atm? especially with network 05:51 < _mukti_> gleki: Not sure. What are you experiencing? 05:52 < gleki> im experiencing livla.js not responding fast 05:52 < gleki> alta: mi jai gau akti 05:52 < mensi> ([FA mi] [CU {jai gau } VAU]) 05:52 < _mukti_> I'll check vrici 05:52 < gleki> alta: mi na akti 05:52 < mensi> ([FA mi] [CU {na } VAU]) 05:52 < gleki> shall i fix this: 05:52 < gleki> alta: mi jai gau na akti 05:52 < mensi> SyntaxError: Expected [,] but "a" found. 05:53 < gleki> hm, now responds fast 05:53 < gleki> some backups were running again probably 05:53 < _mukti_> I annotated "prami simbi'o": http://genius.com/Djemynai-prami-simbio-lyrics/ 05:54 < gleki> en: simbi'o 05:55 < gleki> wth 05:55 < gleki> again slow! 05:55 < gleki> _mukti_: see? it's not replying fast 05:55 < mensi> simbi'o [< simxu binxo ≈ Mutual become] = s1=b1 (mass/set) integrate / unite with each other, displaying quality s2. 05:55 < mensi> |>>> Cf. simxu, jmaji, jorne. |>>> tijlan 05:57 < zipcpi> It breaks because gi'e/jecu depends on copying all the sumti before the main selbri, including all sumtcita, not so useful if one statement is "ergative" but the other isn't. The alternative is using {.i le jai gau go'i}, oftentimes 05:57 < gleki> let's try the fix 05:59 < gleki> alta: mi jai gau na akti 05:59 < mensi> ([FA mi] [CU {jai gau na } VAU]) 06:00 < zipcpi> alta mi broda jai akti 06:00 < zipcpi> alta: mi broda jai akti 06:00 < mensi> ([FA mi] [CU {broda } VAU]) 06:01 < gleki> i just allowed {jai tag? (tag / NA_clause)* broda} 06:01 < gleki> wait 06:01 < gleki> what could {jai na akti} mean? 06:01 < _mukti_> Hmm... seeing slowness logging into vrici 06:01 < gleki> _mukti_: oh nice! very nice! 06:01 < zipcpi> gleki: Dunno, but {jai na fenki}? 06:01 < zipcpi> Though that could just be {na jai fenki} 06:02 < gleki> how is {na} expanded? 06:02 < gleki> jbo: na 06:02 < mensi> na = [NA] tcita lu'e lo bridi le du'u sinxa le natfe be la'e ri |>>> bridi natfe; ja'a |>>> 06:02 < mensi> xorxes 06:02 < zipcpi> Not sure... jitfa? natfe? 06:03 < gleki> ko'a natfe lo du'u broda? 06:03 < zipcpi> I don't know how you can simulate right expanding scope of NA and sumtcita though 06:03 < zipcpi> That is just a little beyond me at this point 06:03 < gleki> arent NA clauses have the same scope as all tags except FA? 06:04 < zipcpi> They didn't use to be, they used to have topmost scope when I was first taught, dunno if that has changed lately as according to that wiki article; I think it makes sense 06:05 < gleki> im tired of fighting over that 06:05 < gleki> and the scope isnt displayed by any parsers anyway 06:05 < zipcpi> Yeah that is a problem... 06:05 < gleki> today i also banned {bai ku} in altatufa 06:05 < gleki> but ... it still works 06:05 < gleki> oh hm, maybe not 06:06 < zipcpi> Why would you want to ban that? It's just {bai zo'e} 06:06 < _mukti_> There's a busy tarsnap process running as root on vrici. Probably a cronjob. 06:06 < gleki> alta: bai gau broda 06:06 < mensi> ([FA ZOhE] [CU { } VAU]) 06:07 < gleki> alta: bai cu broda 06:07 < _mukti_> Yeah, it's a backup. 06:07 < mensi> ([{bai ZOhE} {FA ZOhE}] [cu {broda NF} VAU]) 06:07 < gleki> _mukti_: why cant we/robin make these backups e.g. slower but consuming so many resources? 06:07 < _mukti_> I expect that after it's done running vrici will become more responsive. 06:08 < gleki> but it happens every day 06:08 < _mukti_> gleki: I'm sure it should be possible to either reschedule the backups and/or make them "nicer" 06:11 < gleki> _mukti_: since it happens every day it'd be of high priority for me since eveyone uses mensi nad naxle 06:11 < gleki> *and 06:14 < gleki> en:coi 06:14 < mensi> coi = [COI] vocative: greetings/hello. 06:15 < gleki> alta: bai fa mi mo 06:15 < mensi> ([{bai ZOhE} {fa mi}] [CU {mo VAU}]) 06:15 < gleki> alta: bai ku fa mi mo 06:15 < mensi> ([{bai ku} {fa mi}] [CU {mo VAU}]) 06:19 < gleki> so i removed KU_elidible everywhere except for the end of sumti 06:33 * nuzba @Saffy: Currently @MariiKusha is obsessed with languages and apparently I taught her some sign - @alaric is getting questions about #lojban [http://bit.ly/1JGYr2Z] 06:34 < zipcpi> mukti: Should I attempt a translation? A semantic one though; not a poetic one or one meant to be sung 06:38 < zipcpi> Does this website even allow translations... 06:55 < gleki> how to translate "hover" in computer sense? 06:55 < gleki> (computing) To place the cursor over a hyperlink or icon without clicking. 06:56 < zipcpi> I thought of klikenu'o but I'm not sure I added that 06:56 < gleki> they should be in a table in the wiki 06:57 < zipcpi> Yeah apparently I didn't add that 06:58 < gleki> there should be a list of all such functions 06:59 < zipcpi> Yeah 06:59 < gleki> hm, maybe this? https://api.jquery.com/category/events/mouse-events/ 07:01 < zipcpi> "toggle"; hm that might go with "turn on" and "turn off" 07:02 < zipcpi> Does {akti} means turned on *and* tolpo'u though, or just turned on? 07:02 < gleki> in Vivaldi i used {vraga} for toggle 07:04 < gleki> en: e'a 07:04 < mensi> e'a = [UI1] attitudinal: granting permission - prohibiting. |>>> See also curmi. |>>> 07:04 < mensi> officialdata 07:04 < gleki> jb: e'a 07:04 < mensi> e'a = e'a [interjection] — you may, I allow, e'a nai — prohibiting 07:04 < mensi> :e'a do klama lo nenri — You may come in. 07:04 < mensi> :zu'e e'a nai fagri — No fire allowed! 07:04 < mensi> :e'a nai crepu lo flora — You may not pick the flowers. 07:04 < mensi> :Related words: permite 07:04 < zipcpi> I've heard of {cikna} for "on", but that might be metaphorical 07:06 < zipcpi> Wuzzy has made some lujvo for these: http://jbovlaste.lojban.org/dict/ca'arcadgau 07:06 < gleki> "on" for what? 07:06 < zipcpi> They're a bit long though, and basing them off {cando} is a bit counterintuitive 07:06 < gleki> en: ca'arcadgau 07:06 < mensi> ca'arcadgau [< cabra cando gasnu ≈ Apparatus idle bring about] = g1 (agent) turns device/apparatus g2=cab1=can1 off. 07:06 < zipcpi> "turned on" 07:06 < mensi> |>>> See also: turn on (=narca'acadgau), turned off (=ca'arcando). |>>> Wuzzy 07:06 < zipcpi> sorry 07:06 < zipcpi> I thought since we were talking about computers 07:06 < gleki> for aktygau then 07:07 < zipcpi> I'm a bit confused though, because something can be turned on *and* spofu 07:08 < gleki> explain 07:08 < zipcpi> Computer is turned on, but displaying BSOD or otherwise not functioning correctly? 07:09 < gleki> it's functioning 07:09 < gleki> en: akti 07:09 < mensi> akti = x1 is running service/in operation/performs functions x2; x1 is working, acting, active, in action |>>> See also 07:09 < mensi> ca'a se pilno, tolspofu |>>> gleki 07:09 < zipcpi> Yeah but I don't think you can say it isn't spofu 07:10 < gleki> akti ko'a gi'e spofu ko'e 07:10 < zipcpi> Hm 07:11 < zipcpi> Is it fine if I add "turned on" to the gloss of {akti} then? 07:11 < zipcpi> Then add {aktigau} and {to'aktigau} 07:12 < zipcpi> Well if I'm gonna shorten toly'akti that way might as well add {to'akti} too 07:14 * nuzba @frecafloros: lojbanの品詞は3つ cmavo brivla cmene [http://bit.ly/1TdTPTB] 07:15 < gleki> once i proposed z- =.- for that. i.e. it'd be {tolzaktygau} 07:15 < zipcpi> Ah 07:15 < gleki> although in my translation i always just used nonlujvo style 07:15 < gleki> na akti 07:16 < gleki> oh you reminded me that i need to localize Skype 07:16 < gleki> although they will probably shut down offline version soon 07:19 < gleki> ei traduki 2847 linji 07:19 < zipcpi> Meh it might be time for some agglutinativity, relrafsi style... akti'yga. Not supported yet though, and needs to be defined properly 07:20 < gleki> en: gigle 07:20 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 07:20 < gleki> en: giggle 07:20 < mensi> [< gigdo gletu ≈ 1E9 fuck] 2 da se tolcri: mi'amli, mlimi'a 07:20 < gleki> en: giggl 07:20 < mensi> 2 da se tolcri: mi'amli, mlimi'a 07:20 < gleki> en: mi'amli 07:20 < mensi> mi'amli [< cmila milxe ≈ Laugh mild] = m1=c1 titters/chuckles/giggles. |>>> Synonym to mlimi'a. Cf. cmila, milxe. |>>> 07:20 < mensi> totus 07:20 < gleki> en: mlimi'a 07:20 < mensi> mlimi'a [< milxe cmila ≈ Mild laugh] = c1=m1 titters/chuckles/giggles. |>>> Synonym to mi'amli. Cf. cisma, mi'afra, 07:20 < mensi> mlimi'afra. |>>> sarefo 07:22 < gleki> .dict cash 07:22 < fenki> cash — noun: 1. Money in the form of notes/bills and coins, as opposed to cheques/checks or electronic transactions, 2. (informal) Money — verb: 1. (trans.) To exchange (a check/cheque) for money in the form of notes/bills, 2. (poker slang) To obtain a payout from a tournament 07:22 < gleki> en: cash 07:22 < mensi> 11 da se tolcri: bancuburu, bancubusu, nargrkaju, bankeve'a, bankevesu, bankevexe, bankexu'a, bankexunu, bankexu'o, 07:22 < mensi> spatrnakardiace, vlalikei 07:22 < zipcpi> certyzegase 07:23 < gleki> marji jdini? 07:23 < zipcpi> Maybe... though checks are marji too... 07:25 < gleki> are they jdini? 07:25 < gleki> not debts? 07:25 < zipcpi> I hand it to someone like I would money... not sure how that would be rendered though 07:27 * nuzba @CheshireOcelot: @heuristics solution: write all laws in Lojban. [http://bit.ly/1CiP731] 07:29 < gleki> 2800 strings to translate 07:30 < zipcpi> ko'a certyzegase ko'e ko'i ko'o -> ko'a sefsi lo ka ce'u certyzega ce'u ko'e ko'i ko'o -> ko'a sefsi lo ka fo'a fo'e ce'ai fo'a gasnu lo ka fo'e certyze ko'e ko'i ko'o -> ko'a sefsi lo ka fo'a fo'e ce'ai fo'a gasnu lo ka fo'e zenba lo ka ce'u certu ko'e ko'i kei ko'o 07:33 < zipcpi> coi jboku'ile zo'o 07:35 < gleki> щр ьню 07:35 < gleki> oh my 07:35 < zipcpi> Means roughly what crezenzu'e / crenzuue means 07:35 < zipcpi> See, even that lujvo isn't regular. Left out the sefsi 07:38 < gleki> which lujvo 07:39 < zipcpi> crezenzu'e 07:41 < zipcpi> So yeah, agglutinativity is powerful in numbers, but can be difficult to use 07:43 < gleki> zukte lo ka zenba lo ka certu 07:43 < gleki> how else shuld it be 07:43 < zipcpi> lo ka ma zenba lo ka ma certu 07:44 < gleki> lo ka ce'u zenba 07:44 < gleki> zukte2 is ka 07:44 < zipcpi> Oh 07:44 < gleki> although with this tag proliferation it's 07:44 < gleki> zu'e zei'a certu 07:44 < zipcpi> u'i 07:46 < zipcpi> One nice thing about relrafsi though is that they are very regular. All words can be modified by -'yga. Gismu can be shortened as brodyga 07:46 < gleki> we need CV equivalents for guaspi \ and / 07:47 < zipcpi> None of the jvova'i shortening like cmibi'o and nurgau that obscure the first particle 07:50 < ctefaho> gleki: what where guaspi \ and / exactly? sounds? tones? 07:53 < ctefaho> looks like tones alright 07:55 < ctefaho> Why not just use them the same way;) 07:57 < noncomcinse> coi 08:08 < latro`a> coi 08:21 < noncomcinse> coi latro'a 08:41 < selpli> coi prenu 08:41 < noncomcinse> coi la selpli 08:41 < noncomcinse> do mo 08:43 < selpli> tatpi ru'e .i caku ctemidju 08:44 < noncomcinse> ua 08:44 < gleki> ctefaho: / is like {be}. \ is the reverse 08:44 < gleki> so it's something like {zukte be zenba be certu} 08:45 < noncomcinse> .i mi ji'a tatpi ru'e ni'i nai lonu la .pavnon.nanc. tcika 08:45 < selpli> zo noncomcinse uanai 08:46 < noncomcinse> .i zo nonpamjveglecinse cu dukse loka barda co .irci cmene 08:47 < noncomcinse> .i lo go'i cu ve jmina zo nonpamcinse zo nonglecinse 08:50 < noncomcinse> jbo: nonglecinse 08:50 < mensi> nonglecinse [< no gletu cinse ≈ No gletu* cinse*] = .i x1 cinse fi loka ce'u gledji noda |>>> 08:50 < mensi> nejni-marji 08:50 < noncomcinse> jbo: nonpamcinse 08:50 < mensi> nonpamcinse [< no prami cinse ≈ No prami cinse*] = .i x1 cinse fi loka ce'u prami noda |>>> 08:50 < mensi> nejni-marji 08:51 < selpli> ua 08:53 < noncomcinse> .i ba'a cu'i lo so'e prenu na se slabu 08:54 < gleki> alta: zukte be zenba be certu 08:54 < mensi> ([FA ZOhE] [CU {<(¹zukte [be {FIhA ZOhE} BEhO]¹) (¹zenba [be {FIhA ZOhE} BEhO]¹)> } VAU]) 08:55 < gleki> not sure this parse should be the correct one 09:06 < Amnadya> oi oi oi i lo soho cfila co itkuhile cu fanza je badryriha mi i mi xebni lonu senpi lodu inda loka setadni 09:07 < Amnadya> i mi sohiroipu lifri lo simsa pe la lojban 09:07 < noncomcinse> y 09:07 < Amnadya> i caronu sanji suho nalprane cu coha senpi rodapelobangu 09:08 < Amnadya> xu suhodo jimpe 09:09 < noncomcinse> bazi 09:09 < Amnadya> xu suho do dainmo 09:10 < noncomcinse> .i mi no'e cilre fi la .itku'ile bangu 09:10 < noncomcinse> .i mu'a dai ma cfila 09:10 < noncomcinse> je'e 09:11 < Amnadya> loseskicu bemi secinmo na steci la itkuhile 09:12 < noncomcinse> y 09:13 < noncomcinse> .i ja'o na jimpe 09:16 < gleki> mi dainmo i kuhi no bangu cu prane 09:16 < noncomcinse> .i ja'a go'i 09:16 < gleki> i kuhi la itkuiile zohu mi senpi lo nu le bangu cu poihi kahe se bangu keha 09:17 < Ilmen> loka troci cu.olkai loka tadji codjuno 09:18 < gleki> i taho ma smuni zohoi gleki la itkuiile 09:18 < gleki> ithkuil: gleki 09:18 < ithkuil> gleki: P1S1/STA-gl-ABS-UNI/ASO/U/DEL/NRM-FNC-IFL/UNFRAMED 09:18 < noncomcinse> .i xu djedi co rirni loka pilno zo'oi .h. .enai zo'oi .'. 09:18 < Amnadya> ko catlu lo vlaste 09:18 < gleki> laha 09:18 < Ilmen> rirni ji rirci 09:19 < Amnadya> mi senandu kihutuhalo fonxa 09:19 < gleki> lo crisa rirni cu rirci crino 09:19 < gleki> i ua 09:19 < Ilmen> coi ru'e ta'o 09:19 < noncomcinse> ie 09:19 < noncomcinse> lo'ai rirni sa'ai rirci le'ai 09:20 < gleki> Amnadya: xu do ca pilno lo pelji cukta lo ka tadni le bankixada 09:21 < Amnadya> laha lonu cikre la itkuhile cu cumki ikuhi lofinti nalaha djica 09:21 < Amnadya> i jahapilno lo pelji 09:24 < gleki> gau do ehu la tohanzu cu gletu la itkuiile 09:25 < gleki> gi'e cupra lo panzi 09:30 < Amnadya> mi nelci lo zabna pagbu be ihy ikuhi lo nu ropagbu cu prane sahenai cusarcu lonu mi panpi 09:31 < Amnadya> i lojbobau cusimsa 09:32 < Amnadya> mi pupensi lonu jbobau itkuhile mixre 09:32 < gleki> oi oi 09:32 < gleki> i lo tarmi be lo ibu valsi cu klina 09:32 < Amnadya> cinri falonudostidi tuha la tohanzu 09:32 < gleki> i kuhi lo tarmi be lo lojbo bridi kahe kalsa 09:32 < Amnadya> ie 09:33 < gleki> i da itkuiile ke valsi morna 09:33 < gleki> sisi rafsi morna 09:36 < gleki> i krinu lo nu mi na nelci lo lujvo 09:36 < mensi> i mi i mi i mi mo i mi na nelci 09:37 < Amnadya> mi na toltugni 09:48 < Amnadya> xusuho cinri buhu la jboguhe caho fasnu 10:21 < gleki> y 10:21 < Amnadya> mo 10:21 < gleki> mi mo'u finti lo zmiku be loka zbasu lo glico bi'o lojbo vlaste 10:21 < gleki> i do pu kucli tua lonuzba 10:21 < gleki> i mi spuda 10:21 < Amnadya> ieie 10:21 < gleki> i jia mi favgau la altatufa 10:22 < gleki> alta: bai fa mi mo 10:22 < mensi> ([{bai ZOhE} {fa mi}] [CU {mo VAU}]) 10:22 < gleki> za'adai zo ku na se setca 10:22 < gleki> i zo zo'e se setca 10:22 < Amnadya> sahenei do zukte zuhinedo 10:22 < Amnadya> sahenai 10:23 < gleki> alta: lo brodi brode ui ne do cu brodi brodu 10:23 < mensi> ([FA {lo <(¹brodi [brode {NF ui}]¹) (¹ne [FIhA do] GEhU¹)> KU}] [cu {brodi } VAU]) 10:23 < gleki> zo'oi NF ce zo'oi VAU sinxa lo tertau genrverbu 10:23 < gleki> i vajni te frica 10:23 < gleki> gloss: lo dansu cu dansu 10:23 < mensi> A(n) dancer(s): dance(s) 10:24 < gleki> zo'oi dancer i ku'i zo'oi dance 10:24 < gleki> i ku'i mi nu'o setca lo vlaste la mensi 10:24 < gleki> i ji'a mi milxe ciska lo ve ciksi be tu'a zo le ce zo dei ce zo pe ce zo co'e ce zo sko'opu 10:25 < gleki> i mu'a lu do'e nai dei lo mlatu cu pinxe lo ladru li'u 10:30 < gleki> alta: coi mlatu 10:30 < mensi> ([coi {LO KU} DOhU] [FA ZOhE] [CU {COhE VAU}]) 10:30 < gleki> zo la zmiku se setca 10:30 < gleki> sa 10:30 < gleki> i zo lo zmiku se setca 10:30 < gleki> alta: ma nuzba coi 10:30 < mensi> ([FA ma] [CU { } VAU]) 10:30 < gleki> zo do se setca 10:31 < gleki> i ca ti lo doi zei sumti na ka'e kunti 10:31 < gleki> i ku'i la cilretci zo'u mi na favgau i bebna tutci 10:37 < Amnadya2> lodrata cumo 10:40 < akmnlrse> gleki: de'a fanva .i lo (to pu zu toi) mo'u se fanva cu cmalu sai gi'e pagbu la'oi toolkit 10:40 < akmnlrse> mensi: doi ilmen|menli|samcinki s/ro/ru/ 10:40 < mensi> akmnlrse: mi ba benji di'u ba lo nu la'o gy.ilmen|menli|samcinki.gy. di'a cusku da 10:40 < akmnlrse> Amnadya2: lo kibro nu vitci zo'u xu do se slabu la'oi mosh ne la'e di'e https://mosh.mit.edu http://dan.drown.org/android/mosh/ 10:41 < Amnadya2> taho mi ca se slabu gaje lo modality rafsi gi lo vocalic aspectual rafsi 10:41 < Amnadya2> na slabu doi 10:42 < Amnadya2> i lo fonxa na seponse mi i lomemimoi nakibro jena android 10:42 < akmnlrse> je'e 10:47 < akmnlrse> .i tu'a lo drata zo'u: clugau lo cmavo canlu 10:47 < akmnlrse> vau ge'e 10:52 < noncomcinse> coi ru'e 10:52 < akmnlrse> condi pruce 10:53 < noncomcinse> coi la .akmnlrse 10:54 < akmnlrse> ze'i ku lo jbogu'e di'a se bangu lo lojbo 10:56 < akmnlrse> ... zo di'a co'e 10:56 < gleki> lo drata poi mo cu mokau 10:56 < akmnlrse> ma ka mo kau 10:56 < gleki> 10:37 < Amnadya2> lodrata cumo 10:57 < akmnlrse> ua 10:57 < gleki> i do de'a fanva i je'e i mi ca lo cabdei pu ze'i fanva zo'ei la skaip 10:58 < noncomcinse> .i ma xu djuno lodu zo kau cu mo kau kei je nu makau klama lo zarci vau vau pau pei 10:58 < akmnlrse> ma ni barda 10:58 < gleki> i la faglo'u zo'u a'o rauva'ei simsa la vivaldis 10:58 < noncomcinse> en: va'ei 10:58 < mensi> va'ei = [ROI] converts number to scalar tag; specifies the value on fuzzy logic scale; to the degree (n) on scale ... 10:58 < mensi> |>>> gleki 10:58 < gleki> i la skaip zo'u ei 2800 linji cu se fanva 10:58 < akmnlrse> kufra ja'o 10:59 < akmnlrse> ru'a 10:59 < noncomcinse> .i ma brauzero gi'e se pilno do doi la gleki 10:59 < gleki> ie na barda 10:59 < gleki> i y mi so'aroi pilno la krom i ku'i mi ru'inai troci la vivaldis 10:59 < gleki> i V nu'o bredi vau li'a 11:00 < gleki> i ku'i mi tcenelci lo ka zgana lo lojbo je se fanva ke uidje linji 11:00 < noncomcinse> .i mi na se melbi vy 11:00 < noncomcinse> .i pilno fyly 11:00 < gleki> vy ca'a se favgau 11:01 < gleki> i a'o ba melbi lo so'imei 11:02 < akmnlrse> pu ze'i ku mi troci lo ka gau ce loi ro uenzi pe lo pa cankyuidje cu gendra .i la'a na ga sarcu gi frili jai banzuka 11:02 < gleki> y srana ma poi proga 11:03 < noncomcinse> .i mi na mutce nelci lo za'e tanbrflato larcu 11:03 < akmnlrse> la faglo'u 11:03 < Amnadya2> lo mi krom cu spofu sai sei cunso pinka 11:03 < Amnadya2> i mi tahe pilno la faglohu 11:03 < gleki> noncomcinse: xu do tavla fi vy 11:03 < noncomcinse> .i go'i 11:04 < gleki> i a'o ba co'a ka'e stika lo morna be lo vy canko 11:04 < noncomcinse> .i da'i ganai vy. pilno la'oi .GTK+. gi mi zukyjdipei co troci co pilno 11:04 < gleki> noncomcinse: xu do pu'i zanru lo preti pe ju'oi http://area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 11:05 < gleki> i ei zanru lo mu preti 11:05 < noncomcinse> .i .i'e ie zanru 11:06 < gleki> do pu'i xu zanru 11:06 < noncomcinse> y 11:07 < gleki> i va'i ru'e je'u pei do pu zanru mu preti 11:07 < Amnadya2> smunambigu fazo zanru 11:07 < noncomcinse> .i mi na se jaspu fi sy.o'y 11:07 < noncomcinse> sa fi sy .e'y 11:08 < gleki> e'o do co'a se jaspu gi'e zanru cuxna lo mu preti 11:08 < noncomcinse> .ai 11:08 < Amnadya2> ijehu lonu zanru nasarcu lonu zanru cohe 11:08 < noncomcinse> vi'o ba 11:10 < noncomcinse> .i ni'o xu zasti fa lo samselpla poi co'e lonu lo se ciska se cusku cu binxo lo voksa se cusku 11:10 < Amnadya2> ispik 11:11 < akmnlrse> espeak -v jbo 11:12 < noncomcinse> .i .ai pilno 11:12 < akmnlrse> lo voksa cu fegli (to ku'i no da feglyme'a toi) 11:12 < noncomcinse> ki'e re .a'y 11:15 < Amnadya> re boi ahy 11:15 < akmnlrse> oi oi oi 11:16 < noncomcinse> .i mi mo'ifli co famyma'o vau 11:16 < Amnadya> ikuhi ehu lu kihe lo re me ahy lihu 11:17 < noncomcinse> .i ku'i lu re.a'y li'u ka'e lerfu fi re boi .a'y 11:18 < akmnlrse> va'o de'u na co'e .i no da ku'antora je cu se sko'opu lo de'u mupli 11:19 < noncomcinse> .i jetnu 11:21 < noncomcinse> .i lo mi skami na ka'e kibro pilno .ije ku'i mi jinvi lodu mi pu samcpacu lo se nitcu be la'oi .eSpeak. 11:23 < akmnlrse> ru'a su'o .linuks. gunma cu voi lo cnano mupli be ki cu se pagbu la .ispik. 11:24 < noncomcinse> .i ma cnino se smuni zo voi 11:24 < akmnlrse> lo me zo poi'i moi 11:24 < akmnlrse> s/se // 11:24 < fenki> akmnlrse meant to say: ru'a su'o .linuks. gunma cu voi lo cnano mupli be ki cu pagbu la .ispik. 11:24 < akmnlrse> ... 11:25 < noncomcinse> .i simsa zo poi'i ba'a nai 11:25 < noncomcinse> en: poi'i 11:25 < Amnadya> ahaha 11:25 < mensi> poi'i = [NU] x1 is such that poi'i abstraction is true; x1 binds ke'a within the abstraction. |>>> Originally proposed 11:25 < mensi> by And Rosta around 2001. |>>> selpahi 11:27 < noncomcinse> exp: ru'a su'o .linuks. gunma cu poi'i lo cnano mupli be ke'a cu pagbu la .ispik. 11:27 < mensi> (ru'a [{su'o BOI} {linuks gunma} KU] [cu {poi'i <(¹lo [cnano {mupli }] KU¹) (¹cu [pagbu {la ispik KU} VAU]¹)> KEI} VAU]) 11:27 < noncomcinse> ua 11:28 < noncomcinse> .i ie 11:31 < noncomcinse> .i .oi sai mi nalmo'i lodu makau tadji co namcu 11:32 < akmnlrse> ta'u mo 11:32 < noncomcinse> .i fu'o pau da'i zasti fa lo ci gerku .i lo re gerku cu batci lo nanmu .i xo lo xo gerku cu batci lo nanmu 11:33 < Amnadya> pehi buho sarcu falonu se slabu ro kau rafsi pela gredile 5moi i lo tadji pelonu re pagbu gredile kupoho se cilre na banzu je jaha xlali ikuhi y laha mi tavla lo nomei 11:34 < Amnadya> re lo ci gerku cu batci 11:34 < noncomcinse> ki'e 11:34 < noncomcinse> fu'e 11:37 < noncomcinse> camxes: .i mi klama feku lo zdani 11:37 < camxes> (i [mi CU] [klama { } VAU]) 11:37 < noncomcinse> camxes: .i mi fe klama lo zdani 11:37 < camxes> (i [{mi } CU] [klama {lo zdani KU} VAU]) 11:37 < noncomcinse> exp: fe klama lo zarci 11:37 < mensi> (CU [fe klama] [lo zarci KU] VAU) 11:37 < noncomcinse> .uat. 11:38 < akmnlrse> xu lo pa moi cu sinxa lo du'u klama ba'ei fi lo zdani 11:38 < akmnlrse> spagetufa: klama fe ku lo zdani 11:38 < spagetufa> (CU [klama {<(¹fe ku¹) (¹FI [lo zdani KU]¹)> VAU}]) 11:38 < Amnadya> la selp puzu stidi ije la tsani pu zanru 11:39 < akmnlrse> je'e 11:39 < noncomcinse> .i mi se cfipa zoi .camxes. [fe klama] .camxes. 11:39 < akmnlrse> zo fe va'o lo cipra gerna ka'e selbritcita 11:40 < akmnlrse> exp: jai fa broda .i fa bo co'e 11:40 < mensi> ([CU {jai fa broda} VAU] [i fa bo {CU }]) 11:40 < noncomcinse> .i ma smuni 11:40 < Amnadya> ithkuil: selp 11:40 < ithkuil> selp: P1S1/STA-s-ABS-DCT/VAR/U/DEL/NRM-EXS-IFL/UNFRAMED 11:40 < akmnlrse> na djuno 11:40 < noncomcinse> .i febu klama pe'a ma 11:41 < akmnlrse> .i ka'e se jalge lo nu lo brirebla cu se tolfa'o tu'a fi bu 11:42 < gleki> loi xelso pu'i cusku zo'oi nai i smuni dunli zo nai 11:42 < Amnadya> koi lasuda finti lo itkuhile vlaste irci bo zmiku 11:43 < gleki> i oise'i 11:43 < gleki> loi xelso pu'i cusku zo'oi nai i smuni to'e dunli zo nai 11:43 < gleki> i dukti 11:44 < noncomcinse> .i se xi fai bu jai cfipa :: fa loka ce'u na ka'e jimpe ke'a kei, fe lo glipre, fai la .lojban. 11:44 < noncomcinse> .i ua pei 11:45 < akmnlrse> .u'i 11:46 < akmnlrse> .i xu lo cfipu ka'e zilkai 11:46 < noncomcinse> (lo'ai cfipa sa'ai cfipu le'ai) 11:47 < akmnlrse> cu'u la gleki: ".i do je'a xu se cfipu lo ka ca ma kau jai gau cfari .ei" 11:47 < noncomcinse> .i go'oi zilkai 11:47 < noncomcinse> .i mi zukyjdipei co finti co menli selkei bau la .lojban. 11:48 < akmnlrse> .i lo .korp. nu sisku zo'u za'a 7 roi tai pilno 11:48 < akmnlrse> .i .a'u 11:49 < noncomcinse> en: tai 11:49 < mensi> tai = [BAI] tamsmi modal, 1st place (like)/(in manner 2) resembling ...; sharing ideal form ... |>>> tamsmi is x1 11:49 < mensi> resembles x2 sharing ideal form/shape x3 in property x4 |>>> officialdata 11:49 < Amnadya> xudo cnino me la tsani i do zo co sohiroi pilno 11:50 < noncomcinse> .i mi se slabu la tsani xoi puzu cfari se'u 11:50 < Amnadya> caho carvi fi lo dilnu 11:51 < noncomcinse> ma poi drata cu te carvi 11:51 < Amnadya> ro te farlu 11:52 < Amnadya> i eipiei mi spuda fi lu lo romei lihu 11:52 < noncomcinse> ie nai ru'e 11:52 < Amnadya> mo zo cavluhi 11:53 < noncomcinse> ua ie 11:53 < Amnadya> nu lumci lo carvi 11:54 < Amnadya> cinri pixra 11:55 < Amnadya> sei mo lo itkuhile tadni cu uhe laznymau lo lojbo 11:56 < Amnadya> noda troci 11:57 < Amnadya> mi bebna laha loka troci 11:57 < Amnadya> nonseka i ma prali 11:59 < gleki> lo nu gau gletu la to'anzu 11:59 < Amnadya> ma prali 12:00 < Amnadya> lo itkuhile cecmu xaho zasti 12:00 < gleki> lo nu lo prane bangu co'a zasti 12:00 < gleki> i ibu na prane ba'a 12:00 < Amnadya> nasaiprane 12:01 < Amnadya> i kuhi ty jiha nasai prane 12:01 < gleki> za'adai ba'anaidai mi pu karbi ibu lo lojban 12:01 < gleki> sisi la lojban 12:01 < gleki> i su'ova'ei cinri 12:05 < Amnadya> zoho vaholonu mi certu la itkuhile kei mi kakne loka ctuca lo fenki vukro 12:05 < gleki> uanai 12:05 < gleki> i fy vy mo 12:06 < Amnadya> vukro gunma lo itkuhile tadni i y do djuno 12:07 < Amnadya> la kixadas pu vitke 12:07 < gleki> i ua 12:07 < gleki> i ju'oru'e co'u tadni 12:07 < Amnadya> lakne 12:09 < Amnadya> jaho oi noda ba se ctuca mi 12:12 < gleki> i ku'i ju'o cumki fa lo nu lanli gi'e jmina lo se claxu la lojban 12:12 < gleki> i mi deia jundi 12:12 < Amnadya> jeie 12:46 * nuzba @bgcarlisle: Sunscreen = {solri jelcagna fanta}? #lojban [http://bit.ly/1evSoR9] 12:51 < jcowan> coi rodo 12:52 < Zearen> .i coi 12:52 < clakre> coi jcowan! 12:54 < zipcpi> http://genius.com/Djemynai-prami-simbio-english-translation-lyrics/# How did I do? 12:58 < Zearen> .i mi sampla gi'e ve skina fo la'o gy. Steven Universe gy. 13:11 < akmnlrse> zipcpi: lu "i tai tau la oz lo carbi'e mi bevri lo nu'o se zvati" li'u zo'u jufra pa mei 13:11 < akmnlrse> .i lo drata pagbu zo'u pe'i zabna 13:11 < zipcpi> Oh 13:11 < zipcpi> za'a mi srera lo ka mensepli le jufra 13:13 < zipcpi> mo'u stika 13:30 * akmnlrse ca'o co'e la'au za'o lo fanmo li'u 13:30 < akmnlrse> .i lu ".i do traji lo ka ce xo kau da pe mi traji" li'u jai nandu 13:31 < zipcpi> u'i 13:32 < zipcpi> nandu mi ji'a 13:45 < zipcpi> OK {aktigau} and {na'aktigau} added 13:58 < zipcpi> vlaste: xaste (gimka) 13:58 < vlaste> zo xaste gimkamsmikezypro no catni gismu .i zo xaste gimkamsmikezypro no cipra gismu 13:58 < zipcpi> No open rafsi there though 14:20 < zipcpi> dijysenta? I think we just follow whatever rules the building has, including skipping "bad numbers" due to superstition or whatnot. It's too tedious to renumber every building for some logical system 14:22 < zipcpi> (about building floors. Personally I like ground = 0; probably due to malkluxlu, but basement = negative numbers make sense 14:33 < Bass> Hello! Could you please help me? I'm new to Lojban and trying translate phrase "Growing human" into Lojban. Is "remna banro" the correct translation? 14:34 < zipcpi> What's the sentence? 14:36 < Bass> There is no sentence. The meaning is: this human in the process of growing, physical and intellectual development 14:37 < zipcpi> lo remna poi banro then, perhaps 14:37 < Bass> thank you! 14:38 < phma> or "lo banro remna" 14:38 < phma> coi .bas. 14:43 < Bass> If I mean a specific person, should I use le instead of lo? 14:44 < phma> yes 14:45 < phma> you can still use lo, there is some overlap 14:45 < zipcpi> lo won't be wrong either, but it's a bit more vague, since it could also mean "people, in general" 14:45 < zipcpi> depending on context 14:48 < Bass> Is there cases when I can omit descriptors? 14:48 < zipcpi> You mean like using {lo} / {le} without a selbri? 14:48 < zipcpi> Right now you can't; you have to use something like {le co'e} 14:49 < Bass> Aha, thanks! 14:49 < zipcpi> I tend to switch to {le} for a specific person, and {lo'e} for people in general, while using {lo} for "a / some people". Some people just stick with {lo} though 14:51 < latro`a> le is "backwards" in its specificity relative to english: {le} means *you* have whatever the thing is in mind, "the" means the *listener* (and probably also you) has it in mind 14:51 < phma> If it's going to be the title of something, you could omit the article, and just say "banro remna". 14:51 < zipcpi> Right that 14:52 < phma> so when you start a story, in English, you say "a thingamajig" the first time, and "the thingamajig" thereafter. 14:52 < phma> in Lojban you say "le comdaidza" the first time and "le comdaidza" thereafter. 14:53 < phma> Whether it's a previously mentioned thing (which is independent of the definiteness of the article) is expressed with "bi'u(nai)". 15:01 < Bass> Thank you so much! It was really helpful. Hope that I'll have enough time to continue my introduction to Lojban 15:01 < zipcpi> I have some crazy ideas for elidable co'e; le ku, le vu ku, le nu ku; but you should probably be careful how much time you spend with me, lest I infect you with all these weird experimental ideas not yet accepted by any parser (well except the alta parser, but the alta parser accepts a lot of things) 15:02 < Amnadya2> All about that Bass 15:02 < zipcpi> Heh 15:04 < Amnadya2> alta: mi nelci ge gi 15:04 < mensi> ([FA mi] [CU {nelci NF} {ge gi } VAU]) 15:08 < zipcpi> alta: do nitcu fe po'o lo nu penmi le ninmu bu'u le vu 15:08 < mensi> ([FA do] [CU {nitcu NF} { <(²FE [le {ninmu NF} KU]²) (²bu'u [le {vu COhE} KU]²)> VAU}] KEI¹) KU>} VAU]) 15:08 < zipcpi> jo'au CKT: do nitcu fe po'o lo nu penmi le ninmu zai le vu 15:09 < zipcpi> Still three syllables compared to English "there", but, meh 15:09 < Amnadya2> zai tu 15:10 < zipcpi> I've been told though that if you can't actually signal that place, {tu} doesn't work. Of course though, Lojbanists differ on how pedantic they want demonstratives to be :p 15:12 < zipcpi> I like {zai} cause I learned Chinese lol 15:12 < Amnadya2> yep. nice cognate 15:14 < Amnadya2> but unlike nau its not a direct borrowing (meh), but simply comes from zvati 15:14 < zipcpi> I know 15:15 < Amnadya2> just saying thats why i like zai 15:16 < zipcpi> I don't even know about {nau} anymore. "here and now"? How to separate the two? For utility I think {nau'u} is better. {ca nau'u}, {zai nau'u}. I almost want to CKTJ it with nau lol 15:16 < phma> "zai tu"? you're switching to yonder alphabet? 15:16 < zipcpi> CKTJ 15:16 < zipcpi> I did label it :p 15:18 < phma> what's CKTJ? 15:19 < zipcpi> "ce ki tau jau"; a dialect, or fork if you will, that switches some common longer cmavo with rarer shorter cmavo 15:19 < zipcpi> http://mw.lojban.org/papri/ce_ki_tau_jau 15:26 < Amnadya2> doi la zipcpi i xu do se cinri lonudo fanva suho lisri 15:27 < Amnadya2> i pagbu loka jbopre 15:27 < zipcpi> u'i 15:28 < zipcpi> Watership Down? 15:29 < phma> xu lo do karce cu xrududu? 15:30 < zipcpi> u'i 15:30 < zipcpi> I expect to use {me'oi} a lot there 15:37 < zipcpi> Oh by the way, how does Lojban handle footnotes? My thinking is that it's a mere typographical convention to move {to ... toi} group out of the main text 15:38 < zipcpi> i.e. when narrating, the narrator can either choose to utter the footnote as {to ... toi}, or leave it out altogether 15:39 < phma> In the Bible translation, I'm using the \f tag, just like any other language. 15:39 < phma> When lettering footnotes, do we count the apostrophe as a letter? 15:40 < zipcpi> u'i 15:45 < phma> lu mi me me'o y'y gi'e me me'o zy li'u pei 15:45 < zipcpi> ki'a 15:46 < zipcpi> ki'abu'onai 15:46 < phma> The Greek says "I am Alpha and Omega", the Aramaic says "I am Alaf and Taw" 15:47 < zipcpi> I would render it something like {mi roda tolfa'o gi'e fanmo}, but I'm horrible at replacing metaphors that way :p 15:48 < phma> Metaphors be with you! 15:48 < zipcpi> Well I suppose my sentence is a metaphor as well; just replacing it with one with a different literal meaning 15:51 < zipcpi> Is ' considered the first letter though simply because of computer sort? :p 15:52 < phma> yes. where would you put it? 15:53 < zipcpi> If it weren't I might put it at the end 15:54 < zipcpi> Sort of where & used to be in the old days 15:54 < zipcpi> Not even sure why they needed to consider it a letter lol 15:54 < zipcpi> & that is, not Lojban-' 16:00 < zipcpi> I might prefer numbered footnotes for Lojban, just cause, but that might not work so well for the Bible, what with verse numbers and all 16:06 < zipcpi> exp: broda brodo brodi cei brodu 16:06 < mensi> (CU [broda brodo {brodi }] VAU) 16:06 < zipcpi> Huh 16:06 < zipcpi> exp: ke broda brodo brodi ke'e cei brodu 16:06 < mensi> (CU [{ke ke'e} {cei brodu}] VAU) 16:06 < zipcpi> Rather heavy-handed 16:07 < zipcpi> alta: ke broda brodo brodi ke'e cei brodu 16:07 < mensi> ([FA ZOhE] [CU { } VAU]) 16:15 < ctefaho> So today I learned that I happen to put [ː] everywhere in my jbo-speech 16:15 < ctefaho> What did you learn regarding jbo today? 16:20 < latro`a> zipcpi, that problem is part of why I suggested VAU CEI a long time ago 16:21 < latro`a> (that and the fact that we realized that cei should really bind pro-bridi, not pro-selbri) 16:21 < zipcpi> ua 16:25 < zipcpi> In Malay the seat of emotion is metaphorically centered on the liver, not the heart 16:26 < zipcpi> {livga} 16:26 < zipcpi> Malay "hati" 17:00 < ctefaho> co'o 17:04 * nuzba @Dank_scriptor: ca lo nu nalbirti kei na zukte gi'enai pensi .i e'usai #lojban [http://bit.ly/1KEJOgL] 17:04 * nuzba @Dank_scriptor: ca lo nu nalbirti kei do na zukte gi'enai pensi .i e'usai #lojban [http://bit.ly/1KEJTB1] 17:32 * nuzba @ftrain: the answer is obviously more Lojban https://en.wikipedia.org/wiki/Lojban [http://bit.ly/1KEMoDv] 17:38 * nuzba @MmichaelLlucy: @ftrain if the answer is Lojban the real answer is that you need a different question [http://bit.ly/1KEMP0w] 17:38 * nuzba @dan_schmidt: @ftrain I love that Lojban was invented because Loglan, the most hardcore invented language ever, was not hardcore enough. [http://bit.ly/1KEMMSk] 17:40 * nuzba @ftrain: if we all spoke https://en.wikipedia.org/wiki/Lojban then it's entirely possible we could eliminate microformats http://t.co/9aOufQboyw [http://bit.ly/1M5g6Pg] 17:48 * nuzba @ohlookyhere: can you make puns in lojban tho [http://bit.ly/1KENXl1] 17:50 < zipcpi> Rhymes... assonances... making silly words like ci'onme'a... :p 18:10 < _mukti_> .i lo xebni na'o xebni .i lo cpacu na'o cpacu .i sa'u mi desku 19:10 * nuzba @kevinmarks: @ftrain json-ld is actually an anagram of lojban [http://bit.ly/1Cff1nS] 19:42 * nuzba @mkmagicannon: @ftrain @lojytan_bot If we all spoke Lojban .inaja su ko na'e troci lo nu tcidu ti [http://bit.ly/1dGzGWj] 20:17 < Mr> Can someone help me learn this? 23:12 < gleki> this? 23:12 -!- Irssi: #lojban: Total of 147 nicks [2 ops, 0 halfops, 0 voices, 145 normal] 23:12 < gleki> oh left 23:13 < gleki> krasi: zvati 23:13 < mensi> zvati = .i zo'oi zai banzuxe'o .i zo'oi at bangenugu .i zo'oi upastit banxe'inu .i zo'oi asist bansupu'a .i zo'oi 23:13 < mensi> prisutstvav banru'usu .i zo'oi xadir bangaru'a 23:14 < gleki> zvati does come from chinese "zai" 23:36 < gleki> en: mister 23:36 < mensi> naun = Mr./Mister --- Day changed Mon Jul 06 2015 00:26 * nuzba @uitki: L17-01 - http://mw.lojban.org/papri/L17-01 by Gleki - /* Space: "to the left", "to the right" */ [http://bit.ly/1H3AOuF] 01:29 * nuzba @uitki: Mumble - http://mw.lojban.org/papri/Mumble by Gleki - /* Servers */ [http://bit.ly/1H9OMON] 02:40 < ctefaho> coi 02:41 < gleki> coi 02:44 < dutchie> coi 02:45 < gleki> xu doi datci do dotco 02:48 < dutchie> na go'i .i mi glico 02:49 < dutchie> zoi gy. Josh Holland gy. cmene mi 02:51 < dutchie> .iseki'ubo bau lo glibau mi se cmene zo'oi dutchie 03:06 < gleki> xe'e 03:37 < teleological> coi .a'e cu'i 03:42 < gleki> ue 03:43 < _mukti_> coi la gleki 03:44 < gleki> lo nu stika lo uitki papri cu masno 03:44 < _mukti_> ue nai ui nai 03:45 < _mukti_> You asked if it might be the same thing as what made tiki slow. I haven't investigated, but I suspect not for this reason: 03:45 < _mukti_> Tiki was slow because it lacked a mechanism (in the version we were using) to send email asynchronously. Mediawiki has that -- it's the queue that got backed up. 03:45 < gleki> https://github.com/lojban/lmw/blob/master/README.md 03:46 < gleki> e'u ma'a jmina lo drata nabmi se skicu 03:46 < _mukti_> ai 03:46 < _mukti_> mi ckire do lo nu finti 03:52 < gleki> oi i mutce masno 03:53 < gleki> i xu le selfu skami ca zbasu lo snura fukpi 03:54 < _mukti_> .a'a mi ba facki 03:55 < _mukti_> ue 03:55 < _mukti_> mutce lo ka tolcando 03:55 < _mukti_> .i tu'a lo snura fukpi cu na fasnu 03:58 < _mukti_> There are a number of busy PHP procs. I will restart the master process. Let's see if that improves performance. 03:58 < _mukti_> Restarted. uo 03:59 < _mukti_> If this approach works, we may need to tune PHP. 04:00 < _mukti_> It may be that we're allowing too many child processes for the amount of available memory, and that old processes may need to be more agressively pruned. 04:01 < gleki> again slow 04:01 < gleki> im trying to save a page gain 04:01 < gleki> well faster than the previous attempt 04:02 < _mukti_> If performance is critical, we may need to revisit the way we're doing internationalization. It's clever, but I can see where it might cause performance issues. 04:03 < _mukti_> Also, we have a ton of extensions installed. Lots of opportunities for drag there. 04:04 < gleki> my, even opening the pages is slow 04:04 < gleki> i cant open lmw atm 04:04 < gleki> in fact it's down 04:06 * nuzba @uitki: proga:LMW - Lojbanic MediaWiki - http://mw.lojban.org/papri/proga:LMW_-_Lojbanic_MediaWiki by Gleki [http://bit.ly/1J0e7J4] 04:40 < lablanu> Is there a PEG grammar for Lojban that follows this form? https://github.com/PhilippeSigaud/Pegged#usage 04:43 < gleki> no but i guess you may ask them in Issues on how to convert it from standard notation 04:44 < gleki> or how it is related to other notations 05:43 < f8l> I knew I know this name! It’s the author of ‘D Templates Tutorial’. 05:44 < dutchie> ki'a 06:28 * nuzba @palkeo: kalama musi Nasa ni li pona tawa mi : https://djemynai.bandcamp.com/album/zao kalama musi ante li kin pona : toki Lojban li kute sama toki Hebrew. [http://bit.ly/1HaxuRp] 06:28 * nuzba @palkeo: kalama musi Nasa li pona tawa mi : https://djemynai.bandcamp.com/album/zao kalama musi ante li kin pona : toki Lojban li kute sama toki Hebrew. [http://bit.ly/1gjkDUk] 06:43 < Amnadya> coi 06:44 < Amnadya> coi la mensi 06:45 < Amnadya> tatoeba: omni 06:53 * nuzba @shukil026: Lojban [http://bit.ly/1HaB2mH] 06:55 < Amnadya> mo lahau losenelci jelomelbi lihu 07:00 < Amnadya> afaict in lojban you can reverse e.g. kahe diha -> diha kahe to reverse the meaning whereas you cant in ithkuil 07:01 < Amnadya> so you'd have to use a subclause and selbri for the other meaning 07:01 < Amnadya> and it seems you cant stack tenses like in lojban 07:03 < Amnadya> i also found a scope error in the grammar 07:03 < Amnadya> unless the grammar is wrong and not the example 07:03 < Amnadya> bbl 07:06 < gleki> Tatoeba: omni 07:06 < mensi> no da se tolcri 07:06 < gleki> Tatoeba: coi 07:06 < mensi> coi do zukte ma [1896005] 07:07 < gleki> i ei ningau lo datni sorcu 07:07 < gleki> i ku'i 07:07 < gleki> le tutci pe la mupli je la mensi cu bebna 07:07 < gleki> i oi la amnadiia mo'u cliva 07:43 < csarven> foaf:name "Sarven Capadisli" . 07:43 < gleki> ызфь 07:43 < gleki> spam 07:45 < csarven> gleki No! It is a statement using the RDF language, which happens to contains the triple: subject, verb, object :) 07:45 < gleki> o'u 07:46 < gleki> i completely forgot RDF 07:46 < gleki> i saw no use of it. 07:46 < gleki> maybe use it in LMW somehow? 07:46 < csarven> LMW? 07:46 < gleki> it's usable together with Semantic Mediawiki isnt it? 07:46 < gleki> LMW = mw.lojban.org 07:46 < gleki> lojbanic mediawiki 07:47 < csarven> I suppose so 09:40 < gleki> en:vimstizu 09:40 < mensi> vimstizu [< vikmi stizu ≈ Excrete chair] = s1 is a toilet for v1 to excrete v2 from source v3 via means/route 09:40 < mensi> v4 09:40 < gleki> but why chair 09:40 < gleki> then what is toilet sink 09:43 < hilmi-can> hey ive been to the chat once until now 09:44 < gleki> ok 09:44 < hilmi-can> i bought the book by cohen a time ago, i very much appreciate everybody trying to keep this language alive thumbs up 09:45 < ldlework> coi hilmi-can 09:45 < gleki> if you want to support lojban pls spend a minute registering here and upvoting questions with rating below 10 area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 09:45 < hilmi-can> getting confused with vocabulary, there are several lists… i could not manage finding the right one the get started really 09:45 < gleki> in reply we'll answer all your questions about lojban :) 09:45 * gleki although we'll answer them anyway 09:46 < ldlework> hilmi-can: there are two sides of the lexicon to study, the predicates, and the articles 09:46 < ldlework> hilmi-can: probably the best way to learn the predicates is to study them on memrise or similar service 09:46 < ldlework> http://www.memrise.com/course/17295/gismu-frequency-order/ 09:46 < hilmi-can> thx 09:46 < hilmi-can> and yes i will check the questions 09:46 < ldlework> The best way to learn the articles is to study the grammar such as in the CLL http://mw.lojban.org/papri/the_Complete_Lojban_Language 09:47 < gleki> if you really want dictionaries then there are two: http://mw.lojban.org/extensions/ilmentufa/ircbot/dumps/lojban-en.pdf and http://mw.lojban.org/papri/La_Bangu:_Lojban_Dictionary_with_Examples 09:47 < hilmi-can> i am doing the book already! 09:47 < ldlework> hilmi-can: but nothing beats just staying involved with the community and just jumping into reading and writing lojban 09:47 < ldlework> and asking questions here 09:48 < gleki> indeed. languages are usually learnt iwht the help of other people. 09:49 < ldlework> gleki: is the la bangu dictionary quite modified from the standard dictionary? 09:49 < hilmi-can> i ver much appreciate your enthusiasm 09:49 < gleki> ldlework: it really tried not to 09:49 < clakre> gleki, what did you mean about what is a toilet sink? 09:49 < gleki> ldlework: some rare places were omitted as the preface says 09:49 < clakre> like a sink for washing hands in the bathroom? English just calls that a sink 09:50 < gleki> clakre: no, i mean toilet pan =lavatory bowl 09:52 < clakre> google seems to think a lavatory bowl is the basin for the sink where we wash hands 09:52 < clakre> this must be a cultural difference 09:52 < gleki> i just meant "toilet" is ambiguous 09:52 < gleki> .dict water closet 09:52 < fenki> water closet — noun: 1. (chiefly UK) a flush toilet, or the room, in a public place, containing one 09:52 < gleki> the key word here is "or" 09:53 < clakre> ok, so the gloss for vimstizu is not specific enough 09:53 < clakre> but the meaning is obvious from the component gismu 09:53 < dutchie> ie 09:53 < clakre> also calling the bathroom the toilet is metonymy at this point 09:53 < gleki> imm wondering what is the unambigous word here. 09:53 < dutchie> vimku'a is a word also 09:54 < clakre> because the older meaning of toilet from before the vimstizu is no longer used 09:54 < gleki> "toilet/lavatory bowl/pan"? 09:54 < clakre> truly "toilet" just means the thing you sit on to excrete, and calling the room "the toilets" or what have you is a secondary meaning 09:54 < dutchie> i would say "bowl/pan" specifically refers to the smaller part of the toilet 09:54 < gleki> which meaning is no longer use? the {vimku'a} one? 09:54 < dutchie> to me, "water closet" == {vimku'a} 09:54 < dutchie> never stuzi 09:54 < clakre> no, the meaning of toilet which is no longer used is "prepare yourself, brush your hair, look pretty" 09:55 < clakre> that is very old. 09:55 < gleki> en:vimku'a 09:55 < mensi> vimku'a [< vikmi kumfa ≈ Excrete room] = x1=k1 is a loo/water closet/[room with toilet], in which x2=v1 excretes x3=v2, 09:55 < mensi> in home/building/structure x4=k2 surrounded by partitions/walls/ceiling/floor x5=k3 (mass/ 09:55 < mensi> jo'u) 09:55 < gleki> again "water closet" 09:55 < clakre> calling a vimku'a a toilet is a euphemism 09:55 < dutchie> originally, maybe, but no longer in my experience 09:55 < clakre> everyone thinks primarily of a vimstizu for "toilet" 09:55 < gleki> .dict lavatory 09:55 < fenki> lavatory — noun: 1. A bathroom; a washroom; a room containing a toilet, 2. A facility for washing hands; a basin — adjective: 1. (dated) Washing, or cleansing by washing 09:55 < ctefaho> mi nitcu lo vimku'a 09:55 * ctefaho runs from tsani 09:56 < gleki> .dict toilet 09:56 < fenki> toilet — noun: 1. (archaic) Personal grooming; washing, dressing etc. [from 17th c.], 2. (now rare) One's style of dressing; dress, outfit. [from 18th c.] — verb: 1. (dated) To dress and groom oneself, 2. To use the toilet, or assist (a child, etc.) in using the toilet 09:56 < gleki> xa'a 09:57 * clakre does not presume to speak for other Englishes though... I don't know what they call what in India 09:57 < ctefaho> sa .I mi Nìtcu lo Toirlèti 09:58 * ctefaho continues running in direction from la tsani 09:58 < dutchie> pe'i the lojban words are clear enough from the constituents 09:58 < clakre> go'i 09:59 < dutchie> everyone knows that a kumfa is a room, and a stizu is a seat (or at least a good dictionary tells you that when you look up vimku'a) 09:59 < gleki> "toalete" means lavatory room in Portuguese. Same for Russian 09:59 < clakre> but you are consulting a lojban->english dictionary 09:59 < dutchie> interesting that so many languages have the same euphemistic metonymy 09:59 < clakre> I was aware of the russian usage, but speaking to the English one 10:00 < gleki> i ja'o zo toirleti na kulnu nutli 10:00 < clakre> pe'i the lojban words are superior to the english and international false friends, which are confusing 10:01 <@xalbo> I'm sort of curious what would be used for a urinal or for a squatting toilet (whatever the term for that is) 10:01 < gleki> oh wait isnt "restroom" unambiguous? 10:01 < ctefaho> hey now that everyone is gathered 10:01 < gleki> although who rests there ... 10:01 < clakre> we would have to survey the world languages about "restroom" zo'o 10:02 < dutchie> gleki: restroom is pretty us-centric 10:02 < ctefaho> How do I pronounce "cupmoi" and "cutnoi" without putting a magic sound in between "pm" and "tn"? 10:02 < gleki> another synonym is "public convenience". now translate that to lojban literally 10:02 < clakre> that could mean anything :) 10:03 < gleki> ctefaho: how do you pronounce "keep me"? as [ki:pmi:], dont you? 10:03 < gleki> Today Mukti and I disabled many extensions ion mw.lojban.org in an attempt to make it faster. 10:03 < gleki> *on 10:04 < gleki> most likely we'll disable visual Editor too 10:04 < ctefaho> [ʃˈup:moj] ? 10:04 < gleki> ":" denoted long sound in my transcription 10:04 < ctefaho> p is not a vowel 10:04 < gleki> i.e. applicable mostly for vowels 10:04 < gleki> [ʃˈupmoj] 10:04 < gleki> i guess 10:05 < gleki> "tn" as in "Whitney" 10:05 < ctefaho> I can not do that "pm" without putting something magic in between 10:05 < clakre> ctefaho where are you from? VpmV presents no problem for me, but my syllable-final p's aren't aspirated 10:05 < gleki> ctefaho: even in English? 10:05 < gleki> oh 10:05 < gleki> do svenska 10:05 < ctefaho> I am just assuming you all do [ʃˈupʔmoj] 10:05 < gleki> si sfenska 10:06 < clakre> would you like a recording? 10:06 < clakre> I could pronounce it 10:06 < ctefaho> clakre: I am swedish 10:06 < ctefaho> and I would love to hear you say exactly [ʃˈupmoj] 10:07 * mumble jorne co'a fa la'o fa ctefaho fa 10:07 < clakre> ok I'll get my headphones and try to find my soundcloud pw 10:07 < ctefaho> or mumble? 10:07 < gleki> i probably need to study phonotactics of Swedish 10:08 * mumble jorne co'a fa la'o fa akti fa 10:09 * mumble jorne co'u fa la'o fa akti fa 10:10 * mumble jorne co'a fa la'o fa akti fa 10:10 < hilmi-can> question on varying the sumti oders with those cmavo words like "se" "te" etc. at first glance there are cmavo up until the 5th sumti place? which gismu or maybe tnaru has the most fixed sumtis?by that i mean for example "tavla" goes asa far as "X4" "in this language"… which gismu/tanru has the "most" sumti places? e.g. "X115"?? Is there a limit? i guess it cant be more than maybe a dozen? 10:11 < ctefaho> en besto 10:11 < ctefaho> hmm how did one do that 10:11 < dutchie> en: besto 10:11 < mensi> besto = x1 dons metaphorical asbestos suit x2 to guard against flames x3 on topic x4 from x5, who disagrees with post x6 10:11 < mensi> for reason x7, not realizing that the post was meant to be sent to x8 rather than all of mailing list x9 (default 10:11 < mensi> jboste) where it was posted in response to email x10, whose author wishes selma'o x11 (default SE) were extended to 10:11 < mensi> concisely express place x12 of brivla x13 (default besto) which has place structure x14 and too many places because 10:11 < mensi> of sadistic whim x15 of brivla-maker x16, who also created brivla x17 which has place structure x18, and so winds up 10:11 < mensi> using too many of cmavo x19 (default zi'o) in order to make the brivla ( x13) more usable by standard x20 and wis... 10:11 < mensi> [mo'u se katna] http://jbovlaste.lojban.org/dict/besto 10:11 < ctefaho> ki'e 10:11 < bigcentaur> hilmi-can gismu only have 5, lujvo can have more, but not many do 10:12 < dutchie> too many places is usually avoided because it's too hard to remember 10:12 < bigcentaur> even 5 is pushing it 10:12 < hilmi-can> yeah its reality sense 10:12 < hilmi-can> thx 10:12 < hilmi-can> keep thinking about it 10:12 <@xalbo> Well, there are no official gismu with more than five distinct places (though there are a few with unlimited places, like {jutsi} and {minli}) 10:12 < dutchie> can you subscript SE like you can FA actually? 10:13 <@xalbo> And yes, you can subscript SE just like FA. 10:13 < ctefaho> xi to the rescue 10:13 < bigcentaur> so is {se} se_0 or is {se} se_1? 10:13 < ctefaho> can someone join mumble me and clakre got some sound problems? 10:14 < dutchie> ei mi cadzu fi'o kansa lo gerku .i co'o 10:14 * mumble jorne co'u fa la'o fa akti fa 10:14 <@xalbo> I wonder whether that's actually been specified anywhere. My instinct is that {SE xi PA+} swaps the PA+th place with the first, so {se} is the same as {se xi re} (and {te xi re}, for that matter). 10:16 * mumble jorne co'u fa la'o fa ctefaho fa 10:16 * xalbo adds {fa xi ka'o vo'a xi pai fi xi pai do se xi ka'o na jimpe} to the ideas for obfuscated-lojbo 10:16 < ctefaho> hilmi-can: instead of using place structures you can use case tags/sumytcita instead 10:17 < clakre> I'm uploading it to soundcloud ctefaho 10:17 < ctefaho> or xoi 10:18 < ctefaho> xoi is like noi except it affects the whole bridi (the "verb" and its arguments) 10:18 < ctefaho> clakre: je'e 10:18 < clakre> sorry, is mumble a bot? I did not know what it was 10:18 * mumble jorne co'a fa la'o fa akti fa 10:18 < ctefaho> clakre: uhm 10:18 * mumble jorne co'u fa la'o fa akti fa 10:18 < ctefaho> you aren't Akti? 10:18 < clakre> er, no 10:18 * mumble jorne co'a fa la'o fa ctefaho fa 10:18 < ctefaho> ... 10:19 < ctefaho> who is akti? 10:19 * mumble jorne co'u fa la'o fa ctefaho fa 10:19 < hilmi-can> member:ctefaho thanks, i just dont get it yet but ive copied your answer and i am sure i will stumble on that inthe grammar book sooner or later 10:20 < clakre> https://soundcloud.com/clakre/cupmoi-cutnoi 10:20 < ctefaho> clakre: I chatted with someone on mumble I thought was you 10:21 < ctefaho> clakre: ki'e 10:30 < ctefaho> hilmi-can: xoi is most likely not in whatever book you are reading 10:31 < hilmi-can> so is it an additional more recent feature? ive heard of xorlo, which hast been created... 10:32 < gleki> {xoi} and others are new words being naturally added to the language by users and usage 10:34 * nuzba @oka_iu_tcan: @crushue そういえばそうでした!ロジバンにはヒュムノスから取り入れた意味の語があるんですよー http://jbovlaste.lojban.org/dict/ra%27i%27au [http://bit.ly/1LSNDze] 10:35 < hilmi-can> member:gleki i read an article by robin powell about setting it free 10:35 < hilmi-can> so is it "set free" then? 10:35 < gleki> setting what free? the language? 10:35 < clakre> it is hard to tell from the wikis whether we are still in any sort of baseline period, heh 10:36 < hilmi-can> yes, or do i missunderstand xoi? because from my view if it s a grammatical feature it would mean yes 10:36 < clakre> there is no overview that I could find of "here's where we are in the process" 10:36 < gleki> it wont be set free as natural languages can do. it can be set free by the continuation of formalizaing existing working tools and adding new tools. 10:36 < gleki> existing working tools are less likely to be diassembled 10:36 * mumble jorne co'a fa la'o fa ctefaho fa 10:36 < gleki> {xoi} adds new features to the language. but it doesn't change existing things. 10:36 < gleki> so it's in tune with CLL 10:37 < gleki> the reference grammar that you are reading 10:37 < hilmi-can> ahhright 10:37 < hilmi-can> then yes and no. haha thx 10:38 < hilmi-can> sound 10:39 < zipcpi> That baseline thing was kinda a failed experiment; people will use what they want to use. What the formalizers have to do is ever so slowly try to keep up 10:41 < gleki> the freeze period was established in order to make people concentrate on actually learning the language as it is prescribed. 10:42 * mumble jorne co'u fa la'o fa ctefaho fa 10:42 < ctefaho> who was akti... 10:43 < ctefaho> I think I can guess 10:51 * mumble jorne co'a fa la'o fa ctefaho fa 11:07 < zipcpi> I was wondering when Curtis will talk about {vujnu} 11:08 < zipcpi> Dunno about the rafsi though. Not sure where it'll be most used 11:10 < dutchie> mi ze'epu na jimpe tu'a zo xoi 11:11 < zipcpi> xu do ca jimpe 11:11 < dutchie> uinai na go'i 11:11 < zipcpi> OK... xoi attaches to the bridi. 11:12 < zipcpi> ke'a is associated with xoi just like it is with poi, but it represents the bridi itself; lo su'u no'a 11:12 < zipcpi> So: 11:12 < zipcpi> le trene cu clani xoi ke'a mitre li pano : The train is ten meters long 11:13 < zipcpi> ke'a represents {lo su'u le trene cu clani} 11:13 < zipcpi> Just like {poi}, if {ke'a} is left out it's assumed to fill the first empty spot 11:14 < dutchie> right, so that makes sense grammatically 11:14 < dutchie> but what is {xoi} actually doing? 11:14 < zipcpi> xoi attaches to the bridi 11:14 < dutchie> just sort of generically "attaching" another birid? 11:14 < dutchie> si bridi 11:14 < zipcpi> It qualifies the current bridi with another bridi 11:15 < dutchie> in any particular way, or just "somehow"? 11:15 < latro`a> perhaps a more specific question may help 11:15 < zipcpi> In the way that wraps the current bridi into an abstraction and makes it the subject of another bridi 11:16 < latro`a> how does the bridi fill ke'a? you certainly don't have {lo su'u le trene cu clani cu mitre li pa no} 11:16 < gleki> isnt {xoi} the same as {noi} in {i [fasnu fa lo nu] broda noi brodi}? 11:16 < zipcpi> latro'a: Why not? I think su'u resolves to {ni} here 11:16 < latro`a> and, once you've figured out what the substitution is, what is the difference between using {xoi} and splitting into two sentences with the substitution? 11:16 < gleki> isnt {xoi} the same as {noi} in {i [fasnu fa lo nu] broda KU noi brodi}? 11:16 < latro`a> zipcpi: that's a type error to me, but my {ni} is weird 11:17 < dutchie> ok, i think i get it now 11:17 < zipcpi> latro'a: I think it makes more sense than mitre3 11:18 < dutchie> it's basically just a way of adding extra information because you can't use {poi} with a selbri, and is more freeform than a tanru 11:18 < latro`a> zipcpi: to me mitre1 are things and mitre2 are numbers 11:18 < latro`a> and you use clatre et al. to specify lengths of axes 11:19 < latro`a> at any rate, I think you'd be better off with an example that doesn't require su'u converting to ni in order to make sense 11:19 < zipcpi> {clatre}? What a weird lujvo - just not extensible to the prefixed ones... 11:20 < latro`a> clatre/gantre/tsutre match up with the old (CLL) jvojva, in the sense that their main structure matches mitre 11:20 < gleki> autorestoring {fasnu fa lo nu} in PEG will require some time 11:20 < gleki> tersmus: broda xoi mo 11:20 < tersmus> Parse error at 6 11:21 < zipcpi> latro'a: Yes, but do we really need an entire slate of these lujvo for ki'otre, miltre, etc? 11:21 < latro`a> if you want you could, but why? just fill the second place 11:21 < zipcpi> The third place, you mean. And I don't know about that; it's not extensible 11:21 < gleki> {ki'otre} is emulating english 11:21 < latro`a> this system doesn't use {lo mitre} almost at all anyway, since {lo mitre} is an object instead of a unit 11:22 < latro`a> no, the second place 11:22 < latro`a> vlaste clatre 11:22 < vlaste> clatre = x1=m1=c1 measures x2=m2 meters in its longest direction x3=m3=c2 by measurement standard x4=m4. 11:22 < latro`a> I coined these words, I know how they work 11:22 < latro`a> :) 11:22 < gleki> i'd use some more generic brivla for length rather than {mitre} 11:22 < latro`a> the point of it was to avoid ever having the length of a thing be a sumti 11:23 < latro`a> because we don't really care about that, it's just a crutch that english uses because it's dependent on nouns 11:23 < latro`a> what we really care about is the relationship between objects and the numerical values of their length, in whatever dimensional system 11:24 < latro`a> what would've been nice would have been a way to get a sumti which represents the dimensional system itself, and then making brivla that take that as an argument 11:24 < gleki> in la bangu i thought that {minli} was the generic any-system-length gismu that we need due to "by standard" place 11:24 < zipcpi> I'm not sure that is the correct approach, because an object can be associated with a unit more than one way 11:24 < zipcpi> Dimension is one of them. Another is "hertz", which is defined as second ^ -1 11:25 < gleki> pi'ai coming in a few seconds 11:25 < zipcpi> Which measures "frequency", but the event that the "frequency" measures could vary 11:26 < zipcpi> And exactly. {pi'ai} only makes all these much more complicated 11:26 < latro`a> pi'ai/te'ai deal with building units without needing to make up words; one of the open problems that I left behind when I set this system down was how to make up reasonable shorthands for commonly used compound units 11:26 < latro`a> e.g. joules, watts, hertz 11:26 < gleki> even {ki'otre} is ugly no matter ow ugly is this in english 11:27 < gleki> *how 11:27 < zipcpi> But even km/h, which aren't combined into a single word but is commonly used 11:27 < latro`a> if your gripe is about the verbosity, that can be dealt with through coinage as needed 11:28 < latro`a> you were expressing a gripe about this system lacking expressiveness 11:28 < gleki> technically there is a perfect solution to "km/h". it's just that we cant agree how to compress it nicely 11:28 < latro`a> I disagree with that, but I don't think a full rant without dialogue is a good way to explain why 11:28 < latro`a> instead I'd rather take examples and explain how to say them in this system 11:28 < latro`a> (frequency is one of the few that I would have trouble with, because that has xorlo problems) 11:29 <@xalbo> "I'm traveling (driving) at 100 km/h" 11:30 < zipcpi> The problem is that mitre3 really sticks out like a sore thumb, and isn't extensible to compound units 11:30 < latro`a> this is just for this specific case of lengths; you don't want it for compound units, which is why these words don't supercede mitre 11:30 <@xalbo> I really like {clatre}; it's a short word for a concept that's sort of awkward to say in English. It feels extremely lojbo to me. 11:31 < zipcpi> Hm 11:32 < latro`a> anyway, the selbri for PA km/h is {pi'ai mitre snidu te'ai li ni'u [pa] be li PA no no no} 11:32 < zipcpi> What about the {vi lo mitre be li re} thing? 11:32 < samcinki> "X is N meters long"? 11:32 < latro`a> that one's easy, X clatre li N 11:32 < gleki> jb: rotsu 11:32 < mensi> rotsu = rotsu — x1(entity) is thick in length/width etc. x2(property of x1) 11:32 < mensi> :lo rotsu — thick. 11:32 < mensi> :le dargu cu rotsu lo ka se pagbu ti — The road is thick here. 11:32 < mensi> :Comment: the third largest dimension. ganra, rotsu and clani specify the three dimensions of objects. 11:32 < mensi> :Related words: barda, cinla, ganra, clani, condi, plana, gutci, minli 11:32 < gleki> ok, gantre, tsutre 11:32 < zipcpi> Er I don't think that 1000 m/s = 1 km/h 11:32 < latro`a> sorry 11:33 < latro`a> I meant cacra but defaulted to snidu 11:33 < zipcpi> And I put in fei'u to make it easier to say than te'aini'upa 11:33 < ldlework> Is there any recommended version we should roll out for the next release? We're about to roll out and wondering if we should go for 1.7 or stick with 1.6. 11:33 < ldlework> oops 11:34 < gleki> it's {vi'i lo mitre be pa ki'o mo'e da ze'a lo cacra be mo'e da} + some cmavo to specify average, not possibly infinitesimal speed 11:35 < zipcpi> Maybe we should just change all of them so that all unit selbri have "in dimension x3", even if we might find it difficult to fill for some units 11:35 < zipcpi> So that combinations would behave properly 11:35 < latro`a> that doesn't always make sense; an event snidu a number, and that's it 11:35 < zipcpi> lo ka clatei 11:35 < latro`a> part of the point of doing it with pi'ai/te'ai is that the structure is more or less nonce 11:36 < zipcpi> Hmm 11:36 < latro`a> x1 is whatever thing you are measuring, x2 is a li, and the remaining places are not built into the definition of pi'ai/te'ai 11:36 < latro`a> they might be necessary, but that needs to be settled at the time of lujvo definitino 11:36 < latro`a> *definition 11:37 < zipcpi> The main problem I suppose is that long lujvo are hard on the breath 11:37 < latro`a> sure 11:37 < zipcpi> ki'otrefrinycacra... x.x 11:37 < latro`a> but you can {cei} if need be for now 11:38 < latro`a> until we have zi'evla that everyone actually knows 11:38 < ldlework> small beautiful words with the perfect definition 11:38 < zipcpi> We don't know what to cei it too, that's the problem 11:38 < gleki> {jbocei} i'd say 11:38 < zipcpi> u'i 11:38 < latro`a> again, broda for now 11:38 < latro`a> getting people to agree on short zi'evla for complex concepts will take time and corpus development 11:39 < latro`a> building a general but highly verbose system is much faster 11:39 < zipcpi> It's like I'd even prefer {pi'ai ki'otre fei'u cacra} over {ki'otrefrinycacra} 11:39 < gleki> i guess you will have the most complaint from those used to "/" and "per" in English 11:39 < zipcpi> More syllables, but gives space to catch breath 11:40 < zipcpi> That's why I added fei'u 11:40 < latro`a> the thing is, the english system is very compact for small compounds but completely confined to print for larger compounds 11:40 < latro`a> unless small words are created 11:40 < zipcpi> Mm. 11:40 < latro`a> it's damn near impossible to speak out the units of the gravitational constant 11:41 < gleki> mekso problem detected 11:42 < latro`a> in english, sa'e 11:42 < ldlework> he's saying in engish 11:42 < ldlework> we should make a system that makes everything equally hard to say 11:42 < ldlework> at least it will be consistent 11:42 < ldlework> take the greatest common demoninator and base the convention around that 11:42 < PrincessBecca> speak babm; everything is hard to say in babm 11:43 < latro`a> one quirk of jbo: we have no word for coulomb in the gimste 11:43 < latro`a> it's {pi'ai xampo snidu} 11:43 < PrincessBecca> for example, would you believe "babm" is pronounced "bo-a-bo-mo" 11:43 < latro`a> .u'i 11:43 < PrincessBecca> essentially babm uses the latin alphabet as a syllabary. 11:43 < PrincessBecca> or however you spell that. 11:44 < zipcpi> Sounds like xorban 11:44 < PrincessBecca> dunno xorban 11:44 * ctefaho jst can't get [ː] and [ˑ] out of his jbo-speech 11:44 < ctefaho> any tips? 11:46 < PrincessBecca> i very recently learned to say "o" the international way. basically you say an english "o" without puckering your lips. wish someone would have just told me that. 11:46 < ctefaho> like in sutra I have a hard time getting the u flow to t without any of the above 11:46 * ctefaho uses [ɤ] for o 11:46 * ctefaho and [ɯ] for u 11:46 < latro`a> tr is tricky for americans; we wind up with tcr 11:47 < latro`a> similarly dr becomes djr 11:47 < ctefaho> xmm well I have no troble with the "tra", just the "sut" 11:47 < latro`a> you can enunciate it as su-tra 11:47 * PrincessBecca has no problem with "tr" or "dr" 11:47 < latro`a> if you want 11:47 < ctefaho> [suːt] 11:47 < latro`a> the syllable split doesn't have to be in between 11:47 < ctefaho> well I want the [ː] to burn 11:48 < latro`a> you mean you make a stop sound when you cut off your breath/ 11:48 < latro`a> *? 11:48 < ctefaho> we swedes put [ː] in everywhere, trying to learn that out 11:49 < ctefaho> actually I think I got it now 11:49 < ctefaho> [sˈɯtra] 11:49 < zipcpi> dutchie: Anyway for a perhaps less controversial example of xoi: {le vorme co'a ganlo xoi ke'a vlile} -> The door closed violently 11:49 < zipcpi> It's a bit like adverbs, modifying the bridi itself 11:51 < ldlework> which is different than {le vorme co'a vlile ganlo} 11:51 < ldlework> because 11:51 < ldlework> something reasonably probably 11:51 < zipcpi> It's more specific than tanru 11:52 < latro`a> I think the question is, how exactly is {.i le vorme co'a ganlo xoi ke'a vlile} different from {.i le vorme co'a ganlo .i lo su'u go'i cu vlile} 11:52 < ctefaho> latro`a, seems like I was trying to start "t" with a pause for some reason 11:52 < zipcpi> latro'a: That I dunno. Maybe replace i with ije 11:53 < dutchie> zipcpi: je'e ki'e 11:53 < ctefaho> latro`a: well xoi expands to be in the same sentence 11:53 < ldlework> sentences don't have logical consequence 11:53 < latro`a> you could do the same thing with ge gi and a variable binding 11:53 < latro`a> if you object to the sentence separation 11:53 < ctefaho> using go'i I don't think the difference is that big 11:54 < zipcpi> I won't argue that there aren't other ways to do it, I just find it convenient 11:54 < ctefaho> I dunno about "go'i cu vlile" though 11:54 < latro`a> no no, I'm not saying it's bad, I'm trying to get a handle on what it means 11:54 < latro`a> ctefaho, the cu closes the su'u 11:54 < zipcpi> Ah... in that case I think you're correct 11:54 < ctefaho> right 11:54 < ctefaho> I still mess up the terminators all the time 11:55 < ldlework> ctefaho: well consider that if it didn't, latro`a would be saying that xoi resolves to a referring description. 11:55 < ldlework> Since it does, it resolves to a proposition 11:55 < ldlework> Which makes a lot more sense 11:55 < ctefaho> should sit down and learn them properly but too busy fixing stuff no one cares about 11:56 < ctefaho> also in my world xoi does not expand with a "lo" 11:56 < ctefaho> but that's just my crazy world 11:56 < zipcpi> I've also imported another very related solution from RLP's blog, {gi'ei}, but I don't think it can supersede {xoi} right now 11:56 < ldlework> ctefaho: he isn't saying it expands with a lo 11:56 < ldlework> the resulting asset isn't a sumti, its a bridi 11:56 < ctefaho> "lo su'u go'i" 11:56 < ctefaho> but well 11:56 < ldlework> yes, which is an argument to a larger proposition.. 11:56 < zipcpi> Might actually be a bit easier for nintadni to understand though. I'm not very sure lol 11:56 < ctefaho> doesn't matter much 11:56 < ldlework> The idea of the previous sentence -is- verb 11:57 < ldlework> the return value isn't "the idea of the previous sentence" 11:57 < latro`a> there's a conflict here between "expansion of {xoi}" and "semantic equivalent which doesn't use {xoi}" 11:58 < latro`a> ldlework is (correctly) saying that xoi does not "expand" in the first sense to an abstractor sumti 11:58 < latro`a> (as I understand) 11:58 < ldlework> yes, it asserts a secondary propoposition 11:58 < ldlework> involving an abstractor sumti 11:58 < ldlework> the abstractor sumti is just a way to refer 11:58 < ldlework> to allow us to say something about the preceeeding sentence 11:58 < latro`a> I was asking whether it can be expressed in a semantically equivalent fashion using an external proposition, which has to absorb the preceding proposition as an argument somehow 11:58 < ldlework> the whole point of xoi 11:59 < ldlework> latro`a: ie 11:59 < ldlework> latro`a: I was combating ctefaho's claim that his expansion doesn't use lo 11:59 < latro`a> and apparently the answer is yes, at least to a good approximation (maybe there is some difference of emphasis or such) 11:59 < ldlework> and I was pointing out that that is kind of missing the point 11:59 < ldlework> latro`a: yeah I think xoi does what you said, I don't think it is a complex semantic 12:00 < latro`a> {lo su'u go'i} is at worst insufficiently specific for the purpose intended 12:00 < ldlework> its pretty generic lol 12:00 < ctefaho> let's just say I think of NU different than you do and is in no mood to defend that viewpoint 12:00 < latro`a> (and I don't think it's that, either) 12:00 < ldlework> ie 12:01 < ctefaho> is/am 12:01 < ldlework> I mean at most, su'u is merely a grammatical transformation 12:01 < latro`a> I think I know what you mean when you object to the use of {lo} there, and I think go'i's "magic" fixes your concern 12:01 < ldlework> I'm not sure there's room for your own interpretation 12:01 < latro`a> namely, up to choice of abstractor type, there's really only one {lo su'u go'i kei ku} 12:01 < latro`a> (note the terminators; I'm explicitly forbidding any sumti from being inserted) 12:02 < latro`a> by contrast there are numerous {lo su'u le vorme co'a ganlo}, but go'i carries referents, not symbols 12:03 < ldlework> I love that observation 12:03 < ldlework> It reflects that idea that the reference acts are performed before the propositional act. 12:04 < ldlework> Does go'i contain the illocutionary force of the sentence from which it came though? 12:04 < ldlework> Or just the proposition? 12:04 < zipcpi> It's kinda why I decided to import {gi'ei}; so that you can add the LE and/or the NU if that semantic is useful 12:04 < ldlework> when I say 12:05 < ldlework> xu do catra 12:05 < ldlework> and you remark on go'i 12:05 < ldlework> are you remarking on my question 12:05 < latro`a> illocutionary force is lost 12:05 < ldlework> or the proposition "do catra" divorced from its interrogative force? 12:05 < ldlework> latro`a: cinri 12:05 < latro`a> the CLL is actually pretty clear on this point, although it doesn't use the phrase "illocutionary force" 12:06 < latro`a> note that this means {.i ma klama lo zarci .i go'i} is grammatical nonsense 12:06 < ldlework> latro`a: the CLL was written by logicians not philosophers if I remember 12:06 < latro`a> ehh...there's a lot there 12:06 < ldlework> And probably when the CLL was first written the modern ideas of language philosophy were being refined at that very time 12:06 < ldlework> late 50's early 60's? 12:06 < latro`a> to my understanding there's background in logic, linguistics, and certain philosophy, but not so much *language* philosophy 12:06 < ldlework> latro`a: je'e 12:07 < latro`a> but quine's ideas are big in the CLL 12:07 < ldlework> does anyone know the date of the cll off hand? 12:07 < latro`a> I wanna say 93 but I think I'm way offf 12:07 < latro`a> *off 12:08 < ldlework> oh 12:08 < latro`a> 97 12:08 < ldlework> oh 12:08 < ldlework> wow 12:08 < ldlework> nevermind? 12:08 < ldlework> :) 12:08 < latro`a> loglan was around the time you're describing 12:08 < ldlework> oh right 12:08 < ldlework> its not like lojban was novel 12:08 < ctefaho> speaking of ancient things can someone explain why "mz" is an illegal cluster? 12:09 < latro`a> I think that might be an american bias 12:09 < ctefaho> the leading theory I have heard so far was that it was done to prevent "James" from lojbanifying well 12:09 < latro`a> because it is easy for {ms} to turn into {mz} and for {mz} to turn into things like {mbz} 12:09 < ctefaho> or that 12:10 < latro`a> at least in american speech 12:10 < latro`a> like, if I said "rhythms" out loud, it would be mz 12:10 < latro`a> not ms 12:10 < ctefaho> just more general american bias then 12:10 < zipcpi> Same for {ns} and {nz} though 12:10 < latro`a> similar, not the same; I'd say it's easier to force that distinction 12:10 < ldlework> zipcpi: have you gotten any further in the book? 12:10 < zipcpi> And {mbz} is banned anyway (but wasn't before, I suppose) 12:11 < zipcpi> ldlework: Nope, didn't buy it yet T.T 12:11 < ldlework> I guess I should stop asking and you'd inform me if it ever happens 12:11 < ldlework> :) 12:12 < ldlework> zipcpi: I'm reading through it again, just finished the chapter on references (3 I think?) 12:12 < ldlework> well not references, but reference, as an act 12:12 < ldlework> ei co'o 12:13 < zipcpi> co'o 12:19 < ldlework> One of the things he does in chapter 2 is make the case that illocutionary acts can be understood in terms of a set of constitutive rules 12:19 < ldlework> And he makes that case by building up the rules that constitute a promise 12:20 < ldlework> One of the most interesting conclusions in my opinion was that by manipulating individual rules of the promise you obtain "defective promises" 12:20 < ldlework> For example, if A promises B to do X, if A doesn't actually intend to X, it becomes a "decietful promise" 12:21 < ldlework> If it is obvious that A will do X anyway irrespective of the promise, then you get a vacuous promise 12:21 < ldlework> if B doesn't want A to do B, you get a threat 12:21 < ldlework> and so on 12:21 < zipcpi> Ah... which is all rolled up into {ai} in Lojban, I believe 12:21 < ldlework> Well, that is lojban's explicit force indicator 12:22 < ldlework> Its possible to invoke illocutionary forces without indicators 12:22 < ldlework> via predicates, or just context 12:22 < ldlework> mi je'a ba za ganzu lo mi kumfa 12:22 < ldlework> can be taken as a promise as long as all the constitutive qualifications for a promise are met 12:23 < ldlework> For example "I promise your family will pay" is a defective promise even though it uses an explictly promise force indicating device 12:24 < ldlework> linguistics and logical calculus are not enough to explain language unfortunately 12:24 < ldlework> though that would certainly make everything easier 12:25 < zipcpi> I hardly use je'a; I think you mean {ja'a} 12:25 < ldlework> I'm just making it unnegatable 12:25 < ldlework> IE "No you wont!" 12:25 < zipcpi> Yeah but {ja'a} fits better there because {je'a} only modifies {ba} 12:25 < ldlework> what 12:26 < zipcpi> {je'a} is in NAhE, not NA 12:26 < ldlework> so I'd need je'a bo? 12:26 < zipcpi> I'm not sure that works 12:26 < ldlework> na'e bo works why not je'a bo 12:26 < zipcpi> What does it attach to? 12:26 < zipcpi> I have no idea how bo works with NAhE 12:26 < ldlework> the sentence? 12:27 < ldlework> Anyway, thanks for the clarification. 12:33 < ldlework> Can anyone lojbanize aristotle for me? 12:35 < zipcpi> la .aristotyl. ? 12:35 < dutchie> what was he called in greek? 12:35 * zipcpi checks 12:35 < zipcpi> la .aristoteles. 12:36 < zipcpi> Aristotélēs 12:36 < ldlework> if I wanted to say "the last broda" would I use moi or mai 12:36 < dutchie> for completeness: Ἀριστοτέλης 12:36 < dutchie> :) 12:37 < ldlework> lo romoi broda ? 12:37 < zipcpi> moi 12:37 < zipcpi> mai is a discursive: "Firstly, secondly, lastly," 12:37 < ldlework> je'e 12:37 < zipcpi> mai essentially tuns PA into UI 12:38 < ldlework> interpret, da du ma ei zo'u da citka le plise 12:39 < ldlework> "should" as UI is annoying 12:39 < zipcpi> Who should eat the apple? (I think?) 12:39 < ldlework> yeah 12:39 < zipcpi> There is {iancu} 12:39 < ldlework> UI predicates are more annoying 12:40 < ldlework> but whatever I don't have it in me to care about ui predicates I guess 12:40 * ldlework uses iancu 12:43 < zipcpi> Sigh... I have a problem. Work has gotten... disorganized as of late and life has thrown me some curveballs. I might need to refocus my priorities somewhat 12:43 < zipcpi> So I might be here a little less often; we'll see 12:44 < ldlework> da du ma zo'u da citka le plise xoi da iancu 12:44 < ldlework> yikes! 12:44 < zipcpi> I think you mean xoi se iancu 12:44 < zipcpi> But ... hm 12:45 < ldlework> no, the ke'a or whatever variable article is simply x2 12:45 < zipcpi> Oh right, forgot 12:45 < ldlework> It would be nice to have an article that meant something like xoi ri'oi iancu 12:46 < ldlework> that didn't express the speaker's intention, but rather expressed a predicate relationship 12:46 < ldlework> "should" is so overloaded... 12:46 < ldlework> I'm looking for the nibli sense 12:46 < ldlework> He should be here by now. 12:47 < zipcpi> That one I use {kanpe} 12:47 < ldlework> He's not late, but it makes sense for him to already arrived given the known details of his departure and route and so on 12:47 < ldlework> But kanpe is an effect of whatever underlying predicate I'd like to actually get at 12:47 < zipcpi> Then {ba'a} / {la'a} 12:47 < ldlework> Those are also expectations though 12:48 < ldlework> I mean 12:48 < ldlework> Given the details of his departure and route and means and so on, he should have arrived by now. 12:48 < zipcpi> Yes, I "expect" him to have arrived 12:48 < ldlework> My mere expectation does not give rise to the entailment of him being here based on those details 12:48 < ldlework> My expectation is the result of it 12:48 < zipcpi> mi kanpe ko'a ki'u ko'e 12:49 < ldlework> That you reiterate your position in face of clarification my part leads me to believe you just disagree, ok. 12:49 < ldlework> It almost the suggests half of the expects conclusion 12:49 < zipcpi> I'm confused as to what you want 12:50 < ldlework> That he left at ko'a time, and traveled along ko'e using ko'i _suggests_ he should be here now, despite the likely entailment of me expecting the same, as soon as I consider those things which suggest it. 12:50 < zipcpi> So you want something like {nibli}, but a bit weaker? 12:51 < ldlework> Or to be able to use nibli without having to take a piece of paper and write some axioms proving my nibli2 12:51 < ldlework> in other words, being able to give partial stories that lead to nibli2 12:52 < ldlework> IE, as a useful word able to be used in contexts other than one concerning logicians and chalk 12:52 < zipcpi> I think it's best not to touch {nibli} right now; it's important for mathematical statements and such 12:52 < ldlework> Maybe words are useful in one sense and useful in another 12:53 < ldlework> Maybe I really do mean nibli, I'm just not providing predicate arguments that would satify a logician as my interlocutor but totally satisfy the person I'm standing there waiting with 12:54 < zipcpi> I'm not sure what's wrong with kanpe ki'u; do you really need to zi'o out {lo kanpe}? 12:55 < ldlework> Because the focus is on the elements giving rise to the expectation 12:55 < ldlework> Nothing about kanpe provides. 12:55 < zipcpi> Then front the ki'u 12:55 < ldlework> for example 12:56 < ldlework> all those things, the departure, route and so on COULD krinu him being here on time 12:56 < ldlework> but I expect that he be late! 12:56 < ldlework> there are two predicates here 12:56 < ldlework> without nibli we only have a word for one 12:57 < zipcpi> {lakne}? Not sure 13:00 < zipcpi> That makes {nibli} even more problematic, because then it indicates that you are expecting against logical implication, which is probably not desirable. What you want is "forensically suggest" 13:01 < zipcpi> Circumstantial evidence 13:03 < zipcpi> My proposed zi'evla: {indika} 13:13 < zipcpi> http://jbovlaste.lojban.org/dict/indika 13:15 < zipcpi> Would be a great word for something like Phoenix Wright or Sherlock Holmes, I assume 13:16 < durka42> psh, sherlock holmes doesn't use circumstantial evidence, he just knows 13:16 < zipcpi> u'i 13:18 < ctefaho> are there any words anywhere on jbo-pronounciation and https://en.wikipedia.org/wiki/Vowel_length ? 13:19 < ctefaho> all I know is that vowel length isn't supposed to matter 13:19 < ldlework> Why don't we just create gismu? 13:19 < zipcpi> I don't know what it should look like 13:20 < ldlework> Also 13:20 < ldlework> What kind of circumstantial evidence is there if not logical evidence? 13:21 < zipcpi> Circumstantial evidence requires interpretation, and may be contradicted by further evidence 13:21 < ldlework> I wish we could divorce the consititive elements of logical calculus and ike, everything else. 13:21 < ldlework> zipcpi: so can a logical conclusion 13:21 < ldlework> that's what a counter example is 13:21 * durka42 pops up, points to {jinvi} and ducks down behind the counter again 13:21 < ldlework> seriously, the assets or particles of concern that arrive out of the calculus of logical systems and their rules that govern are so totally seaprate from everyhting outside that domain 13:22 < ldlework> its frustrating 13:22 < ldlework> durka42: Everything points to him being here on time, but I expect him to be late. 13:22 < ldlework> How can I constrast my jinvi against the thing that it is constrated against 13:23 < ldlework> What is the propositional content from which my opinion contrasts 13:23 < zipcpi> to'eki'unai 13:23 < durka42> well 13:23 < durka42> it's the jinvi4 isn't it 13:23 < durka42> you expect him to be late based on something 13:23 < durka42> presumably the fact that he's always late 13:23 < ldlework> Yes, but what suggests he should be here on time? 13:23 < durka42> hmm 13:23 < zipcpi> {lo indika} 13:23 < durka42> heh 13:23 < ldlework> How are all the effects that promote his timeliness related, propositionally, to arise the implication from which my opinion contrasts 13:24 < zipcpi> You still haven't given us enough to construct a sentence with 13:24 < ldlework> zipcpi: hehe 13:24 < durka42> yes I got the question 13:24 < durka42> you don't have to ask it again with longer words :) 13:24 < ldlework> I have though, but you can deny it in the "I don't understand anything you say so its your fault" thing 13:24 < ldlework> But I have no tools for that 13:24 < zipcpi> No, you just vaguely point to things that suggest that he's on time 13:24 < ldlework> durka42: for everyone's benefit 13:25 < ldlework> He left from a place close enough, on a train that left on time that travels fast enough, and demonstrated knowledge of the route, and his legs work and so on 13:25 < zipcpi> Thus all I could do right now is {lo indika cu indika} 13:25 < durka42> so a possible rephrase is "[a reasonable person] would expect him to be on time because of X Y Z, but I think he will be late because I know him" 13:26 < ldlework> I think its beyond agentitive consideration 13:26 < ldlework> Think of it as a simulation 13:26 < ldlework> With constraints that suggest an outcome 13:26 < ldlework> it doesn't matter if there's no life in the universe 13:26 < ldlework> given an incomplete description, the description promotes one outcome over another 13:27 < durka42> if all evidence suggests a tree fell in the forest, but nobody saw the evidence, did it make a sound? 13:27 < ldlework> <3 13:27 < ldlework> I think not! 13:27 < durka42> well so then there is an agent, isn't there? 13:27 < ldlework> I was making fun of the idea that there has to be 13:27 < zipcpi> lo nu la .djon. cu jibni cu indika lo du'u la .djon. ca darca .iku'i ki'u lo nu la .djon. cu jai cafne fai lo ka lerci kei mi jinvi lo na se go'i 13:27 < ldlework> I was playing god, using that conclusion 13:28 < ldlework> preventing the sound whimsically 13:28 < ldlework> zipcpi: I think making nibli only usable in contexts where logical caluclus is taking place is like saying anytime you use a referring description you reduce the universe to only one thing existing and things liek that 13:28 < ldlework> I think its the same conflation 13:29 < zipcpi> Ugh... I'm not gonna be roped to that argument again 13:29 < durka42> I also want to reach for {nibli} but maybe you guys have already discussed that 13:29 < durka42> but it's too definitive 13:29 < ldlework> And I think that, zipcpi, you would stamp your approval on lines of nibli1 for which a counter-example could be shown that you didn't expect 13:29 < ldlework> and then you'd be like, well nibli1 was incomplete! 13:29 < ldlework> and that's all I'm asking you allow me to do 13:30 < ldlework> in all the contexts where the entire picture doesn't arrive from a set of axioms 13:30 < ldlework> but context 13:30 < ldlework> like we're here waiting for a guy 13:30 < ldlework> nothing explains his tardiness 13:30 < ldlework> (even though, objectively something does - IE the missing clause from nibli1) 13:30 < durka42> "logic entails that he will arrive in 2 minutes, but I expect that he won't" 13:30 < durka42> but that's different from the original sentence isn't it 13:31 < ldlework> "all the consequences of the details of this travel that I have observed all cooperate to" entail him arriving in 2 minutes, but I expect that he wont 13:31 < ldlework> What ties all of those aspects of his travel together if not rational implication, IE logic 13:32 < zipcpi> ro lo se zgana be mi 13:32 < ldlework> Do we have to sit down and prove the truth-value of whether he will indeed arrive in 2 minutes to state as such? 13:32 < zipcpi> cu indika 13:32 < ldlework> If only emphatically? 13:33 < zipcpi> The point is we are talking about circumstantial evidence now 13:33 < ldlework> Take the counter example. 13:33 < ldlework> We take a few axioms 13:33 < ldlework> I ask, does this nibli the nibli2 13:33 < ldlework> you rub your chin and declare yes! 13:33 < ldlework> and I then I say, ah but what about this counter example 13:33 < ldlework> you go drats! the nibli1 must be incomplete 13:33 < zipcpi> circumstantial evidence is different from an axiomatic system 13:33 < ldlework> But suppling this extra clause we make room for the counter example and the nibli-ness is restored 13:34 < ldlework> zipcpi: no, circumstantial just means there isn't enough to entail the conclusion directly 13:34 < ldlework> because there is missing information 13:34 < zipcpi> Exactly, because you still expected him to be late 13:34 < zipcpi> What is left unimplied 13:34 < ldlework> The same kind of missing modeling that lead you to believe the unaltered nibli1 entailed the nibli2 13:34 < zipcpi> Is that the pattern of him being late 13:34 < ldlework> which doesn't matter 13:34 < zipcpi> Is to you a stronger {indika} than everything else 13:34 < ldlework> Because things that we leave out of nibli1 should all be treated equally 13:34 < ldlework> nibli doesn't care that you're talking about people's habits 13:35 < ldlework> why should it, it is a predicate 13:35 < ldlework> What you're saying is that 13:35 < ldlework> Noone should ever utter nibli statements until they have exhaustively explored all possible variants of the nibli1 to objetively prove that the following nibli2 is entailed 13:35 < ldlework> this isn't possible 13:35 < ldlework> academically or otherwise 13:36 < ldlework> I'm saying, let me make the same nibli conclusions, wihtout all the inofrmation 13:36 <@xalbo> The point is that {nibli} only works in very limited formal structures, where you actually *can* prove that all primes greater than 2 are odd. 13:36 < ldlework> let me make the same mistake or characterizations of _GIVEN_ nibli1's 13:36 < ldlework> that logicans do 13:36 < ldlework> xalbo: except the cases where you find a counter example 13:36 < ldlework> then the nibli1 is defective 13:37 < ldlework> let me make conclusions given the same statements with defective models 13:37 <@xalbo> You can do that, but that's completely unrelated to {nibli}. 13:37 < ldlework> the assumption is always that we're considering enough nibli1 to entail the nibli2 13:37 < ldlework> xalbo: please 13:37 < ldlework> use the example I gave or tell me why it is false 13:37 < ldlework> I show you a nibli1 a set of axioms and arguments 13:37 < ldlework> and I claim that it entails nibli2 13:38 < ldlework> I give you 6 months and you cannot prove that the nibli1 doesn't entail the nibli2 13:38 < zipcpi> Then the first logician has just made a mistake. However, the statement "ko'a indika ko'e" is not false because "ko'i indika lo du'u na me'au ko'e" 13:38 < ldlework> Yes precisely! 13:38 < ldlework> the logican made the conclusion of nibli1 given a constrained an incomplete nibil2 13:38 <@xalbo> "I give you 6 months and you cannot prove that the nibli1 doesn't entail the nibli2" -- that's not how logical entailment works. 13:38 < zipcpi> No, he made a mistake 13:38 < ldlework> omg 13:38 < zipcpi> He made a statement that was false 13:38 < ldlework> xalbo: not until we know it is false! 13:38 < ldlework> IE a counter example is shown 13:38 <@xalbo> You don't fail to prove that it doesn't entail...and therefore conclude it does. Or at least, you don't call that logical entailment. 13:39 < ldlework> Actually you do 13:39 < ldlework> We conclud things logically entailed as long as no counter example to the nibli1 or model, is shown 13:40 < ldlework> If I was to produce a counter example to the models we have 'proven' then we get to retroactively go back and say all the things we've been saying are logically entailed, are not 13:40 <@xalbo> I don't remember "Proof by giving up before we found a counterexample" in any of my classes. 13:40 < ldlework> the nibli1 was an incomplete model 13:40 < ldlework> xalbo: say the same thing you're trying to insinuate 13:40 < ldlework> with a direct assertion 13:40 < ldlework> rather than an implication 13:40 < ldlework> and we'll see if it sounds ridiculous or makes a lot of sense 13:40 < ldlework> I think your implication is hiding the ridiculousness of your actual conclusion 13:40 < ldlework> so say it frankly 13:41 <@xalbo> I directly assert that {nibli} isn't what you're looking for. It is a statement about how we manipulate symbols in formal systems. It's not useful for talking about whether someone will show up for dinner or not. 13:41 < ldlework> No 13:41 < ldlework> Please actually say what you mean above 13:41 < ldlework> About logical entailment 13:41 < ldlework> That if we prove something logically entails 13:41 < ldlework> and are satisfied 13:41 < ldlework> for say 1000 years 13:41 < ldlework> have we not been saying "that is logically entailed" 13:41 < ldlework> for a thousand years 13:41 < ldlework> until someone produces a counter example? 13:41 < ldlework> Surely, everyone was wrong 13:41 < ldlework> But they still concluded it was nibli 13:42 < ldlework> So there is an effect of incomplete models 13:42 < ldlework> at play, in all considerations of whether something is nibli or not 13:42 < dutchie> if someone pulls out a counterexample, it wasn't proved properly in the first place 13:42 < ldlework> dutchie: exactly 13:42 < ldlework> So when I say 13:42 * dutchie wades in without having fully read scrollback 13:42 < ldlework> anything nibli anything 13:42 < ldlework> I'm doing just as good of a job 13:42 < ldlework> as logicians saying nibli anything nibli 13:42 < ldlework> Because we both are saying things that are potentially NOT entailed 13:43 < ldlework> So when I say a bunch of stuff that you say "no that's just jinvi" 13:43 < dutchie> i would say that "x=2" nibli "x^2 = 4" 13:43 < ldlework> No I'm saying, this partial model impies, or entails the conclusion I'm providing 13:43 < dutchie> which is certainly a logician/mathematician's statement 13:43 < ldlework> well that can't be true 13:43 < ldlework> x=2 doesn't _entail_ anything 13:44 < ldlework> You may say, I disagree that it is nibli, because what if "z which you haven't considered in your model is true? then your entailment is false." 13:45 < ldlework> I'm saying the same goes for logicans but we don't stop them at every turn to point out there might be missing parts of their model 13:45 < ldlework> We merely use the predicate to understand that the logician believes there is entailment 13:45 < ldlework> And we then investigate the model and conclusion to see whether we agree 13:45 < zipcpi> axiomatic systems != modeling 13:45 < ldlework> I'm pleading to let me make these same mistakes as a speaker and same demands of my interlocutor for _arbitrary sets of incomplete models_ 13:45 < ldlework> zipcpi: that's exactly what modeling means 13:46 < ldlework> a model is a set of axioms 13:46 < dutchie> um 13:46 < dutchie> not in model theory it isn't 13:46 < dutchie> but that's probably not relevant 13:46 < dutchie> actually it might be 13:47 < ldlework> co'o for a while 13:48 < dutchie> i agree, however, that insisting on total proof for {nibli} is likely to make the word useless outside of formal logic 13:48 < ldlework> and as I've just proven, useless there too 13:49 < zipcpi> I dunno, someone fire up the gismu co'e for indika, and maybe we should make BAI for it too 13:49 < ldlework> I think my point is hidden in my question to durka 13:50 < ldlework> If not logic, what are the things that relate the details of one's departure that leads others to make conclusions about them 13:50 < zipcpi> Circumstantial / forensic evidence. 13:50 < ldlework> I'm still using logical entailment to arrive at my purported nibili2 13:50 < zipcpi> The English word "logic" is overloaded 13:50 < ldlework> just using a very incomplete model 13:50 < ldlework> for which many counter-examples may be produced 13:51 < ldlework> but likely, my listener will make an immediate consideration of whether they agree with my conclusion based on the incomplete model 13:51 < ldlework> we will both go on, having concluded one way or another, given the incomplete model 13:51 < ldlework> resulting in either a "Yeah what the hell we're going to be late!" 13:51 < ldlework> or "Nah, I heard there was maintenence on his tram line" 13:52 < ldlework> All we're trying to do with nibli is say that the conclusion is rationally derrived from the givens 13:52 < ldlework> as opposed to say, whimsy, preference, malace, goodwill or anything else 13:52 < ldlework> without having to say 13:52 < ldlework> here's my givens, here's my conclusion, my conclusion is racli, y'know! 13:53 < ldlework> as three different steps, or even 2 13:53 < ldlework> since saying "something concludes from something else" is a pretty fundamental assertion 13:54 < ldlework> zipcpi: I believe the only difference between nibli and your thing, is whether the model is incomplete and whether the truth-value of the conclusion can be proven without a doubt by peers far removed from context given purely only the model set forth in nibli1 13:54 < ldlework> I argue that classifying certain models as complete and universally nibli is very ivory tower, and makes a useless predicate from the word 13:57 <@xalbo> Formal systems exist, though, in which absolutely proof actually is possible. And since Lojban was sort of built around FOPL, having a predicate that specifically and precisely encapsulates one of the fundamental operations of those systems doesn't seem that horrible. 13:57 < ldlework> Except that its not 13:57 < ldlework> We just covered the case of counter examples 13:57 <@xalbo> No, you just asserted that there are counter examples. O 13:58 < ldlework> counter examples exist and until you show me a history of logic where counter examples are always shown right as part of the provisioning of logical systems 13:58 <@xalbo> I'm saying that there can be no such thing. Period. 13:58 < ldlework> right, so you place certain systems you personally like into an ivory tower, great 13:59 < ldlework> Disregard that logic is shattered into a million different types to account for the counter-examples and paradoxes that arise in each one 13:59 <@xalbo> Show me a proof that was held to be true in an axiomatic system for any length of time before a counterexample was discovered. 13:59 < ldlework> No counter-examples! is exclaimed, an all systems of logic were immediately unified. 13:59 < ldlework> And then on the second day... 14:00 < ldlework> xalbo: point to me at system for which no counter example or no derrivative formalization was required 14:00 < ldlework> Since you challenge me to provide even 1, I'm sure you have the benefit of reaching into the limitless bounty of self-consistent useful systems of logic for which no problems were ever discovered. 14:01 < dutchie> well gödel kind of ruined our fun as far as that goes 14:01 < ldlework> And then prove to me no counter-examples exist for that system 14:01 * ldlework snickers. 14:01 < ldlework> dutchie: hehe 14:02 < dutchie> but still, if you believe e.g. peano arithmetic is inconsistent, then we have bigger fish to fry 14:02 < ldlework> I'm not going to make claims about specific systems. 14:02 < zipcpi> Darn that Godel lol... still though, it's not as if someone could provide a "counterexample" to the prime thing. There can be different axiomatic systems, like Euclidean vs non-Euclidean geometry, but nothing in one system "disproves" the other 14:03 <@xalbo> dutchie: Who knows if it's consistent? Maybe we'll find a counter example tomorrow! 14:03 < ldlework> I'm just saying, that there's nothing inherent about a system that makes it universally consistent forever, as systems we've made that conclusion about, have shown not to have that property. 14:03 < dutchie> no 14:03 < ldlework> lol 14:04 < dutchie> but i'm sure i do remember that certain weaker systems of arithmetic can (using external methods) to be consistent (i.e. no counterexamples) 14:04 < dutchie> "This means it is possible to algorithmically determine, for any sentence in the language of Presburger arithmetic, whether that sentence is provable from the axioms of Presburger arithmetic" 14:04 < ldlework> Also the x3 allows for the same arguments considered under different rules. 14:05 < dutchie> although it's definitely in "ivory tower" land 14:05 < ldlework> So maybe my x3 is "given the incomplete intuitional rationality that drives us in any context besides those formally defined outside of cultural intuition" 14:05 <@xalbo> Creating a consistent axiomatic system is really, really easy. What's hard is to create a consistent axiomatic system that models reality. 14:05 < ldlework> so there 14:06 < dutchie> xalbo: indeed. people often forget the qualification of GFIT that it applies to systems *in which you can model arithmetic* 14:07 <@xalbo> dutchie: Well, that and the fact that it doesn't mean that there aren't plenty of true statements that *can* be proved in the system. 14:08 < dutchie> xalbo: yes, but that's irrelevant to the current "you can't prove there aren't counterexamples" discussion 14:08 < ldlework> I'll just use fi zo'e from now on 14:08 < dutchie> ldlework: lol 14:08 <@xalbo> It's just that the system can't prove all true statements. That doesn't mean it *can* prove statements that *aren't* true. 14:08 < ldlework> xalbo: but you go further 14:09 < ldlework> I mean, given a specific logical system I actually agree with your extremism with related to entailment 14:09 < ldlework> I don't fundamentally agree 14:09 < ldlework> disagree* 14:09 <@xalbo> I assert that a system that is not consistent is not a valid nibli3. I further assert that your "given the incomplete intuitional rationality that drives us in any context besides those formally defined outside of cultural intuition" is not a consistent axiomatic system. 14:09 < ldlework> The givens must actually give rise to the conclusion 14:10 < ldlework> But its not really a question of whether the logical system is complete 14:10 < ldlework> Its whether the nibli1 is complete enough to entail the nibli2 under a given logical system 14:10 < ldlework> Also the predicate doesn't say anything about consistent axiomatic system, it just says rules for logic 14:11 < ldlework> Your disagreement is that, saying, his train left on time and that its close enough to the destination isn't enough facts or givens to actaully conclude they should or will arrive on time 14:11 < ldlework> your problem isn't with what logical system is used, its that those givens don't actually entail 14:11 < ldlework> right? 14:12 <@xalbo> My problem is that I consider it a huge type error. 14:12 < ldlework> Another thing is 14:12 < ldlework> where's the type error 14:13 < Amnadya> May i see the sentence under discussion? 14:13 < dutchie> i'm not sure there is one 14:13 < ldlework> Amnadya: its about whether you can use nibli if you have an incomplete nibli1 14:14 < ldlework> IE a nibli that doesn't fully entail nibli2 14:14 < ldlework> IE, say the nibli1 provides a bunch of facts, but you could add one more that would make the nibli2 false, or true, depending on the nature of the missing given 14:14 < ldlework> IE, the given nibli1 isn't enough to fully consumate an entailing conclusion 14:15 < zipcpi> But axiomatic systems don't work that way. You don't prove "all primes greater than 2 are odd" by adding one more fact or one more circumstantial evidence 14:15 <@xalbo> I think we're defining "entail" to mean drastically different things. 14:15 < zipcpi> Formal logic is not a Phoenix Wright courtroom; there are no "turnabouts" 14:15 < ldlework> x is a man => x is mortal 14:16 < ldlework> x is mortal can be true or not true, the givens don't fully consumate the entailment of the conclusion 14:16 < ldlework> But by adding in 14:16 < ldlework> all men are mortal 14:16 < ldlework> or 14:16 < ldlework> all men are immortal 14:16 < ldlework> we fully consumate the truth value of the conclusion 14:16 < ldlework> and decide whether the things on the left fully entail the conclusion on the right 14:17 < zipcpi> *you don't disprove, sorry 14:17 < dutchie> i wouldn't say that is really an aspect of nibli1 but of the full "entailment in some system" predicate 14:17 < ldlework> what? 14:18 <@xalbo> dutchie: I claim that {nibli} *is* the full "entailment in some system" predicate. 14:18 < dutchie> xalbo: and i agree 14:18 <@xalbo> je'e 14:18 < Ilmen> {loduu co'e gi'e remna fa ko'a cu nibli loduu ko'a mroka'e} 14:18 < Amnadya> coi la ilmen 14:18 < ldlework> I agree as well 14:19 < dutchie> well, nibli1 in itself can't be fully "complete" or "incomplete" 14:19 < dutchie> it depends on your nibli3 14:19 < dutchie> well, this is where formal logic gets a bit hairy wrt godel 14:19 < dutchie> and the complicated interaction between consistency and completeness 14:19 < ldlework> Okay I guess I'll conclude iwth this question 14:20 < ldlework> Is there value, or enough value to disrupt otherwise cromulent discussions, conversations or any discourse, to stop all interlocutors to discuss whether nibli is justified in unformal discourse given that the speaker didn't provide nibli3 14:20 <@xalbo> nibli1 can be sufficient to prove the nibli2 (for instance, "human(x) ∧ ∀x: human(x) → mortal(x)", or it can be insufficient for that purpose (e.g., just "human(x)") 14:20 < ldlework> xalbo: right 14:21 < dutchie> perhaps what we disagree on is what nibli3 should mean in informal discourse when unspecified 14:22 < dutchie> the obvious choice of "normal human intuition/experience" doesn't fit in well with formal logic 14:22 < ldlework> but we are never using lojban for formal logic 14:22 < ldlework> by never I mean, I have yet to see it on accident where Iw asn't looking for examples 14:23 < dutchie> but that's what we agreed nibli is for? at least that's what i'm interpreting it as 14:23 < ldlework> So 14:23 < dutchie> well, not necessarily formal formal logic, but certainly a strong implication 14:23 < ldlework> the argument places of a predicate are predetermined 14:23 < Ilmen> I wonder about the differences between {nibli} and {banzu}. 14:24 < ldlework> dutchie: anyone saying nibli in the way I'd like to use it, is trying to emphazise the rationality of the purported conclusion 14:24 < ldlework> IE, its not formal formal logic, but they are invoking its name for a reason 14:24 < dutchie> right 14:24 < ldlework> I do like the distinction between sufficient and nessesary conditions 14:24 < ldlework> Like what banzu provides 14:24 < dutchie> in that case, then they must at least subconsciously have some system in their mind 14:25 < ldlework> that he left on time, and the station is close enough, is sufficent for him being here on time 14:25 < ldlework> dutchie: there's also {ko'a je zo'e je zo'e nibli} where ko'a is the only substantive reason I've given :) 14:26 < Ilmen> ko cliva lo mi vacri doi ka glare dukse 14:26 < dutchie> my impression of {banzu} is that it's material sufficiency; you {banzu} enough food to avoid starving 14:26 < ldlework> it should be as general as possible 14:26 < ldlework> so you should interpret it as the most general form of sufficiency as you're able 14:27 < ldlework> its the compliment of nitcu 14:27 < dutchie> ok agreed 14:27 < Ilmen> je isn't appropriate here; {joi} would be preferable 14:27 < Ilmen> pe'i 14:27 < ldlework> oh because je multiplies the proposition? 14:27 < ldlework> je'e 14:27 < ldlework> ki'e 14:27 < dutchie> possibly je is appropriate? but they certainly have a different meaning 14:28 < dutchie> if the implications are true separately, then je works; if not then joi 14:28 < ldlework> dutchie: I think he means ko'a je ko'e nibli really means 14:28 < ldlework> ko'a nibli AND ko'e nibli 14:28 < Ilmen> .e/je blindly create two different bridis 14:28 < ldlework> which is probably false 14:28 < dutchie> indeed 14:28 < dutchie> just checking, i'm still only new to lojban 14:29 < dutchie> but anyway, {ko'a joi zo'e joi zo'e} is the speaker not providing enough information, not nibli being wrong 14:30 < zipcpi> Example sentence: lo nu le degypri cu mapti cu indika lo du'u la .tom. cu catra 14:30 < Ilmen> He leaves information to context, probably because it's sufficiently clear not to bother superfluously adding it explicitly 14:32 < zipcpi> le nu le degypri cu mapti cu indika lo du'u la .tom. cu catra :: The fact that the fingerprints match implies that Tom is the killer. 14:33 < dutchie> nu or du'u for indika1? 14:35 < ldlework> nu implie du'u 14:35 < ldlework> so I'd say du'u 14:35 < durka42> is {indika} just {ve jinvi fo zi'o}? 14:36 < ldlework> durka42: he should be here on time, but I expect he'll be late 14:36 < Amnadya> Yeah, implies to whom? 14:36 < durka42> is that supposed to be an answer to my question? 14:36 < zipcpi> durka42: Not sure. There is a possibility of tying it to Bayesian evidence 14:36 < ldlework> durka42: its to distinguish opinion from weak entailment 14:36 < ldlework> I don't believe he will be here 14:37 < ldlework> But all facts point to him being here 14:37 < durka42> zipcpi: mm. like, x2 is more likely to be true given that x1 is true 14:37 < ldlework> You know, except the ones that would _entail_ him being here. 14:37 <@xalbo> durka42: Not completely. You rarely jinvi both x and ¬x, but it's not uncommon for there to be evidence toward both. 14:37 < durka42> true 14:37 < ldlework> Also, banzu seems to give a "he could be here on time" and we're looking for something else 14:38 < durka42> but the negative evidence would convince someone who didn't see the positive evidence 14:38 < durka42> and vice versa 14:38 < zipcpi> Yes, think Phoenix Wright and the constant turnabouts :p 14:38 < ldlework> We need a predicate that x1 allows for x2 to be true but might be false 14:38 < durka42> zipcpi: I dunno what that is 14:38 < ldlework> (only beause it doesn't entails) 14:39 < zipcpi> durka42: Courtroom drama game 14:39 < Amnadya> "na natfe" 14:39 < ldlework> also what happens when nibli2 is statistical in nature? 14:39 <@xalbo> Amnadya: Yes, that's it exactly. 14:40 < ldlework> again that's just "he could be here on time" 14:40 < ldlework> "circumstances allow for his prudence" 14:40 < ldlework> perhaps that's what english speakers mean by "should" ? 14:41 <@xalbo> That's among the 50 ba'u things they mean by "should". 14:41 < ldlework> right, but in the context of what we've been talking about 14:41 < ldlework> I'm thinking more like, "all externalizing factors X encourage or promote the outcome Y" 14:41 < ldlework> IE they do more than allow for it 14:41 < dutchie> {kanpe fi so'e}? 14:42 < ldlework> The road was wet, they were going to fast, the brakes were broken, and they were on their phone." 14:42 < dutchie> er, {piso'e} i think 14:42 < ldlework> None of this -entails- that they crash 14:42 < ldlework> But it doesn't merely allow for it either 14:42 < ldlework> And I feel like all of this would be true 14:42 < ldlework> If the only person alive in the universe was the person who crashed 14:43 < ldlework> those factors still did more than allow for the crash 14:43 < ldlework> even though no one opines or expects any given outcome 14:43 < Amnadya> laknyzenriha, 14:44 <@xalbo> What I've wanted for a while, and {indika} sounds pretty close, is "p(x2|x1) > p(x2|¬x1)". That is, knowing x1 we believe x2 to be more likely than we would have believed it to be not knowing x1. x1 is evidence in favor of x2. 14:44 < ldlework> We're going to far to blame any given effect for the true causation though 14:44 < ldlework> unless the only causation you mean by rinka is the lakne zenba 14:45 < ldlework> But how can something be evidence of something that hasn't happened? 14:46 < ldlework> nibmli :) 14:46 < zipcpi> ko'a can be statistical/circumstantial evidence for ko'e, but ko'e might still not be true 14:46 < ldlework> I meant the other way around 14:46 < ldlework> How do we think about entailment when the outcome is statistical 14:46 < ldlework> IE, X entails Y might happen 14:47 < ldlework> When we don't even purport to show that X entails Y _will_ happen 14:47 < Amnadya> cumnihi 14:47 < dutchie> i don't think nibli is appropriate for that 14:47 < ldlework> dutchie: its not possible to entailment the conditions for possibility? 14:48 < ldlework> to entail* 14:48 < ldlework> in this case we only care to entail that the state of things that supports some outcome is true 14:48 < ldlework> I don't see why that doesn't fall under the same purview 14:49 < dutchie> you could entain the possibility, but not the possible event 14:49 < ldlework> That seems to create a distinction where one isn't demanded 14:49 < dutchie> well, it depends on how strict you want your semantics for nibli to be i suppose 14:50 < ldlework> Mmm I think "proving the circumstances that make some event possible, true" 14:50 < Amnadya> ithkuil: ilmen 14:50 < ithkuil> ilmen: P1S1/DYN-lm-ABS-UNI/COA/M/DEL/NRM-EXS-IFL/UNFRAMED 14:50 < ldlework> is the same as saying 14:50 < ldlework> "proving that some event is possible" 14:50 < ldlework> is the same semantic data 14:50 < dutchie> yes 14:50 < zipcpi> Amandya: lol what does it mean? 14:50 < Ilmen> .a'u 14:50 < dutchie> but it's different to "proving that some event happens/will happen" 14:50 < ldlework> Right but that's what I mean by statistical 14:50 < Amnadya> something about music 14:50 < ldlework> In that truth-value becomes meaningless 14:50 < Ilmen> .u'e 14:50 < durka42> ithkuil: durka 14:50 < ithkuil> durka: P1S1/STA-d-IND-AGG/ASO/U/DEL/NRM-EXS-IFL/UNFRAMED 14:51 < zipcpi> -lm- "music" 14:51 < ldlework> you don't say either or, you can only entail likelyhood or mere potential 14:51 < ctefaho> how about a word for "x may or may not be the case" 14:51 < ctefaho> and make it a UI 14:51 < durka42> that's cumki 14:51 < ldlework> I mean, isn't this the very problem that fuzzy logic addresses? 14:51 < ldlework> Where the conclusions are not binary 14:51 < dutchie> or an abstractor? "x1 is the possibility of [bridi] with probability x2" or something 14:51 < ctefaho> cumki is too specific 14:51 < dutchie> i have no idea if that's useful though 14:51 < ctefaho> we need something vaguer 14:52 < ctefaho> +zo'o 14:52 < zipcpi> zilkanpe? 14:52 < ldlework> I like the idea of 'crippled entailment' 14:52 <@xalbo> cu'o? 14:52 < zipcpi> ua zo cu'o 14:52 * durka42 can't think of anything less specific than cumki :) 14:53 < ldlework> durka42: lakne 14:53 < ldlework> oh that's more specific 14:53 * ldlework burps. 14:53 < zipcpi> And {zilpa'a} is already defined 14:53 < ldlework> "defective entailment" 14:54 <@xalbo> faurjavnarfau: x1 either happens or it doesn't. 14:54 < zipcpi> latyjavnarmlatu 14:54 < ctefaho> durka42: cumki cai 14:54 <@xalbo> .u'i 14:54 < ldlework> IE, nothing is entailed, but everything involved could potentially be involved in such an entailment 14:54 < ldlework> IE 14:54 < ldlework> Nothing provided contradicts, and works towards entailment but is insufficent for real entailment 14:55 <@xalbo> latyjvenarmlatu 14:55 < ldlework> IE, its not counter-evidence, and its not irrelevant evidence 14:55 < durka42> na'e natfe? 14:55 < ldlework> And evidence, for whatever word means evidence for things that haven't happened yet 14:56 < ldlework> like the role gravity plays in support beams breaking 14:56 < ldlework> IE, gravity doesn't break beams 14:56 < ldlework> but it pulls other things down 14:56 < ldlework> the weight of those things might create force enough to break the beam 14:56 < ldlework> IE mere gravity doesn't entail the support beam breaking 14:57 < ldlework> Neither does the weight of what is supported 14:57 < ldlework> Neither does the crack in the beam 14:57 < ldlework> But leaving out gravity "the crack" + "the weight" something something the support breaking 14:57 < ldlework> Anything that allows us to demonstrate our rational intuitions of causation 14:57 < ldlework> For realized and unrealized events 14:57 < ldlework> Or truths 14:58 < ldlework> IE I didn't state everything that entails the breaking of the support beam 14:58 < ldlework> I can't prove that it will break, but I want to convey that it logically follows, and that you should agree with me not becaue I'm popular and have opinions people follow 14:58 < dutchie> does {rinka} fit there? 14:58 < ldlework> But because you can use the same rationalization to come to the same conclusion 14:59 < ldlework> dutchie: go back to the guy who should be here on time because everything has gone in their favor 14:59 < ldlework> But opined not to show up on time (for whatever reason) 15:00 < durka42> I'm slowly being convinced that {indika} is necessary and that it means P(x2 | x1 is true) > P(x2 | x1 is unknown) 15:00 < ldlework> I'm trying to discover the universal consitutive semantic contained within english "should", in this context 15:01 < dutchie> i think i agree with durka42 15:01 < durka42> so it will be defined in terms of {zilkanpe} I guess 15:02 < durka42> actually {nilyla'e} is better than {zilkanpe} because y'all are afraid to use {zi'o} 15:02 < zipcpi> u'i 15:02 < ldlework> I still think its closer to banzu 15:03 < ldlework> There's no consideration or valuation 15:03 < ldlework> its not a matter of expectation, rather than an observation of the rationalaity of things 15:03 < durka42> hmm 15:03 < durka42> example sentence with banzu? 15:03 < ldlework> IE, how the simulation or system ought to behave based on its observed characteristics 15:03 < durka42> banzu2 is supposed to be a "purpose" 15:04 < ldlework> There's too much hand-wavy "meaning" as in personally-identifying meaning being injected here :D 15:04 < ldlework> even with "purpose" 15:04 < durka42> I am just wondering what form the banzu2 would take 15:04 < ldlework> Does the law assume entailment? 15:04 < ldlework> Like nibli assumes entailment? 15:04 < durka42> which law? 15:05 < ldlework> mortal law 15:05 < ldlework> What is the verb that characterizes that which Sherlock Holmes does the most. 15:05 < ldlework> Beyond expectation 15:05 < ldlework> Does Holmes conclude based on "logical entailment" 15:05 < durka42> evidence 15:05 < durka42> what's mortal law? what are you asking, lol 15:05 < dutchie> deduction? 15:06 < ldlework> In that, the facts imply the conclusion 15:06 * durka42 still wonders how banzu can fit into this 15:06 < ldlework> dutchie: okay so what characteristic of a bunch of isolate facts, allow us to use them collectively to deduce conclusions 15:06 < dutchie> en:didni 15:06 < mensi> didni = x1 deduces/reasons by deduction/establishes by deduction that x2 is true about x3 from general rule 15:06 < mensi> x4 15:06 < ldlework> IE, what is the ka broda 15:06 < ldlework> of the facts themselves 15:06 < ldlework> that allows one to deduce some conclusion that would hold up in any court 15:06 < ldlework> rather than being an expectation, valuation, or otherwise 15:07 < durka42> actually what sherlock does is induction, but yeah 15:07 < ldlework> IE, the facts entail the conclusion to everyone's satisfaction 15:07 < ldlework> but not entail in the way we've covered 15:07 < dutchie> durka42: i am aware, but everyone (including holmes/conan doyle) calls it deduction 15:07 < ldlework> "I deduced that you" 15:07 < ldlework> ie 15:08 < ldlework> I think that I have finally found a way of enunciating my question that seems satisfactory to myself, I'm not sure I care what the eventual answer is 15:08 < dutchie> haha 15:08 < ldlework> But I feel like I've arrived at its proper utterance 15:08 < ldlework> Sorry xalbo 15:08 < ldlework> I want to know the property of those facts. 15:08 < ldlework> Not the property of anything to do with the deducer 15:08 < dutchie> so why not {didni}? 15:09 < ldlework> That characterizes Holmes 15:09 < durka42> {nusna} 15:09 < ldlework> Not the collection of facts 15:09 < dutchie> ah yes 15:09 < ldlework> An really 15:09 < durka42> you sure like the sound of your own voice :p 15:09 < ldlework> I'd love just a general observation 15:09 < ldlework> Not just a lojbanic one 15:09 < ldlework> it seems a very subtle semantic 15:09 < ldlework> especially in face of what it really meants to logically entail 15:09 < durka42> I'm still not sure what you were getting at 15:09 < durka42> with all those rhetorical questions 15:10 < durka42> were you trying to get us to say "evidence"? 15:10 < ldlework> dutchie: do you feel that you do? 15:10 < ldlework> just polling, not insinuating 15:10 < ldlework> In the same way that some facts, by their own nature, can entail some logically cromulent conclusion 15:10 < ldlework> What can we say about facts that allows a detective to create deductions based on them 15:11 < ldlework> We say that the facts entail the conclusion, we don't speak of the logician 15:11 < dutchie> i think you've finally figured it out, and it is a valid question 15:11 < ldlework> But here, we say that Holmes uses these facts to deduced 15:11 < ldlework> What can we say about Holmes' facts, that are sufficient to allow him to use them for deduction 15:11 < ldlework> If not "entailment" 15:11 < durka42> he doesn't deduce, though 15:11 < durka42> he induces 15:12 < durka42> he just uses patterns 15:12 < ldlework> durka42: sure we already covered that distinction 15:12 < dutchie> natlang pedanticism that i think is irrelevant 15:12 < durka42> right but it's inaccurate to say "what allows him to deduce" 15:12 < ldlework> durka42: okay regardless of the verb we use to describe the behavior of the agent 15:12 < durka42> if something allowed him to use deduction, then he would nibli 15:12 < ldlework> moving on from that 15:12 < ldlework> what verb can we use for the facts 15:12 < ldlework> what verb makes facts valid for induction that allows for the prosecution of law and so on 15:12 < dutchie> ve nusna? 15:12 < durka42> "suggest" in english 15:12 < ldlework> IE so sufficient that everyone is satisfied in their conclusion they lead do 15:12 < dutchie> suggest is probably too weak 15:13 < dutchie> maybe there isn't an english word 15:13 < ldlework> The only word I know that even gets close to the semantic we're exloring is suggest 15:13 < ldlework> I also thought entail 15:13 < ldlework> In the way that we've proved who comitted the crime 15:13 < dutchie> implies 15:13 < ldlework> But logic is in a tower we can't access here 15:13 < dutchie> these words are (unsurprisingly) all bound up with formal logic sematics 15:13 < ldlework> exactly... 15:14 < ldlework> Maybe it is logic that Holmes is using. A difference sense, but one that involves the rational implications of givens that arrive at conclusoins etc etc 15:15 < ldlework> A sense of which is enough to put people to death and so on 15:15 < durka42> a different nibli3 then 15:15 < ldlework> Even though it can never be _entailed_ 15:15 < dutchie> there is certainly a difference between pure/mathematical/formal logic and everyday/detective/legal logic 15:16 < dutchie> perhaps nibli3 is enough? it definitely bears further thinking about 15:16 < ldlework> Well if you ask me 15:16 < ldlework> I think that people sayin nibli in normal conversation should be addressed as it is with Holms 15:16 < ldlework> only when a nibli3 is provided 15:16 < dutchie> maybe i'll post to the mriste 15:16 < ldlework> should you start attacking people based on specific rules of logic and so on 15:16 < ldlework> I still think that xalbo's wider point about what entailment means is true 15:16 < ldlework> fundamentally true 15:16 < ldlework> objectively true 15:17 < ldlework> but true in a way that doesn't address at all how we talk to each other 15:17 < durka42> it kinda has to be context dependent 15:17 < ldlework> IE, the truer true 15:17 < ldlework> language is not true or false most of the time 15:17 < durka42> some lujvo like {cmacynibli} and {holmes zei nibli} can help to disambiguate 15:17 < ldlework> haha 15:17 < ldlework> Lojban will eventually become a mockery of itself.. 15:18 < durka42> well obviously I don't mean literally {holmes zei nibli} 15:18 < ldlework> durka42: sure, but I enjoyed the joke 15:19 < dutchie> e'u sa'u tolcmacyni'i 15:20 < dutchie> si zo tolcmacyni'i 15:20 < durka42> maybe {cmacyni'i} and {jijyni'i} 15:20 < durka42> la .cirlak. cu jijyni'i certu 15:21 < dutchie> {jijyni'i} is good 15:21 < zipcpi> http://jbovlaste.lojban.org/dict/biglogji?bg=1;langidarg=2 ????!?!?!?? 15:21 < durka42> .w deontic 15:21 < fenki> "Deontological ethics or deontology (from Greek δέον, deon, 'obligation, duty'[1]|) is the normative ethical position that judges the morality of an action based on the action's adherence to a rule or rules." - https://en.wikipedia.org/wiki/Deontic 15:21 < dutchie> that's a lot of big words 15:22 < zipcpi> Not sure about logji. Maybe flese is better; but flese's place structure... oi 15:22 < ldlework> We could create a gismu 15:22 < ldlework> IE a fundamental predicate 15:22 < ldlework> So we can create all sorts of colored implications 15:22 < zipcpi> flese is already a gismu; the problem is that it's a Curtis gismu 15:22 < durka42> indeed 15:23 < ldlework> We could create a gismu 15:23 < durka42> do we need a gismu? 15:23 < ldlework> Its pretty close to nibli 15:23 < ldlework> IE, its pretty damned fundamental 15:23 < ldlework> and would allow us to create vocabulary on top it 15:23 < ldlework> that thing I just said 15:23 < zipcpi> Er, {indika}? 15:24 < durka42> {jijyni'i}? 15:24 < ldlework> fu'ivla 15:24 < zipcpi> No, do you mean you want a gismu for {indika}? 15:24 < ldlework> So nothing deserves a gismu 15:24 < ldlework> Just polling for the underlying valuation here... 15:25 < zipcpi> I did leave a comment on {indika} for possible gismu-ization 15:25 < zipcpi> And maybe even a BAI 15:25 < durka42> nah, you were making a blanket ad-ridiculum statement, which makes it really hard to talk to you :) 15:25 < ldlework> durka42: thanks, I was worried this conversation had gone too long without making a comment on my personality 15:25 < dutchie> we were doing so well 15:26 < ldlework> Oh I forgot my smiley face 15:27 < durka42> whatever "underlying valuation" means, I have a high bar on gimy'inda personally, so if the lujvo works I don't see a need to gismuize it 15:27 < durka42> but I could be convinced 15:27 < durka42> that's why I asked 15:27 < zipcpi> Well crap... didni vs nusna 15:28 < ldlework> I said "How about a gismu" 15:28 < ldlework> and you guys made non-gismu suggestions 15:28 < ldlework> And I said "Okay so no gismu ever." 15:28 < ldlework> As to tease out your valuations on what deserves a gismu and what doesn't 15:28 < zipcpi> I was asking what you wanted to make a gismu 15:28 < ldlework> zipcpi: I know I'm replying to durka42's implication that none of my speech makes any sense 15:28 < ctefaho> ldlework: I for one love gismu 15:29 < durka42> yeah it's just weird that "maybe not a gismu for this" is equivalent to "no gismu ever" 15:29 < ctefaho> but the ones I add keep getting vandalized 15:29 < durka42> maybe it's just case-b-ycase 15:29 < ldlework> durka42: really because you identified it as taking your position to the extreme 15:29 < durka42> case-by-case* 15:29 < ldlework> exactly what I intended to do 15:29 < durka42> but why would you do that 15:29 < durka42> it's rude 15:29 < ldlework> Why? only if I'm trying to be rude 15:29 < ldlework> I did it as something ti distinguish against 15:29 < durka42> that is _not_ how rudeness works, llol 15:29 < ldlework> I didn't then go and make your personality about not wanting gismu ever 15:29 < ldlework> I didn't characterize you 15:29 < ldlework> note what is really rude 15:30 < durka42> this is a stupid conversation 15:30 < ldlework> Agreed 15:30 < ldlework> can you please stop prompting it? 15:30 < ldlework> daily? 15:30 < ldlework> sometimes more frequently than that? 15:30 < ldlework> Stop making conversation about people and then saying "omg lets stop doing this" 15:30 < ldlework> just skip to the "don't do it" part 15:30 < durka42> I'm not going to dignify that with a response, other than "I'm not going to dignify that with a response". 15:30 < ldlework> regardless of how tempted you are 15:30 * ctefaho goes back to making awesome jbo-stuff alone 15:30 < ldlework> durka42: yes, you're very dignified as is 15:32 < durka42> so should we jbovlastegau {cmacyni'i} and {jijyni'i}? 15:32 < durka42> or do we need a gismu? I'd be content to base these on {nibli} and call that the fundamental concept, but perhaps someone has a different opinion 15:32 < zipcpi> What's wrong with {indika}? 15:32 < durka42> well, can it be defined in lojban? 15:33 < durka42> if not, we need some more words :) 15:33 < zipcpi> If we want to define it on Bayesian logic, we can 15:34 < durka42> only kind o' logic there is 15:34 < ctefaho> hey how about merging all brivla into one word form 15:34 * durka42 looks around furtively for any frequentists hiding in the shadows 15:34 < durka42> I've never seen a frequentist in here, so the probability of finding one is zero 15:34 < durka42> xo'o 15:34 < ldlework> I don't see how nibli is the foundation for this went we went so far to show that this isn't nibli 15:35 < durka42> I thought we eventually circled back around to saying it _was_ nibli but with a different nibli3 15:35 < ldlework> And it is some isolate concept not well represented in lojban or english 15:35 < ldlework> Let's ask xalbo 15:35 < durka42> ctefaho: we could just number all the gismu and say {mi rodypavoreci la .bastn.} 15:35 < ctefaho> ldlework: propose a gismu with def and I can put it somewhere 15:35 < ctefaho> durka42: brodaxiPA 15:35 < ldlework> ctefaho: I don't know how gismu generation works at all 15:36 < dutchie> ctefaho: uisai 15:36 < dutchie> si u'i 15:36 < dutchie> sai 15:36 < ctefaho> ldlework: well the consonants must not conflict but I don't know where that list is online 15:36 < ctefaho> I have the gismu generator, lemme see if I can find them rules 15:36 < ldlework> Also it seems other's are opposed to gismu for some reason 15:36 < ctefaho> ...right I hadn't installed python 15:37 < ldlework> I think durka42 because he has established it is nibli based 15:37 < ldlework> durka42: is that fair? 15:37 < ldlework> Why gismu if we've established its derrivative? 15:37 < zipcpi> lo ni'ai ke'a nilyla'e febu va'o fabu cu dubzma lo ni'ai ke'a nilyla'e febo va'onai fabu 15:37 < durka42> you think that I what? 15:37 < ctefaho> ldlework: I for one do at least have an idea what should be in zi'evla/lujvo/gismu 15:37 < durka42> en:ni'ai 15:37 < mensi> ni'ai = [NU] x1 is a number/value such that the abstraction is true, under mathematical system x2 // x1 binds to ke'a 15:37 < mensi> within the abstraction. 15:37 < ldlework> durka42: don't want a gismu, but a lujvo because we decided the concept is derrivative of nibli? 15:37 < zipcpi> Numeric poi'i 15:37 < ctefaho> ...but not sure if I can define it well right now 1 minute before going sipna 15:38 < ctefaho> others just doesn't seem to care about the distinction 15:38 < durka42> zipcpi: looks good for {indika} 15:38 < durka42> the only real advantage of gismu is they can have short rafsi, pe'i 15:38 < ctefaho> see logbanu for where that's gonna go 15:39 < ldlework> ctefaho: where what's going to go? 15:39 < durka42> ldlework: well, I thought that's what we decided 15:39 < ctefaho> durka42: yeah gismu should be the only ones with short rafsi imo 15:39 < ldlework> durka42: yeah I was just reiterating to make sure that was also your reasoning for lujvo 15:39 < ctefaho> ldlework: https://newlojban.wordpress.com/ 15:39 < ldlework> I swear I haven't been trying to be snide or impolite at any point in this conversation 15:39 < ctefaho> merging zi'evla and lujvo which selpai is doing 15:40 < ldlework> ctefaho: I don't understand how that's possible 15:40 < ldlework> does he have an article? 15:40 < ldlework> I don't mean cmavo :D 15:40 < ctefaho> well logbanu tinkers with the morphology a bit 15:40 < ctefaho> https://newlojban.wordpress.com/root-words/ 15:40 < ctefaho> err wait 15:40 < ctefaho> wrong 15:40 < ldlework> oh god "new lojban" 15:40 < ctefaho> https://newlojban.wordpress.com/morphology-2/ 15:40 < ldlework> ctefaho: selpahi is responsible for this? 15:40 < ctefaho> (ASSUMING selpai hasn't droppped logbanu completely) 15:41 < ctefaho> ldlework: yep, his work 15:41 < ldlework> I wish that selpahi would hear me out about some of my ideas related to descriptions but it seems he is resolute in his own ideas about them 15:41 < durka42> it's hard to know what selpahi is and isn't working on, .u'i 15:42 < ldlework> I don't mind inventing a new lojban under selpahi since I trust him but I lament that I cannot meaningfully contribute to his model. 15:42 < ctefaho> selpai hasn't been on irc much though so I haven't had time to talk with him either 15:42 < ctefaho> ldlework, what ideas? 15:42 < ldlework> ctefaho: basically speech acts and how reference works. 15:42 < ctefaho> I am kind of tinkering with my own dialect but maybe I will be the only one to speak it, who knows 15:42 < ldlework> ideas that are valiantly partially expressed by what's going on with gadganzu 15:43 < ctefaho> what are these ideas, in brief? 15:43 < ldlework> Basically I think we've made a mistake in trying to strangle the caclulus of fopl until the point where we can speak it outloud 15:43 < ldlework> Rather than starting from language and trying to create a human language informed by the principles of fopl 15:43 < ldlework> And I think a lot of problems arise from doing the first 15:43 < Amnadya> merging lujvo and zihevla is the same as removing rafsi assignments 15:44 < ldlework> And that more and more strange bandaids arise from not getting the fundamental modeling of language quite right 15:44 < dutchie> ldlework: is it perhaps a bit late to change that with lojban? 15:44 < dutchie> though possibly the best place to recruit people to your cause 15:44 < ctefaho> ldlework: Is this kinda like "semantics before logic"? 15:44 < ldlework> dutchie: I think that mostly all that needs to change is how we perform reference 15:44 < ldlework> And I think that reference has been traditionally hard for lojban because 1, of trying to interpret existential variables as the same thing we mean by "linguistic reference" 15:44 < ldlework> And the fact that the ideas of reference are still evolving actively 15:45 < dutchie> je'e 15:45 < ldlework> I mean in the wider philosophical and linguistic fields 15:45 < ldlework> I think this is why lojban's description system changed in the pas too 15:45 < ldlework> Because it really was, any time you make a description, you make an existential assertion 15:45 < ldlework> which we all eventually agreed is innane for an actual language 15:45 < ldlework> with xorlo 15:46 < ldlework> I think lojban's centricity around the proposition is ideal. I think this is lojban's major achievement. 15:46 < dutchie> ie 15:46 < ldlework> It just so happens that fopl is also central around the proposition, so by abusing it for language we just got lucky 15:46 < ldlework> It doesn't work the same when you open the fopl box to explain reference, rather than the language philsophy box. 15:47 < ctefaho> sorry what is "proposition" in this context? the bridi? 15:47 < ldlework> ctefaho: yeah, but minus any illocutionary force 15:47 < ldlework> so just the references and the predicates 15:47 < ctefaho> also sorry again "illocutionary force"? 15:48 < ldlework> the recent discovery in the philosophy of language as it pertains to analyzing speech with the tools of logical analytics is that reference is actually a separate act from the propositional act 15:48 < ldlework> uh 15:48 * ctefaho actuallt jsut googles everything 15:48 < ldlework> John smokes. 15:48 < ldlework> Does John smoke? 15:48 < ldlework> John smoke! 15:48 < ldlework> all contain the same proposition but the eventual illocutionary force is different for each 15:49 < ctefaho> I think this is kinda what some of my own insights are I just don't really know the terms 15:49 < ldlework> What I have been reading is that if you want to analyze these statements logically, you must first seaprate the act of reference from the propositional one 15:49 < ldlework> IE 15:49 < ldlework> the thing that is being predicated by the verb "smokes" isn't the description "john" 15:49 < ldlework> Its whoever the speaker meant by that description 15:49 < ldlework> IE the logical expansion doesn't involve the name at any point 15:49 < ldlework> IE when I say 15:50 < ldlework> The dog is old 15:50 < ldlework> The argument to 'is-old' isn't some existentially dog quantified thing 15:50 < ldlework> reference is a seaprate act 15:50 < dutchie> i'm curious; are you a linguist by education/training? 15:50 < ldlework> all there is to fall under the purview of logical analytics is da old 15:50 < ldlework> dutchie: I'm absolutely nothing by education or training 15:50 < ldlework> I never attended school beyond grade school 15:51 < dutchie> fair enough 15:51 < ldlework> I just read the literature 15:54 < ctefaho> Not sure if I understand you correctly, but this "referential" acts vs. the propositional, is that "lo broda" vs "[zo'e] broda"? 15:56 < ldlework> I'm not going to argue which existential variable cmavo you want to use to represent x in Ex 15:56 < ldlework> But the point is, in fopl, the x doesn't _mean_ anything, it only gains meaning through predication 15:57 < ldlework> The mistake lojban made is that our descriptions are involved in this logical predication 15:57 < ldlework> rather than a utility in speech to get our listeners to identify what is inside the Ex, so we can subsequently go on to say things about whatever is inside that x 15:58 < ldlework> if you have to model it sensibly any given utterance passes through several stages 15:58 < ldlework> first it is uttered 15:58 < ldlework> then its referents are resolved 15:58 < ldlework> rather 15:58 < ldlework> its description terms are resolved into referents 15:58 < ldlework> then _referents_ partake in a proposition along with some predicate 15:59 < ldlework> then an illocutionary force uses the proposition to perform some speech act 15:59 < ldlework> Wow, that's amazing! 15:59 < ldlework> then we resolve what 'that' refers to 15:59 < ldlework> Wow, Ex amazing 15:59 < ldlework> drop the illocutionary forces 15:59 < ldlework> Ex amazing 15:59 < ldlework> "there exist something for which it is amazing" 16:00 < ldlework> Let's do the same thing to a description using thing 16:00 < ldlework> oh btw, the "Wow" comes back, to provide the illocutionary force 16:00 < ldlework> Wow, Bob is amazing! 16:00 < ldlework> Wow, Ex is amazing => amazes(x) 16:00 < ldlework> Wow, That bike is amazing 16:00 < ldlework> Wow, Ex is amazing => amazes(x) 16:01 < ldlework> Predicate logic is inherently devoted to the task of discovering logical entailment between predicates 16:01 < ldlework> If we come up with some axioms about predicate A 16:01 < ldlework> And some axioms about predicate B 16:01 < ldlework> We can then make conclusions about things that are A based on their qualities of being B 16:01 < ldlework> This is different than when I say "the dog is hungy" 16:01 < ldlework> "the dog" does not exist as a predicate in the logical entailment 16:02 < ldlework> When I say 16:02 < ldlework> These, which are dogs, are hungry because of their dogness. 16:02 < ldlework> I have imported the description into the logical entailment 16:02 < ldlework> But I don't do that, all I have said is 16:02 < ldlework> hungry(x) 16:02 < ldlework> Dog does not fall under the purview of whether the statement is true 16:02 < ldlework> Because the truth of dog(x) must be considered seprately 16:03 < ldlework> and modern language philosophers will tell you that we don't ever do this when we are communicating with speech 16:03 < ldlework> We never import the truth-value of descriptions into the truth-value of propositions involing the referents of those descriptions 16:03 < ldlework> It may be false that that referent is a man, but nevertheless, I care to say that referent is noisy. 16:04 < ldlework> The man is noisy. 16:04 < ldlework> The truth-bearingness of descriptions has everything to do with the utility of the description to refer. 16:04 < ldlework> Not to assert. 16:04 < ldlework> And in certain contexts false-beariness is -more- important 16:04 < zipcpi> Well it almost as if we're rediscovering what was somewhat problematically referred to as "non-veridicality" 16:05 < ldlework> zipcpi: its more of the observation that all speech is potentially non-veridical 16:05 < ldlework> and that the default referring act is veridical 16:05 < ldlework> And that specifically denoting whether the description is truth-bearing upon the referent is rarely, if ever, adding value to speech or logical analytics 16:06 < ctefaho> (also hey zipcpi did you see these https://newlojban.wordpress.com/novel-cmavo/ ) 16:06 < ldlework> under the purview of logical analytics, xorlo show's the absurdity of the conclusion 16:06 < ctefaho> ldlework: I think I understand, somewhat, just need some time to process 16:06 < ldlework> under the purview of language philosophy, philosophyers are telling its na'e 16:06 < ldlework> its "irrelevant" 16:06 < zipcpi> Don't you mean na'i 16:06 < ldlework> yeah thanks 16:06 < ldlework> na'i 16:06 < ctefaho> It sounds kinda compatible with what I am doing too 16:07 < ctefaho> I have been striving to "preserve the truth" of bridi's in my rules for expansion 16:07 < ldlework> Its a conflation of language and logical calulus where we usually say something like 16:07 < ldlework> all men are mortal 16:07 < ldlework> x is a man 16:07 < ctefaho> with default NA as ja'a for *every* bridi 16:07 < ldlework> therefore x is a mortal 16:07 < ldlework> x is a man does NOT REPRESENT NOUN PHRASES 16:08 < ldlework> and yet 16:08 < ldlework> when we "logically expand" most lojanic utterances 16:08 < ldlework> we include the description predicates in those expansions 16:08 < ldlework> so either the original logicians make a simple excusable conflation 16:08 < ldlework> or the philosophy of language as a whole is headed in the wrong direction 16:08 < ldlework> the ego of lojbanists will have a lot to say on that question 16:09 < ldlework> But I argue its far simplier 16:09 < ctefaho> in my dialect that just becomes "That which wasn't really a man after all, IS INDEED smoking" 16:09 < ldlework> and the correction is largely a simplification of lojban's gadri system 16:09 < zipcpi> Well we made {le'elkai} to "shield" the referring description from the predication; so maybe all we need is another magic predicate for xorlo 16:09 < zipcpi> {xorlokai} 16:09 < ldlework> I've never heard of {le'ekai} so I have no idea what you mean. 16:09 < ldlework> Also I believe the correction can be made at the level of the existential variable 16:09 < zipcpi> Ilmen made it while watching me talk about {le} 16:09 < ldlework> rather than the gadri 16:10 < ldlework> and the result is that a self-consistent system of gadri arises 16:10 < ldlework> but it will never be self-consistent if we must base it upon the logical conflation 16:10 < ctefaho> ldlework: would love to discuss this more but I really got to to sleep now. did save all you said though. 16:10 < ldlework> ctefaho: I've been talking about this singular topic for weeks maybe going on multiple months now 16:10 < zipcpi> I'm not sure I can agree with making it impossible to refer to the existential variable 16:11 < ctefaho> I have been a bit busy fixing...stuff last weeks to consider these things fully 16:11 < ldlework> zipcpi: you never refer to the existential variable 16:11 < ldlework> the existential variable contains the referent of your description 16:11 < ldlework> Its like a map 16:11 < ldlework> descriptions are a map to the city 16:11 < ldlework> the predicate of your propsoition is the thing you exclaim when you first see the city 16:11 < ldlework> "melbi!" 16:11 < ldlework> the map is not beautiful 16:11 < zipcpi> And some existential variables don't have *any* descriptions attached to them 16:11 < ldlework> the map only helped us identify what is beautiful 16:11 < ldlework> zipcpi: right like, I argue, zo'e 16:12 < ldlework> or da 16:12 < ldlework> they both just represnt the thing to which you want to refer 16:12 < ldlework> IE, you're just supposed to know where the city is 16:12 < ldlework> you don't get a map 16:12 < ldlework> but I'll tell you about that city anyway 16:12 < ldlework> da melbi 16:12 < ldlework> infact here 16:12 < ldlework> I'm not even telling you it is a city 16:12 < ldlework> I'm just telling you something exists and it is beautiful 16:12 < ldlework> and even in the sentence 16:13 < ldlework> "The city is beautiful" 16:13 < ldlework> I only mean to say "some thing is beautiful" 16:13 < ctefaho> ldlework: Does and if so how does your "da melbi" and "zo'e melbi" differ? 16:13 < ldlework> Whether or not the variable refers to a definite reference or not 16:13 < ldlework> definite referent* 16:14 < zipcpi> But now you have added a qualifier to that thing. You may not be veridically referring it as a city, but you are trying to restrict it from a mere 'something" 16:14 < ldlework> zipcpi: but only in the reference act 16:14 < ldlework> not in the propositional one 16:14 < ldlework> I only refer, so that you know what to stick into melbi1 16:14 < ldlework> once you have identified my referents 16:14 < ldlework> the descriptions decay 16:14 < ldlework> they become irrelevant 16:14 < zipcpi> da poi le'elkai lo ka tcadu zo'u: da melbi 16:14 < ldlework> it is only important that the refernce was consumated 16:14 < ctefaho> ugh zipcpi pls 16:15 < zipcpi> exp: da poi le'elkai lo ka tcadu zo'u: da melbi 16:15 < mensi> ([{da KU} VAU]¹) KUhO>} zo'u] [da {CU }]) 16:15 < ldlework> If you think about it 16:15 < ldlework> the citiness isn't important to the content of the message at all 16:15 < zipcpi> That's what le'elkai is for 16:15 < ldlework> whether that thing is beautiful isn't *predicated* on whether it is a city or not 16:15 < ldlework> zipcpi: I'm telling you, language philosophy is telling us it is never important 16:16 < ldlework> if it was important, the predicate "city" would appear in the proposition 16:16 < ldlework> Well it is also important to confused logicians 16:16 < ldlework> Who are trying to udnerstand langauge *only* through logical calclulus 16:16 < ldlework> Wherein everything must be veridical 16:16 < ldlework> and included in the proposition 16:16 < ldlework> (what you're doing) 16:16 < ldlework> (and then making a narticle the n"unimport" the predicate, lol) 16:17 < ldlework> just don't import the description predicate 16:17 < ldlework> and be done 16:17 < ldlework> lo broda cu brode => da brode 16:17 < zipcpi> le'elkai solves that problem. It abstracts {lo ka tcadu} from any requirement of veridicality, and just means "That which is described by this predicate" 16:17 < ldlework> There isn't a problem 16:17 < ldlework> to solve 16:17 < ldlework> don't import the description into the quantified proposition 16:17 < zipcpi> That which is imputed "cityness" for the purposes of description, regardless of whether it is a city 16:17 < ldlework> then you don't need to include it just to say it isn't included 16:18 < zipcpi> But the semantic of "cityness" is relevant somewhere 16:18 < ldlework> to the act of referring 16:18 < zipcpi> Otherwise we'd have no need for gadri at all, and just refer to {da} all the time 16:18 < ldlework> not to the act of determining whether the predicate is true of the referents 16:18 < ldlework> wrong 16:18 < ldlework> gadri has the specifc purpose of allowing us to refer 16:18 < ldlework> da doesn't allow us to refer 16:18 < ldlework> it only allows us to make existential assertions 16:18 < zipcpi> I'm saying the semantic belongs somewhere, even if we aren't asserting the predicate 16:18 < ldlework> It doesn't 16:19 < ldlework> what is the goal in expanding the utterance to a lgoical form? 16:19 < ldlework> for which including description predicates serves? 16:19 < ldlework> To prove whether my referent is true of the description? or to prove the referent is true of the verb? 16:19 < ldlework> seriously 16:19 < ldlework> consider that 16:19 < zipcpi> da = "something/ some object", da poi broda then restricts it to those that broda 16:19 < ldlework> don't 16:19 < ldlework> just answer the utilitarian question 16:19 < ldlework> what is your end goal to expand anything 16:19 < zipcpi> But we don't want {le tcadu} or even {lo tcadu} to be {da poi tcadu} 16:19 < ldlework> what? 16:20 < ldlework> its na'i 16:20 < ldlework> reference is a separate act than any of the illocutionary ones 16:20 < ldlework> John is beautiful. 16:20 < ldlework> The city is beautiful 16:20 < ctefaho> ldlework: well unless I am completely confused, the predicate is obviously about whether the argument fits the predicate rather than the argument fitting its referent? 16:21 < ldlework> No information that helps us determine the truth-bearingness of the statement, whether the referent behind the John-curtain is beaitufil, by knowing the referents name 16:21 < ldlework> No imfortaion that helps us determine the truth-bearningness of the statement, whether the referent behind the City-curtain is beautiful, by knowing that the thing is a city. 16:21 < zipcpi> You are, by your descriptions, qualifying, restricting what the sumti could refer to, even if you aren't necessarily describing it veridically 16:21 < ldlework> Yes but my statement doesn't say 16:21 < ldlework> "Only cities are beautiful" 16:21 < ldlework> which is what logicans do 16:21 < ldlework> when they say 16:22 < zipcpi> {po'o} is not implied 16:22 < ldlework> use the transitive property to relate the predicate city and the predicate beautiful 16:22 < ctefaho> {lo prenu cu jmaji jecu jukpa jecu citka} - The truth of them gathering and cooking and eating is obviously separate from whether the "lo prenu" semantically fits "prenu"? 16:22 < ldlework> "propositional logic" 16:22 < ldlework> it deals with the entailment of -predicates- 16:22 < ldlework> If one predicate entails the other 16:22 < ldlework> And x is one predicate 16:22 < ldlework> x must be the other predicate 16:23 < ldlework> this has nothing to do with language, beyond what is determined -in the proposition-! 16:23 < ldlework> IE, the referents predicated through some predicate 16:23 < ldlework> When you quantify something in logic you are not saying 16:23 < ldlework> I'm referring to some dogs 16:23 < ldlework> Some dogs are beautiful 16:23 < ldlework> that's now how it works 16:23 < ldlework> yo uare trying to establish relationships betweeen -predicates- 16:24 < ldlework> You are making qualitative rules for what something is, if it is also something else 16:24 < ldlework> When I say 16:24 < ldlework> The city is beautiful 16:24 < ldlework> I don't mean to entail either of those predicates with each other 16:24 < ldlework> in any way 16:24 < ldlework> they are not related in any form 16:24 < ldlework> nothing about citieness has anything to do with beautifulness, as predicates and that's the only thing fopl will ever address 16:24 < zipcpi> That which I describe as a "city" is beautiful 16:24 < ldlework> This is why reference is a seprarte act 16:24 < ldlework> yes 16:25 < ldlework> so in your logical expansion 16:25 < ldlework> there is one predicate 16:25 < ldlework> and one referent 16:25 < ldlework> nothing in fopl has anything to say about 16:25 < ldlework> things which you describe as cities 16:25 < ctefaho> I think I agree with you as far as I understand. How is this controversial? 16:25 < ldlework> and their logical entailment to beauty 16:25 < ldlework> ctefaho: because logicans, who work on language, *must* as a measure of their own applicability, conflate this 16:25 < ctefaho> .u'i 16:26 < ldlework> fopl is a calculus of predicates 16:26 < zipcpi> The {ka} predicate is merely to show how the speaker describes it, which, under {le'elkai}, restricts {da} to that referent, and not a mere "something that exists" 16:26 < ctefaho> well I am happily tearing shit out in STT 16:26 < ldlework> zipcpi: the only reason you ever do fopl is to prove a relation between two predicates 16:26 < ctefaho> I just want a language to think in and this...logical weirdness, I am allergic to it 16:26 < ldlework> what is the relationship between is-a-city and the predicate is-pretty 16:26 < ldlework> that is established in the sentence, "that city is pretty" 16:27 < zipcpi> No, now the relationship is is-the-referent-described-by-the-speaker-by-citiness 16:27 < ldlework> what? 16:27 < zipcpi> da poi le'elkai lo ka tcadu 16:27 < ldlework> stop inventing things to explain your confusion 16:27 < ctefaho> but I am going to sleep now, will happily discuss more tomorrow, you and zipcpi can probably have a constructive debate meanwhile 16:28 < ldlework> ctefaho: co'o 16:28 < ctefaho> so co'o rodo 16:28 < ldlework> zipcpi: if we're going to expand lojban to statements that can be formally analyzed you don't get to embed speech conventions into your analysis 16:28 < ldlework> all you have is DA PA and BRODA 16:28 < ldlework> that's all fopl has 16:29 < ldlework> fopl can't tell you anything about the statement if you use anything beyond what it itself haves 16:29 < ldlework> Given this constraint, you will invent pre-xorlo lojban 16:29 * mumble jorne co'u fa la'o fa ctefaho fa 16:29 < ldlework> Because that's all mere fopl can give us, as far a linguistics go 16:29 < ldlework> you -can't really use fopl as language- 16:29 < ldlework> its a lie 16:30 < ldlework> What you can do is invent a language, that easily resolves down to fopl assertions, and analyze those 16:30 < ldlework> I'm telling you, that in order to realize _what is really being asserted_ in language 16:30 < zipcpi> I am just not convinced by your assertion that the description adds nothing to the predication 16:30 < ldlework> you must perform reference as a separate act 16:31 < ldlework> zipcpi: in certain forms it does 16:31 < ldlework> IE when you're talking about predicates themselves 16:31 < ldlework> IE with indefinite descriptions 16:31 < ldlework> BAM now we know why indefintie descriptions have never been formalized logically 16:31 < ldlework> outside of strange exceptional rules of scope that no one agrees on 16:32 < ldlework> Because indefinite descriptions refer to predicates not truth-bearing referents 16:32 < ldlework> But I'm getting ahead of myself since I don't want to introduce this formalization of indefinite descriptions yet 16:32 < ldlework> zipcpi: also, definite statements formed existentially also refer to predicates 16:32 < ldlework> All men are mortal 16:32 < ldlework> Nothing is referred. 16:33 < ldlework> The man who killed Smith, is insane. 16:33 < ldlework> Nothing is referred! 16:33 < ldlework> Except predicates. 16:33 < ldlework> In which we create logical entailments between predicates, but not referents. 16:33 < ldlework> The goal of definite reference is to apply a predicate to that which is not a predicate. 16:33 < ldlework> IE a definite reference. 16:34 < ldlework> If you think deeply about this, you will suddenly understand generalizations. 16:34 < ldlework> And how "I like cats" only has one referent. 16:34 < zipcpi> And what's wrong with something meaning "is the definite referent", as opposed to some other city that you were not referring to with "the city"? 16:34 < ldlework> what? 16:36 < ldlework> zipcpi: because you never want the opposite 16:36 < ldlework> Or rather, at least when trying to apply logical analytics to speech. 16:36 < ldlework> Otherwise you run into the same problem everyone else has in attempting to do just that. 16:37 < zipcpi> But you are qualifying it. It may be impractical to qualify it to mean one/some particular cities, but you mean to do that with the description and the context 16:37 < ldlework> You say I'm qualifying, and then you say something I don't understand. 16:38 < zipcpi> "The city" could mean, say, Houston. If it is, then you are referring to Houston, not another city, or a cat 16:38 < ldlework> Yeah, exactly 16:38 < ldlework> So the only thing involved with the assertion is 16:39 < ldlework> whatever is behind the curtain of my description 16:39 < ldlework> and whatever predicate I asserted about it 16:39 < ldlework> therefore if you wanted to logically analyze my speech 16:39 < ldlework> the predicate contained in my description -has no bearing on that truth value- 16:39 < zipcpi> Neither does {lo se le'elkai} 16:39 < ldlework> zipcpi: Show me the specific fopl calculus that has a similar construct 16:40 < ldlework> zipcpi: Then explain to me how including the description, even with your superflous article, contributes to logical analytics of the assertion? 16:40 < zipcpi> It's exactly why {le'elkai} is "magic", it's meant to restrict {da} to be whatever you may be talking about at the time 16:40 < ldlework> No 16:40 < ldlework> It does *nothing* 16:40 < ldlework> literally nothing 16:40 < ldlework> except to add noise to that which you're trying to analyze 16:41 < ldlework> My goal is to understand the truth-bearingness of some expression 16:41 < ldlework> But now I have it interprete /then disregard/ this clause 16:41 < ldlework> just because.... you invented it? 16:41 < ldlework> Certainly not because it -contributes to the goal of analytics- 16:41 < zipcpi> The description means something. It's not the main assertion, but the things you typically describe as {le tcadu} would be quite different from the things you'd describe as {le mlatu} 16:41 < ldlework> okay you repeated your position 16:41 < ldlework> which means I have had no effect 16:41 < ldlework> the repetition of your position doesn't do anything to convince me 16:42 < ldlework> so I'd say we're at an impasse 16:42 < ldlework> which is okay, and not to the detriment of anyone's character 16:42 < ldlework> so let's just decide reasonable people can disagree on this point, as they have in the past 16:42 < zipcpi> I could say the same thing about you; you keep asserting that descriptions add nothing 16:42 < ldlework> I ask you to meditate on what the goal, the actual attempted accomplishment is deisred when applying logical analytics to some statement. 16:42 < ldlework> zipcpi: I also give tons of texts to supplant them 16:42 < ldlework> that are not merely tautologies 16:43 < ldlework> in fact I've gone to the extents of various directions in attempts to justify that which I assert 16:43 < ldlework> Giving reasoning about the purpose of the thing you wish to conflate 16:43 < ldlework> Consider what fopl is *for* 16:44 < ldlework> And consider what people who actually study language (not me, but those who I invoke on my behalf) are saying about how descriptions relate to logic 16:44 < ldlework> If you can answer the question what fopl is for 16:44 < ldlework> You will realize that including referring descriptions doesn't contribute to that goal 16:45 < ldlework> making your article and your justification for it superflous 16:45 < ldlework> in that it doesn't address the underlying goal that you're trying to serve 16:45 < ldlework> you're trying to expand something in the language to form which is perceptable to a certain kind of analytics 16:45 < ldlework> but you ignore the goal of that analytics, skip right over it 16:46 < ldlework> Also its very abstract so its also possible that its just hard to overcome how closely related but not isomorphic fopl and language are. 16:46 < ldlework> When we apply fopl to speech, we only care about the propositional content. 16:46 < ldlework> Since that's the only think fopl can analyze 16:46 < ldlework> If we inclde our descriptions into the fopl we get cll lojban 16:46 < ldlework> if we don't, we get post-reference propositions like I sugggest 16:47 < ldlework> If we listen to you 16:47 < ldlework> we get fopl statements that include terms that don't actually add any information that fopl cares about so you end up "ignoring it" 16:47 < ldlework> Ie, keep em in, but then tell the analyzer to ignore it 16:48 < zipcpi> But I do care when you refer to Houston and not anything else 16:48 < ldlework> I wish it could go *ping* in your mind, that it is only the da, which is true of the central propositional predicate 16:48 < ldlework> or not 16:48 < ldlework> and that's the only thing fopl cares about 16:48 < ldlework> We use single variables, like da, and multiple predicates, to establish logical entailment between those predicates that the da shares. 16:48 < ldlework> When we only do fopl 16:49 < ldlework> and we establish entailment relationships among predicates, we don't actually say what in the world is true about those predicates beyond "things which are ture about this predicate" 16:49 < ldlework> IE when we say 16:49 < ldlework> A implies B, Ex where A(x) therefore B(x) 16:50 < ldlework> We're not making assertions about what is contained in the x beyond that it is both an A and a B 16:50 < ldlework> in the statement 16:50 < ldlework> The city is beautiful 16:50 < ldlework> We *do not* wish to entail the predicates beauty and city 16:50 < ldlework> We do wish to assert that beautiful(x) 16:50 < ldlework> logically 16:50 < ldlework> and as a SPEECH ACT 16:50 < zipcpi> But we are entailing the predicates beauty and "the thing that I refer to" 16:50 < bigcentaur> coi ro do 16:50 < ldlework> I hope that you know what I mean is beautiful 16:51 < ldlework> zipcpi: no that's false 16:51 < ldlework> there's nothing about beautiful that entails being referred to by me 16:51 < ldlework> and there's nothing about being referred ot be mea that entails being beautiful 16:51 < ldlework> not in the same way that 16:51 < ldlework> man and mortal are entailed together 16:51 < ldlework> IE, the kind of entail fopl cares about 16:52 < zipcpi> There is an x, where x is magic-referring-predicate *and* x is beautiful 16:52 < ldlework> That's -----speech-------- 16:52 < ldlework> Not ----------logic----------- 16:52 < ldlework> To get what you want 16:52 < ldlework> that magic referring predicate _always entails beauty_ 16:53 < ldlework> because fopl discerns entailment _between predicates_ 16:53 < ldlework> it does not discover invisible referents for which two predicates are true 16:53 < ldlework> When we say 16:53 < ldlework> all men are mortal 16:53 < ldlework> x is a man 16:53 < ldlework> therefore x is mortal 16:53 < ldlework> we are not making conclusions about X! 16:53 < zipcpi> For all x where x is a man, x is mortal 16:53 < ldlework> we are making conclusions about the inherent entailment between two predicates! 16:53 < ldlework> So if what you're saying is true 16:54 < ldlework> we must entail some logical entailment of beautify by the nature of being referred to by me 16:54 < ldlework> can't you see how innane and completely wrong this is?! 16:54 < zipcpi> No, it only means for the x which you refer to 16:54 < ldlework> no 16:54 < zipcpi> x is beautiful 16:54 < ldlework> that's not what fopl does 16:54 < ldlework> to get what you want 16:54 < ldlework> it would just be 16:54 < ldlework> x is beautiful 16:54 < zipcpi> But which x? 16:54 < ldlework> It doesn't matter. 16:54 < ldlework> Because my goal in fopl 16:55 < ldlework> is not to identify what x is 16:55 < ldlework> but to entail predicates! 16:55 < ldlework> I can introduce other predicates 16:55 < hilmi-can> i really feel dumb reading that. ill try my best to catch up. 16:55 < ldlework> to entail relationships between the predicates 16:55 < ldlework> such as 16:55 < ldlework> if x is beautiful 16:55 < ldlework> x must be pretty 16:55 < ldlework> or whatever entailment you want 16:55 < ldlework> this doesn't tell us 16:55 < ldlework> that 16:55 < zipcpi> hilmi-can: Sorry, this is some rather abstract/obtuse stuff; if you have a question, please ask in #ckule 16:55 < ldlework> "a pretty" is the thing to which we refer in fopl 16:55 < ldlework> there is no reference in fopl 16:56 < ldlework> It only establishes a relationship between the predicate beautiful and pretty 16:56 < ldlework> There is no reference in fopl 16:56 < ldlework> So 16:56 < ldlework> In your attempt to expand some lojban out 16:56 < ldlework> to a form that could be analyzed 16:56 < ldlework> you are creating the WRONG PROPOSITION 16:56 < ldlework> IE, the proposition that the speaker never claimed 16:56 < ldlework> by injecting the referring description into the proposition that will eventually by analyzed 16:57 < ldlework> All cities are beautiful 16:57 < ldlework> ^THIS is an exitential statement 16:57 < ldlework> Even though it uses a referring description the description refers to a predicate! 16:57 < ldlework> not an actual city 16:57 < zipcpi> When we say da poi remna cu mroka'e; all we mean is there is an x, which is a remna, and which also is mroka'e 16:57 < ldlework> _here_ an entailment is established 16:57 < ldlework> Right 16:57 < zipcpi> There is nothing about {nibli} or whether remnaness entails mroka'eness 16:57 < ldlework> poi dodges the fopl 16:58 < ldlework> because poi has *nothing to do with analyzing the truth-bearingness of a statement* 16:58 < ldlework> poi is how you encode your original descriptions without changing the truth value 16:58 < ldlework> this is precisely why I say *all descriptions expand to the same poi ---if they must---* 16:59 < ldlework> Because to analyize the truth bearingness, to apply the principles of fopl /I don't need to know your descriptions/ 16:59 < ldlework> I only need to know what you eventually referred to 16:59 < ldlework> Because it is _that thing_ which will be true of the main predicate or not 16:59 < ldlework> It doesn't matter how you got me to identify _that thing_ as the argument of the proposition 16:59 < ldlework> the proposition is not true or false of your descriptions, it is true or false of your referents 17:00 < ldlework> the description can be additionally true or false of the referent 17:00 < ldlework> but neither in logic would you care 17:00 < ldlework> or in speech 17:00 < ldlework> logicians don't care because 17:00 < ldlework> when someone says 17:00 < ldlework> The city is pretty 17:00 < dmiles_afk> most of the time we are creating assertions in logic .. i feel mostly we are defining or adding to the language for further description 17:00 < ldlework> no logican wants to formalize whether there is some entailment between those specific predicates 17:00 < ldlework> because only an insane person would think so 17:00 < ldlework> language philosophers don't care 17:00 < ldlework> Because reference is a totally separate act fromp redication. 17:01 < dmiles_afk> yet i see that often we prettend we are creating new entailments (ok we are doing that sure) but we are more involved in defining a language 17:01 < ldlework> dmiles_afk: I agree, when we make existential statements like "The man who murdered smith is insane" 17:01 < ldlework> Or "All men are mortal" 17:01 < ldlework> But not "The apple is poisonous" 17:02 < ldlework> We do not expand the semantics of the langauge when we describe external referents 17:02 < ldlework> We expand the language when we predicate predicates themselves. 17:02 * dmiles_afk thinks for a moment 17:02 < dmiles_afk> "All men are mortal" we are defining mortal based on manhood? 17:03 < dmiles_afk> hehe .. i know the other way arround .. but i thik both ways 17:03 < ldlework> We are defining an entailment between manhood based on mortality. 17:03 < ldlework> IE, the predicates is-a-man is intrinsically related to is-a-mortal in a way that is logically entailed 17:04 < ldlework> If we happen to be referring, not to the predicate is-a-man, but to _a specific man_ we create no such entailment. 17:04 < ldlework> We don't expand the semantics of "mortality" 17:04 < ldlework> And we don't even comment on the predicate is-a-man 17:05 < ldlework> Outside, perhaps, that if this man is indeed a man, that humanity may afford mortality. 17:05 < dmiles_afk> i am glad you are pointing out "this is authorial form" not logic 17:05 < ldlework> But it no longer entails it. 17:05 < ldlework> You could argue that when you make a definite reference 17:05 < ldlework> you create a nonce-predicate 17:06 < ldlework> to-be-that-thing-to-which-I-refer 17:06 < ldlework> but since it is nonce 17:06 < ldlework> you will never observe anything meaningful 17:06 < ldlework> you will never arrive at a conclusion by way of logical analytics that you couldn't arrive at without that predicate 17:06 < ldlework> it has no inherent characeristics by which it could be *inherently* related to some other predicates 17:06 < ldlework> IE, they have no logical entailment by definition 17:07 < ldlework> because that nonce verb has no meaningful definition by which you would be able to establish any entailments with any other predicates 17:07 < ldlework> except perhaps synonmous tautologies 17:08 < ldlework> anything referred to by me is a referrent 17:08 < ldlework> x is referred to by me, therefore x is a referent 17:08 < ldlework> you're never going to be able to entail is-referred-by-me, in the way the fopl was designed to, to the other predicate "beauty" 17:08 < dmiles_afk> there is this set of variables and things that i secretly thinnk have to be comoinged in 17:08 < ldlework> Maybe reference is a concept that fopl never intended t ocare about 17:09 < dmiles_afk> for example "thereExists a me in which refers to an x" 17:09 < ldlework> and *does not become defective* by not addressing it! 17:09 < ldlework> Because the goal of fopl is not something that depends on addressing reference 17:09 < zipcpi> I don't see what's the difference in form between "all humans are mortal" and "all the cities are beautiful", except that one directly refers to things that are human, while the other points to a particular set that you describe as "cities" 17:09 < ldlework> No 17:09 < ldlework> you just cheated 17:09 < dmiles_afk> "there exists an assertion maker who assert all humans are mortal" 17:10 < zipcpi> Then it's just {ro da melbi}? 17:10 < bigcentaur> in terms of aristotelean logic, they're of the same form, "All S is P" 17:10 < ldlework> "All men are mortal" is different than "All the cities are beautiful" and "All the men are mortal" 17:10 < ldlework> bigcentaur: right, as all logic really is 17:10 < ldlework> because logic, specifically predicate logic, is the logic of predicate entailment 17:10 < ldlework> not referents 17:10 < bigcentaur> how tautological 17:10 < bigcentaur> :P 17:10 < zipcpi> I'm just trying a different approach, to show that there is a difference between qualified {da} and unqualified {da} 17:11 < ldlework> zipcpi: there isn't 17:11 < ldlework> the real difference is referring descriptions 17:11 < ldlework> and not referring descriptions 17:11 < ldlework> IE, when we refer to predicates as to establish new entalments between them 17:11 < ldlework> If you understand the existential variable in fopl as not addressing reference at all 17:12 < ldlework> therefore giving rise to all the problems of applying fopl to indefinite reference 17:12 < zipcpi> You just keep asserting that there isn't, but that doesn't explain why all humans is {ro da poi remna}, while all the cities is not {ro da poi *something*} 17:12 < ldlework> and then you say okay 17:12 < ldlework> lojban is not quite fopl 17:12 < ldlework> and its language trying to be fopl 17:12 < ldlework> then *two* fundamental variable types is exceedingly intuitive 17:12 < ldlework> when I predicate upon da based referents, I mean to assert, plainly, predicates upon non-predicate referents 17:13 < dmiles_afk> *two* fundamental variable types are egain not universal vs existenriall qunatified... they are... ? 17:13 < ldlework> when I predicate upon zo'e referents, I mean to assert, plaining, upon predicates themselves 17:13 < ldlework> I mean to establish an entailment between two predicates, rather than a single predicate and a single referent 17:13 < ldlework> this perfectl explains 17:13 < ldlework> That man is mortal 17:13 < ldlework> and 17:13 < ldlework> Men are mortal 17:13 < ldlework> one creates an entailment one does not 17:13 < ldlework> it makes everything so easy to explain and understand 17:13 < dmiles_afk> (this is sometimes called a-box vs t-box?) 17:13 < zipcpi> There is an entailment; just not in the "A causes B" sense 17:14 < zipcpi> More as a "Something is A and B" sense 17:14 < ldlework> zipcpi: not in any sense that anyone but you uses 17:14 < ldlework> seriously 17:14 < zipcpi> That is also how "All humans are mortal" works. All that are human, are also mortal 17:14 < ldlework> Nope 17:14 < dmiles_afk> (secretly i was wondering if i was goign to discover that a-box vs t-box infernce methods need to be differnt) 17:14 < ldlework> It creates a relationship between two predicates. 17:15 < zipcpi> Says nothing about whether mortality is "in" humanity... seriously, this is why we have gg-lo'i in the first place 17:15 < zipcpi> Or the alternatives {kairpau} and {kai'a} 17:15 < ldlework> I love how your answer to everything isn't a substantial build up of justification 17:15 < ldlework> its just "that's why we have magic article that doesn't advance our ability to apply analytics to natural speech" 17:15 < ldlework> y'know, one of the core goals of the experiment 17:16 < ldlework> I seriously just built up a self-consistent framework that obliges all quirks 17:16 < ldlework> Instead of stepping out of your position 17:16 < ldlework> and just exploring the consequence of what I'm saying 17:16 < ldlework> seeing if the shoe fits 17:16 < zipcpi> No, you want to make {roda} mean {ro da poi co'e} 17:16 < ldlework> its just "no, magic article" 17:16 < ldlework> Its like, in this channel, I never expect everyone to just agree 17:17 < ldlework> but to atleast conceptualize what I'm talking about, explore it on its own foundations, then throw it away 17:17 < ldlework> Rather than, well yes, the big thing you just built up isn't isomorphic with hwere I was trying to go, so you're just making shit up and not talking about anything that's important to lojban. 17:17 < dmiles_afk> i serrously am amzed about this channel.. i been on irc for a long time as sad it's not until today i knew of it's existence 17:17 < bigcentaur> hi dmiles_afk 17:17 < bigcentaur> :) 17:18 < dmiles_afk> i started being interesting lojban arround 1998.. i was afraid it was going to just be interligua (thoughi'd been happy if that was all it was) 17:18 < zipcpi> So "All that I refer to as "cities", is beautiful" says nothing about beauty being "because" I refer to it 17:18 < ldlework> zipcpi: if we conceptualize lojbanic logical expansions in the way I've expressed, that linguists suggest we do (obviate referential predication), 1.you don't need your magic article, 2. nothing about lojban's ability to be logically expanded is lost 2. the ability to apply analytics to lojban becomes easier 17:19 < zipcpi> All it does is restrict what {roda} refers to 17:19 < ldlework> zipcpi: All Men are mortal is not an enumerating referential description in any writing I've read I'm sorry 17:19 < bigcentaur> dmile_afk: nice! i've only been around since oct '14 ha 17:19 < ldlework> Moreso, it falls in to ALL the writing I've read describing some english definite forms as creating indefinite descriptions 17:19 < ldlework> you know why they observe that? 17:19 < bigcentaur> dmiles_afk: nice! i've only been around since oct '14 ha 17:19 < dmiles_afk> but lojban is about building a dialog system in which a computational inner voice can think in? 17:20 < ldlework> Not because its going to make lojban easier to define 17:20 < ldlework> because of the -logical entailment of such statements- 17:20 < ldlework> If you simply create this man je that man je that man je that man je that man 17:20 < ldlework> is mortal 17:20 < zipcpi> Nor does "All men are mortal" say that they are mortal *because* they are men 17:20 < ldlework> you do not create an entailment between humanity and mortality 17:20 < ldlework> you simply say 17:21 < ldlework> each person who happens to be man, also happens to be mortal 17:21 < ldlework> that the predicate morality, happens to be true of all these referents for which the predicate humanity is also true 17:21 < zipcpi> So each thing that happens to be what I refer to as a city, also happens to be beautiful. What's the difference between the logical forms? 17:21 < ldlework> but there might be a man out there we didn't count, or doesn't exist yet, that becomes mortal 17:21 < ldlework> entailment! 17:22 < dmiles_afk> when i worked on SUMO they did not let me add entails/2 17:22 < ldlework> one *never refers to any man but the predicate is-a-man* then establishes a new truth about that predicate, not its referents, as it pertains to the other predicate mortality 17:22 < ldlework> the other refers and creates no entailment! 17:22 < zipcpi> It creates an entailment between the things you happen to be referring to at the time 17:22 < ldlework> ro da poi remna cu ka'e morsi, merely says, everything that happens to be a man, happens to be mortal 17:22 < ldlework> it does not say that, if you are a man, you must be mortal 17:22 < ldlework> because all you've done is 17:22 < dmiles_afk> when you say .. entailment you mean a logical prol,o neck like ":-" ? 17:23 < ldlework> gone to each man and say 17:23 < ldlework> you are human 17:23 < ldlework> you are mortal 17:23 < ldlework> that's it 17:23 < ldlework> such as 17:23 < ldlework> you are tall 17:23 < ldlework> you are hungry 17:23 < ldlework> two things completely not entailed 17:23 < ldlework> you do that for each man 17:23 < ldlework> you create no entailment 17:23 < ldlework> since there may come a day when a tall satiated man is born 17:23 < ldlework> if it were _entailed_ 17:23 < zipcpi> My point is that "All the cities are beautiful" are the same, except that you may only be referring to a subset of cities 17:23 < ldlework> no such out come is possible, ever 17:24 < ldlework> that's because you're using the indefinite article 17:24 < ldlework> All _the_ cities 17:24 < ldlework> which is different than 17:24 < ldlework> All cities are beautiful 17:24 < ldlework> it just is, that's how the english works 17:24 < ldlework> And I suggest we solve the problem with not pragmatics 17:24 < ldlework> a very messy and impossible unofrmal study of this problem 17:24 < ldlework> and define definite and indefinite reference as the foundataion from which all our descriptions are derrived! 17:24 < ldlework> you either create assertions on referents 17:24 < zipcpi> The point is that {da} is still being restricted. If you say it isn't you have to say {ro da melbi} makes sense there 17:25 < ldlework> or you create entailment relationshisp 17:25 < ldlework> no 17:25 < ldlework> there is no da 17:25 < ldlework> and there is nothing to quantify 17:25 < ldlework> there is zo'e 17:25 < zipcpi> Outer quantifiers quantify 17:25 < ldlework> that in this case reprersents a predicate 17:25 < zipcpi> It would be {ro le tcadu} 17:25 < ldlework> zipcpi: I believe sometimes you just say things 17:25 < ldlework> words, and phrases that you've heard at similar times in simiar discussions 17:25 < ldlework> without connecting them to what you're saying or what you're replying 17:25 < zipcpi> {ro le tcadu} != {le ro tcadu} 17:26 < ldlework> Yes, and tell us how 17:26 < ldlework> And then 17:26 < ldlework> How it is directly applicable to the point you used this observation to reply to 17:26 < zipcpi> ro le tcadu: All the cities; le ro tcadu: the cities, which number all 17:26 < ldlework> Nope now you're making semantic interpretations 17:26 < ldlework> You're not going to lojban's definition 17:26 < ldlework> And expanding them existentially 17:26 < zipcpi> mu le tcadu: Five of the cities; le mu tcadu: the cities, which numbers five / the five cities 17:26 < ldlework> So you're just making shit up here 17:26 < ldlework> Based on your conceptualization of how reference works in lojban 17:27 < zipcpi> This is how outer quantifiers work vs inner quantifiers 17:27 < zipcpi> If you're just gonna swear at me then there is no point in continuing 17:27 < ldlework> "the 5 of what cities" 17:27 < ldlework> or do you argue there is no default inner quantifier 17:27 < ldlework> and that 17:27 < ldlework> mu le tcadu, actually means something substantially isolate from le mu tcadu 17:28 < ldlework> other than just rearranging your english and declaring "there!" 17:28 < ldlework> expand them 17:28 < ldlework> to fopl 17:28 < ldlework> as all lojban should be able to do 17:28 < ldlework> In neither case are you making an indefinite description like you hope 17:28 < ldlework> You are iterating 5 cities of some larger set, potentially all of them 17:28 < zipcpi> mu da poi le'elkai lo ka tcadu / da poi le'elkai lo ka mumei je lo ka tcadu 17:28 < ldlework> or you're iterating 5 cities 17:28 < ldlework> in both case neither creates an entailment 17:29 < ldlework> such as 17:29 < ldlework> All cities are beautiful 17:29 < ldlework> zipcpi: by expanding to da 17:29 < ldlework> you only get to use da 17:29 < ldlework> you don't get to use your magic particles, sorry 17:29 < ldlework> I can't apply logic to your magic particles 17:29 < ldlework> and therfore cannot conclude anything meaningful from your expansions 17:29 < zipcpi> mu da poi ke'a mi skicu ... whatever 17:30 < ldlework> when you can expand mu lo broda and lo mu broda in a way that doesn't simply enumerate some actual broda and claim something about those broda, but actually establish a new entailment about what it means to broda 17:30 < ldlework> then you will have my ear 17:30 < ldlework> but I *know* you can't do that 17:30 < ldlework> In the same way you can't humm a number. 17:30 < ldlework> You've missed the point. 17:31 < dmiles_afk> hrrm google translate does nto tranlate lojban? 17:31 < zipcpi> Now you want to overturn outer and inner quantifiers as well. I can't follow you there 17:31 < ldlework> zipcpi: don't say things that are obviously false 17:31 < zipcpi> Which one, the first one or the second one 17:32 < ldlework> That I want to overturn outer an inner quantifiers. I'm showing you, that neither inner nor outer quantifiers do what you want them to do 17:32 < ldlework> IE, create entailment between two predicates 17:32 < ldlework> No amount of quantification can create the kind of entailment you'd like. Ever. 17:33 < ldlework> You're either saying something about each individual broda, which may be true but says nothing about brodaness 17:33 < ldlework> Or 17:33 < dmiles_afk> is zipcpi using quantification to means something synonmous with referants? 17:33 < ldlework> You're saying something about some subset of some larger subset (potentially all broda) 17:33 < ldlework> But never says anything about brodaness 17:33 < zipcpi> You are saying something about each individual thing that you refer to as {le tcadu} 17:33 < ldlework> yes 17:33 < ldlework> exactly 17:33 < ldlework> No matter inner or outer quantification 17:33 < ldlework> neither creates enatialment _as defined_ 17:34 < ldlework> not your characterization of me trying to destroy lojban or some bullshit 17:34 < ldlework> or misunderstanding it so badly I wish to redefine it 17:34 < zipcpi> So what's the difference between saying something about each individual thing that {remna}, and saying something about each individual thing that {me le tcadu}, however you want to define {le}? 17:34 < ldlework> zipcpi: Either to be human is to be mortal. 17:34 < ldlework> Or merely, each man that exists today is mortal. 17:34 < ldlework> You get one. 17:34 < ldlework> Pick. 17:35 < zipcpi> The second, except nothing about tense is specified 17:35 < ldlework> That doesn't matter 17:35 < ldlework> There could be a man right now that is not mortal that you didn't count 17:35 < ldlework> because you did not entail that maness is mortal 17:35 < ldlework> you merely went to each one and told them 17:35 < ldlework> you are tall 17:35 < ldlework> you are hungry 17:35 < ldlework> no where did you establish that tallness entails hunger 17:35 < ldlework> and never will you with this strategy 17:36 < ldlework> I have to assume you don't mean that man-ness entails mortal-ness because that's not what you said 17:36 < zipcpi> And I don't. That is dealt with {nibli} or {kairpau} or whatever 17:36 < ldlework> You implied it statistically which is different. 17:36 < ldlework> zipcpi: right the only way it is nibli 17:36 < ldlework> is if we entail 17:36 < ldlework> IE we say 17:36 < ldlework> All men are mortal. 17:36 < ldlework> Where "All men" 17:37 < ldlework> isn't "ro da poi remna" 17:37 < ldlework> but some construct lojban doesn't support yet 17:37 < ldlework> IE indefinite descriptions 17:37 < ldlework> that allow us to express entailment between predicates, directly. 17:38 < ldlework> I want a lojban that affords me both. 17:38 < ldlework> In a way that both are core features of my utilization of that language 17:38 < ldlework> not wrapped up in some external preidcate 17:38 < dmiles_afk> by practicing logban practically.. was it discovered there were parts of logic that had been seemingly neglated to be taught in school but were probably very important? 17:38 < ldlework> favoring one description type over the other 17:39 < zipcpi> dmiles: I'm not sure I'm the right person to ask 17:39 < dmiles_afk> (i am wondering if those are practicalities that ldlework and zipcpi are hashing out) 17:39 < zipcpi> Nah 17:39 < ldlework> dmiles_afk: my overall lesson is that lojban conflated somethings that exist in fopl as tools of language, and fopl and language are so nearly related that we will never be able to illuminate the important distinctions 17:39 < ldlework> such as reference and existential variables 17:39 < ldlework> Where existential variables exist to allow us to relate predicates (not refer) 17:40 < ldlework> And references allow us to apply predicates to actual things, not create propositional entailments (to actually refer) 17:41 < dmiles_afk> i wonder if the logicans that is using a machine like CYC is actualyl using more than one variable type (and he didnt know it) 17:41 < ldlework> Except for in the cases where the form of the statement is existential itself, where-in other actual language philosphers say in english the definite form is abused to create indefinite descriptions sometimes, where-in I argue, being lojban we should just distingusih it as explicitly as possible, at the level where it is concerned 17:42 < dmiles_afk> ok your answer wass yes 17:42 < dmiles_afk> oops you meant the english speaker 17:42 < ldlework> When I say "My friend is human. Some philosophers say to be a man is to be mortal. I conclude my friend is mortal. 17:42 < dmiles_afk> i think logicains do tthose same evils in fopl sometimes 17:42 < ldlework> There are several levels going on 17:43 < ldlework> There is the logical structure of my actual utterance, IE, the propositions I used to formluate my speech. 17:43 < ldlework> There is the illocutionary acts performed by that which is not propositional content in my speech. 17:43 < dmiles_afk> same evils = definite form vs indefinate forms .. even in fopl 17:43 < ldlework> Then there is the logical structure of the claims inside the propositional content of my speech. 17:43 < ldlework> Its like 17:44 < ldlework> We just don't care to ever reveal that perhaps, you can't encompass language with the facilities of fopl. 17:44 < ldlework> IE, you can't take what consititutes fopl, give each thing a new "language-y meaning/purpose" and then call this new interpretation of fopl, "a logical lanaguage" 17:45 < ldlework> That instead, language is far more complex than fopl, that you must first create a language, and consider the results from the pearch an of informed logical history 17:45 < ldlework> That abusing existential variables, which allow us to denote that which preidcates share, as the same thing as a referring description, is so inherent to our understanding is ... well I didn't meant to characterize that observation as something, just to make the observation 17:45 < ldlework> it exists. 17:46 < ldlework> But its wrong 17:46 < dmiles_afk> (secretly i think logicans are *dialoging* with folp .. sometimes trying to redfine its very nature.. yet it can happen something by accident) 17:46 < ldlework> And leads many people to say "You can't say that, that's false" 17:46 < ldlework> Or similiar strangeness. 17:46 < ldlework> Of varying forms. 17:46 < dmiles_afk> (are *dialoging* with folp implemntaiton designed by the greeks) 17:47 < ldlework> Yeah that's just not possible outside of saying speech in the indefinite. 17:47 < ldlework> But that's all that logicians are interested in. 17:47 < ldlework> So it makes perfect sense to me. 17:49 < dmiles_afk> defiantley this is really opening up my mind :) 17:49 < latro`a> unfortunately I've seen a lot of places where we abuse existential variables and in the process *lose* understanding, especially in mathematics 17:50 < latro`a> and not just in mathematics where logic is the focus; real analysis is the main place I have in mind 17:50 < dmiles_afk> when a logician "authors" a logic statemnt "a(X) => b(X) & c(X,Y) & d(Y). " 17:50 < dmiles_afk> they are creating "if a(X) was proven to be true, then this d(Y) is true at least when i have framed (and logically (and can legally) produce) the other things like b(X) where c(X,Y) are what i meant to relate them" 17:51 < dmiles_afk> "futher i would like that a(X) implies b(X) .. but only if you havent already refuted c(X,Y) and d(Y)" 17:51 < latro`a> relatedly, shadowing local variables creates lots of confusion outside of relatively trivial contexts 17:52 < dmiles_afk> "i want not b(X) to be entailed by not c(X,_). " 17:52 < latro`a> this is part of the struggle of getting a student to first grasp algebra; "but you just said x=6, how can you now have x=4?" 17:52 < dmiles_afk> "i want not b(X) to be entailed by not c(X,_). but only when a(X) is proven" 17:53 < dmiles_afk> "oops not c(X,Y). is only solvable when we have a legitimate d(Y). " 17:54 < dmiles_afk> "we can skip a legitimate d(Y). though if we for sure have a legitimate a(X)" 17:55 < dmiles_afk> " a(X) (that is any refernt) can only be legitimized if we feel confidant we can legitimatize d(Y). even without the presence of referrents" 17:56 < latro`a> one instance of this problem: in analysis you often want to take an error and split it up into pieces, where you have a way of guaranteeing that the pieces separately can be made arbitrarily small. so sometimes you wind up proving something like "for every epsilon ... the error is less than 3 epsilon", when what you wanted to prove was "for every epsilon ... the error is less than epsilon" 17:57 < dmiles_afk> (in essence i am saying i am sttarting to think that quantification and referancy are differnt even in Fopl !) 17:58 < dmiles_afk> (i have always pretended quantification and referancy were the same in fopl) 17:58 < latro`a> not really, but this issue is part of why building set theory in first order logic is hard 18:00 < dmiles_afk> well this is helping me undersand te changes i mayneed to make in my fopl to prolog code translator 18:02 < ldlework> zipcpi that thing I said about you saying stuff that doesn't connect was bullshit I shouldn't have said that nor do I believe it 18:02 < ldlework> I'm sorry about that. 18:03 < dmiles_afk> perhaps if i added more wrappings on the outside of fopl to add veriabel liek (all ?QUERY (thereExist ?PROP (thereExist ?WORLD (implies (in ?WORLD ?QUERY) ?PROP)))) 18:04 < dmiles_afk> (that is adorning exome extra variables to be present durring normal for transitions) 18:04 < dmiles_afk> (to be present durring normal form transitions) 18:05 < dmiles_afk> for exmaple.. i used to when translating NL to logic.. I *had* to add in the ?SPEAKER ?ADRESSEE 18:06 < dmiles_afk> or else many statment like "Hello friend" was not translateble to fopl 18:07 < ldlework> I don't think fopl statements really refer 9r even speak about the real world or imagined, in that the formalization that includes "there exists" has anything to do with asserting that anything exists except a variable that will allow us to denote predicate relationships 18:07 < dmiles_afk> even "All 5 cities are beautiful" i can only translate when i know the ?SPEAKER 18:08 < ldlework> anything else is abuse 18:08 < dmiles_afk> (logically i agree with you ldlework) 18:09 < dmiles_afk> oh sorry yes i was leaving out the preds.. naughty me 18:11 < ldlework> after all you can take any fopl statements and replace all the predicates with nonce predicates and still do fopl! 18:14 < ldlework> it truly is a calculus for proving certain things about an abstract network of dependency and entailment. we just often use predicates from the real world but that's not a requirement at all 18:15 < dmiles_afk> one new trick ii plan to test is: to convert (love joe sally) ==> (exists ?pred (exists ?arg1 (exists ?arg2 (implies (and (refersToAs ?world ?pred loves) (refersToAs ?world ?arg1 joe) (refersToAs ?world ?arg1 joe) ) (and (imagined-now ?world (?pred ?arg1 ?arg2)))))) 18:15 < ldlework> Lol I can't parse that on my phone 18:16 < dmiles_afk> nrlaly we mihgt assert: (love joe sally) 18:16 < dmiles_afk> normally 18:16 < ldlework> is this using rpn? 18:16 < dmiles_afk> CYC 18:16 < ldlework> dunno what that is 18:17 * dmiles_afk checks ot see if its compatable withj rpm 18:17 < dmiles_afk> rpn 18:17 < ldlework> I just mean where the operator comes first 18:17 < dmiles_afk> oh yes 18:17 < ldlework> ie, is love the predicate? 18:17 < ldlework> oh OK 18:17 < dmiles_afk> right prefixed all operator 18:17 < ldlework> I guessed "reverse polish notation" 18:18 < dmiles_afk> so .. (implies (and (refersToAs ?world ?pred loves) (refersToAs ?world ?arg1 joe) (refersToAs ?world ?arg1 joe) ) (and (imagined-now ?world (?pred ?arg1 ?arg2))))) 18:18 < dmiles_afk> oops i messsed sally there 18:19 < ldlework> heh you should start by building your grammar so we know what you mean by it 18:19 < dmiles_afk> i also imple a precial predicate (i left out per arg places) 18:20 < dmiles_afk> (i left out per arg place there is one implicit arity one predicate!) 18:20 < uuZIT> Holy shit, Idlework, did you just talk for 7 hours straight? o_O 18:22 < dmiles_afk> (loves joe sally) is actualyl 4 statments "it is now poissiklbe for joe to love someone" "sally is lovable" "loves is a predicate" and "joe loves sally" 18:22 < ldlework> it's a difficult subject 18:23 < dmiles_afk> if i use (loves joe sally) in a anteceedant i say that all 4 statemtns have to be true 18:23 < dmiles_afk> this means my CNF must see all 4 litterals 18:24 < dmiles_afk> and conversion to Horn clause entailments must be aware of all 4 18:25 < dmiles_afk> if i use negation (i havfe to durring refutation strategy) not all 4 statemtns are based on the same type of logic 18:25 < dmiles_afk> "it is now possiklbe for joe to love someone" is modal logic 18:25 < ldlework> I 18:25 < ldlework> ii 18:26 < dmiles_afk> "not all 4 statemtns are based on the same type of logic" <- ignore that since that is a side trip :) 18:28 < dmiles_afk> what i am saying i am starting to be aware of is that producing logical entialment (that is so i can answer queries and prove antecedancts) is based on many factors that i wasnt told were part of logic 18:29 < dmiles_afk> that is based on the process of almost having a diolog with an infernce engine 18:29 < dmiles_afk> that starts to become as nasty as english 18:30 < dmiles_afk> only because i forget that is what is going on 18:30 < dmiles_afk> (as the logician) 18:31 < dmiles_afk> as the implemntor of the logical infernce engine .. i almost have to guess at what the logican meant.. even though they think they areusing plain simple fopl 18:32 < dmiles_afk> i have to see them as they add logical statements as interatively builing up a world 18:32 < dmiles_afk> iteratively* 18:33 < dmiles_afk> oops i meant incrementally 18:34 < dmiles_afk> 1/2 way thru the dialog (i dont get a measure of what 1/2 way is) but i know they are gojng to be underspecific... and all the way they thru they are thinking they are fully specicific 18:35 < dmiles_afk> at the end of this.. they will still have been grossly nderspefic to the meanings of mortals and humans 18:35 < ldlework> dmiles_afk: you're building an inference engine?! 18:35 < dmiles_afk> later on they will tell that being mortals means dying one day 18:36 < dmiles_afk> (but before they told us that we have to accept that we have enough to start proving things tfor them) 18:36 < dmiles_afk> ldlework: yes 18:36 < dmiles_afk> i keep trying to revise the impl to make it work! 18:37 < ldlework> dmiles_afk: what is it implemented in? 18:37 < dmiles_afk> Prolog 18:37 < ldlework> feck 18:37 < dmiles_afk> it started out in 1999 with Mark Stickel's PTTP 18:38 < dmiles_afk> which is the fist and only sound and compete reasoniner i have found in 15 years! 18:38 < dmiles_afk> (though CYC almost counts) 18:39 < dmiles_afk> but even to pureness of PTTP chokes on our world 18:39 < dmiles_afk> becasue we cant communicate or logic sometimes 18:39 < dmiles_afk> well that was a lie.. sorry 18:40 < dmiles_afk> i feel like we have been able to oversikmlify our logic 18:40 < dmiles_afk> i dont know who's fault it is 18:41 < dmiles_afk> i am just saying plain old mathmatical logic seem to be hard.. when it should be the only thing not difficult 18:41 < dmiles_afk> i cant even blame people 18:41 < dmiles_afk> (sorry to keep saying "i") 18:41 < ldlework> Probably because most logic, that takes place within us, takes place under our nose so to speak 18:43 < ldlework> IE, it is the exception that we stop to actually think about logic directly. We're otherwise pretty intuitive creatures, from our standpoint anyway. The brain is probably doing all sorts of sophisticated stuff :) 18:43 < dmiles_afk> i didnt want to design the language for my infenrce ingine i so borrowed Cycorps lanuage CycL 18:44 < dmiles_afk> CycL (and maybe logban)( are the only two semicomplete logical languages that i know of 18:45 < dmiles_afk> Cyc is the program in 1999 i was supposed to implment in Prolog 18:45 < ldlework> I think an inference engine would have to do exactly what we do honestly 18:45 < dmiles_afk> i took a 10 year break and now in the last year working on it again 18:45 < ldlework> It would have to separate the reference act from the propositional act 18:45 < dmiles_afk> ahah.. now you saying things i need to hear! 18:46 < ldlework> If I say "A ball is blue." "The gift is very important to me." 18:46 < ldlework> How can you inference engine do *anything* with real speech unless it does exactly that? 18:46 < ldlework> When it encounters "the gift" what will it do? 18:47 < ldlework> Will it try to create conclusions based on the other things here which are some how logically related to giftness? 18:47 < ldlework> Or will it simply attempt to assign the description to some external object? 18:47 < ldlework> Like the one introduced in the first sentence? 18:47 < dmiles_afk> "How can you inference engine do *anything* with real speech unless it does exactly that?" <- for sure! but maybe this is true? "How can you inference engine do *anything* with First Order Predicate Logic unless it does exactly that?" 18:47 < ldlework> With reference as a separate speech act, the inference engine will have to pick something in context. 18:48 < ldlework> dmiles_afk: well 18:48 < ldlework> You can treat all speech as non-veridical 18:48 < ldlework> And interpret 18:48 < ldlework> "The gift is very important to me" 18:48 < ldlework> as relating the predicates "important" and "giftness" 18:48 < ldlework> But I'd say you are wasting your time since that's now how speech works 18:48 < ldlework> But probably the direction most would have you go in. 18:49 < ldlework> sorry, "You can treat all speech as veridical*" 18:49 < ldlework> Actually 18:50 < ldlework> Even veridical definite descriptions have no bearing on the core proposition 18:50 < ldlework> Its just a totally unrelated assertion made in addition to the main one 18:51 < ldlework> Unless the goal is to characterize 18:51 < ldlework> And the information that x was a Ball in addition to being Blue is going to somehow lead you to conclusions that depend on *characterizing* balls. 18:51 < ldlework> In that your inference engine might conclude, well the only ball I've ever encountered was blue, so maybe all balls are blue? 18:52 < ldlework> But seriously, this isn't the purview of fopl 18:52 < dmiles_afk> (*nod*) 18:52 < ldlework> dmiles_afk: but it might be the purview of inference engines 18:53 < ldlework> for sets of inference engines that you don't care it is wrong, as long as it makes an educated guess 18:53 < ldlework> IE, using characterizations 18:53 < ldlework> Its interesting to me to think of characterization in this way 18:54 < ldlework> For example if I say 18:54 < ldlework> My dad is a great guy. 18:54 < ldlework> You're supposed to understand that there is someone out there that is a great guy. 18:54 < ldlework> And use "My dad" to figure out which one 18:54 < ldlework> The description and the eventual proposition are not supposed to meet. 18:54 < ldlework> But what do we do when we make characterizations? 18:55 < ldlework> Well ldlework's dad is awesome, and this guy is your dad, so maybe your dad is awesome. 18:55 < ldlework> Instead of attributing the assertion to just the referent, the guy, who happens to be my dad. 18:55 < ldlework> You're attributing the assertion to the description. 18:55 < ldlework> But this perfectly aligns with the thinking that indefinite descriptions refer to predicates. 18:56 < ldlework> Because if I remove the relativization 18:56 < ldlework> Dads are awesome. 18:56 < ldlework> I get the characterization, the generalization 18:56 < ldlework> Its not only wrong to ignore the reference act as separate than the propositional one, it leads to bad thinking. 18:57 < ldlework> I wonder if it can be formalized to 18:57 < ldlework> Anytime the perlocutionary effect is the application of a root predicate to the description rather than the referent, you get a generlization. 18:58 < ldlework> This makes sense if we model our understanding of indefinite descriptions as having no referent. 18:58 < ldlework> Or referring to predicates themselves. 18:58 < dmiles_afk> (all this is true of FOPL you thi9nk .. not just English .. correcT?) 18:59 < ldlework> no 18:59 < dmiles_afk> (i am interested in understand english to) 18:59 < ldlework> I don't believe reference is a thing that takes place in fopl at all 19:00 < ldlework> the variables used in fopl are only useful for the calculus, to objectify the relationship between two predicates 19:00 < dmiles_afk> ok .. right that just transation of speech when you say refernce 19:00 < ldlework> but fopl variables don't nessecitate that they actually refer to anything 19:00 < ldlework> they just represent the connection between men and mortal 19:00 < ldlework> they don't represent men or mortals 19:00 < dmiles_afk> ah! 19:01 < ldlework> There exists should be rephrased 19:01 < dmiles_afk> (see secretly I use the vars to pass arround referants) 19:01 < ldlework> If there ever was such a thing, 19:01 < ldlework> dmiles_afk: okay let me explain 19:01 < ldlework> in lojban we have a grammar for producing fopl statements trivially 19:01 < ldlework> for example 19:01 < ldlework> da represents the existential variable x in fopl (let's say) 19:02 < ldlework> so instead of 19:02 < ldlework> F(x) 19:02 < ldlework> I can just say 19:02 < ldlework> x froda 19:02 < ldlework> the structure is 19:02 < ldlework> 1 predicate 2 3 4 5 19:02 < ldlework> so if we had another variable 19:02 < bigcentaur> .i .u'i do'o za'o casnu zo'e 19:02 < ldlework> we'd put it after the predicate 19:02 < ldlework> so if we wanted 19:02 < ldlework> F(x, y) 19:02 < ldlework> we might say 19:02 < ldlework> da froda de 19:03 < ldlework> make sense so far? 19:03 < dmiles_afk> so far .. with you 19:03 < ldlework> if I wanted to create an entailment between two predicates 19:03 < ldlework> I can do 19:04 < ldlework> da zo'u gajanai da remna gi da kamymorsi 19:04 < ldlework> For x, if x is human, then, x is mortal 19:05 < ldlework> traditionally you would formulate this as 19:05 < dmiles_afk> (implies (human ?X) (mortal ?X)) 19:05 < dmiles_afk> (how i write it in fopl) 19:06 < ldlework> ro da poi remna ku'o kamymorsi 19:06 < ldlework> which says 19:06 < ldlework> for all x which is human, is mortal 19:06 < ldlework> here's where the conflation happens 19:07 < ldlework> in fopl speak, when we say "for all x if A then B" all we mean to do, is establish, or denote an entailment between A and B 19:07 < ldlework> but in -language- when we say "for all x" we _refer_ to things we then assert things about. 19:08 < ldlework> This is bad, because in the goal of fopl we wish to explore the relationships between predicates 19:08 < ldlework> In language we wish to explore the nature of individual things and their characteristics. 19:08 < ldlework> Notice that in the language, we don't get the entailment nessecarilly. 19:09 < ldlework> For every human that human is mortal. 19:09 < ldlework> All this really does 19:09 < ldlework> Is line up a bunch of objects in the world into a line 19:09 < ldlework> and then visits each object placed in the line 19:09 < ldlework> and says 19:09 < ldlework> "Hello, you are a human. You are also mortal." 19:09 < ldlework> It does this for every human, until the line is exhausted. 19:10 < ldlework> The only entailment created is *statistical* 19:10 < ldlework> "Well since all the humans are mortals, then, *probably* to be human, implies mortality." 19:10 < ldlework> fopl does something else entirely 19:10 < ldlework> it says, for M(x) to be true, then H(x) must also be true 19:11 < ldlework> It creates logical entailment between the definition of the predicates. 19:11 < dmiles_afk> still following 19:11 < ldlework> Regardless if there are no humans anywhere, at anytime, and Human and Mortal are fake predicates we just made up and don't mean anything. 19:12 < ldlework> I argue (really, I'm parroting) that when we state things, in language, we really care to assert predicates upon our referents. 19:13 < ldlework> When I say, "The hostess is nice." I don't mean to create any entailment between host and nice. 19:13 < ldlework> I do wish to assert niceness upon the person to whom I refer. 19:13 < ldlework> In fact, no entailment between niceness and host-ness could ever be useful 19:13 < ldlework> Except for a generalization 19:13 < ldlework> And if that's what I really meant to do 19:14 < uuZITxirel> Whaaat? 8 hours and still talking? You're insane. ;-) Anyways, I say good-bye to all. 19:14 < ldlework> As a speaker that's what I would do 19:14 < uuZITxirel> co'oroda 19:14 < ldlework> I would create that entailment 19:14 < ldlework> via 19:14 < ldlework> Hosts are nice. 19:14 < dmiles_afk> (peek at this .. https://github.com/TeamSPoon/PrologMUD/wiki/E2C ) are the statements after the words entailmnets? 19:15 < ldlework> dmiles_afk: Are you creating a relationship between what it means to be a shopper and what it means to drink? 19:15 < ldlework> Or are you imbuing a specific drinker with some quality of drinking-ness? 19:15 < ldlework> I argue it is obviously the latter 19:15 < dmiles_afk> 2nd 19:16 < ldlework> dmiles_afk: now what if it was an indefinite description? 19:16 < dmiles_afk> dang i was hoping i had an exmaple of that 19:16 < ldlework> Shoppers drink the pepsi? 19:16 < dmiles_afk> (none of those on that page are!) 19:16 < ldlework> Now we are creating an entailment between what it means to be a shopper and to drink 19:17 < ldlework> It even seems strange to us, to have a whole class of things drinking the same definite referent of the description "the pepsi" 19:17 < ldlework> as if to be a shopper you must share the great single can of pepsi 19:17 < ldlework> or that our speaker is russian 19:17 < ldlework> an included the definite article on the indefinite description "Pepsi" 19:18 < ldlework> "Shoppers drink Pepsi" 19:18 < ldlework> Even though it is false, it no longer seems strange. 19:18 < dmiles_afk> (implies (shopper ?x) (pepsiDrinker ?x)) 19:18 < dmiles_afk> is (implies (shopper ?x) (pepsiDrinker ?x)) an entailmnet? 19:18 < ldlework> It seems strange in the first case because since the first description is indefinite, we refer not a or some or all the shoppers 19:18 < ldlework> but to the predicate to-shop itself 19:19 < ldlework> so we seem to define the very act of shopping in regards to drinking a single pepsi 19:19 < ldlework> dmiles_afk: I don't know youammar so I can't tell what's definite and what's indefinite 19:19 < ldlework> your grammar 19:20 < dmiles_afk> "All shoppers drink pepsi sometimes" -> (implies (shopper ?x) (pepsiDrinker ?x)) 19:20 < ldlework> well 19:20 < ldlework> remember 19:20 < ldlework> All shoppers isn't always the same as "Shoppers" 19:20 < ldlework> Just like if we tell each human that they are mortal 19:20 < ldlework> We haven't created a relationship between men and mortality 19:20 < ldlework> beyond the statistical 19:20 < dmiles_afk> "shoppers drink pepsi sometimes" -> (implies (shopper ?x) (pepsiDrinker ?x)) 19:21 < ldlework> so we should denote explicitly when we mean indefinite and definite 19:21 < ldlework> dmiles_afk: what's the definite form? 19:21 < dmiles_afk> "a shopper drinks pepsi sometimes" -> (exists ?x (implies (shopper ?x) (pepsiDrinker ?x)) ) 19:22 < ldlework> dmiles_afk: and the indefinite-with-definite form? 19:22 < dmiles_afk> whats the english of indefinite-with-definite form? 19:23 < ldlework> Shoppers drink the Pepsi. 19:24 < ldlework> I feel like you're making the same conflation 19:24 < ldlework> what is the ?x in (shopper ?x) in the purely indefinite form? 19:24 < ldlework> the predicate "shopper" 19:24 < ldlework> then what's the point? 19:25 < ldlework> You need to different forms of reference. 19:25 < ldlework> You need existential statements, and you need characterizing statements. 19:25 < ldlework> And english is going to give you nothing but trouble 19:25 < ldlework> Because some indefinite forms are definite. 19:25 < ldlework> And some definite forms are indefinite. 19:26 < ldlework> Lojban has the problem that it is trying to solve indefinite reference with quantification. 19:26 < ldlework> Rather than a denoting article 19:26 < dmiles_afk> (existsExistsAtLeast (2 ?x) (implies (shopper ?x) ( (and (isa ?drink DrinkingEvent ) (doer ?x ?drink) (objectConsumed ?y ?pepsi) (pepsi ?pepsi))) 19:26 < ldlework> Like in english 19:26 < ldlework> Aha, you are going to solve it with quantification also I see. 19:26 * dmiles_afk is a slow typer 19:26 < dmiles_afk> right 19:26 < ldlework> In my opinion, all you've created is statistical implication, not logical entailment. 19:27 < dmiles_afk> but i am worried that one size quatification solution is worng maybe 19:27 < ldlework> dmiles_afk: well of course it is 19:27 < ldlework> how do you accomplish the two following goals: 19:27 < ldlework> entailing that humanity is inherently mortal by way of definitional entailment 19:28 < ldlework> or merely observing that each human is also mortal without asserting entailment beyond the statististical 19:28 < dmiles_afk> ok .. good question! lets see 19:28 < dmiles_afk> how do we encode the difernc3e in fopl ? 19:28 < ldlework> it doesn't exist. 19:28 < ldlework> this is my point 19:29 < dmiles_afk> so there are things that fopl cant represent? 19:29 < dmiles_afk> (secretly i am under the assumption that anything under the sun can be written in fopl) 19:29 < ldlework> No, fopl can express reference but not in the sense that you can just abuse a mechanic of fopl, the existential variable, actually talk about referents. 19:30 < ldlework> You must predicate reference, as any other predicate 19:30 < ldlework> when you wish to encodify what someone said and how they said it 19:30 < ldlework> Because the existential variable is not isomorphic with those things we use descriptions to refer 19:30 < ldlework> When I say 19:30 < ldlework> My dad is nice 19:30 < ldlework> the predicate "dad" never shows up in fopl 19:31 < ldlework> because the proposition here has nothing to do with dads and how they relate to niceness 19:31 < ldlework> it has to do with *some guy* being nice. 19:31 < ldlework> all we get from the fopl in this case is 19:31 < ldlework> Nice(x) 19:31 < ldlework> that's it! 19:31 < dmiles_afk> we also relate that guy to dadness and who's dad it is 19:31 < ldlework> But that's not the content of the proposition 19:31 < ldlework> I'm not asserting that something is a dad 19:31 < ldlework> I'm asserting something is nice 19:32 < ldlework> "My dad" is a referring expression, so that in the context of language and communication, you know what is nice. 19:32 < ldlework> But if we start to conflate the existential variable, with /that guy/ 19:32 < ldlework> Then we start to create entailments between dad-ness and nice-ness 19:32 < ldlework> Beyond the statistical! 19:32 < dmiles_afk> in fo i do that as (and (nice ?dad) (fatherOf me ?dad)) 19:32 < ldlework> Certainly there are non-nice dads. 19:33 < ldlework> Right, but 19:33 < ldlework> If your goal is to understand the truth of the statement 19:33 < dmiles_afk> or (exists ?dad (and (nice ?dad) (fatherOf me ?dad))) 19:33 < ldlework> the truth with depends on the referents, and the predicate nice 19:33 < ldlework> but not of dad 19:33 < ldlework> can you see that? 19:33 < dmiles_afk> "goal is to understand the truth of the statement" (query (and (nice ?dad) (fatherOf me ?dad))) 19:33 < ldlework> it depends on whether /that guy/ is nice. the predicate "dad" never even comes into play 19:34 < dmiles_afk> result is.. ?dad = ldleworkDad 19:34 < ldlework> I don't understand your grammar, sorry 19:34 < ldlework> But you should be able to agree or disagree without it I think 19:35 < ldlework> Take for example 19:35 < ldlework> We want to teach your bot about dad-ness 19:35 < ldlework> And we want it to arrive at the most platonic understanding of dad-ness possible 19:35 < ldlework> Not our personal favorite idealness, but the ideal that encompasses all instances for which we would say is true of the predicate dad 19:36 < ldlework> IE, niceness and terribleness are both equally valid 19:36 < ldlework> And to teach your bot, it must consume natural speech 19:36 < ldlework> Your bot will have two goals of course 19:36 < ldlework> 1. To know what I mean by what I say to it 19:36 < ldlework> 2. To take some statements as providing additional definition for what the predicate to-dad means 19:36 < ldlework> Now, if you don't carefully divorce these two goals 19:37 < ldlework> You will end up learning many many things about dad-ness for which many dads will prove false under your eventaul learned model 19:37 < ldlework> For example 19:37 < ldlework> My dad is nice 19:37 < ldlework> He plays the baritone 19:37 < ldlework> My dad and Jeremy's get along together 19:37 < ldlework> I wish all dads could spend unlimited time with their kids 19:38 < ldlework> Once troy's wife gives birth, he's going to be a great father 19:38 < ldlework> and so on 19:38 < dmiles_afk> right i have to know so much to understand dadness 19:38 < ldlework> well more importantly 19:38 < ldlework> a lot of this information is irrelvant to dadness 19:38 < ldlework> In most cases 19:38 < ldlework> dad appears in a description clause 19:39 < dmiles_afk> http://logicmoo.org:3602/cgi-bin/cg?cb-c-all-asserted-knowledge&c5501 1000+ statments about Dad.. yes not sure if the robot undersands dad-ness yet 19:39 < ldlework> in most cases it does NOT appear as the core predicating phrase 19:39 < ldlework> do understand what I mean? 19:39 < dmiles_afk> i think you saying that dadness needs defining in fopl ? 19:39 < ldlework> no I'm saying 19:40 < ldlework> when I use the predicate "dad" inside of a referring description 19:40 < ldlework> the larger proposition does not help you understand what it means to be a dad 19:40 < ldlework> IE compare 19:40 < ldlework> My dad is nice. 19:40 < dmiles_afk> oh anaphoric resolution 19:40 < ldlework> All Dads have had sex. 19:40 < ldlework> or rather to be clear 19:40 < ldlework> Dads have had sex. 19:41 < ldlework> So you don't think I'm making an enumerating definite reference 19:41 < ldlework> In the first, we can't learn anything about dad-ness 19:41 < dmiles_afk> to the collection of all dads yeah 19:41 < ldlework> Some guy is a dad, and Some guy is nice 19:41 < dmiles_afk> buty not to a individual dad 19:41 < ldlework> But that doesn't tell us anything about what it means to dad. 19:41 < ldlework> Its not the collection of all dads. 19:41 < ldlework> All dads could be nice. 19:41 < ldlework> But that may not be what it _means to dad_ 19:42 < dmiles_afk> All Dads have had sex which produced a child. 19:43 < ldlework> Don't use All 19:43 < ldlework> It means that you're thinking about specific dads 19:43 < dmiles_afk> Dads have had sex which produced a child. 19:43 < ldlework> in this case all of them 19:43 < ldlework> yes 19:43 < dmiles_afk> still may not be what it _means to dad_ :) 19:43 < ldlework> now we're using the predicate it self 19:43 < ldlework> dmiles_afk: sure, but there we can make reasonable disagreements 19:43 < ldlework> but we wont have a reasonable disagreement is one of my axioms is that "well, shit, MY dad is nice" 19:43 < ldlework> So 19:43 < ldlework> to go back 19:44 < ldlework> definite descriptions do not provide definitional predicatory extension in the same way that indefinite descriptions do and definitional statements do 19:44 < ldlework> My dad is cool. There's no need to include the description in the fopl statement. 19:45 < ldlework> Only if you care to understand what it means to be _my dad_ 19:45 < ldlework> Further 19:45 < ldlework> the assertion is /some guy/ is cool 19:45 < ldlework> and the truth of this statement depends on whether this guy is cool 19:45 < zipcpi> exp: ko'a soi racli se'u krinu ko'e 19:45 < mensi> ([ko'a {soi se'u}] [CU {krinu }]) 19:45 < ldlework> and not whether he's my dad 19:45 < zipcpi> exp: ko'a soi racli cu krinu ko'e 19:45 < mensi> ([ko'a {soi SEhU}] [cu {krinu }]) 19:47 < ldlework> I think that language is far more complex than the trivial way we translate it to fopl statements. 19:47 < ldlework> I think that with an inference engine you have to start breaking down speech in terms of language, *then* fopl 19:47 < ldlework> Because any given statement tells you multiple things 19:48 < ldlework> You may have to track multiple completely separate lessons from any given statement. 19:48 < dmiles_afk> ldlework: Oh.. my infernce engine doesnt use english :) I have to talk to it in fopl 19:48 < ldlework> If you only care to use the machine to learn the true meaning of predicates, you have to filter some parts of language. 19:48 < ldlework> If you don't really care what it means to dad in the universal sense but want to learn about actual things, like _my dad_, then you have to filter others. 19:48 < zipcpi> .i fabu soi racli cu krinu lo nu da zenba lo ka birti febu <-- Another proposed def for {indika} 19:49 < ldlework> If you just say, any given utterance in any given language, is just some singular self-coherent fopl expression 19:49 < ldlework> I feel that will always be a bot that never gets finished 19:49 < zipcpi> x1 is a rational reason for someone to be more certain of x2 19:49 < ldlework> For example 19:49 < ldlework> My dad is cool. 19:50 < ldlework> zipcpi: why is it rational 19:50 < ldlework> its their opinon? 19:50 < zipcpi> No, because the probability of something, with this new piece of knowledge, is higher than without that knowledge 19:51 < dmiles_afk> when i talk to my bot.. it only use fopl (if i want to speak engish it has to translate) 19:51 < ldlework> What determines the probability of any given event? 19:51 < ldlework> How do I use pieces of knowledge to derrive such a value? 19:51 < dmiles_afk> to say My dad is cool ? 19:52 < zipcpi> I'm not studied enough to argue Bayesian mechanics 19:52 < ldlework> dmiles_afk: no sorry, talking to zipcpi 19:52 < ldlework> The Bayesian interpretation of probability can be seen as an extension of propositional logic that enables reasoning with hypotheses, i.e., the propositions whose truth or falsity is uncertain. 19:53 * ldlework coughs quietly. 19:55 < dmiles_afk> " I feel that will always be a bot that never gets finished" <- it never is finished. FOPL is alwasy a rough geenralization 19:55 < ldlework> dmiles_afk: I don't see why it has to be 19:55 < dmiles_afk> when i am speaking to it in FOPL.. it knows 30 seconds lkater i might say something else 19:56 < ldlework> We just need better mechanisms for translating speech into fopl and I don't think it is like how we are doing it right now. Or at least as I've seen it in the lojban community and peripherally in studying lojban. 19:56 < dmiles_afk> "i see you" produces 15 differnt FOPL assertions 19:56 < dmiles_afk> "bill loves sally" makes 4 fopl assertions 19:56 < ldlework> dmiles_afk: I do like that you're digging further than the surface assertion 19:57 < ldlework> I like that you derrive things like "You are visible" from "I see you" 19:57 < ldlework> And probably depends on no other knowledge 19:57 < dmiles_afk> to me when i am in the REPL with a theorem prover i pretend i am having a dialog 19:57 < ldlework> Just the fact that its being use as the object in that predicate 19:58 < dmiles_afk> ah yeah that i see as a builtin part of fopl .. is well formedness 19:58 < ldlework> But there's only so much you can glean right? 19:58 < dmiles_afk> well-formedness means that we assign malness to your father 19:59 < dmiles_afk> maleness 19:59 < ldlework> Like you can't glean that dad-ness is defined in terms of niceness, right? 19:59 < dmiles_afk> i can glean has a level of meaness vs niceness 19:59 < ldlework> did you skip a word? ^ 19:59 < dmiles_afk> and is capable of actions 19:59 < ldlework> what? 20:00 < dmiles_afk> "my dad is nice" 20:00 < ldlework> right 20:01 < dmiles_afk> i can glean that its possible 20:01 < ldlework> what is 20:01 < dmiles_afk> i can glean that its possessable 20:01 < ldlework> oh dadness? 20:01 < ldlework> sure ok 20:01 < dmiles_afk> and that it realates to other beings in which they may enjoyiit and refer it to nice 20:01 < ldlework> haha no 20:01 < ldlework> your thing doesn't do that 20:01 < ldlework> but you mean 20:02 < ldlework> can be percieved as nice 20:02 < ldlework> or can be nice 20:02 < dmiles_afk> and other beings have things like dads.. but have no idea what dads are 20:02 < ldlework> I thought you covered that in the possesive 20:02 < dmiles_afk> but at least one dad can be percieved as nice 20:02 < ldlework> can and has 20:03 < dmiles_afk> if you say "my whatnot" that meant being may have whatnots 20:03 < ldlework> right I get you 20:03 < ldlework> dmiles_afk: if your thing wasn't written in prolog I'd definitely be down to help you help it understand lojban 20:04 < dmiles_afk> what languages do you program in ldlework? 20:04 < ldlework> dmiles_afk: python, javascript, C# and nim, mostly 20:04 < dmiles_afk> (right now the bot also understands ACE ) 20:05 < ldlework> I don't quite understand how to parse with parsers so I've never toyed with this stuff 20:05 < dmiles_afk> i have a C# bot called cogbot but stopped working on it :( 20:06 < dmiles_afk> none of the AI code is in the C# part though 20:06 < dmiles_afk> well ecpec thte pathfinder i wrote 20:07 < dmiles_afk> i was appalled it that no one had ever wrtten an opensource 3d pathfiinder in C# 20:07 < dmiles_afk> so i published one 20:14 < dmiles_afk> btw i was going to make it speak/thinnk in lojban :) 20:14 < dmiles_afk> but i decied fopl keeps me focused 20:15 < dmiles_afk> (i been working on AI for 25 years.. and most of those years are me trying to code something distracting the AI needs to avoid the real challenge) 20:16 < ldlework> dmiles_afk: would you characterize your bot as an inference machine or semantic reasoner 20:17 < dmiles_afk> its a theorem pver/ semantic reasoner 20:18 < dmiles_afk> i am trying to prove that Mark Stickels CNF resolution method is possble 20:18 < dmiles_afk> even though in everyone has declared it too combinatoric 20:19 < dmiles_afk> for 15 years people been writting papers about how to lose tautological data as a performance boost 20:20 < dmiles_afk> (about canceling out p v ~p ) 20:20 < ldlework> dmiles_afk: do you mind if I ask a completely unrelated question 20:20 < dmiles_afk> but they lose quantification 20:20 < dmiles_afk> go ahead 20:21 < dmiles_afk> i now think the new SAT techniques only proliferate that methodolgy 20:22 < ldlework> what do you think about applying genetic algorithms to assist with the design of neural networks? 20:22 < dmiles_afk> genetic algorithms will improve most anything ... 20:23 < ldlework> haha 20:23 < dmiles_afk> but neural networks? 20:23 < dmiles_afk> (what about them that need improvemnt? ) 20:23 < ldlework> dmiles_afk: generally its pretty tough to create a topology that works best for your problem 20:24 < dmiles_afk> i started using GA to decide some heuristics.. that was a real speed up 20:24 < ldlework> dmiles_afk: I use GA to create virtual life-forms 20:24 < ldlework> probably the most boring use-case academically 20:24 < dmiles_afk> though secretly i know the GA (how i implmented will bite me when my uses change)( 20:25 < dmiles_afk> when my usage of the system that it optimises 20:25 < ldlework> right 20:26 < ldlework> dmiles_afk: I use them to basically evolve goal-driven behavior graphs 20:26 < ldlework> well, behavior tables 20:26 < dmiles_afk> see that is 100% right on 20:26 < ldlework> behavior graphs would be amazing why haven't I thought of that 20:26 < ldlework> well 20:26 < ldlework> techincally it is a graph I suppose 20:26 < dmiles_afk> (not a NN right?) 20:26 < ldlework> nope 20:26 < ldlework> imagine a grid 20:26 < dmiles_afk> thank god 20:26 < ldlework> on the grid you place some agents 20:26 < ldlework> the agents can 'see' the 4 squares around them 20:27 < ldlework> at each turn, the agent takes what they see, and look up what they should do in some table 20:27 < ldlework> everytime an agent runs over a piece of food I get them some fitness 20:27 < ldlework> I breed them and eventually they are good at finding food 20:27 < ldlework> then I introduce a second species that gets points for eating the first species :P 20:29 < ldlework> I have tinkered with ideas about making it a continuous simulation rather than a generative one 20:29 < ldlework> after 500 turns I stop the simulation and do cross-breeding among the highest ranked and then start the simulation over again with the new population 20:29 < ldlework> doing it continuously would be tricky because you need some method of breeding 20:29 < ldlework> birth 20:29 < ldlework> and so on 20:29 < dmiles_afk> what are the genes dictating? 20:29 < ldlework> the genes are the look-up table 20:30 < ldlework> the table has a key for every possible combination of what they see 20:30 < dmiles_afk> what they see to how they react? 20:30 < ldlework> mapping to some action 20:30 < ldlework> yes 20:30 < ldlework> actually its a bit more nuanced than that 20:30 < ldlework> there's an additional input 20:30 < ldlework> "what state am I in?" 20:30 < ldlework> which is just an integer 20:30 < ldlework> the mapping includes what to do 20:30 < ldlework> but what the next state is too 20:31 < ldlework> this drastically improves the complexity of the behavior 20:31 < dmiles_afk> "how abundent or scarce is food?" "do i even care about abundancy?" 20:31 < ldlework> since now they can 'plan' and 'orchestrate' action 20:31 < ldlework> like turns and such 20:31 < dmiles_afk> "do i care if dad sees me eat his baby?" 20:31 < ldlework> heh 20:32 < ldlework> I just realized, I never change the mapped action 20:32 < dmiles_afk> so what happens if you increase the rules of their world? 20:32 < ldlework> increase the rules? 20:32 < dmiles_afk> (like for instance the previous genration protects their young from being eaten) 20:32 < dmiles_afk> or agression for pleasure 20:33 < dmiles_afk> (things to distract the easing process) 20:33 < ldlework> remember there's no intelligence 20:33 < ldlework> its just look up tables 20:33 < dmiles_afk> easing/eating 20:33 < ldlework> the GA finds that look up table that is best for walking over food squares 20:33 < ldlework> in the search space of possible beahvior tables 20:35 < dmiles_afk> sounds neat though 20:35 < dmiles_afk> well can always add desise 20:35 < dmiles_afk> erm what desiese 20:36 < dmiles_afk> see the theory is.. if you add some weird unrealted incidents in tehir life that follows some course.. their GA will adappt arround it 20:36 < dmiles_afk> (even thugh they dont do anyhting other than exactly what they do) 20:37 < dmiles_afk> for example.. you add a lake where hey might drown in the middle of the tile world 20:37 < dmiles_afk> will will breed whatnots that will somehow avoid the lake.. yet we cant really see how they did that 20:39 < dmiles_afk> even though that was supposed to be evoluiton requireing special functionalty such of the kind only a NN would provide 20:52 < ldlework> oh sure 21:28 * nuzba @464161niftg: pilin pona. a, jan mute li toki e ni: toki Lojban li kute sama toki Slav (Sula? Lawe? La?¥ [http://bit.ly/1RhyFGw] 21:56 < kmir> coi io 21:57 < kmir> ma jundi 22:23 < spollei> coi la kmir 22:23 < kmir> ui coi do 22:24 < kmir> xu do slabu mi .i'unai 22:24 < spollei> na slabu 22:25 < kmir> je'e .i ko skicu do a'u 22:25 < spollei> mi zu tavla 22:26 < kmir> tavla ca ma 22:27 < spollei> masti li so'o 22:28 < kmir> .i xu do lojbo tadni puzu masti be li so'o 22:30 < spollei> mi di'inai tadni 22:31 < kmir> ua je'e 22:31 < spollei> mi di'inai tadni lo lojbo bangu 22:32 < kmir> .i caku do ca'o se cinri ma 22:34 < spollei> zoi url http://blogs.perl.org/users/drforr/2015/07/perl5-to-perl6-conversion-a-la-perlcritic.html url 22:36 < kmir> .ua .i mi na jimpe no da poi me la perl u'i 22:39 < spollei> mi zi sipna 22:40 < kmir> .i ko pluka senva 22:41 < spollei> ki'e la kmir co'o 22:42 * kmir sanga lu .i ku'i no da tirna .i no da tirna do~~ 22:59 * nuzba @uitki: ¡Bienvenido!/Español - http://mw.lojban.org/papri/%C2%A1Bienvenido!/Espa%C3%B1ol por Gleki [http://bit.ly/1Cm51JY] 23:01 * nuzba @uitki: Добро пожаловать!/Русский - http://mw.lojban.org/papri/%D0%94%D0%BE%D0%B1%D1%80%D0%BE_%D0%BF%D0%BE%D0%B6%D0%B0%D0%BB%D0%BE%D0%B2%D0%B0%D1%82%D1%8C!/%D0%A0%D1%83%D1%81%D1%81%D0%BA%D0%B8%D0%B9 by Gleki [http://bit.ly/1Cm53S7] 23:03 * nuzba @uitki: 已更新页面:Bienvenue !/fr - http://mw.lojban.org/papri/Bienvenue_!/fr 由 Gleki [http://bit.ly/1Cm5agq] 23:05 * nuzba @uitki: 已更新页面:欢迎! (中文) - http://mw.lojban.org/papri/%E6%AC%A2%E8%BF%8E%EF%BC%81_(%E4%B8%AD%E6%96%87) 由 Gleki [http://bit.ly/1Cm5gEP] 23:49 < keegi> MEM: 16202 MB - 25% (4050 MB) free --- Day changed Tue Jul 07 2015 00:14 < gleki> LIVLABOT: sutysisku now able to search for glosswords (currently only direct match works) http://mw.lojban.org/extensions/ilmentufa/i/en/index.html#sisku/tanru%20and 04:07 * mumble jorne co'a fa la'o fa ctefaho fa 04:08 < ctefaho> coi 04:17 < ctefaho> ldlework: so I think what you said in yesterday is basically one of the things I am doing in my dialect, or at least trying to do 04:18 < ctefaho> as part of the "containment of truth" 04:19 < ctefaho> /assertion 05:03 < ctefaho> ldlework: To me *all* bridi have a default NA as ja'a, including main-bridi and sub-bridi having their own NA. The truth of all NA are considered separately, and formal expansion never adds, removes, alters, morphs or moves NA around (like merging 2 bridi with 2 NA into 1 tanru with 1 NA). The truth of noi/"adjectival" clauses' assertion (which the gadri are still based on) are always separate 05:03 < ctefaho> from the truth of the assertion of the bridi it appears in. and there is none of the "default-predicate-entailment" you speak of. In "lo/le/da/pa/whatever Tcàdu cu [ja'a] Mèlbi" The assertion of the resulting noi-clause may be true or false and the assertion the main bridi may be true or false, independent of each other. Do tell if this is in-compatible with your view. 05:03 * ctefaho bbiab 05:19 * nuzba @oka_iu_tcan: 誰かロジバンで"vowel"ってなんていうか教えて e'o jungau mi lodu'u fu makau fanva fe zoi gy. vowel gy. lo jbobau [http://bit.ly/1glcte0] 05:24 < gleki> it's a well known bug/feature of Lojban that ja'a is two syllables otherwise {cu} would be useless 05:25 < gleki> mensi: doi zipcpi i think i fixed relevance in sutysisku search. now it outputs words for "and" first. see what else you need to make this app perfect http://mw.lojban.org/extensions/ilmentufa/i/en/index.html#sisku/and 05:25 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.zipcpi.gy. di'a cusku da 05:33 * nuzba @mochirial: @oka_iu_tcan jbovlaste の訳ではダメなんですか? http://jbovlaste.lojban.org/lookup?Form=lookup.pl1&Strategy=*&Database=en%3C-%3Ejbo&Query=vowel&submit=Search [http://bit.ly/1gldHpF] 05:37 * nuzba @Space2O2: I think it is a little difficult to talk about elementary particles in Lojban. That is because the words related to them are incomplete yet. [http://bit.ly/1gle2bT] 05:39 < gleki> and when they become complete we'll know everything about the universe 05:39 < zipcpi> What's the best way to say a clock is "fast" or "slow"? 05:39 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: i think i fixed relevance in sutysisku search. now it outputs words for "and" first. 05:39 < mensi> see what else you need to make this app perfect http://mw.lojban.org/extensions/ilmentufa/i/en/index.html#sisku/and | 05:39 < mensi> 2015-07-07T12:25:37.878Z 05:41 < zipcpi> zo bercribe na se jbovlaste 05:52 < ctefaho> gleki: I made ja'a to ja. I use .a .e .o .u for the logical connectives in the simplified system leaving ja/je/jo/ju free 05:52 < ctefaho> so say hello to "mi ja klama" and "mi klama .ecu broda" 05:53 < ctefaho> I honestly can't think of anything else that would suit .a/.e/.o/.u better than the logicals 05:54 < gleki> hahaha 05:54 < gleki> tell that selpahi 05:54 < zipcpi> Tell what to selpahi? 05:55 < ctefaho> I assume selpai is gonna use them for some tonal-like feature 05:56 < gleki> he wanted A to go away 05:56 < gleki> although i once proposed A as this is how it was in Loglan originally 05:57 < zipcpi> Oh simplified connectives? 05:57 < ctefaho> well I don't always agree with selpai 05:57 < gleki> en: selpai 05:57 < ctefaho> (it is a pun on senpai) 05:57 < zipcpi> se pajni 05:57 < gleki> backup time 05:58 < gleki> ditcu lo nu lo se snura fukpi 05:59 < gleki> use {li} for {ja'a}. like in tok pisin 05:59 < ctefaho> gleki: so is he gonna use .a/.e/.o/.u for something or did he just want to nuke them? 05:59 < gleki> he never told for what he could use them 05:59 < gleki> and i dont care 06:00 < ctefaho> je'e 06:00 < gleki> please everyone test new relevance search in sutysisku. 06:00 < gleki> relevance is now the priority 06:22 * nuzba @okau_junpe: 説明文翻訳案に訂正に訂正を重ねて管理者の方には申し訳なかったけれど…でも本当にいい辞書だと思う。 la sutysisku 。ロジバンについてちょっと調べたいという時に感動的なくらいに便利だ。 http://mw.lojban.org/extensions/ilmentufa/i/ [http://bit.ly/1CkvTJZ] 06:24 * nuzba @okau_junpe: La sutysisku. Bona vortara ilo, mi opinias. Ĝi estas rave uzebla, kiam oni serĉas iom da informo pri loĝbanaj vortoj. http://mw.lojban.org/extensions/ilmentufa/i/ [http://bit.ly/1CkwjA5] 06:43 < gleki> lojxbanaj 06:43 < gleki> ne logxbanaj 06:44 < gleki> sed la sutysisku ne estas nur lojxbana vortara 06:48 < zipcpi> What's the best way to say a clock is "fast" or "slow"? 06:48 < gleki> je'esai 06:49 < gleki> i zo'onai i dont know any ways, not even the best :D 06:49 < zipcpi> :p 06:50 < zipcpi> (of a clock or watch) showing a time ahead of the correct time. 06:50 < gleki> start from wiktionary probably 06:50 < gleki> another vision of this is that it counts time intervals too fast 06:51 < durka42> junla be lo tordydu'e? 06:51 < zipcpi> Yeah I think that would be the most literal interpretation of {sutra} / {dukse sutra}; but generally we mean that it's ahead of the correct time when we say "fast" 06:51 < gleki> lo junla cu te tcidu lo ba tcika 06:52 < zipcpi> I'm thinking {jai clira} or {jai lerci}, but that might be vague or confusing 06:52 < gleki> y 06:52 < gleki> lo junla cu te tcidu lo tcika be lo lerci 06:53 < zipcpi> A clock that reads a time of late 06:54 < zipcpi> Hm that = fast? 06:54 < zipcpi> But {jai clira} seems like that = fast 06:54 < zipcpi> Hence the confusion :p 06:56 < durka42> well as you said there are two interpretations 06:56 < durka42> either it's actually fast, or it's just set to the wrong time 06:56 < durka42> the distinction ought to be "found in translation" to lojban :p 06:56 < zipcpi> I believe the "wrong time" sense is more common; measuring "actually too fast" requires more observation 06:57 < zipcpi> And of course, the latter would cause the former 06:58 < durka42> right 06:58 < durka42> once I had a microprocessor clock that definitely ran fast, and I hooked it up to a computer overnight to gather data 06:59 < durka42> in the morning it showed 0% inaccuracy 06:59 < durka42> and after that it never ran fast again 06:59 < gleki> "goes fast" is that it inevitably will show late time. you will notice it sooner r later. 06:59 < gleki> another meaning of "fast" is that it is simply set wrong initially 06:59 < durka42> early 06:59 < durka42> oh no wait, late 07:00 < durka42> gah 07:00 < zipcpi> Well it's "early" in that it shows a time earlier than it should 07:00 < zipcpi> Erm 07:00 < zipcpi> {shows a time} earlier than it should 07:00 < zipcpi> graaah 07:00 < gleki> dang, now i cant understand my lojbna translation but im completely confused by the english metaphor 07:00 < durka42> scooooope 07:01 < durka42> lo'e snidu be ma'i le junla cu tordu 07:01 < gleki> lo junla cu te tcidu lo tcika be lo lerci - clock is a "book (a thing that we read" on which we read the timestamp of a later event 07:01 < durka42> (or {be fi} but let's pretend ro ma'i zei terbri cu spoja vau zo'o zo'onai) 07:02 < gleki> http://tatoeba.org/jbo/sentences/show/3734313 07:02 < gleki> .title 07:02 < gleki> translation challenge 07:02 < fenki> [ mupli lo uenzi be lo bangenugu fa "zoi zoi Tom was right after all. zoi" (to la'oi Tatoeba toi) ] 07:02 < zipcpi> The problem is that {clira} and {lerci} are somewhat subject to interpretation. Either a "fast" clock shows a {lerci} time, or it is {clira} in showing time 07:03 < gleki> not in my translation imo 07:03 < gleki> 2. (used of timepieces) indicating a time ahead of or later than the correct time 07:04 < gleki> it's from WordNet 07:05 < zipcpi> You can say that it shows a time, earlier than it should show it 07:05 < gleki> "That clock is one minute fast." 07:05 < zipcpi> = The clock reads one minute later than it should, or The clock reads a time, one minute earlier than it should 07:05 < gleki> exp: lo va junla cu te tcidu lo tcika pe baza lo mentu be li pa 07:05 < mensi> ([lo {va junla} KU] [cu {te tcidu} {lo KU}] GEhU¹)> KU} VAU]) 07:06 < gleki> WordNet has its opinion 07:06 < durka42> le junla cu jarco lo'e tcika pu ri zi lo mentu be li pa 07:06 < zipcpi> That's WordNet's definition, but "early" and "late" can have different interpretations depending on how you want to expand them, and that's the problem with defining something like {julclira} 07:07 < gleki> i prefer {jai gau zgana fai} 07:07 < gleki> and i think you forgot {pe} somewhere 07:08 < gleki> no word for "alarm" and "alarm" clock" in lojban 07:08 < zipcpi> You're here early - The clock shows "1:00" earlier than it should 07:08 < zipcpi> = The clock is fast 07:09 < zipcpi> fanza zo'o 07:09 < gleki> sinxa junla probably for alarm clock 07:09 < zipcpi> Nah... we need a word for alarm/siren 07:09 < gleki> lujvo mean what you make them mean 07:09 < zipcpi> But we'd need a word for alarm/siren, without the clock 07:10 < gleki> lo tutci be lo nu jungau fi lo ckape a lo na se djica 07:13 < zipcpi> By my own intuition, I believe {julclira} = "fast" while {julylerci} = "slow", but some others might disagree, thinking that describing the time read is more important than the sense of whether the *clock* is early or late 07:14 < zipcpi> Also seems awfully specific for a lujvo, but no idea how else to condense the concept 07:14 < gleki> i now feel that i really fixed everythin i could and should in sutysisku 07:15 < gleki> what isnt fixed is the ugliness of jvs db itself but that's another problem 07:15 < zipcpi> By the way thanks for the glossword fix 07:16 < gleki> pls report of anything else you want 07:16 < gleki> i only noticed that some glosswords needed fixing in jvs itself 07:16 < zipcpi> lo junla cu dukse lo ka jai clira fai lo ka arco lo tcika 07:17 < gleki> jbo: arco 07:17 < mensi> 18 da se tolcri: jarco, bilma, cmila, junla, kindi'u, mipri, ri'e, sa, salci, simlu, skami, ta, ti, tigni, tu, vidnyja'o, 07:17 < mensi> vidnyxra, xendo 07:29 < gleki> en: arco 07:29 < mensi> arco = x1 shows/demonstrates x2 (object/event) to audience x3 |>>> Where jarco is for showing/displaying a property, 07:29 < mensi> this word is to be used for showing/demonstrating an object. |>>> spheniscine 07:29 < gleki> =zgagau? 07:30 < gleki> gasygau even 07:30 < zipcpi> I'm not sure 07:30 < zipcpi> I jumpt between sajgau, zgagau, and gasygau until I just said, screw it, let's just define arco 07:31 < gleki> then mentioning jarco is useless here because then jarco would be just {arco lo ka} 07:32 < zipcpi> lo se ckaji be ma 07:33 < gleki> ko'a arco lo ka ce'u broda = ko'a arco lo su'u ko'a broda 07:33 < zipcpi> So we want arco2 to be able to be both an object and a property? 07:33 < zipcpi> That's just how {jarco} was used before 07:34 < zipcpi> The recommendation of the gimste to use {tu'a} didn't quite make sense 07:35 < gleki> for me jarco is about manifestation not about showing anything 07:36 < zipcpi> jb: jarco 07:36 < mensi> jarco = jarco — x1(entity) demonstrates or shows x2(property of x1) to x3(entity) 07:36 < mensi> :lo se jarco — demonstrated, manifested property. 07:36 < mensi> :fi mi'a fa la Alis cu jarco lo ka stati — To us Alice appeared to be smart. 07:36 < mensi> :la Kevin cu jarco lo ka gidva — Kevin acts as a guide. 07:36 < mensi> :Related words: tigni, cipra, zgana, jvinu, lanli, mipri, simlu 07:36 < gleki> as for abstraction non-ka places in La Bangu any such place can be used as a ka-place and ce'u will refer to x1 unless specifed explicitly otherwise 07:36 < zipcpi> Specified how? {ka} doesn't have an x2 07:38 < gleki> x2(property of x1) <-- specified 07:38 < zipcpi> I meant your nonce-ka places 07:39 < gleki> they are just e.g. "x2 (clause)". this is where it's not specified 07:41 < zipcpi> Oh you mean {mi arco lo nu salci} vs {mi arco lo ka salci}? 07:42 < gleki> yes 07:42 < zipcpi> Just like {ma djica lo nu cenba} vs {ma djica lo ka cenba} 07:42 < gleki> any nu place can always be converted to ka 07:43 < zipcpi> Not sure about always. {lo ka broda cu fasnu}... 07:44 < gleki> here it's nonce place 07:44 < gleki> there are some examples of it 07:44 < durka42> o_O 07:44 < gleki> jb: ckilu 07:44 < mensi> ckilu = ckilu — x1(property of nonce place with kau) is a scale of units for measuring x2(property of nonce place with 07:44 < mensi> kau) 07:44 < mensi> :lo ckilu — scale. 07:44 < mensi> :lo ka mitre ma kau cu ckilu lo ka te sepli ma kau ma kau — Metric scale measures the distance between two points. 07:44 < mensi> :Related words: gradu, merli, cimde, manri 07:44 < zipcpi> Maybe your jarco is to arco what kaidji is to djica 07:45 < gleki> because in ckilu someone deleted gradu1 so gradu1 is no longer accessible 07:45 < durka42> mi na'e jimpe 07:45 < durka42> zo ckilu 07:45 < gleki> no, my jarco isnt related to zgzgau, sajgau, gasygau at all 07:45 < gleki> durka42: ckilu is basically {zi'o gradu} 07:45 < zipcpi> You still use the word "show", and it still has an audience place 07:46 < gleki> "to manifest" is a better gloss 07:46 < gleki> english doesnt have good glosses for that 07:47 < zipcpi> Yeah, I'm starting to become more sure that we are really stretching to save {jarco} here 07:49 < gleki> ca lo vanci la barak obamas cu jarco lo ka rirni lo tixnu be vo'a vau lo lanzu be vo'a e la'a lo drata prenu 07:49 < gleki> i ku'i ca lo djedi cu jarco lo ka presidente 07:49 < gleki> sa 07:49 < zipcpi> x1 fi x2 fe x3 sajgau gi'a visygau gi'a zgagau gi'a selsaugau gi'a me'au lo simsa 07:49 < gleki> i ku'i ca lo donri cu jarco lo ka presidente 07:50 < gleki> remove visygau and replace with gasygau 07:50 < zipcpi> That's what la a'ernai defined it as 07:50 < gleki> things can be shown to blind people 07:50 < zipcpi> Yeah... 07:51 < cliva> vi'o 07:51 < gleki> ue 07:51 < gleki> coi la cliva mi'e lo co'a cliva 07:51 < cliva> coi co'o 07:52 < cliva> ta'o 07:52 < cliva> spagetufa: bi'ylga 07:52 < spagetufa> SyntaxError: Expected [.\t\n\r?!\u0020] or end of input but "b" found. 07:52 < cliva> y 07:52 * cliva cu xruti lo ka zvati lo ctino 07:52 < zipcpi> What would you want that to mean? 07:53 < cliva> cnino ke zi'evla tarmi vau sa'u 07:53 < zipcpi> u'i 07:54 < zipcpi> ro vlatai cu ei se snigau vau zo'o 07:55 < cliva> .ei zasti fa su proga poi ki lo cunso valsi cu cupra je cu snigau 07:57 < zipcpi> markenofi'e: x1 is the destroyer of worlds x2 containing cats x3 owned by x4 under conditions x5 with epistemology x6 07:57 < cliva> mu'a ie .u'i 07:57 < cliva> .i lo nu ba'e cupra zo'u la'e di'e zasti http://zbaga.ax.lt/thethat.html?big .i ku'i pilno ge nai lo gerna gi li malrxaka 07:58 < zipcpi> si'au na akti 07:59 < cliva> spofu va'o la romge 07:59 < cliva> .i ri'a bo mi lazni lo ka pilno la strelka fancu ku ne la djavaskripti 08:00 < cliva> .i la faglo'u zo'u na spofu 08:00 < zipcpi> ie 08:00 < zipcpi> za'a 08:17 < zipcpi> ei co'o 08:24 < zipcpi> Other idea on fast/slow clocks; build it off {xa'o/za'o}. fast = "It's already showing 1:00 (when it's only 0:55)"; slow = "It's still showing 1:00 (when it's 1:05)" 08:25 < zipcpi> Dang this dimension confusion 08:27 < zipcpi> xa'o has no rafsi though so we have to cheat by using wrong hyphens 08:30 < zipcpi> Maybe just {julbavja'o}; {julpruja'o} 08:32 < zipcpi> ca'e ra'oi ja'o co'a rafsi zo arco jenai zo jarco iau zo'oru'e 08:45 < gleki> interesting that Interlingua has many more sentence in Tatoeba but it has only 4 contributors. We have 18 but still we lag behind. 08:56 < gleki> the second day LMW is up and nothing is happening to it 09:00 * nuzba @uitki: Ложбан и самогоноварение - http://mw.lojban.org/papri/%D0%9B%D0%BE%D0%B6%D0%B1%D0%B0%D0%BD_%D0%B8_%D1%81%D0%B0%D0%BC%D0%BE%D0%B3%D0%BE%D0%BD%D0%BE%D0%B2%D0%B0%D1%80%D0%B5%D0%BD%D0%B8%D0%B5 by Gleki [http://bit.ly/1HbhsDY] 11:29 * mumble jorne co'u fa la'o fa ctefaho fa 11:29 * mumble jorne co'a fa la'o fa ctefaho fa 12:22 * ctefaho just ran into the Swedish version of lojban.org 12:23 < ctefaho> it's... 12:23 < ctefaho> weird 12:34 < bigcentaur> .i ma cizra do 12:36 < ctefaho> primarily the "Vid fyra års ålder dog Anders mormor" ambiguity example which is *not* ambigious at all in Swedish 12:37 < ctefaho> ...which I think is supposed to be instead of the "He's left" example 12:38 < bigcentaur> .i mi na pu zgana de'u 12:38 < ctefaho> maybe "Anders mormor dog vid fyra års ålder" would be a bit ambigious 12:38 < ctefaho> ambiguous* 12:38 < ctefaho> actually no lol that example sentence is complete fail 12:40 < ctefaho> unless you are speaking very bad swedish 12:41 * ctefaho pokes gleki with a stick 13:46 < Amnadya> ma g 13:50 < Amnadya> ma nuzba 14:05 < zipcpi> gleki: {kajde} might be good for "alarm" 14:05 < Amnadya> coi. attàwîl. 14:05 < zipcpi> coi 14:08 < zipcpi> More crazy CKTJ ideas: {ru'e} -> {rei}. Let {xei} be hex digit fourteen; we don't need to book two monosyllabic cmavo for it :p 14:12 < zahlman> what's CKTJ? 14:13 < zipcpi> "ce ki tau jau", a dialect/fork that swaps certain common, longer cmavo with rarer, shorter cmavo 14:14 < zipcpi> http://mw.lojban.org/papri/ce_ki_tau_jau 14:14 < zahlman> named for certain of the swaps, I presume? 14:14 < zipcpi> Yep 14:17 < Amnadya> la cektaujbo 14:17 < zipcpi> I defined {tcekitaujau} and {tcekitaujei} 14:18 < zipcpi> zei: cektaujbo 14:18 < mensi> creka zei tanru zei lojbo 14:18 < zipcpi> u'i 14:20 * nuzba @uitki: ce ki tau jau - http://mw.lojban.org/papri/ce_ki_tau_jau by Spheniscine [http://bit.ly/1HecA3U] 15:00 < durka42> o hai zipcpi I'm in ur definishunz fixin ur linkz 15:03 < zipcpi> Hi 15:03 < durka42> oh there aren't actually links in the notes field 15:03 < durka42> that's lame 15:03 < zipcpi> Yeah 15:04 < durka42> it only calls Wiki::mini() 15:04 < durka42> not Wiki::interpret() 15:05 < durka42> I wonder if it strips tags... 15:07 < durka42> o_O it does not 15:08 < durka42> fuck 15:08 < durka42> zipcpi: http://jbovlaste.lojban.org/dict/tcekitaujei :) 15:08 < zipcpi> Yeah I saw 15:08 < durka42> yeahhhhh 15:08 < zipcpi> That's kinda dangerous if a user can just inject arbitrary JS 15:08 < durka42> it should strip all HTML tags 15:09 < durka42> man JVS is so insecure 15:09 < durka42> were you around when I managed to log in as officialdata by accident 15:09 < zipcpi> No lol 15:09 < durka42> I just constructed my own session cookie 15:09 < zipcpi> Mm cookies 15:09 < durka42> mmhmm 15:09 < zipcpi> No idea how that idiom came about 15:27 < durka42> so after all that we could have used
for linebreaks in the notes, heh 15:27 < durka42> but $\\$ will still work after I submit a patch to strip tags 15:27 < zipcpi> Scale-breaking madness: http://www.quantum-munich.de/research/negative-absolute-temperature/ 15:29 < zipcpi> +0 != -0. Bye bye math :p 15:31 < durka42> psh, already true in IEEE794 floating point 15:31 < zipcpi> lol 15:33 < f8l> Integer underflow. :-Þ 15:34 < zipcpi> Yep, they integer underflowed the universe 15:46 < f8l> ‘exited state’. I guess I’ve found a typo. 16:14 < ctefaho> camxes: cfladri 16:14 < camxes> SyntaxError: Expected [.\t\n\r?! ] but "c" found. 16:14 < ctefaho> :( 16:14 < Amnadya> slinkuhi 16:16 < ctefaho> how exactly? 16:16 < Amnadya> because pacfladri is a lujvo 16:17 < ctefaho> right 16:18 < ctefaho> fitting how that became evil law causing sadness 16:27 < ctefaho> camxes: cflaidrei 16:27 < camxes> (cflaidrei VAU) 16:27 < ctefaho> ui 16:33 < Amnadya> eqtal ki. do gleki 16:38 < Amnadya> (eqtalûb) 17:32 * nuzba @redstatist: lmao remember lojban [http://bit.ly/1gmHCxG] 17:32 < durka42> mi morji tu'a la .lojban. 17:49 * mumble jorne co'u fa la'o fa ctefaho fa 19:34 * nuzba @fotono: lo #lojban zantufa tricu cu banro http://guskant.github.io/gerna_cipra/zantufa_farvi.svg .i e'u te ciski lo cnino ke jonma'o ciste fo la'e di'e http://mw.lojban.org/papri/zantufa_jonma%27o [http://bit.ly/1HdQRpI] 19:42 * nuzba @fotono: su lo #lojban zantufa tricu cu banro http://guskant.github.io/gerna_cipra/zantufa_farvi.svg i e'u te ciksi lo cnino ke jonma'o ciste fo la'e di'e http://mw.lojban.org/papri/zantufa_jonma%27o [http://bit.ly/1HdSx2s] 22:58 * nuzba @oka_iu_tcan: @oka_iu_tcan ロジバン [http://bit.ly/1J3sfRQ] 23:15 < cliva> zantufa: gijetigitigiti 23:15 < zantufa> ([{gi je} ti { } GIhI] VAU) 23:15 < cliva> ketitibroda 23:15 < cliva> zantufa: ketitibroda 23:15 < zantufa> ([ke ti ti KEhE] CU [broda VAU]) 23:15 < cliva> zantufa: ketitike'ejeputa 23:15 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "j" cu se tolcri 23:16 < cliva> zantufa: ketitike'ejeketa 23:16 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "j" cu se tolcri 23:16 < cliva> zantufa: pukujeca 23:16 < zantufa> ([{pu ku} {je }] VAU) 23:16 < durka42> oh dear 23:16 < cliva> nu tinke sai 23:16 < durka42> .u'i 23:17 < durka42> xu do sanji lo nabmi pe lo nu zo gi gau za'upamei 23:17 < durka42> zantufa: gije gija ti gi ti gi ti 23:17 < zantufa> minajimpe: .i kanpe tu'a [uU] .onai [yY] .i ku'i "i" cu se tolcri 23:18 < cliva> iep 23:18 < durka42> je'e 23:18 < cliva> spagetufa: +cktj su mei'e ti ta tu mi do di'e cu mo 23:18 < spagetufa> ([FA { KU}] [cu {mo VAU}]) 23:19 < durka42> ma nau zbusufukai 23:20 < cliva> si'au G zbusufukairai 23:23 < cliva> lo zunle nu jmaji zo'u tu'a PEG boi po'o na banzu lo nu pagbu lo te pruce 23:23 < cliva> spagetufa: tijetajatu 23:23 < spagetufa> ([FA { ja tu}] VAU) 23:23 < cliva> ^ ti se gasnu se pi'o lo djavaskripti 23:25 < durka42> .i'e 23:25 < durka42> mi to'e djica lo nu PEG setca zo fa po'onai 23:25 < durka42> lo nu setca lo vrici cu te sampla pe'i 23:26 < cliva> ie .i lo me PEG co krasi be la .altatufa cu to'e jinsa ri'a ku 23:26 < durka42> ba'asai 23:27 < cliva> (to ku'i lo me la spagetufa moi na ba'e mutce lo ka jismau toi) 23:27 < durka42> .u'i --- Day changed Wed Jul 08 2015 00:01 < gleki> {grafu} and {fardu} are useful 00:48 < gleki> "middleware" 00:48 < gleki> ma xe fanva 01:02 * nuzba @patrickDurusau: Today's Special on Universal Languages #Lojban #logic http://ow.ly/PiJv9 [http://bit.ly/1IICBuE] 01:20 * nuzba @464161niftg: la xarju .iu sei fasybau fanva xamsku cu ctuca fo la lojban [http://bit.ly/1G6zvdu] 02:20 * nuzba @okau_junpe: pe'i xamgu fa la sutysisku .i lo djica co facki be lo sampu datni be lo valsi cu se prali lo se manci lo ka pilno http://mw.lojban.org/extensions/ilmentufa/i/ [http://bit.ly/1NNIeam] 02:58 < PrincessBecca> au mi klama lo gusta poi se cusku fo la'o zoi http://makeupbycharada.com.au/?p=500 zoi 03:08 < gleki> poi se skicu 03:08 < gleki> jb: skicu 03:08 < mensi> skicu = skicu — x1(entity) tells about, describes x2(entity) to x3(entity) with description x4(proposition) 03:08 < mensi> :lo skicu — describer. lo se skicu — described, told about. lo ve skicu — description. 03:08 < mensi> :le gidva pu skicu fi mi'a fe lo nu mi'a cliva ca lo lerci vau lo nu lo aftobuso pu'o tsuku — The guide explained to us 03:08 < mensi> that we would leave later in that the bus hadn't yet arrived. 03:08 < mensi> :mi pu skicu lo cinri lisri lo mi panzi — I told an interesting story to my child. 03:08 < mensi> :Comment: skicu is to tell, cusku is to say, tavla is to talk, cusku lo notci is to notify. 03:08 < mensi> :Related words: lisri, tavla 03:09 < gleki> skicu is to tell, cusku is to say, tavla is to talk, 03:09 < PrincessBecca> je'e 03:09 < gleki> cusku lo notci is to notify. 03:10 < gleki> i noticed that some esperanto speakers dont see the difference between "diri" and "paroli". well, in this la bangu dictionary i added the explicit annotation 03:13 < PrincessBecca> i think it's less about not seeing the difference, more about trying to talk spontaneously and messing up while learning. 03:14 < PrincessBecca> a lot of komencantoj de Esperanto don't realize that "paroli" exists, or that "diri" exists, and generalize. the same sort of happened with me. 04:11 < Amnadya> cahe rinsa - attàwîl. 04:12 < gleki> ma nuzba lo nu tinkygau ibu 04:12 < gleki> vau coi la amnadiia 04:13 < Amnadya> pamai ba lonu mohu xrukla lomi zdani cu retsku sohi da la kixadas 04:14 < gleki> do ko'oiru'e na forgese si co'u morji zo'ei la mupli 04:14 < Amnadya> i aho ky tugni filoduhu spofu 04:14 < Amnadya> mina forgese 04:14 < gleki> gi'e zukte noda zo'o 04:14 < gleki> sa i go'oi tugni gi'e zukte no da 04:19 * mumble jorne co'a fa la'o fa ctefaho fa 04:22 < Amnadya> fanza falonu ky na pu tugni filoduhu ihy claxu zo coha je zo cohu vuho poi sumtcita to genrkasu toi 04:22 < Amnadya> ikuhu jaha claxu 04:23 < gleki> i ja'o mo i ja'o fa ei lo selbri bangu cu jinga 04:24 < gleki> i kuhi do jinvi fe laha lo duhu lo itkuiile cu ei se pilno bangu i kuhi xu jaha gohi ei 04:24 < Amnadya> jaho bahepehi ei jmina lo re genrkasu 04:25 < gleki> la lojban ca spoja ki'u tu'a ma'oi bai 04:25 < Amnadya> ri na pagbu ihy 04:25 < Amnadya> ina nabmi ihy 04:26 < gleki> lo rafsi poi simsa ma'oi bai cu pagbu kuhi 04:26 < Amnadya> ie ikuhi nohe simsa lo lojbo loka spofu 04:27 < gleki> fau lo nu claxu so'o lo rafsi pe ma'oi bai vau na ku xu plixau da'i fa lo nu vimcu ro lo simsa rafsi gi'e pilno lo drata tutci 04:28 < Amnadya> li genrkasu cu vajni je jicmu pagbu la itkuhilu 04:28 < Amnadya> lo* 04:29 < Amnadya> i jiha frili pagbu 04:30 < Amnadya> lonu mulgau noi nu jmina jihi ci cnino genrkasu cu xagmau lo nu vimcu roda 04:30 < gleki> i xu naku cumki fa lonu ba sarcu fa lo nu jmina lo pa ki'o simsa rafsi 04:30 < Ilmen> .u'e ti'e la .valman. noi bangu zo'u lo panra be zo je cu selbri 04:31 < gleki> la valman na slabu mi 04:31 < Amnadya> xu lo preti cu srana lu fiho kibystu lihu 04:32 < gleki> na go'i i srana lu fi'o gerku li'u 04:32 < Amnadya> lahasai tau li kiho na sarcu 04:46 < gleki> mi pu zo'onmo stidi lo nu zo li basti zo cu i a'o jai gau spofu fai no da 04:46 < gleki> si zasti 05:14 < gleki> en: tcanaca 05:14 < mensi> tcanaca = x1 is the node of present concern/is the current position in graph x3 (not necessarily oriented; notice the 05:14 < mensi> terbri number), connected (possibly indirectly) to/with other nodes x2 (possibly incomplete set) via (oriented) paths x4 05:14 < mensi> |>>> "Current position" is in thought. See also: grafu, tcanaba, tcanapu. |>>> krtisfranks 05:14 < gleki> en: tcanapu 05:14 < mensi> tcanapu = x1 (node/vertex/station) is backward of/along from x2 in oriented graph x3 (graph with orientation) using 05:14 < mensi> oriented edge path x4 (ordered sequence of ordered pairs; oriented edges). The path from x2 to x1 along the 05:14 < mensi> (now-unoriented version of the) edges used in x4 is counter/against/upstream of the orientation of x3 (and/or given by 05:14 < mensi> the oriented version of x4). |>>> Orientation is given from/by x3, so only an edge need be submitted if x3 is fully 05:14 < mensi> specified; if x3 is not fully specified, x4 can take the burden of specifying the orientation and subgraph of particular 05:14 < mensi> interest (namely, the two vertices x1 and x2, intervening vertices along the path, and the orientation of the given 05:14 < mensi> edges connecti... 05:14 < mensi> [mo'u se katna] http://jbovlaste.lojban.org/dict/tcanapu 05:15 < gleki> looks like someone is living in some alien world and is desperately trying to explain it using human words 05:22 < Amnadya> xu do kakne loka stika la mabka tutci jahe lonu lo sance vreji cu pagbu 05:22 < gleki> oi oi oi 05:23 < gleki> i i'a i ko skicu 05:23 < gleki> i tai ma pagbu 05:23 < Amnadya> muha lo preti cu se pagbu lo se bacru jufra 05:24 < Amnadya> i jahebo kahe tirna jimpe ecre 05:26 < gleki> ua 05:26 < gleki> xm 05:26 < gleki> en: xm 05:26 < mensi> 48 da se tolcri: bixma'u, caxmati, gaxmolgle, laxma'i, rokybixmivborvelvi'u, sexmti, tixmamta, xadbixma'u, zansaxmu'e, 05:26 < mensi> banbuxumu, bangaxumu, bankuxumu, banmuxumu, banpuxumu, bansuxumu, bantuxumu, banxumu'a, banxumubu, banxumucu, banxumudu, 05:26 < mensi> banxumu'e, banxumufu, banxumugu, banxumuje, banxumuju, banxumuke, banxumuku, banxumulu, banxumumu, banxumunu, banxumu'o, 05:26 < mensi> banxumupu, banxumuru, banxumusu, banxumutu, banxumu'u, banxumuve, banxumuvu, banxumuxe, banxumuxu, banxumuzu, banzupuzu, 05:26 < mensi> baurnuxumu, besmamta, citnau, kanbyma'i, mivyvelvicyvauro'i, ro'irbi'o 05:26 < Amnadya> valsi - valsi, vs sance - valsi 05:27 < gleki> mi ba pensi 05:27 < Amnadya> i liha mi ja makaunoi pilno cu bilga loka finti lo sance vreji 05:27 < gleki> i mi lazni 05:27 < Amnadya> iha 05:27 < Amnadya> kihe pensi 05:28 < gleki> i ca ti mi na morjilo du'u lo krasi be la cilretci i ie ba'e mi pu finti 05:29 < gleki> i ua mi pu pilno la'oi JQuery 05:29 < gleki> i mi nelci i ie 05:29 < valsi> 38 results: sance, valsi, bangu, cmavo, cmene, gadri, gismu, jufra, rafsi, slaka… 05:33 < Amnadya> sihau lo prosumti cu nandyrai pagbu la itkuhile 05:33 < gleki> uesai 05:34 < gleki> i mi pu jinvi lo du'u rai frili 05:35 < Amnadya> lo sampu me ra cu frili ikuhi lo rafsi javni cu cfipu 05:35 < Amnadya> i na panra lo fadni selbri 05:36 < Amnadya> gaha do ma nandyrai 05:39 < gleki> lo ni rafsi 05:39 < gleki> e lonu lo rafsi cu cnegau lo tarmi be lo drata rafsi 05:39 < Amnadya> juho nandu 05:41 < Amnadya> ilaha la gredile 5 moi zohu lo nahe lazni remna cu nitcu tau zeha lo masti be li ci loka mohicli mohuku 05:42 < gleki> la'a krinu lo nu mi na ze'a lanli je tadni ibu 05:44 < Amnadya> y oi laha li ci na drani i li suho xa 05:44 < Amnadya> to dohe lo fadni remna toi 05:44 < gleki> xm 05:44 < gleki> i zo ze'a mabla 05:44 < gleki> i au ma'oi roi 05:44 < Amnadya> i rafsi fa 1728 da 05:45 < gleki> rafsi i'a i ku'i lo rafsi cu cnegau lo rafsi 05:45 < Amnadya> 1728 / 100 = 18 ca ro djedi 05:46 < Amnadya> i jaho jaha cumki fa tau zeha li ci 05:50 < Amnadya> i xamgu cipra i lo snada be tau la gredile mumoi cu lahasai snada tau lo drata gredile noi plujymeha 05:53 < Amnadya> i mani cinri do falosiho mi ciska pa lojbo brafiha jecu vecnu ri 05:55 < gleki> ainai mi te vecnu i sa'u mi te vecnu so'u da 05:55 < Amnadya> mi djuno 05:55 < gleki> i ku'i lo uenzi zo'u au no loujvo cu se pilno 05:55 < Amnadya> lol 05:55 < gleki> i sa'u mi na te vecnu lo cukta 05:55 < gleki> i mi rusko 05:55 < gleki> i lo'e rusko cu pilno lo torent 05:56 < gleki> i lo nu mi jmive zo'u pa proga pu'i se vecnu fi mi 05:56 < gleki> ije ze'a lo barda mi te vecnu no cukta 05:58 < gleki> i mi bilga lo ka stika la'e la alis i ku'imi na co'ei oi se'i 05:58 < Amnadya> lo preti pu nohe srana do noi pamei prenu kuho je srana fe rahu lonu da orginale lojbo brafiha je kahe se vecnu i nuho fasnu 05:58 < Amnadya> i sihau nohe cinri do 06:00 < gleki> so'i cinri cfika nu'o se tcidu mi i mi na krici lo nu do ka'e finti lo cinri i fau lo nu do kakne vau y zo'u y na djuno u'i 06:00 < Amnadya> ko pensi do pohonai ko pensi lonu kahe cinri lo se cinri be lo jbobau beho poi zukpei loka tadni 06:01 < Amnadya> i lonu zasti fa suho brafiha kahe sidju 06:01 < Amnadya> fi loka jdice 06:02 < Amnadya> i mi caho gunka ci brafiha i mi friti loka pa lo ci brafiha cu se ciska bau lo jbobau 06:33 < Amnadya> noda frati... 06:35 * nuzba @Rodericus: @lirondos Por cierto, por si te interesa… Hay al menos una lengua en que esos juegos de palabras son imposibles: https://es.wikipedia.org/wiki/Lojban [http://bit.ly/1Hg7hhq] 06:38 < Ilmen> nodamafrati 06:48 < Amnadya> mi nakahe copypaste lo pu selsku sepiho lo fonxa 06:51 < niftg> .u'i zo copypaste ka'e se smuni zo'ei zo fukpu'i 06:52 < Ilmen> je'e .i ja'o mo'urívbi lokagánse fami 06:54 * nuzba @mochirial: http://jbovlaste.lojban.org/lookup.pl?Form=lookup.pl2&Database=*&Query=dinso 試験的gismuながら恐竜、正確には古代の大型爬虫類、ってありますね。中国語で恐竜は恐龙kǒnglóngなので英語のdinosaurまんまですね。 [http://bit.ly/1HMIkAs] 07:00 < gleki> Amnadya: re'i re'i 07:00 < gleki> Amnadya: ke'u mu'a zo brafi'a zo'u ki'u ma finti lo lujvo 07:01 < gleki> Amnadya: i au do jvocaufi'i 07:01 < niftg> .uo kavbu sepiho lo degji tahi lo nu degjai ku lo ckuliki sediho lo ri tuple .isekihubo mi na kakne co samciha yhybu noi cmamau 07:02 < Amnadya> pehi cizra se stidi doi la gleki 07:02 < Amnadya> tau lo gismu ku poho na banzu i tokpona simsa 07:03 < gleki> zo brafi'a zo'u ki'u ma 07:03 < gleki> y 07:03 < Amnadya> steci klesi lo cfika 07:04 < gleki> zo bracfika e'u 07:04 < gleki> zo bramlatu e zo brazamsi i'a ru'e e zo bracfika 07:06 < gleki> sa'ai zo braxamsi le'ai 07:07 < niftg> di'a apostrof zei samci'a kakne .i .a'o na pu'i raktu zunti lo ralju nu casnu 07:07 < niftg> fa ro tu'a mi 07:08 < xajmi> coi .i mi cmevla xajmi .i mi tolsti lojban 07:08 < Ilmen> coi ckulikykalte 07:09 < niftg> coi za'u do mi'e pu'o ckulikycatra 07:10 < gleki> do famyma'o si terminatore 07:11 < samcinki> coi jai se tepsne be lo ro ckuliki 07:11 < gleki> Amnadya: e'uzo'oru'e do fanva la'e lu lo certu ce la margarita li'u lo lojbo 07:11 < gleki> i bracfika 07:12 < Ilmen> .a'e nai doi ka glare kei ko lo mi vacri cu cliva 07:12 < xajmi> mi tersmu nai "ckulikykalte" 07:13 < Ilmen> ckuliki'ykalte 07:13 < gleki> xu do na te smuni lu ckulikykalte li'u 07:14 < gleki> i li'a lo litki cukta cu rirci 07:14 < gleki> i na frili fa lo nu kalte lo litki cukta 07:15 < gleki> i e'e do troci 07:15 < niftg> .u'i snuti lujvo 07:16 < durka42> zo fardu gimkamsmikezypro zo pardu 07:17 < gleki> ie jungau 07:17 < gleki> sisi 07:17 < gleki> ei jungau 07:17 < durka42> vi'o 07:19 < gleki> i lo gismu canlu cu barda 07:19 < durka42> je'u 07:19 < niftg> .ua ma'i la camxes lo nu vimcu ybu noi na jai sarcu cu na sarcu lo nu lujvo vau vau ta'o 07:20 < durka42> ku'i xu roroi sarcu fa lo nu lo cnino gismu na janli ba'e ro nalselpilno gismu 07:21 < gleki> y 07:21 < gleki> ei le cecmu cu jdice 07:21 < durka42> go'i 07:21 < durka42> ta'o xu zo pardu cu se pilno 07:21 < gleki> i mu'a pe'i lo valsi be zo fardu ka'e vajni ja plixau 07:21 < durka42> la korpora cu .arco no da 07:22 < gleki> zo pardu puzi jai se stidi 07:22 < gleki> i pe'i zo pardu jai frili fai lo ka jai se morji 07:22 < gleki> i ku'i zo fardu cu lobykai valsi 07:22 < gleki> i seki'u bo pe'i na nabmi fa lonu cuxna lo drata tarmi 07:22 < gleki> i zo fardi mu'a 07:23 < gleki> vlaste: fardi (gimkai) 07:23 < vlaste> no results. http://vlasisku.alexburka.com/fardi 07:23 < gleki> y 07:23 < gleki> ta'i ma 07:23 < gleki> zukte lo cipra 07:23 < gleki> se pi'o la vlaste 07:25 < gleki> vlaste: fardi (gimka) 07:25 < vlaste> zo fardi gimkamsmikezypro no catni gismu .i zo fardi gimkamsmikezypro no cipra gismu 07:25 < gleki> vlaste: farde (gimka) 07:25 < vlaste> zo farde gimkamsmikezypro no catni gismu .i zo farde gimkamsmikezypro no cipra gismu 07:25 < gleki> lo canlu cu barda 07:27 < xajmi> xodor: https://chudleycannons.files.wordpress.com/2011/12/missinghodor.jpg 07:32 < xajmi> cmigau porpi 07:32 < xajmi> coi? 07:32 < xajmi> co'o 07:33 < gleki> co'oi 08:43 < ctefaho> Question of the day: do you guys pronounce "iancu" with or without denpa-bu before it? 08:59 < quintus> How would you translate english "even" in the sense of "emotive for exaspiration" as in "I don't even know what you're talking about"? I'm split between .u'inai and .ainai and .a'enai and .o'unai and .uanai 09:00 < quintus> of course any attitudinal would apply to the entire sentence 09:00 < quintus> not any individual word 09:01 < dutchie> quintus: attitudinals apply to the word 09:01 < dutchie> mi na djuno .uinai 09:01 < quintus> would apply to the whole sentence 09:02 < quintus> in {mi na djuno .uinai}, would .uinai not refer to the entire bridi? 09:02 < dutchie> no, that'd be {.uinai mi na djuno} or {mi na djuno vau .uinai} 09:02 < quintus> I thought vau+UI and central bridi+UI were identical 09:03 < dutchie> https://lojban.github.io/cll/13/9/ 09:03 < quintus> reading currently 09:03 < dutchie> "the indicator (or group) attaches to the word immediately to its left, and indicates that the attitude is being expressed concerning the object or concept to which the word refers" 09:04 < Ilmen> ctefaho: I'd usually pronounce it without leading glottal stop, as the "i" would already be a consonant (glide) 09:05 < ctefaho> Ilmen: je'e. and in "lo jai bai iancu"?;) 09:06 < dutchie> as for your actual question, i'm not sure. possibly {ba'e} maybe? 09:06 < Ilmen> camxes: baiiancu 09:06 < camxes> SyntaxError: Expected [.\t\n\r?! ] but "b" found. 09:07 < quintus> I don't remember where I read it, but I feel pretty sure that {.i mi dunda .ui ti do} and {.i .ui mi dunda ti do} and {.i mi dunda ti do vau .ui} were identical. I think the justification given was that since the main bridi denotes the relationship between the sumti, indicating emotions about the connection between the sumti is equivalent to indicating emotions about the entire bridi. 09:07 < quintus> Especially considering the meanings of "bridi" 09:07 < quintus> But yes this is not particularly relevant to my initial question 09:07 < Ilmen> Hm, I don't see why there would be a problem pronouncing it [baɪjánʃu], but otherwise you can add a ʔ here 09:07 < quintus> camxes: ba'e 09:07 < camxes> SyntaxError: Expected [.\t\n\r?! ] but "b" found. 09:07 < Ilmen> @ ctefaho 09:07 < quintus> er, what's the bot command for jbo->en? 09:08 < dutchie> en:ba'e 09:08 < mensi> ba'e = [BAhE] forethought emphasis indicator; indicates next word is especially emphasized. 09:08 < ctefaho> Ilmen: well if [j] is legal for both bai and ian we get a double consonant 09:08 < quintus> .ua 09:09 < ctefaho> so I am gonna put denpa bu everywhere again now;) 09:09 < Ilmen> ctefaho: same with {ii}. [j] and [i] are allophones of the /i/ phoneme, you can use [j] to avoide double -ii- being pronounced as a long [i] 09:09 < dutchie> .u'i 09:09 < quintus> I feel like ba'e would be relevant in "So easy, even an end-user could use it!" but less so in "I don't even know what it is!" 09:10 < Ilmen> I guess you also should avoid pronouncing "ii" as a double [j] :) @ctefaho 09:10 < ldlework> quintus: dunno I think bu'u djuno is similar 09:10 < Ilmen> (Anyway I usually don't pronounce {ai} as [aj]) 09:11 < ctefaho> Ilmen: I more prefer [aɪ] and [aʊ] myself 09:11 < ldlework> quintus: though, it does seem like there is a sense of "sufficiency" with "even" in the english 09:11 < ldlework> Like "I don't even know what it means much less what it is for!" 09:11 < quintus> "much less" is another interesting construction 09:12 < ldlework> I wonder if ba'u is relevant 09:12 < quintus> I wonder if "even" and "much less" are used often enough to denote a construction 09:12 < ldlework> mi na ba'u djuno 09:12 < ldlework> That I don't know, would be an understatement! 09:12 < Ilmen> en: .o'ocu'i 09:12 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 09:12 < quintus> z.b. "...don't even...much less..." 09:12 < ldlework> err rather, ba'unai 09:12 < Ilmen> en: o'ocu'i 09:12 < mensi> o'ocu'i = [UI*1] attitudinal: patience - mere tolerance - anger. |>>> officialdata 09:12 < ldlework> en: ba'u 09:12 < mensi> ba'u = [UI3] discursive: exaggeration - accuracy - understatement. |>>> See also satci, dukse. |>>> 09:12 < mensi> officialdata 09:12 < quintus> "neither...nor..." 09:13 < ldlework> .i ba'u nai mi na djuno "I don't even know" 09:13 < ldlework> Still not quite the same 09:14 <@xalbo> quintus: I'm fairly certain that an attitudinal after the main selbri just binds to the selbri itself. Certainly in practice that's how it's used, although the difference is usually pretty subtle. It's possible that CLL says otherwise, but I doubt it. 09:14 < ldlework> What about 09:14 < ldlework> en: be'u 09:14 < mensi> be'u = [UI5] attitudinal modifier: lack/need - presence/satisfaction - satiation. |>>> See also claxu, nitcu, mansa. 09:14 < mensi> |>>> officialdata 09:14 < ldlework> .i be'u mi na djuno 09:14 < ldlework> I don't know, and that's not enough. 09:14 <@xalbo> I've seen {ji'a} or {ji'a sai} used for other meanings of "even", and this might be the same, really. 09:15 < quintus> xalbo: I have no idea where I saw/read that; perhaps it was a non-xuncku more practical tutorial 09:15 < durka42> {mau} might be useful for the "much less" part 09:15 < quintus> but I'm certain I have 09:15 < ldlework> xalbo: I really think "I don't even know" is measure of sufficiency for some other thing 09:15 < ldlework> "How could I convince him. I don't even know him." 09:16 < quintus> I feel like that last one is slightly different 09:16 < gleki> I don't even know what you're talking about <-- do'a would be fine here imo 09:16 <@xalbo> {.i mi na djuno ji'a sai} being "Among the things I don't do, we can even go so far as to add knowing to them.", or "I don't even know." 09:16 < ldlework> "How can I give a sufficient answer? I don't even know what you're talking about." 09:16 < ldlework> Its totally a measure of sufficiency, pe'i 09:16 < quintus> My initial question was closer to gleki's use 09:16 < durka42> I don't even know him => I don't know him, and if I did it wouldn't help much 09:16 < quintus> it turns out words have multiple uses in natlangs and that it is hard to distinguish between them 09:17 < quintus> !quotable 09:17 < durka42> true fact :) 09:17 <@xalbo> Supercollider? I barely know 'er! 09:17 < ldlework> durka42: I don't see the context in which both of your sentences are swappable 09:17 < gleki> almost any English "cmavo" has numerous meanings 09:18 < dutchie> often as a "gismu" in its own right! 09:18 < dutchie> is/have etc 09:19 < ldlework> well at least brivla 09:20 < quintus> what's the difference between a brivla and a gismu? 09:20 < ldlework> brivla is the class of words that gismu belong to 09:20 < quintus> brivla can be combinations of gismu? xuncku? 09:20 < ldlework> along with lujvo and fu'ivla and so on 09:20 < ldlework> they are all predicates 09:21 < ldlework> of different morphological form 09:21 < quintus> gismu is the subset of brivla, as is lujvo and fu'ivla? 09:21 <@xalbo> {xuncku} is a lujvo, which is a brivla but doesn't gismu. The other class of brivla that aren't gismu are the fu'ivla, like {.irci} 09:21 < ldlework> gismu, lujvo, fu'ivla, zi'evla each denote a morphologocal type that belong to the class called "brivla" or "predicate words" 09:21 <@xalbo> quintus: Yes. 09:21 < quintus> okay thanks, I'm a bit rusty 09:22 * gleki and we havent even gotten to selbri/brivla distinction/confusion 09:23 < ldlework> That's not too hard 09:23 < ldlework> when we say "predicate words" 09:23 < ctefaho> ENTER BRODA 09:23 * ctefaho hides 09:23 < ldlework> we mean that, they produce grammatical "predicate terms" 09:24 < quintus> brod[aiueo] 09:24 < ldlework> So the words all share the property that when used, they produce "predicate terms" or "selbri", in the grammar, rather than the morphology. 09:24 < quintus> okay Saussure 09:25 < gleki> are you saus-sure? 09:25 < quintus> saus-positive 09:25 <@xalbo> quintus: "brivla" is a class of words based on how they're spelled (end in a consonant, have a least one consonant cluster). "selbri" is a what a word or a construct is doing in a sentence (acting like a verb). 09:25 < ldlework> Which can be shown by the fact that selbri can be composed from multiple words 09:26 < quintus> lo brivla cu brivla .i lo selbri cu se brivla 09:26 <@xalbo> The words {xagji} is always a brivla. The word {go'i} isn't. But in {mi xagji}, {xagji} acts as a selbri. In {mi go'i}, {go'i} acts as a selbri. 09:26 < dutchie> xalbo: you mean ending in a vowel 09:27 <@xalbo> dutchie: I do indeed. .o'a nai ro'a 09:27 < quintus> I'm out to lunch but I'll be back 09:28 < quintus> o/ 09:31 < gleki> all DerSaidjin's issues found in la bangu dictionary are important, have to be and will be dealt with 09:34 < Ilmen> en: selbrivla 09:34 < mensi> selbrivla [< se bridi valsi ≈ 2nd conversion predicate word] = x1 is a lexically defined predicate word (predicate 09:34 < mensi> particles included), signifying relation x2 (n-ary property) in language x3 |>>> Encompasses gismu + lujvo + fu'ivla + 09:34 < mensi> brika'i. |>>> Ilmen 09:35 < gleki> jb: selbrivla 09:35 < mensi> selbrivla [< se bridi valsi ≈ Se* bridi* valsi*] = selbrivla — x1(text) is a morphologically defined predicate word, 09:35 < mensi> verb word 09:35 < mensi> :lo selbrivla — verb word. 09:35 < mensi> :zo mlatu selbrivla — mlatu is a selbrivla. 09:35 < mensi> :Related words: lujvo, gismu, zi'evla, cmavo 10:13 < durka42> whoa, curtis used {ko'au} 10:17 < ctefaho> ^^ 10:18 < durka42> http://jbovlaste.lojban.org/comments.html?valsi=29243;natlangword=0;commentid=2203;definition=0 10:54 < gleki> en: ko'au 10:54 < mensi> ko'au = [BAI] kosmu modal, 1st place with purpose... |>>> See seko'au, kosmu, tezu'e, tepi'o, tefi'e, ki'u |>>> 10:54 < mensi> ctefaho 11:50 * nuzba @s_i_d_a: ca'o zvafa'i .i a'o lo danmo cu bevri loka badri #lojban #tanka [http://bit.ly/1KTSUoq] 11:51 < gleki> ue 11:51 < durka42> mo 11:52 < durka42> ma uenmo do 11:53 < gleki> lo se bevri cu se ckaji 11:54 < gleki> https://github.com/pxg/Node-Rss-to-Twitter 11:55 < gleki> a'o ca da le cecmu co'a pilno la'e di'u lo nu benji lo uitki nuzba la tu'itrer 11:55 < gleki> si tu'iter 11:57 * nuzba @s_i_d_a: ca'o zvafa'i .i a'o lo danmo cu bevri loka badri #lojban #haiku [http://bit.ly/1KTTsuC] 11:57 * nuzba @AscensionArcana: prelife: Before the start of life ۞ zmadu: (Lojban) more; x1 exceeds/is more than x2 in property/quantity x3 (ka/ni) by amount/excess x4. [http://bit.ly/1KTTsL4] 12:07 * nuzba @oka_iu_tcan: mo'u finti / 書けました #はじロジ #ロジバン #lojban 18. 命令 http://cogas.github.io/hajiloji/article/nunctu/18.html 19. 関係詞 http://cogas.github.io/hajiloji/article/nunctu/19.html 20. 縛位詞(BE, BEI類) http://cogas.github.io/hajiloji/article/nunctu/20.html [http://bit.ly/1HhBzAh] 12:51 < Amnadya> coi 12:55 < Amnadya> lo mi cmene mi melbi 12:57 * nuzba @uitki: web Pages in Lojban - http://mw.lojban.org/papri/web_Pages_in_Lojban by Sorpaas [http://bit.ly/1eGKhRO] 13:01 < Amnadya> do mo doi cupleuup 13:09 * nuzba @oka_iu_tcan: えっ、すごい。http://tatoeba.org からjbo-en を検索できる sutsis ができている! http://mw.lojban.org/extensions/ilmentufa/i/muplis/index.html#sisku/ [http://bit.ly/1eGNv7T] 13:09 < durka42> coi Amnadya 13:09 < Amnadya> coi 13:13 * nuzba @oka_iu_tcan: http://mw.lojban.org/extensions/ilmentufa/i/ja/index.html えええええ la sutysisku やばいなwww lujvo 入れたらrafsi分解して gismu列を提示してくれるのかよww [http://bit.ly/1eGOy7L] 13:14 < Amnadya> do bazi klama la doitland 13:14 < Amnadya> iepei 13:14 < mensi> ei mi tugni 13:14 < Amnadya> doitclant 13:16 < Amnadya> la doihotcyguhe 13:16 < durka42> go'i 13:16 < durka42> cliva ca lo xavyplini djedi 13:16 < Amnadya> xu snada tau la duuolingos 13:17 < durka42> toldukse co jundi uinaise'i 13:17 < durka42> mi ka'e cusku lo se du'u mi na pinxe lo birje 13:17 < Amnadya> uinaidai i wo ist die toilette 13:18 < Amnadya> ko cusku tau lo birje 13:18 < ctefaho> Ich habe kein Kartoffeln 13:18 * ctefaho end of german knowledge 13:18 < Amnadya> potatis 13:18 < durka42> Ich das Bier trinkst nicht 13:18 < durka42> trinken* 13:19 < durka42> Mein Deutsch ist kaput 13:19 < Amnadya> ich trinke kein bier 13:19 < durka42> zo'oi kein mo 13:19 < Amnadya> zo no 13:20 < Amnadya> no tricu = kein Baum 13:20 < Amnadya> ija keine Bäume 13:20 < durka42> ua 13:22 < Amnadya> pehipei la duholingos cu zabna 13:24 < Amnadya> tahonai doi la cirko do klama xu la franfurt 13:25 < durka42> mi nelci la duuolingos 13:30 < Amnadya> ma ni lo duhu dahi mi cisj 13:30 < Amnadya> si 13:31 < Amnadya> ciska lo lojbo brafiha je cu vecnu ri 13:31 < Amnadya> cu cinri do 13:33 < durka42> da'i mi .erve lo'e lojbo brafi'a vau .u'i 13:33 < Amnadya> i vahi ruhe xu kanpe lonu lonu suda orginale ke lojbo brafiha je kahe se erve kahe xamgu la jboguhe 13:34 < Amnadya> uhi ui jehe 13:38 < Amnadya> mi xaho pu cusku fi la gleki fe lo seduhu mi caho gunka ci brafiha jecu friti loka pa lo ci brafiha cu se ciska bau lo jbobau 13:39 < Amnadya> i kuhi la gleki na pu se cinri 14:15 < Amnadya> smaji temci 14:28 < durka42> mi sutra sampla 14:28 < durka42> ei mulgau so'i da pu lo nu cliva 14:29 < Ilmen> de'u zo'u jijyni'i .e'u pei lodu lo drata poi jbopre cu kansa fi lo ka ciska 14:30 < Ilmen> ku soi se cpedu 14:30 < Ilmen> coi jundi .i ju'inai tolju'i 14:32 < Amnadya> na platu lonu lo drata cu kansa i puhi lifri lonu lo nu kansihu tau lo cfika cu seljahe lonu masno 14:32 < Ilmen> .i se'i lo nu fa gunka ca pa da fe ci cfika zo'u so'i mei 14:34 < Ilmen> .i je'e do'u lakne fa lonu mi fliba loka jimpe lodu makau .itlokuti lo ba'o se cusku 14:34 < Amnadya> kahe simlu i kuhi naku ro nu gunka cu nu ciska i sohi nu platu cu fasnu i ca lonu mohu platu lo lisri cu coha ciska 14:35 < Amnadya> i lonu capada ciska ci cohe cu nu mutce ikuhi lonu platu ci cohe cu cumki je na dukse loka nandu 14:35 < Ilmen> je'e .aivli 14:36 < Ilmen> di'ai pu'o ju ca'o finti 14:41 < Ilmen> .i mi toltce jeku'i ckaji loka xenru lonu lo pu nu lo jbopre za'umei cu kansi'u loka cfika finti cu se dicra .i ku'i faumlu mi fa lonu lo drata kansa puku ga co'u se cinri gi zukcfu dukse .i pu nandu fa tau lo nu ca ku za'u re da cu kansa sei frimo'i 14:44 < Amnadya> sihau 14:46 < Amnadya> ikuhi lo sidbo pu jalge lo miho nu cfika kansihu 14:46 < Amnadya> i mi pilno su lo sidbo 14:47 < Amnadya> zo maha* 14:48 < Amnadya> i frilymau faloka mi nahe se sidju mulgau 14:48 < Ilmen> lonu sugrute cusecupra cubanzu lonu punapalcau 14:49 < Amnadya> ie 14:50 < Ilmen> to 14:51 < Amnadya> zo to itkuhile 14:52 < Amnadya> i lu gau mi lihu smuni 14:52 < Ilmen> je'e cligau .i'o 14:54 < Ilmen> ni'o .e'i mi livmau seju canci nau 14:54 < Ilmen> co'o balgau 15:06 * nuzba @oka_iu_tcan: lojban、なんだかんだ見てると情報系の人は長続きしてないイメージある [http://bit.ly/1dMT9Vo] 15:11 < ldlework> zo dibdai for treasure? 15:15 < durka42> sure 15:17 < ldlework> je'e 15:18 < ldlework> lujvo in children's books, yay or nay? 15:19 < phma> depends 15:20 < phma> on how advanced the child is at learning the language and if the lujvo is one he's likely to have heard 15:20 < ldlework> its a children's book, so it would be read by multiple children 15:20 < ldlework> which do I use to guage? 15:20 < ldlework> but yeah, I guess you have to for some concepts 15:21 < ldlework> like find, in the sense of sispe'i 15:21 < ldlework> sisku penmi is .. just strange 15:21 < phma> what's it about, and how old are the children? 15:21 < ldlework> but somehow makes perfect sense as a lujvo 15:21 < ldlework> phma: think, like, one of the first book a child "reads" 15:21 < ldlework> so like See Spot Run, or similar 15:25 < phma> I wouldn't have a problem with sispe'i, would a child that age have problems pronouncing or hearing in tolcri? 15:37 * nuzba @kaleidic: @eraserhd You must view Lojban as trying as hard as possible to take fun out of the world. [http://bit.ly/1HcjQeA] 15:48 < Amnadya2> xu la ilmen pu cusku suda mi ba tau zo grute 15:49 * mumble jorne co'u fa la'o fa ctefaho fa 15:49 < ctefaho> co'o 15:51 < jenca> how do i say first, broda. then, brode. finally, brodi? 15:51 < Amnadya2> You can use "mai" 15:51 < Amnadya2> pamai remai romai 15:52 < phma> broda ce'o brode ce'o brodi 15:53 < ldlework> woah 15:53 < ldlework> does that work on bridi tails or just selbri or what 15:54 < jenca> ki’e 15:55 < Amnadya2> With selpahi connectives you can use it on briditails with ceho cu. ceho works as any other connective 15:55 < ldlework> its too bad there's no MOI-ish version 15:55 < ldlework> or some forethought version? 15:56 < ldlework> How would you represent the english speech, "First we gotta find some cash." 15:56 < Amnadya2> pamai 15:57 < ldlework> first utterance ordinal? 15:57 < ldlework> en: pamai 15:57 < mensi> pamai = [MAI*] discursive: first utterance ordinal. |>>> officialdata 15:57 < ldlework> en: mai 15:57 < mensi> mai = [MAI] utterance ordinal suffix; converts a number to an ordinal, such as an item or paragraph number. |>>> 15:57 < mensi> officialdata 15:57 < ldlework> I don't think that's the same 15:58 < ldlework> in the english, the force of "first" is to mark the sentence as logical dependency of some antecedent, I think 15:58 < ldlework> "Let's get something to eat." "First we gotta find some cash." 15:58 < ldlework> Its not really, "First, you're wrong. Second, I'm right!" 16:01 < xunlohu> I don't know the context, but I have seen the second statement used to make different points 16:04 < ldlework> yeah I can see that 16:30 < ldlework> strange no good word for "top" 16:35 < ldlework> Logically: "only X does Y" == "all Y are X" 16:35 < ldlework> is this true? 16:36 < ldlework> oh yeah 17:55 < cizypij> coi ro do 17:55 < durka42> coi 17:56 < cizypij> .i casnu ma 17:56 < durka42> so'u da za'a 17:58 < cizypij> .i zo'o ru'e pau nai da mo 18:01 < cizypij> ni'o xo ma mi ma zasti 18:11 < noname756> Where can i learn lojban? 18:14 < durka42> noname756: http://mw.lojban.org/papri/Welcome!/en#Learning_Lojban 18:16 < noname756> ty ty 18:27 < phma> coi 18:27 < bigcentaur> coi 18:27 < phma> do mo pei 18:28 < bigcentaur> .i zo'o mi co'e 18:29 < phma> mi nitcu lo nu me'andi skagau .iku'i co'e 18:29 < durka42> coi 18:29 < phma> coi la durka 18:30 < durka42> coi zo bo zei retsku 18:31 < noname756> coi 18:33 < phma> zo bo zei retsku ki'a 18:33 < durka42> do pu retsku lo srana be zo bo 18:34 < phma> ua 18:35 < phma> bobobo-bo bo-bobo? 18:37 < durka42> jbonobo 18:40 < phma> xu lo jboia cu gunta lo jbonobo 18:40 < ldlework> Is there a reference of the cmavo related to liteary writing? 18:40 < ldlework> paragraphs and so on 18:40 < phma> like ni'o and no'i? what else? 18:41 < ldlework> That's what I'm asking 18:41 < phma> mai? 18:41 < durka42> make one! 18:42 < ldlework> Well I'm not really sure how they're used. Have we enumerated them just now? 18:42 < phma> ni'o and no'i both start paragraphs; no'i resumes the topic of a previous paragraph. 18:44 < durka42> yeah I guess so 18:44 < durka42> {ni'o pa mai} is common I think 18:44 < durka42> also you can iterate {ni'o} to make different levels of headings 18:44 < durka42> ni'o ni'o ni'o pa mai 18:44 < ldlework> can you mark a title? 18:44 < ldlework> would "$the_title zo'u:" be a neat convention for chapter titles? 18:45 < durka42> could work 18:45 < durka42> you can also just use the first jufra after ni'o 18:45 < phma> I've put just "ni'o " in bold 18:45 < ldlework> I like that ni'o actually denotes the paragraph 18:46 < ldlework> whereas I might style a chapter heading above 18:47 < ldlework> phma: are you saying you style that as a title above the text? 18:47 < ldlework> or inline with the text? 18:47 < phma> on a line by itself, this is a section heading, not a chapter heading 18:48 < ldlework> What about chapter headings? 18:48 < phma> This was on Wikipedia; there is no chapter in an article. 18:48 < durka42> the other question, is {ni'o ni'o} a larger or smaller heading? 18:49 < phma> a ni'oni'o section is bigger than a ni'o section 18:49 < ldlework> I can't bring myself to merely reiterate an article like that 18:49 < phma> of course you can say "ni'ono'ini'o" and confuse people 18:50 < ldlework> it seems brutish 18:50 < phma> what do you mean, reiterate an article? 18:50 < ldlework> ni'oni'o 18:51 < phma> "ni'o" isn't an article 18:51 < ldlework> oh boy 18:51 < phma> it's a new topic marker 18:51 < durka42> it isn't 18:51 < durka42> but also, why not? 18:52 < phma> "le" is an article 18:52 < durka42> it seems better than defining separate cmavo for h1 .. h9 18:52 < phma> "ni'o gerku cu blabi ni'o mlatu" is ungrammatical 18:53 < phma> "le gerku cu blabi le mlatu" is grammatical 18:53 < ldlework> durka42: except to get h1 you need 9 repetitions? 18:53 < ldlework> phma: I get it... 18:54 < ldlework> I know what a noun-phrase determiner is and what it does 18:54 < durka42> ldlework: well it's just relative to the other sections in the same document 18:56 < ldlework> I just can't seem to take the idea of saying ni'oni'oni'o seriously 18:56 < durka42> ok, don't 18:56 < durka42> nobody is making you do it 18:56 < ldlework> I didn't say anyone is 18:56 < ldlework> I'm remarking, as one does in a place like this 18:56 < ldlework> Did I remark wrongly? 18:59 < ldlework> phma: bad habit of using 'article' to mean non-predicate word 19:06 < phma> coi clakre .i xu do cmima la me ly.y'ybucy 19:13 < ldlework> is "fraug" the closest one can get for a lojbanization of "frog"? 19:14 < durka42> I'd say {frag} 19:16 < ldlework> "tod" for "toad" ? 19:16 < durka42> ie 19:24 < ldlework> what if 19:24 < ldlework> lu broda li'u se cusku ko'a fa lu brode li'u 19:25 < durka42> ii 19:25 < durka42> xu do djica lo nu lo munje cu spoja 19:25 < ldlework> lu ko cikna li'u se cusku ko'a fa lu do ba zi jai lerci li'u 19:26 < durka42> mi na jimpe 19:26 < ldlework> oh well thought it might be intuitive 19:27 < durka42> you're using the interpretation where FA-ing an already filled place inserts a {je}? 19:27 < ldlework> yeah 19:27 < durka42> I mean I figured it out 19:28 < ldlework> Is this an established interpretation? 19:29 < durka42> it's one of them 19:30 < ldlework> Do you know if any others are more popular? I haven't really noticed much about FA-ing filled stuff. 21:40 * nuzba @AscensionArcana: I Was Born On The Day Before Yesterday. https://youtu.be/G3U0wpL-CEg prulamdei: (Lojban) yesterday; x₁ is the day before x₂ in day-standard x₃ [http://bit.ly/1NUDuAq] 22:26 * nuzba @jbojufbotjpn: lo cmene pe lo prije pendo be mi cu smuni zo zifre bau lo jbobau / 私の賢い友人の名前はロジバンで自由という意味です。 (770535) [http://bit.ly/1NPYmb9] 23:55 < gleki> Amnadya2: coi mi'e me lo lojbo jatna i pei LLG pleji fi do fo lo nu do favgau CLL --- Day changed Thu Jul 09 2015 00:03 < Amnadya2> xu do basti la lojbab loka jatna lylygy 00:03 < Amnadya2> xu la and cohu djica 00:04 < Amnadya2> iji poho do cmina lo ralju kamni zaho 00:05 < ldlework> melbi https://play.spotify.com/track/4BLXtLRfhp6GFT9M6aa4Kd 00:06 < Amnadya2> i xuna cizra falonu do jatna i do jinvi lodu lylygy palcau 00:07 < gleki> Amnadya2: mi me lo jatna i naku su'e pa jatna cu zasti 00:09 < gleki> i lylygy claxu lo prali be mi i ku'i lo nu LLG pleji fi do zo'u mo 00:09 < Amnadya2> iepei lonu do cmima lo ralju kamni na nuzba 00:09 < mensi> mi tugni i ie mi tugni 00:09 < gleki> ie na nuzba 00:09 < gleki> i ta'onai i'apei mi cpedu lo nu LLG pleji fi do 00:10 < Amnadya2> mi na djuno i bahanai lylygy na kakne 00:10 < gleki> i sa'u do puzi djica lo nu vecnu i ja'o do se cinri lo kate pleji 00:10 < Amnadya2> gohi 00:10 < gleki> i da'i do cpedu tu'a xo rupnu 00:11 < Amnadya2> uhe i mi puzisai coha cikna 00:11 < gleki> fi'i cikna 00:12 < gleki> i do me si du lo BPFK jatna i ja'o do bilga lo ka kurji i pe'i LLG kakne lo ka su'ova'ei jdini sidju do 00:13 < Amnadya2> cylyly xi papipa na jibri bypyfyky 00:13 < Amnadya2> i mi ma gunka dahi 00:14 < Amnadya2> i pleji fo ma 00:14 < gleki> lo nu do cikre lo mabla pagbu be CLL gi'a vimcu lo mabla pagbu i pe'i mutce srana BPFK 00:15 < Amnadya2> lonu cikre je lonu vimcu na pagbu tau cylyly papipa 00:16 < gleki> i ku'i lo nu do me BPFK jatna zo'u na mutce vajni i sa'u do jbopre gi'e jbocre gi'e kakne lo ka gunka i ku'i do ca'a na gunka i ja'o mo i ja'o cumki fa lo nu lo nu LLG pleji fi do cu rinka lo nu CLL zenba lo ka prane 00:16 < gleki> i ie na srana la'a CLL papi pa 00:16 < gleki> i ja'o CLL re pi no zo'u mo i da'i xu do tugni 00:16 < Amnadya2> mi na na gunka i gunka soho lojbo noi iha na srana lylygy 00:17 < gleki> da'i xu do bredi lo ka te pleji fo CLL 00:18 < Amnadya2> cumki vahlonu djuno loduhu makau ve pleji jeloduhu makau pleji morna 00:18 < gleki> xm 00:18 < Amnadya2> i kuhi uhi mi zaho senpi lodu lylygy kakne 00:18 < gleki> i ma jdice lo du'u makau ei se vimcu gi'a se cikre 00:19 < Amnadya2> bypyfyky 00:19 < gleki> Amnadya2: na vajni fa lo nu do senpi i mi ji'a su'ova'ei senpi i ku'i mi na nupre i mi djica lo ka troci 00:19 < Amnadya2> jehe 00:20 < gleki> mu'a lo nu cikre lo srana be lo sko'opu ce la xorlo zo'u LLG pleji bazi lo nu do mo'u cikre zo'u ei pleji ma 00:20 < gleki> i ku'i xu na ku do jdice lo du'u makau cu jai sarcu fai lo ka pamoi se cikre 00:20 < Amnadya2> iha pei mi zeha pensi 00:21 < gleki> iha iha 00:21 < gleki> i ta'o mi ca zgana la terminator 00:21 < gleki> xi y 2015 00:21 < gleki> i mabla 00:21 < Amnadya2> ue do catlu suho skina 00:22 < gleki> i no'e cinri 00:22 < Amnadya2> uenai na cinri 00:22 < gleki> i lo se lisri cu rapli 00:37 < gleki> i u'i la cilretci ku noi mabla tutci cu cinri be mi zmadu la terminator 00:37 < Amnadya2> uhi 00:38 < gleki> i sa'u mi turni la cilretci i mi na turni lo mabla skina 00:41 < Amnadya2> la famymaho zoho 00:41 < gleki> i la'a lu la cilretci cu cinri zmadu do li'u ka'e lobykai te rekto 00:51 < Amnadya2> ithkuil: ňato 00:52 < gleki> na jundi 00:52 < Amnadya2> zaha 00:54 < Amnadya2> xu na ambigu 00:55 < Amnadya2> i laha mi na jimpe fi lo itkuhile prosumti 00:57 < gleki> na djuno 00:57 < gleki> i xu do ca'o surla ne'a lo xamsi 00:59 < Amnadya2> vreta lo ckana i vlile brife ca lo cabdei 01:00 < gleki> e'u do surla gi'enai jbozu'e ja itkuiile'yzu'e 01:00 < Amnadya2> ňatowa 01:01 < Amnadya2> mi na kahe zukte noda 01:01 < Amnadya2> i loka itkuhile tadni cu se surla 01:03 < gleki> fenki 01:04 < Amnadya2> liha mi lazni jiha i cadzu i zutse lo canre i liho 01:06 < Amnadya2> i jihi noda fasnu i mi nelci 01:07 < gleki> e'e doi la mensi ko kansa la amnadiia 01:48 < tsani> coi 01:52 < ldlework> coi banli foldi be lo glare vacri 01:53 < ldlework> hehe, long time 01:56 < ldlework> well you were not actually hot today, surprisingly 01:56 < ldlework> first drizzle in a while 02:06 < gleki> ko fanva fi lo lojbo ca ti 02:06 < gleki> i lo me'oi drizzle cu mo 02:08 < gleki> .dict drizzle 02:08 < fenki> drizzle — noun: 1. Light rain, 2. (physics, weather). Very small, numerous, and uniformly dispersed water drops, mist, or sprinkle. Unlike fog droplets, drizzle falls to the ground. It is sometimes accompanied by low visibility and fog — verb: 1. (trans., intransitive) To rain lightly; to s[...] 02:08 < gleki> xu lo linto poi carvi 02:09 < gleki> i sicpi lo linto vau la'a 02:10 < gleki> phma: http://jbovlaste.lojban.org/comments.html?valsi=8442;natlangword=0;commentid=0;definition=15066 02:38 < gleki> i think i should write another scriptfor mass uploading real sentences to Tatoeba 02:38 < gleki> this is how we can quickly increase the number of lojbna sentences there 03:15 < tsani> gleki: .i .ei zo'o la cadgu'a cu pilno zo drisle 03:17 < tsani> ldlework: .i zo'o ko vitke fi lo kadno .i lo bu'u tsani ca lo crisa na gladu'e vau ui ui nai 03:17 < tsani> .ije mi tolmo'i lo du'u do makau xabju .i la .kolorados. xu 03:26 < ctefaho> coi lo tsani 03:27 < gleki> coi lo blanu galtu 03:31 < latro`a_> coi 03:32 < Amnadya2> coi lo jbini be lo terdi joi lo kensa 03:38 * nuzba @uitki: ASCII table keys names - http://mw.lojban.org/papri/ASCII_table_keys_names by Gleki [http://bit.ly/1HkfRvr] 04:03 < gleki> Pirahã is considered a "primitive" language where it is very difficult to express many concepts (it's more restrictive than toki pona). For example, the Pirahã can't count as they have no concept of quantity. They have no tenses, few kinship terms, and only have color terms for "light" and "dark". Ithkuil on the other hand is an incredibly "efficient" language where a speaker can express many things withou 04:03 < gleki> t difficulty. 04:03 < gleki> Pretty much, a Pirahã-Ithkuil creole would be like trying to merge a canoe and an aircraft carrier. 04:05 < gleki> </quote> 04:31 < Ilmen> I think I've read that Pirahã has no lexical word for colors other than light/dark, but that there are other means to express other colors (with several words? I don't remember). 04:35 < Ilmen> In Japanese, a good deal of colors names are made with the name of a thing having this color plus a -iro suffix meaning "color"; for example, "mizuiro" (water color) and "kitsuneiro" (fox color); so I wouldn't be surprised it they use something similar 04:35 < gleki> i suppose prototyping is what is always used 04:35 < gleki> Homer used the same system 04:36 < gleki> they mixed rainbow colors with other properties of objects 04:36 < gleki> pure rainbow system is a pretty late invention 04:36 < Ilmen> je'e 04:36 < gleki> i perceive {skari2} as a prototype 04:37 < Ilmen> ni'o mi de'a jundi 04:37 < gleki> jb: skari 04:37 < mensi> skari = skari — x1(entity) is of colour x2(entity) as perceived by observer x3(entity) 04:37 < mensi> :lo skari — coloured, colored. lo se skari — colour, color. 04:37 < mensi> :le dacti cu skari lo xunre — It is colored red. 04:37 < mensi> :le flora cu skari lo tsani mi — To me the flower has the color of the sky. 04:37 < mensi> :Comment: x2 is the prototype having the color, e.g. lo tsani (sky) or lo xunre (red) 04:37 < mensi> :Related words: blanu, bunre, cicna, cinta, crino, grusi, narju, nukni, pelxu, xunre, zirpu, carmi, kandi, xekri, 04:37 < mensi> blabi 04:54 < gleki> heh, semantical list of roots in ithkuil https://www.reddit.com/r/Ithkuil/comments/3btk55/list_of_roots_by_categorytheme/ 04:54 < gleki> much better 04:55 < gleki> both for upper ontology dictionary and for learningithkuil 04:56 < Amnadya2> claxu lo 80 cnino gismu 04:57 < gleki> e'e do jungau la kerest 04:57 < Amnadya2> ca lo xavdei 04:57 < gleki> to la gleki co'a ta'e pilno zo e'e toi 04:59 < gleki> xm. i dont like this list since it doesnt explain patterns and stems with their meanings 04:59 < Amnadya2> ie i lo stem liste cu plixauvdahi 04:59 < Amnadya2> plixau dahi 05:02 < gleki> maybe i dont understand something but the list is very short 05:02 < Amnadya2> ie 05:02 < gleki> lo gimste ci barda zmadu 05:03 < Amnadya2> laha na se pagbu lo nahe tanxe gismu 05:03 < gleki> ua 05:03 < Amnadya2> ikuhi mi nasai birti 05:05 < Amnadya2> i liha lo gimste cu brazma 05:05 < Amnadya2> jihi 800 da itkuhile gismu 05:05 < gleki> i ie so'i lo ralrafsi si me'oi root su se claxu 05:06 < Amnadya2> noda itkuhile uitki 05:06 < Amnadya2> i nandu faloka kansihu loka jajgau lo datni 05:06 < Amnadya2> i gugykgredile poho 05:07 < Amnadya2> gugyl 05:07 < gleki> e'a do pilno lo lojbo uitki 05:08 < gleki> i ku'i lo uitki cu mabla lo ka jajgau lo se stura datni 05:08 < gleki> i lo gugle gredile cu pluka zmadu 05:10 < Amnadya2> na pluka mi i kuhi laha praktikale traji 05:15 < gleki> pluka be mi zmadu lo uitki 05:44 < Amnadya2> ei suda itkuhile panra la vlasisku 05:46 < gleki> y 05:46 < gleki> na nabmi 05:46 < gleki> i la sutysisku ca vlipa dunli la vlasisku 05:46 < gleki> i dunli ja zmadu 05:46 < gleki> i ku'i mi nitcu lo datni 05:46 < gleki> i au da zbasu lo XML vreji e lo simsa 05:46 < Amnadya2> jehe 05:46 < gleki> i ca ti lo itkuiile vlaste na zati gi'a mabla 05:46 < gleki> *zasti 05:48 < gleki> exp: na zasti gi'a mabla 05:49 < gleki> a'o zo na sko'opu na'e citka zo mabla 05:49 < mensi> ([CU {na zasti} VAU] [gi'a {CU <mabla VAU>} VAU]) 05:49 < gleki> za'a na citka i ku'i la ilmentufa na kaken tu'a lo sko'opu 05:49 < Amnadya2> na citka 05:49 < gleki> *kakne 05:49 < gleki> i ja'o dunli lu zasti na gi'a mabla li'u 05:53 * nuzba @mochirial: ロジバンだと「吹く」は「風」brifeを使って表します。 x1はx2(起源)・x3(速度)の風; x1はx2から吹く; x2はx1を吹く x2は方角も。 http://jbovlaste.lojban.org/lookup.pl?Form=lookup.pl2&Database=*&Query=brife [http://bit.ly/1NVBWpA] 05:58 < Amnadya2> taho zahu 12kiho da stem 05:59 < gleki> xe'e'e 05:59 < gleki> i sampu bangu 06:01 < Amnadya2> dohanai na cunso 06:41 * la_kristan may become disconnected at any time, without warning (to you or herself) 06:42 < la_kristan> coi 06:43 < la_kristan> .i xu su'o lo prenu cu zvati 06:44 < la_kristan> :-( 06:46 < ctefaho> I'm kinda zvati 06:47 < la_kristan> kinda zvati, ey? 06:48 < ctefaho> I had forgotten someone had made kinda a gismu 06:48 < la_kristan> well, that's more response than I'm getting from #tokipona 06:48 < la_kristan> oh, is it? 06:48 < ctefaho> and sorta 06:48 < ctefaho> lel 06:49 < ctefaho> well they are "joke gismu" 06:49 < la_kristan> like besto? 06:50 < ctefaho> yes 06:50 < la_kristan> ah. 06:50 < la_kristan> very useful. 06:50 < ctefaho> mi kinda sorta klama 06:51 < la_kristan> mi kinda sorta besto 06:52 < ctefaho> you win 06:52 * ctefaho hands la_kristan the price of the day 06:52 < la_kristan> whatever 06:54 < ctefaho> mi kinda sorta besto cribe 06:54 < la_kristan> but at least the definitions of those words don't take up a whole page... 06:55 < la_kristan> djuno the capital of Alaska? 06:57 < la_kristan> sorry, now that I've finished the exercises in L4B, I don't know what to do with myself. 06:59 * la_kristan pu mo'u tcidu fi la'o gy. Lojban for Beginners .gy .e la'o gy. The Complete Lojban Language .gy. 07:00 < ctefaho> have you read le cmalu noltru? 07:00 < ctefaho> it could be a good "target" to be able to read it 07:01 < ctefaho> (is at http://mw.lojban.org/papri/le_cmalu_noltru ) 07:02 < la_kristan> okay, I'm looking... 07:03 < ctefaho> you haven't read The Little Prince before have you? 07:03 < la_kristan> no, I haven't 07:04 < ctefaho> hmm 07:04 < ctefaho> what about The Metamorphisis? 07:05 < la_kristan> haven't heard of that one... 07:06 < ctefaho> one of Kafka's works 07:06 < gleki> {kinda} is an important gismu in my opinion no matter what *kind of* history it has 07:07 < ctefaho> COUGH MALGLI COUGH 07:07 < la_kristan> I just sit around reading djemynai lyrics, which may not be beneficial... 07:07 < ctefaho> very beneficial if they make you adapt ce-ki-tau-jau ;) 07:07 < la_kristan> I don 07:08 < la_kristan> ah! stupid phone! 07:08 < la_kristan> I don't consider that benefit, really... 07:09 < gleki> well, you are free to rephrase {kinda} using other words. xu do kakne 07:09 < ctefaho> well I am the guy making his own dialect so you probably shouldn't listen to me too much:p 07:09 < la_kristan> and when I can't figure out what he's saying, because his cmavo are weird... 07:10 < gleki> that's the problem 07:10 < gleki> with these splinter dialects 07:11 < la_kristan> I speak my own dialect of English, and everyone understands me... 07:11 < la_kristan> but English was vague anyway. 07:13 < la_kristan> Except, maybe the guy from Belgium, whom I PMed the other day; I'm not sure my message was intelligible. 07:13 < la_kristan> oh well. 07:14 < la_kristan> how do you say "I read about [something]"? 07:15 < menli> la_kristan: {mi tcidu lo srana} 07:17 * ctefaho co'oru'e 07:17 * nuzba @uitki: ಲೊಜ್ಬಾನ್ - http://mw.lojban.org/papri/%E0%B2%B2%E0%B3%8A%E0%B2%9C%E0%B3%8D%E0%B2%AC%E0%B2%BE%E0%B2%A8%E0%B3%8D by Gleki [http://bit.ly/1KRarPQ] 07:17 * nuzba @uitki: लोज्बान - http://mw.lojban.org/papri/%E0%A4%B2%E0%A5%8B%E0%A4%9C%E0%A5%8D%E0%A4%AC%E0%A4%BE%E0%A4%A8 by Gleki [http://bit.ly/1SaUToP] 07:18 < la_kristan> how does the something fit in? 07:18 < la_kristan> ctefa'o: see ya 07:18 < menli> Oh, I thought you wanted it to be left to context. 07:18 < menli> {tcidu lo srana be X}, {tcidu tu'a X} 07:19 < la_kristan> ki'e 07:19 < menli> "about X" as a predicate argument is usually {tu'a X} 07:22 < la_kristan> .i mi tcidu tu'a la'o gy. dotside .gy. 07:24 < xajmi> coi... 07:24 < la_kristan> coi 07:25 < xajmi> I would like to translate the "little red riding hood" tale as a practice in lojban... and i want to know if this is how it is said: 07:25 < xajmi> xunre sedgai lo muvzva cmalu ku 07:26 < la_kristan> I don't know; I'm barely out of the beginner course. 07:27 < xajmi> really? me too 07:27 < Ilmen> Le Petit Chaperon Rouge? 07:27 < xajmi> that same 07:27 < xajmi> although that is in french 07:28 < xajmi> as a guide here are the meanings of each word:xunre = x1 is red 07:29 < xajmi> sedgai = x1 is a hood (headwear) covering x2 07:29 < xajmi> muvzva = x1 wanders in area x2 07:30 < xajmi> cmalu = x1 is smal in dimension x2 compared to x3 07:30 < xajmi> lo ... ku transforms selbri into sumti 07:30 < xajmi> I think it is well translated but i'm not sure 07:31 < Ilmen> en: muvzva 07:31 < Ilmen> vlaste: muvzva 07:31 < vlaste> muvzva = x1 wanders within an area x2. 07:31 < la_kristan> .i mi cilre ledu'u zo le cu lidna zo du'u kei la .lojban. la'o gy. Lojban for beginners .gy lenu mi tcidu 07:31 < mensi> muvzva [< muvdu zvati ≈ Move at] = x1 wanders within an area x2. |>>> muvdu zvati |>>> 07:31 < mensi> gusnikantu 07:32 < Ilmen> I'd translate the French title as {la cmalu ke xunre ke muvzva sedgai}, maybe 07:32 < la_kristan> trying to fill all the places of "cilre"; I think I got it... 07:32 < Ilmen> Although I'm unsure about how to best translate "chaperon" 07:34 < xajmi> chaperon means "hood (headwear)" right? 07:34 < Ilmen> I speak French, but "chaperon" is really a rare word, so I'll check on Witkionary 07:35 < Rodericus> Besides English, other languages in which I've seen the title only have: (article +) a combination of a diminutive of “hood” and “red”. Is there another version similar to the English one? 07:35 < Ilmen> Yeah, originally it means "hood"/"head covering" 07:36 < la_kristan> I'd translate the title as "ruĝkapuĉulineto", because that's how I roll. 07:37 < Ilmen> How about simply {la cmalu ke xunre se sedgai}? 07:37 < Ilmen> "The small red-hooded" 07:37 < la_kristan> which reminds me (don't ask me why), what does "balnema" mean? 07:38 < Ilmen> Whale. 07:38 < Ilmen> vlaste: balnema 07:38 < vlaste> balnema = x1 is a baleen whale of species x2 07:38 < la_kristan> whale! 07:38 < la_kristan> fascinating song title that! 07:39 < xajmi> then "chaperon (french) = hood = sedgai (lojban)" 07:40 < la_kristan> it looked to me like a description of something with a tendency to balni... but I figured that wasn't it. 07:40 < Ilmen> In modern French "hood" is "capuche", as a side note. I've never heard the word "chaperon" outside this story :) 07:40 < la_kristan> oh! 07:41 < la_kristan> so that's where "kapuĉo" comes from! 07:41 < xajmi> LOL thanks limen 07:42 * la_kristan does not speak French... 07:42 < xajmi> in spanish "caperusa" is the word used in the tale but the most comon word is "capucha" in modern spanish 07:43 < Ilmen> ua 07:43 < la_kristan> a cognate, then... 07:43 < la_kristan> mi español es muy malo. 07:44 < xajmi> spanish is my mother tongue... so... yeah 07:44 < xajmi> *mother language 07:45 < Ilmen> Mio esta muy oxidado (ay, no puedo añadir el tilde ahora) 07:45 < la_kristan> english is mine, much to my dismay... 07:45 < xajmi> not to worry... 90% of the internet is in english so... you can learn almost anything if you already know english 07:45 < Ilmen> Very nice, I haven't seen a native Spanish speaker visiting #lojban for quite a while ^^ 07:46 < xajmi> ki'e do (thank you) 07:47 < la_kristan> mi volas forĵeti la anglan entute, kaj paroli Esperanton dum la resto de mia vivo! 07:47 < Ilmen> unfortunadamente la mayoria de los materiales lojbanes no estan taducido al español, es muy probablemente la razon principale 07:48 < la_kristan> toki Inli li ike mute mute mute tawa mi!!! 07:50 < durka42> mi se bangu lo spano .iku'i na'e pamoi co bangu 07:50 < xajmi> It is also because lojban is really geeky (I'm just being honest)... 07:51 * Ilmen wants their Linux back to be able to type the acute accent 07:51 < la_kristan> ми уиле е ни: ми токи е токи пона кепекен сителен Лоси. 07:52 < xajmi> I don't understand a single word in russian 07:52 < la_kristan> сителен Лоси ли пона тауа ми! 07:52 < gleki> once our russian group translated Waves to Russian and the japanese subcommunity translated them to Japanese. Also one French guy translated them to French. I guess that's mostly it 07:52 < la_kristan> it's not Russian... 07:53 < gleki> im talking about Wave Lessons 07:53 < xajmi> ukranian? 07:53 < la_kristan> it's just toki pona in Cyrillic letters. 07:53 < gleki> ми ша тавла бау ла ложбан 07:54 < la_kristan> .и ми го'и 07:54 < xajmi> what on earth is "toki pona" 07:54 < xajmi> ? 07:54 < durka42> another conlang 07:54 < xajmi> oh... weird 07:54 < Ilmen> A minimalistic-issim philosiphical conlang 07:54 < Ilmen> With about 100 words 07:55 < la_kristan> "the language of good" 07:55 < la_kristan> hence the name; 07:55 < xajmi> oh... amazing 07:55 < la_kristan> toki = talk, language 07:55 < gleki> toki: Lojban 07:55 < mensi> lojban = toki Losupan. en la toki pona. taso la toki pona ala 07:56 < durka42> toki: language 07:56 < mensi> 11 da se tolcri: toki, toki Inli li pona., toki Kanse, sona e toki, sona pona e toki, toki li pona tawa mi., toki lon 07:56 < mensi> toki, toki nanpa wan, toki pona e toki, toki pona li toki pona., toki sitelen 07:56 < la_kristan> pona = good, simple 07:56 < xajmi> da faq??? 07:57 < xajmi> jajajaja 07:57 < durka42> ? 07:57 < Ilmen> .u'i 07:57 < clakre> jajaja is not what you want in here u'i 07:57 < xajmi> u'i 07:57 < clakre> u'i or xaxaxa instead :) 07:58 < gleki> jajaja. jejeje. je'ije'ije'i. jojojo. jujuju 07:58 <@xalbo> Now I need to try to come up with a context where {jajaja} would be grammatical. I'm not sure that's possible without magic... 07:58 < xajmi> i'm sorry, but "toki nanpa wan, toki pona e toki, toki pona li toki pona" isn't what we want either, and yet here it is... u'i 07:58 < clakre> la gleku pu spofu la mensi! 07:58 < clakre> zo'o 07:59 < Ilmen> zo spogau .e'u 07:59 < gleki> mi jai gau spofu fai la mensi 07:59 < Ilmen> zo gleki .e'u ji'a 07:59 < gleki> i gau mi spofu fa la mensi 07:59 < gleki> i gau mi la mensi co'a spofu 07:59 <@xalbo> (magic being something like {me zo ja ja ja zei co'e}) 07:59 < clakre> u'i 08:00 < gleki> me zo ja ja ja zei xajmi 08:00 < la_kristan> speaking of which, djemynai's connectives look weird to me. 08:01 < gleki> what were they? 08:01 < durka42> he is using this set http://selpahi.weebly.com/lojban/how-to-substantially-simplify-the-lojban-connective-system-my-connective-system 08:01 < gleki> {je cu} probably? 08:01 < Ilmen> {ga ja} too 08:01 < la_kristan> yeah, je cu, and stuff. 08:01 < Ilmen> I have the feeling of having heard {gu ja} too, but I hope I just misheard 08:01 < gleki> je cu can be replaced with {gi'e} 08:02 <@xalbo> Ilmen: I think {gu ja} may be forethought tanru-internal, but I'm not sure. 08:02 < durka42> Ilmen: guja is {gu'a}, right? though {gu} seems to have been missed when JACU was added to camxes 08:02 < Ilmen> Ah 08:02 < Ilmen> Didn't remember of the existence of GUhA. 08:03 < durka42> possibly the least used connective? 08:03 < durka42> fighting for that title with ZIhE, ba'a 08:03 < Ilmen> li'a 08:03 < xajmi> sorry. I got lost here... I don't know that much lojban... but perhaps now someone could help me? 08:03 < durka42> sidju do ma 08:04 < Ilmen> Sure. No vacila en preguntar 08:04 < Ilmen> se tienes preguntas 08:04 < Ilmen> *si 08:04 < xajmi> I want to translate the "little red riding hood" tale to lojban as practice, and I want to know if ot is well said "xunre sedgai lo muvzva cmalu ku" 08:04 * Ilmen has their Esperanto interfering with Spanish again >_< 08:04 < xajmi> u'i limen 08:05 < durka42> A red hat covers the wandering small-thing. 08:05 < xajmi> so what would be a better translation? 08:06 < Ilmen> {la cmalu ke xunre se sedgai}, I suggest. 08:06 < durka42> are you going for a sentence or a noun-phrase? 08:06 < Ilmen> "The small red-hooded one" 08:06 < durka42> .i'e Ilmen 08:06 < xajmi> A noun-phrase "the little red riding hood 08:06 < gleki> i think {je} instead of {ke} can also work there but we are delving into details here 08:07 < Ilmen> gleki: with "se sedgai", I think it would mean that the hood is small 08:08 < Ilmen> {lo cmalu se sedgai be lo xunre} ---> {la cmalu ke xunre se sedgai} 08:08 < Ilmen> ie nai pei 08:08 < mensi> ei mi tugni 08:08 < gleki> i invented my own cekitaujau dialect. and already implemented it in a parser. 08:09 < Ilmen> xajmi: {la + selbri} = the one(s) named [selbri...] 08:09 < gleki> it's called .......... 08:09 < gleki> li ! 08:09 < Ilmen> xajmi: la se sedgai = The hooded one(s) 08:09 < gleki> alta: mi li monsuta 08:09 < mensi> ([FA mi] [li {monsuta NF} VAU]) 08:09 < Ilmen> hmm 08:10 < Ilmen> maybe {la cmalu se ke xunre sedgai} would be better actually, if we're nitpicking 08:11 < xajmi> lo cmalu se sedgai be lo xunre = that which is small and wears a hood, and is red? 08:11 < Ilmen> se sedgai be lo xunre = to be hooded with something red (with a red hood) 08:11 <@xalbo> {la cmalu se ke xunre sedgai} seems good, yeah. Small, (red hood)ed. 08:12 * xalbo wonders whether mapku includes hoods. Probably not. 08:13 < Ilmen> xajmi: "ke" is a tanru-grouping operator; I think you probably haven't learn about it yet 08:13 < gleki> http://tiki.lojban.org/tiki/la+cmalu+je+xunre+mapku 08:14 < Ilmen> cmalu zdani gerku = (cmalu zdani) gerku = dog of small home 08:14 < Ilmen> er 08:14 * la_kristan finishes reading selpa'i's proposal. 08:14 < Ilmen> "small dog of home" 08:14 < Ilmen> cmalu ke zdani gerku = cmalu (zdani gerku) = small (home dog) 08:14 < xajmi> aparently not... sorry 08:15 < Ilmen> so adding "ke" changes what "cmalu" applies to, in this case 08:15 < xajmi> wow thanks gleki 08:16 < Ilmen> cmalu nixli ckule = (cmalu nixli) ckule = [small girl] school 08:16 < xajmi> ui... got it thanks Ilmen 08:16 < Ilmen> cmalu ke nixli ckule = small [girl school] 08:16 < Ilmen> so here it's the school that's small 08:16 < gleki> just reformatted in the new wiki: http://mw.lojban.org/papri/la_cmalu_je_xunre_mapku 08:16 < la_kristan> well, this connective proposal is interesting, but... 08:17 < la_kristan> it kinda messes things up for backward compatability, ĉu ne? 08:18 < gleki> no 08:18 < durka42> actually, it's backwards compatible 08:18 < Ilmen> lol, what are those ".:" before the {i}? 08:18 < gleki> it doesnt break any existing usage 08:18 < durka42> though once you have such a simple and wonderful connective system, the temptation is high to repurpose the old connective words :) 08:18 < gleki> i dont know wy those ":" are there. In Tiki those are used for centering text although maybe it's just stylistic 08:18 < la_kristan> well, it's confusing to me... 08:19 < gleki> centering didnt work anyway 08:19 < gleki> the new connective system doesnt break the old one 08:19 < gleki> exp: mi klama je cu citka i do klama gi'e citka 08:19 < mensi> ([mi {CU <klama VAU>} {je <cu (¹citka VAU¹)> VAU}] [i {do <CU (¹klama VAU¹)> <gi'e (¹CU [citka VAU]¹) VAU>}]) 08:19 < Ilmen> {gaJA} and {guJA} do break past usage. {JAcu} doesn't, except some very rare and tricky sentences involving {lu}, maybe 08:20 < Ilmen> er 08:20 < durka42> they do? 08:20 < durka42> GA JA was never allowed, was it? 08:20 < xajmi> well I apreciate your help guys... I hope we can talk some other time... have fun!!! 08:20 < xajmi> co'o 08:20 < durka42> co'o di'ai 08:20 < Ilmen> maybe they don't, durka42. I guess I'm too hot to think properly 08:21 < durka42> .u'i ko na xenru .i mi ka'e na'e drani 08:21 < jsgrant> .i coi la terdi 08:21 < durka42> coi grant 08:21 < Ilmen> coi 08:21 < Ilmen> co'o la xajmi 08:21 < la_kristan> all I know is, that learning a language that is still subject to so much tinkering wasn't exactly what I was looking for... 08:21 < jsgrant> Are there any non-one off comics for/in Lojban? Seems like a fun project to do while learning in-tangent. 08:22 < jsgrant> la_kristan: What, regarding Lojban? 08:22 < la_kristan> where I come from, tinkerers aren't taken seriously. 08:22 < la_kristan> jsgrant: yeah. 08:22 < durka42> where you come from = esperanto? 08:22 < gleki> la_kristan: the use the normal system 08:23 < jsgrant> la_kristan: You could always go for Loglan, or something; I don't think that's been touched for sometime as far as I'm aware. 08:23 < la_kristan> durka42: exactly 08:23 < gleki> neither Lojban was touched. 08:23 < durka42> well, lojban is much younger than esperanto 08:24 < Ilmen> la_kristan: {jeCU} is backward compatible, and the experimental parser both GIhA and JACU coexist without any problem. 08:24 < durka42> the point stands, though 08:24 < Ilmen> Anyway nothing force you to use JACU, you can always use the standard GIhA 08:24 < jsgrant> Generally, Lojban having a notable number of individuals in the community wherein they are more willing/likely experiment with some of the primitive to complex structures in the lang, is a big selling point to me. 08:24 < gleki> {jecu} is an addition to the language. it doesnt change the existing language 08:24 < gleki> tinkering is usually when you break existing things like with cekitaujau 08:25 < durka42> that's disingenuous to say. if it catches on, you'll hear it in the speech of others 08:25 < durka42> backwards compatibility doesn't mean we aren't tinkering :) 08:25 < durka42> I don't think there's as much tinkering going on as it seems, but I don't want to misrepresent it 08:25 < durka42> jsgrant: :) 08:26 < la_kristan> but the coexistence of both merely necessitates the learning of both systems. 08:26 < gleki> lojban develops and new words are accumulated 08:26 < gleki> and new constructs 08:28 < la_kristan> like iĉismo would make things simpler, if it has been there from the beginning, but at this point it only confuses people and increases the learning load. 08:29 < la_kristan> because it's one of the few reform proposals that's gained enough traction to be a problem. 08:29 * Ilmen ja uzas i^cismon 08:30 < la_kristan> then Ilmen is part of the problem. 08:30 < gleki> in the Crash Course i only mentioned that {je} is allowed instead of {e} but i dont mention {je cu} 08:30 < gleki> i hope this doesnt increase any learning load since {je} already exists 08:30 < Ilmen> Nu, simple imagu, ke "i^c" estas simpla vortoradiko, kaj ne sufikso. ^Gi ne estus kontraufundamenta, ^cu ja? 08:31 < gleki> i can only see that "sxli" is used normally. 08:31 * la_kristan may be disconnected at any time... 08:31 < gleki> viricxo, virino, patricxo, patrino. okej. kio estas "patro" nun? :D 08:32 < la_kristan> jen la problemo. 08:32 < Ilmen> Mi tamen ne uzas "patri^co", tial, ke "patro" estas bedaurinde baze i^ca vorto 08:33 < la_kristan> homoj kiuj uzas na "iĉ" estas kontraŭfundamentaj! 08:33 < la_kristan> zo'o 08:34 < Ilmen> Eble io kiel "gepatrero" povus funkcii, kvankam ^gi estas malbela 08:34 < la_kristan> ! 08:35 < la_kristan> mi vidas iufoje "gepatro" 08:35 < la_kristan> tio ne estas /tro/ fia, sed... 08:37 < Ilmen> Well, I think I'll take my leave now, instead of talking off-topic with ugly diacritic-less Esperanto :P 08:37 < Ilmen> co'o 08:37 < la_kristan> lol 08:38 < la_kristan> co'o 08:38 < jsgrant> Ilmen: co'o 08:38 < durka42> co'o di'ai 08:39 < la_kristan> tiuj, kiuj ne uzas ĉapelajn literojn nek h-sistemon... 08:40 < la_kristan> estas KONTRAŬFUNDAMENTAJ!!! 08:42 < la_kristan> .i ta'onai na pluka mi fa lenu lo prenu cu cizra pilno lo cmavo 08:42 < gleki> je'e je'e 08:42 < gleki> i nabmi 08:42 < gleki> i ku'i la lojban ca'o zasti iju lo prenu cu troci lo ka cnegau ly 08:43 * gleki tiuj, kiuj uzas x-sistemon estas ... kontrauXfundamentalistoj :D 08:44 < la_kristan> fiacxe! 08:45 < gleki> "Zamenhofo" sutysisku needs fixes for being able to search both in standard, x- and h- systems 08:47 < la_kristan> hm, not familiar with that... 08:47 < gleki> http://mw.lojban.org/extensions/ilmentufa/i/zamenhofo/index.html#sisku/Lojban 08:48 < gleki> i used some dictionary that uses x-system but probably standard system as well 08:48 < gleki> a mess 08:48 < la_kristan> but what I'm saying is, if some people use that taujau whatever-it-is, and others don't, then we've got a split! 08:49 < gleki> taujau is a problem because {mi broda lo du brode} can mean two different things in the two dialects 08:50 < la_kristan> yeah. 08:51 < la_kristan> btw what does "ni" mean in this system? 08:51 < gleki> i guess it doesnt change 08:51 < la_kristan> I'm still trying to decipher rap lyrics... 08:51 < gleki> i guess you need to just mass replace all words ki=>ke'a, ce=> ce'u 08:52 < la_kristan> du -> du'u 08:52 < gleki> se'e => zu'ai 08:52 < gleki> si'u=> si'au 08:53 < gleki> http://mw.lojban.org/papri/ce_ki_tau_jau 08:53 < gleki> i dont know which of that was used 08:55 < la_kristan> oh, su'o -> su! 08:55 < la_kristan> I thought maybe he just wanted to erase his song! 08:57 * nuzba @uitki: la cmalu je xunre mapku - http://mw.lojban.org/papri/la_cmalu_je_xunre_mapku by Gleki [http://bit.ly/1To1mzo] 09:07 * nuzba @uitki: Learning materials. Secondary links - http://mw.lojban.org/papri/Learning_materials._Secondary_links by Gleki - /* Talking to Lojbanists. Live support */ [http://bit.ly/1NW9h3W] 09:08 < la_kristan> But, of course people are attracted to shorter words... 09:09 < la_kristan> so this could attract a following. 09:09 < durka42> that's why I go back and forth in my feelings about CKTJ 09:09 < la_kristan> and once you've got the big shot musician using it... 09:10 < la_kristan> then we need to know it, to understand him. 09:10 < la_kristan> and then what? 09:10 < durka42> the album certainly made more people aware of JACU and CKTJ 09:11 < la_kristan> it's a subversive diabolical plot! 09:11 < gleki> jacu saves no syllables most of the time 09:11 < durka42> not everything is about syllables 09:11 < la_kristan> fdk 09:11 < la_kristan> uh, i mean 09:11 < la_kristan> afk 09:12 < la_kristan> if you need me, highlight me. 09:15 < gleki> je'e je'e i ai mi pilno lo bombila 09:15 < ctefaho> what exactly is the "'" called in lojban? 09:15 < gleki> en: y'y 09:15 < mensi> y'y = [BY2] letteral for '. |>>> officialdata 09:15 < ctefaho> I did not ask for the letteral 09:16 < ctefaho> is there a lujvo/tanru/whatever for it? 09:16 < gleki> y'y is a pro-sumti 09:16 < gleki> i.e. it works as a noun or pronoun 09:16 < ctefaho> I know 09:16 < gleki> you may say {zo y'y} 09:17 < durka42> not sure if there is a selbri defined for it 09:17 < durka42> like "." is denpa bu, "," is slaka bu... 09:17 < ctefaho> yeah durka42 that was what I wondered 09:17 < gleki> then i need a selbri for z" 09:18 < ctefaho> you are very helpful now gleki 09:18 < durka42> karsnaseibu'i 09:19 < ctefaho> actually nvm;o 09:26 < la_kristan> vasxu bu, maybe? 09:31 < cliva> spagetufa: +cktj lo remna cuvoilonuzoko ze'u zmadu zodolokamakaunilonu cusku cekicu cafne ka'ese jalge lonuki fengu 09:31 < spagetufa> ([FA {lo remna KU}] [cu {<voi (¹[FA {lo <nu (²[FA {zo ko}] [CU {<ze'u zmadu> <(³[FE {zo do}] [FI {lo <ka (⁴[FA {ma kau}] [CU {<ni (⁵[FA {lo <nu (⁶CU [cusku {<(⁷FE ce⁷) (⁷FI ki⁷)> VAU}]⁶) KEI> KU}] [cu {cafne VAU}]⁵) KEI> VAU}]⁴) KEI> KU}]³) VAU>}]²) KEI> KU}] [CU {<ka'e (²se jalge²)> <(²FE [lo {nu <(³FA ki³) (³CU [fengu VAU]³)> KEI} KU]²) VAU>}]¹) KEI> VAU}]) 09:31 < cliva> la_kristan: ^ cekitaujau parser, might be useful 09:32 < la_kristan> cliva: it could be, if I could make any sense of it... 09:34 < la_kristan> perhaps a simpler example might be in order? 09:34 < cliva> je'u lo te pruce be lo barda jufra ka'e simlu lo ka kalsa 09:35 < cliva> spagetufa: +cktj lo nu su da kakne tau lo .itku'ile na lakne 09:35 < spagetufa> ([FA {lo <nu (¹[FA {<su BOI> da}] [CU {kakne <(²FE [tau {lo itku'ile KU} LU'U]²) VAU>}]¹) KEI> KU}] [CU {<na lakne> VAU}]) 09:38 < la_kristan> vlaste: lakne 09:38 < vlaste> lakne = x1 (event/state/property) is probable/likely under conditions x2. 09:38 < la_kristan> vlaste: itku'ile 09:38 < vlaste> itku'ile = x1 pertains to Ithkuil/Illaksh language in aspect x2 09:40 < la_kristan> that anyone can do anything with Ithkuil is not likely? 09:40 < la_kristan> true enough, I guess. 09:42 < la_kristan> but anyhow, this seems to be a major overhaul of the cmavo. 09:42 < la_kristan> and not compatible with the standard usage. 09:43 < la_kristan> and not really necessary... 09:44 < la_kristan> I'm not really very concerned with reducing syllables. 09:44 < ldlework> what is a major overhaul? 09:44 < ldlework> oh 09:45 < la_kristan> In Esperantujo one becomes accustomed to the multiplicity of syllables. 09:45 < cliva> around 15-20 out of all 500 or so cmavo are affected, but the changes are highly visible, je'u 09:46 < la_kristan> In English, I'd say my friend is a lefty; in E-o she's a "maldekstramanulino". 09:46 < la_kristan> no shortage of syllables there. 09:48 < la_kristan> is there a difference between "gi je" and "je cu"? 09:48 < cliva> same meaning (though with je cu you have two places to attach attitudinals) 09:49 < la_kristan> and what do the two "gi"s mean in this line? 09:50 < la_kristan> .i ca gi do mrobi'o gi mi jmive za'o 09:50 < cliva> "ca gi" also works as a gek: "while you're dying, I'm still alive" 09:51 < la_kristan> ah 09:52 < la_kristan> in both connective systems, or just this newer one? 09:53 < cliva> both of them 09:53 < la_kristan> okay. 09:53 < jdjdjdjsjyjyj> coi la cliva 09:53 < cliva> coi la .u'enmo 09:53 < Amnadya> ihe 09:55 < Amnadya> kaheku ku lu jai riha manci lihu ja lu nunmanci kihu 09:55 < Amnadya> si lihu 09:56 < Amnadya> i pehi loremoi 09:56 < Amnadya> i ambigu kuhi 09:56 < cliva> je'e .i lo .itku'ile ga'a mi za'o .itku'ile 09:56 < cliva> y 09:57 < cliva> je'e .i lo .itku'ile ga'a mi za'o ba'e *zi'e* .itku'ile 09:58 < Amnadya> sahe ka nunmanci dahi 09:58 < Amnadya> saheruhe 09:58 < la_kristan> hĥhĥh 09:59 < Amnadya> i astrato 10:00 < Amnadya> ithkuil: nadya 10:00 < ithkuil> nadya: P1S1/STA-n-OBL-UNI/CSL/A/PRX/RPV-EXS-IFL/UNFRAMED 10:00 < Amnadya> ie 10:01 < Amnadya> dahi zei ka zei sucti 10:01 < Amnadya> A = ka/siho 10:02 < Amnadya> RPV = dahi 10:02 < Amnadya> saheruhe 10:02 < cliva> da'i xamgu fa lo nu ka'e ku lo ditfaulte rafsi cmene na pagbu lo gloso 10:03 < Amnadya> ie i tcaci mi 10:04 < la_kristan> don't you agree that having two mutually incompatible cmavo systems competing for usage in a language whose userbase is already rather small could be a problem? 10:04 < gleki> i do 10:05 < gleki> shanoxilt seems to be of the same opinion 10:05 < gleki> but that's all a long story 10:46 < Amnadya> ithkuil: hata'bawëla'n 10:46 < ithkuil> hata’bawëla’n: CNF-P1S1/STA-CTX/PPS/ASR-a’-b-OBL-ASR/FAC-UNI/CSL/M/DEL/NRM-EXS-ASU-IFL/UNFRAMED 10:46 < Amnadya> ro lo ditfaulte 10:46 < Amnadya> ikuhi lo romoi na cohe 10:48 < Amnadya> ithkuil: hata'bawëla 10:48 < ithkuil> hata’bawëla: CNF-P1S1/STA-CTX/PPS/ASR-a’-b-OBL-ASR/FAC-UNI/CSL/M/DEL/NRM-EXS-IFL/UNFRAMED 10:48 < Amnadya> ca drani 11:08 < Amnadya> oiho mi se fanza lo correlative case i mi na jimpe fi lo jai tai pilno 11:10 * nuzba @oka_iu_tcan: http://vodka-pomme.net/projects/tengwar-for-lojban/lojteng 調べてみると結構開発してるひといるっぽい [http://bit.ly/1J6wjkc] 11:12 * nuzba @oka_iu_tcan: 色んな人がデザインしてるんだけど、テングワール素人だから、どれがイイ感じなのかがわからんのよね…。ほら、なんか、ひらがなをロジバンで使えるようにデザインされて、「ち」をティの音に充てられたりすると嫌やん?そんな感じよ http://mw.lojban.org/papri/tengwar [http://bit.ly/1J6wV9w] 11:14 * nuzba @Kohelang: テングワールには色々なモードや流派があるのでarihatoさんがどの書き方を選んだのかは不明ですが、音価:西方語/読み順:クウェンヤで lojban lo logji bangu と一応読めます。これでロジバン的に意味は通りますか? [http://bit.ly/1RmAZfC] 11:16 * nuzba @oka_iu_tcan: ロジバン ~論理的な言語~ っていうロジバンwikiのサブタイみたいなやつやな [http://bit.ly/1J6xkJb] 11:49 < gleki> Amnadya: la'oi correlative zo'u zo fancu pei 11:49 < Amnadya> y 11:50 < ldlework> The name Penny results in an unfortunate lojbanization. 11:50 < Amnadya> i sahu mi tesimlu loka malgi sai 11:50 < gleki> i ku'i zoi gy 11:50 < gleki> As will be described in Section 5.4.2.3, the CORRELATIVE is used with verbal case-frames to convey the equivalent of simple relative clauses in Western languages. 11:50 < gleki> gy 11:50 < gleki> ge'e 11:50 < gleki> i na djuno i mabla sidbo vau pe'i 11:50 < Amnadya> cizra mi 11:51 < gleki> i simsa zo pe i simsa zo fancu i simsa zo poi 11:51 < gleki> The CORRELATIVE case (discussed in Section 4.5.25) is used to create case-frames which are semantically equivalent to the English phrase ‘that/which/who is/are…’ Such a case-frame would be used in conjunction with specific Functions (see Section 5.1) to convey whether the relationship of the relativized clause to the main clause is one of description, copula identification, etc. 11:52 < gleki> i lo rusko jbopre soi sisi itkuiile'ypre pu finti la'e di'e 11:53 < gleki> lo xamsi poi blanu => framed+correlative 11:54 < ldlework> Is there any other consonant I can use to terminate "peni" (lojbanization of Penny) 11:54 < ldlework> other than "s" I mean 11:54 < gleki> <translate who="gleki">thus we have three stages of getting a "~=property": 1. the light one using correlative 2. the heavy one using decriptive function 2. the total one using framed relation + correlative 11:55 < gleki> the sea is blue = stative/ descriptive 11:55 < gleki> the sea that is blue = framed/ correlative 11:56 < gleki> and no way to say "blue sea"</translate> 11:56 < ldlework> blanu xamsi 11:56 < ldlework> the english is a tanru also 11:56 < ldlework> so nothing of value is lost 11:57 < gleki> oh oh 11:58 < durka42> ldlework: yes. any other consonant. :p 11:58 < ldlework> even z? 11:58 < durka42> sure 11:58 < ldlework> neat 11:59 < gleki> one more quote: as far as i understand Correlative sets relation between formatives which is identical between oblique-content and oblique-verb. art-obl sex-obl - sex shows itself in art (sex is art-izing itself) 11:59 < ldlework> now I can referr to my cat iwthout feeling shame 11:59 < ldlework> ki'e 12:02 < noncomcinse> gleki: .i xu la sutysisku pu kakne loka djuno fi lo veljvo 12:05 < gleki> noncomcinse: xu ca na kakne 12:05 < Amnadya> jehe la gleki 12:05 < noncomcinse> gleki: go'i 12:06 < Amnadya> i rahu sohi ja sohe lo mupli jufra cu cizra mi su da 12:06 < Amnadya> au mi jimpe 12:06 < gleki> ldlework: the newest rule is that words are lojbanized to the gismu nearest in pronunciation. Thus Penni becomes {pinji}. Feel free to use </zoho> 12:07 < noncomcinse> .i zoi .urli. http://mw.lojban.org/extensions/ilmentufa/i/en/#sisku/nerkla .urli zo'u jungau mi noda 12:07 < ldlework> gleki: u'i 12:07 < ldlework> Why are people using h? 12:07 < gleki> this time because ' can't be a part of html tags 12:09 < Amnadya> because itd easier on this phone 12:10 < gleki> noncomcinse: cinri i na nabmi mi http://snag.gy/sDoox.jpg 12:12 < noncomcinse> gleki: mi pilno lo la plise ku fonxa 12:13 < noncomcinse> .i ku'i zoi .urli. http://mw.lojban.org/extensions/ilmentufa/i/en/#sisku/ta'anai mu'a .urli cu tolspofu 12:14 < noncomcinse> s/ta'anai/ta'a%20nai/ 12:14 < fenki> noncomcinse meant to say: .i ku'i zoi .urli. http://mw.lojban.org/extensions/ilmentufa/i/en/#sisku/ta'a%20nai mu'a .urli cu tolspofu 12:18 < noncomcinse> .i zo .rai,en. zo'u xu nitcu tu'a me'o slaka bu 12:27 < durka42> zo raien du zo ra,ien 12:27 < zahlman> zo rai'yn pei 12:28 < durka42> ne'au 12:28 < noncomcinse> xu me'o .ybu zo'u cmevla curmi 12:29 < durka42> go'i 12:29 < noncomcinse> ua 12:29 < noncomcinse> .i mi ckule finti bau la .lojban. 12:29 < zahlman> i'e 12:33 < zahlman> mi puzi co'e la lajda'arxivos lo mi me'oi cidysys .i zdile .i ku'i gasnu co cortu lo mi degji oi 12:33 < noncomcinse> .i finti bau lo glico ji'a 12:33 < noncomcinse> .i .ei cliva co'o .irci 12:33 < zahlman> oiru'e co'o 12:34 < noncomcinse> .oi ja'ai ji'a 12:34 < noncomcinse> co'o 12:53 < Amnadya> ithkuil: penis 12:53 < ithkuil> penis: P1S1/STA-p-ABS-UNI/COA/M/DEL/NRM-PLA_1/1-EXS-IFL/UNFRAMED 12:53 < Amnadya> oi 12:54 < durka42> are all of these actually valid or is it just GIGO 12:54 < durka42> does ithkuil have any phonotactics, .u'i 12:54 < Amnadya> they are real possible words 12:54 < ctefaho> camxes: penis 12:54 < camxes> penis 12:54 < Amnadya> see the gloss 12:54 < ctefaho> ++ 12:56 < Amnadya> ithkuil: camxes 12:56 < ithkuil> camxes: P1S1/STA-c-OBL-CPN/ASO/N/DEL/NRM-PLA_1/3-EXS-IFL/UNFRAMED 12:56 < camxes> SyntaxError: Expected [.\t\n\r?! ] but "P" found. 12:58 < durka42> .u'i 12:58 < durka42> I don't see any gloss.. 12:58 < durka42> ithkuil: vlaste 12:58 < ithkuil> vlaste: P1S1/STA-vl-OBL-COH/ASO/M/DEL/NRM-RPS-IFL/UNFRAMED 12:58 < vlaste> no results. http://vlasisku.alexburka.com/P1S1%2FSTA-vl-OBL-COH%2FASO%2FM%2FDEL%2FNRM-RPS-IFL%2FUNFRAMED 12:59 < Amnadya> the output 12:59 < Amnadya> is what it means 12:59 < durka42> ithkuil: vvvvvvvvvvvvvvvvvvvvvvv 12:59 < ithkuil> error 12:59 < durka42> so there are phonotactics, but they are quite loose :) 12:59 < durka42> possibly a strict superset of lojban 12:59 < durka42> ithkuil: mitavladofolologjibangu 12:59 < ithkuil> error 13:00 < Amnadya> thats not how ithkuil works 13:00 < ctefaho> ithkuil: bananas 13:00 < ithkuil> bananas: P1S1/STA-b-OBL-UNI/COA/M/DEL/NRM-AGC_1/5-PLA_1/5-EXS-IFL/UNFRAMED 13:00 < ctefaho> ^means what? 13:01 < durka42> just what it says 13:01 < durka42> li'a 13:01 < Amnadya> look up all the abbreviations 13:02 < Amnadya> b + a + n + an + as 13:02 < Amnadya> 5 rafsi 13:03 < ctefaho> ugh 13:03 < ctefaho> are they all cases? 13:04 < ctefaho> all I know about Ithkuil is the semantics behind their cases not how words work 13:04 < ctefaho> I assume that's a verb? 13:05 < Amnadya> can be verb or noun 13:05 < Amnadya> the second a is the case 13:05 < Amnadya> first* 13:06 < Amnadya> a word can only have one case 13:06 < Amnadya> the different rafsi occupy different slots 13:08 < ctefaho> but ok a case only goes into 1 word 13:08 < ctefaho> a/1 13:08 < Amnadya> hm? 13:09 < ctefaho> I had skipped the morphology part completely 13:09 < Amnadya> ithkuil is 97% morphology 13:10 < ctefaho> that's why I am not into it;) 13:10 < Amnadya> almost zero syntax 13:10 < ctefaho> well that's one way to do I guess 13:13 < ctefaho> besides jbobau I am more interested in Taoq Dzu now tbh 13:13 < ctefaho> sleek and lean 13:13 < ctefaho> compared to bulky jbobau 13:13 < ctefaho> I do like my jbobau bulky though 13:13 < Amnadya> Toaq Dzu huh 13:14 < ctefaho> ...Toaq Dzu 13:14 < ctefaho> but ok done with some stuff, can focus on bananas now 13:15 < Amnadya> b is the root 13:15 < ctefaho> "P1S1" means what? 13:15 < Amnadya> everything else is derivation 13:15 < Amnadya> pattern 1 stem 1 13:16 < ctefaho> man you weren't kidding when you said this language was all about morphology 13:20 < ctefaho> ...I just don't know what I am to look for 13:20 < Amnadya> its overwhelming 13:20 < ctefaho> je'u 13:21 < Amnadya> but it gets better 13:21 < ctefaho> b is a "formative"? 13:22 < Amnadya> no 13:22 < Amnadya> b is the root 13:22 < Amnadya> the whole word is the formative 13:23 < ctefaho> jimpe 13:24 < ctefaho> ITM 13:24 < ctefaho> Intermittent 13:24 < ctefaho> b 13:24 < ctefaho> is that it? 13:24 < Amnadya> no 13:25 < Amnadya> b means desire. its in the lexicon 13:26 < ctefaho> right 13:26 < ctefaho> I just don't know where to look, the morphology is...overwhelming 13:27 < ctefaho> sigh looks like I have to read everything from the beginning 13:27 < Amnadya> you need to learn to recognize which slot a rafsi is uccupying 13:27 < Amnadya> occupying 13:30 < ctefaho> well the introduction wasn't very helpful 13:31 < ctefaho> ok found what I need 13:32 < Amnadya> it took me several reads through the book until i started to understand something 13:32 < Amnadya> be patient 13:32 < ctefaho> you could at least point me to where to start 13:33 < Amnadya> everywhere 13:33 < Amnadya> what are you looking for 13:34 < Amnadya> the website doesnt have any easy intro. its only a grammar reference 13:38 < ctefaho> well I need to know what I am looking for to find it 13:38 < Amnadya> the gloss tells you 13:39 < Amnadya> cryptically 13:39 < Amnadya> and experience 13:39 < ctefaho> ...but i have no idea what the gloss means 13:39 < Amnadya> the website tells you 13:39 < ctefaho> ... 13:39 < Amnadya> each 3letter code 13:40 < Amnadya> is a 13:40 < Amnadya> affix value 13:40 < ctefaho> where do I find the corresponding chart? 13:40 < ctefaho> or charts 13:41 < zipcpi> vlaste: torso (gimka) 13:41 < vlaste> zo torso gimkamsmikezypro no catni gismu .i zo torso gimkamsmikezypro no cipra gismu 13:41 < zipcpi> xu gimy'inda 13:41 < ctefaho> I am trying to understand the word structure but my god what a beast 13:41 < Amnadya> http://www.ithkuil.net/abbreviations.html 13:41 < zipcpi> mi pu pensi tu'a zo insa pe la tokpona ge'u ji'a 13:42 < ctefaho> ua ki'e la'oi .Amnadya. 13:42 < Amnadya> nahi gimyhinda i zo gimyhinda valsi lo se oinmo 13:43 < zipcpi> u'i 13:43 < Amnadya> lo gismu na vajmau lo drata brivla 13:43 < ctefaho> za'a do na nelci .y'y. 13:43 < mensi> i mi i mi i mi mo i mi na nelci 13:44 < Amnadya> lo fonxa na nelci 13:44 < mensi> ji'a mi mutce nelci i ie 13:44 < ctefaho> je'e 13:44 < ctefaho> some part of me still instantly thinks "fonxa" means fox 13:45 < Amnadya> ithkuil: fonxa 13:45 < ithkuil> fonxa: P1S1/STA-f-ERG-nx-EXS-IFL/UNFRAMED 13:45 < ctefaho> so P1S1 in that chart is? 13:45 < ctefaho> ... 13:45 < ctefaho> P1S1 doesn't seem to be in the chart nor P or S 13:46 < Amnadya> they are the vowel before the root 13:46 < Amnadya> part of the root meaning 13:46 < ctefaho> ithkuil: bananas 13:46 < ithkuil> bananas: P1S1/STA-b-OBL-UNI/COA/M/DEL/NRM-AGC_1/5-PLA_1/5-EXS-IFL/UNFRAMED 13:46 < ctefaho> root as in "Cr"? 13:46 < Amnadya> bananas = abananas 13:47 < Amnadya> yes 13:47 < ctefaho> = ? 13:47 < ctefaho> it is elliptical? 13:47 < Amnadya> the a can be omitted 13:47 < ctefaho> right 13:47 < Amnadya> also 13:48 < Amnadya> look at the gloss again. there are dashes and slashes 13:48 < Amnadya> dashes separate different rafsi 13:48 < ctefaho> dashes separate the word components I presume? 13:48 < Amnadya> slashes are different categories 13:48 < ctefaho> ..yeah 13:49 < ctefaho> yeah so far I am with you 13:49 < Amnadya> some rafsi contain information about multiple categories 13:49 < ctefaho> and the same sound can be in different categories too 13:49 < ctefaho> (s) 13:49 < Amnadya> sure 13:50 < Amnadya> Cr means root Consonant 13:50 < ctefaho> but ok whatever the first a does is P1S1? 13:50 < Amnadya> C is always consonant 13:51 < zipcpi> rafsi: tos 13:51 < mensi> y no da se tolcri 13:51 < zipcpi> rafsi: zma 13:51 < mensi> zo zmadu se rafsi zo'oi zma 13:51 < zipcpi> tosymabru 13:51 < Amnadya> the a before the b 13:51 < Amnadya> is the stem and function 13:51 < Amnadya> its invisible because it was elided 13:52 < Amnadya> defaults are usually elided 13:52 < zipcpi> va'o da'i lo nu ra'oi .tos. rafsi zo torso kei zo tosymabru co'a uo se smuni 13:52 < ctefaho> ...is "a" part of the I Cv? 13:52 < ctefaho> is there a consonant elided before it too? 13:53 < Amnadya> a lot has been elided but you dont have to care 13:53 < ctefaho> je'e 13:54 < Amnadya> Vr is it yes 13:54 < zipcpi> ku'i tau zoi zoi torso mammal zoi .y. 13:54 < Amnadya> cant be Cv because its a vowel 13:54 < zipcpi> lo mabru poi se torso go u'i 13:55 < ctefaho> Pattern 1 Stem 1 (a) 13:55 < ctefaho> gotcha 13:55 < Amnadya> yes 13:56 < ctefaho> b is Cr then? 13:56 < Amnadya> yes 13:58 < ctefaho> and (a) is simply "stative manifestation"? 13:58 < ctefaho> or do the 3 stem things mess around with that 13:58 < Amnadya> which a 13:58 < zipcpi> roda je noda gimy'inda 13:59 < Amnadya> the a marks three things. stem 1 pattern 1 and stative function 14:00 < ctefaho> yeah the first elided a in bananas 14:00 < Amnadya> yes 14:01 < ctefaho> and the Cr-b simply means "want/desire/request" 14:01 < ctefaho> ? 14:02 < Amnadya> yes. and the preceding a makes it more specific 14:02 < ctefaho> so, stative manifestation of desire 14:03 < ctefaho> ...whatever that means 14:03 < Amnadya> state of desire 14:03 < ctefaho> je'e 14:03 < ctefaho> just looked up the word "stative" 14:03 < ctefaho> kinda sorta helped 14:03 < ctefaho> camxes: kinda sorta 14:03 < camxes> ([kinda sorta] VAU) 14:03 < ctefaho> ++ 14:03 < ctefaho> better gismu have never been coined 14:04 < ctefaho> next a is Vc then I believe 14:05 < Amnadya> yes. case 14:05 < Amnadya> oblique case 14:05 < ctefaho> content 14:05 < ctefaho> that one I at least know what it means 14:06 < zipcpi> mi uenmo ru'e lo nu zo mroxadni punai se jbovlaste 14:06 < Amnadya> the next affix is a Ca affix and combines five categories at once 14:07 < ctefaho> State of desire, that which is desired? 14:07 < ctefaho> the word is what is desired in some state of desire? 14:07 < ctefaho> hmm 14:07 < Amnadya> zahi djica 14:07 < ctefaho> but OBL? 14:08 < Amnadya> the participants are not specified. theyd be the nouns 14:08 < Amnadya> verbs take oblique case 14:09 < zipcpi> lo'e zazdumu lo'e nundumu frica ma 14:09 < zipcpi> .y. 14:09 < ctefaho> is this a verb then? 14:09 < Amnadya> unless they are in framed relation, in which case they take any noun case 14:09 < zipcpi> lo'e zazdumu lo'e nundumu cu frica ma 14:09 < Amnadya> tosmabru 14:09 < zipcpi> za zdumu 14:10 < ctefaho> well what does the Vc-a do to the word then? 14:10 < Amnadya> we cant be sure its a verb 14:10 < zipcpi> li'a zozdumu cnino gismu 14:10 < Amnadya> but lets just say yes 14:10 < ctefaho> when did it stop being a verb then? 14:10 < ctefaho> or when will it 14:10 < Amnadya> it didnt 14:10 < Amnadya> its just both 14:10 < Amnadya> dont worry bout it 14:10 < ctefaho> okay 14:11 < Amnadya> and oblique case does nothing with verbs 14:11 < Amnadya> just a place holder 14:11 < zipcpi> How does a state differ from an event? 14:11 < ctefaho> I see 14:11 < ctefaho> n is COA ? 14:12 < Amnadya> see above about Ca 14:12 < Amnadya> the n marks five categories 14:13 < Amnadya> separated by slashes 14:14 < Amnadya> configuration 14:14 < ctefaho> the "UNI/COA/M/DEL/NRM"? 14:14 < Amnadya> yes 14:14 < Amnadya> affiliation 14:14 < ctefaho> one letter marking 5 categories at once? 14:14 < Amnadya> etc 14:14 < Amnadya> yes 14:14 < Amnadya> the a also marked three 14:15 < ctefaho> ok found the tables 14:15 < ctefaho> are you a computer or something 14:15 < ctefaho> seriously how do you speak something like this 14:15 < Amnadya> the Ca affix has 1728 values 14:16 < ctefaho> and one is supposed to learn them? 14:16 < Amnadya> yes. but you could learn them naturalky 14:16 < Amnadya> through exposure 14:16 < ctefaho> no 14:17 < ctefaho> just no 14:17 < Amnadya> uhi 14:17 < ctefaho> that seems totally unfit for a human language 14:17 < Amnadya> wanna bet money on it? 14:17 < ctefaho> I rather learn a hundred more BAI than that 14:17 < PrincessBecca> what is this about, is this about places again 14:17 < Amnadya> BAI suck tho. 14:18 < Amnadya> most of them anyway 14:18 < ctefaho> agreed 14:18 < ctefaho> I cut a whole lot of them out in my dialect 14:18 < Amnadya> tenselike BAI are good 14:18 < ctefaho> and some other changes 14:18 < ctefaho> I am now sitting on 100 free non-experimential cmavo that I don't know what to do with 14:18 < zipcpi> PrincessBecca: It's about Ithkuil 14:18 < PrincessBecca> oh 14:18 < ctefaho> oh well usage will probably be found 14:19 < ctefaho> as for Ithkuil 14:20 < Amnadya> zoho we are only at the n so far 14:20 < ctefaho> Uniplex configuration, Coalescent affiliation, Monadic perspective, Delimitive extension, Normal essence 14:20 < Amnadya> correct 14:20 < ctefaho> ...is it possible to explain that in a short and simple way? 14:20 < Amnadya> not for me on a phone 14:21 < ctefaho> but it means some crazy massive semantic stuff 14:21 < ctefaho> gotcha 14:22 < ctefaho> next a is "Suffix: Degree of Agency" 14:22 < Amnadya> yes 14:22 < ctefaho> plus 1 14:22 < Amnadya> the a marks the fifth degree 14:22 < Amnadya> a-n 14:23 < Amnadya> the n is the suffix 14:23 < ctefaho> plus Existential context, Informal designation 14:23 < ctefaho> just more semantic specifiers then? 14:23 < Amnadya> the context is default so ignore it 14:24 < Amnadya> the gloss shouldnt show it 14:24 < ctefaho> cliva said he would fix it 14:24 < Amnadya> informal designation comes from the penultimate stress 14:25 < Amnadya> which tells you which stem to pick 14:25 < ctefaho> I get the idea that it puts a massive amount of semantics in simple words 14:26 < ctefaho> /in a small space 14:27 < zipcpi> exp: tu'e broda .i brode tu'u 14:27 < mensi> (tu'e [{CU <broda VAU>} {i <CU (¹brode VAU¹)>}] tu'u) 14:27 < ctefaho> so uniplex is some holistic state of desire? 14:27 < zipcpi> exp: (tu'e [{CU <broda VAU>} {i <CU (¹brode VAU¹)>}] tu'u) goi ko'a 14:27 < mensi> SyntaxError: Expected [yY] but "o" found. 14:28 < zipcpi> exp: tu'e broda .i brode tu'u goi ko'a 14:28 < mensi> SyntaxError: Expected [yY] but "o" found. 14:28 < zipcpi> exp: tu'e broda .i brode tu'u cu'e ko'a 14:28 < mensi> SyntaxError: Expected [yY] but "o" found. 14:28 < Amnadya> the formative makes little sense i think 14:28 < zipcpi> exp: tu'e cu'e ko'a broda .i brode tu'u 14:28 < mensi> (tu'e [{<cu'e ko'a> <CU (¹broda VAU¹)>} {i <CU (¹brode VAU¹)>}] tu'u) 14:28 < Amnadya> there is still another suffix 14:29 < Amnadya> -as 14:29 < ctefaho> I thought you said to ignore it? 14:29 < Amnadya> thats not it 14:29 < zipcpi> exp: cu'e ko'a tu'e broda .i brode tu'u 14:29 < mensi> SyntaxError: Expected [yY] but "u" found. 14:29 < ctefaho> ah 14:30 < zipcpi> I've heard of proposals to expand the grammar of tu'e...tu'u but I'm not sure how 14:30 < Amnadya> PLA suffix 14:30 < Amnadya> fifth degree 14:30 < zipcpi> What I want to do is associate a tu'e...tu'u group to a sumti 14:30 < ctefaho> some setting 14:30 < ctefaho> and some 5th degree specifier? 14:30 < zipcpi> I made {goi'i} to associate a bridi to a sumti 14:31 < ctefaho> place/state 14:32 < Amnadya> place where one keeps or shelters 14:32 < ctefaho> I can't seem to find where PLA is defined 14:33 < Amnadya> suffixes 14:33 < ctefaho> suffixes? 14:33 < ctefaho> ah je'e 14:33 < ctefaho> "place where one keeps or shelters" right 14:35 < ctefaho> is all of "5-EXS-IFL" relevant? 14:35 < Amnadya> 5 is part of the suffix 14:35 < Amnadya> fifth degree 14:36 < ctefaho> what is EXS? 14:36 < ctefaho> can't find the def 14:36 < Amnadya> EXS is irrelevant here. existential context 14:36 < ctefaho> oh there 14:37 < Amnadya> its not even in the word 14:37 < Amnadya> the vowel is elided 14:37 < Amnadya> it would be bananasa 14:38 < ctefaho> so it is another word then 14:38 < Amnadya> which is the same 14:38 < ctefaho> ? 14:38 < Amnadya> but default elided 14:39 < Amnadya> existential context is default 14:39 < zipcpi> exp: tu'e cu'e ko'a broda .i brode tu'u coi ko'a 14:39 < mensi> (tu'e [{<cu'e ko'a> <CU (¹broda VAU¹)>} {i <CU (¹brode VAU¹)>}] tu'u [coi ko'a DOhU]) 14:39 < ctefaho> asa is the same as "abanan"? 14:39 < Amnadya> eh? 14:39 < ctefaho> what is the same then? 14:40 < ctefaho> the elided first a? 14:40 < Amnadya> as what 14:40 < ctefaho> <Amnadya> which is the same 14:40 < ctefaho> ->? 14:40 < Amnadya> bananas = bananasa 14:40 < ctefaho> ah 14:40 < Amnadya> the final is is the context 14:40 < Amnadya> final a 14:40 < ctefaho> ahh 14:40 < ctefaho> now I get what you meant 14:40 < ctefaho> "its not even in the word" 14:41 < Amnadya> phew 14:41 < ctefaho> thought you meant sa falls off 14:41 < ctefaho> ...as 14:41 < Amnadya> i cant speak well on the phone sry 14:41 < ctefaho> nah I am just tired 14:41 < ctefaho> so sa 14:41 < ctefaho> sa is IFL? 14:41 < Amnadya> as-a 14:42 < Amnadya> no 14:42 < Amnadya> IFL is informal designation. it comes from the penultimare stress 14:42 < ctefaho> then I am lost 14:44 < ctefaho> if it is neither of AGC_1/5-PLA_1/5-EXS-IFL/ 14:44 < ctefaho> or is the sa that 5? 14:44 < ctefaho> AS 14:45 < Amnadya> as i said the 5 is part of the suffix 14:45 < Amnadya> an is AGC_1/5 14:46 < Amnadya> as is pla1/5 14:46 < ctefaho> right 14:46 < Amnadya> fifth degree of the suffix 14:46 < Amnadya> type 1 14:46 < ctefaho> I just mixed things up again 14:47 < ctefaho> "one who does/is made to do" 14:48 < ctefaho> "place where one learns to" 14:48 < Amnadya> no 14:48 < Amnadya> five 14:48 < ctefaho> sa "place where one keeps or shelters" 14:48 < Amnadya> yes 14:49 < ctefaho> so "State of Desire" + "Uniplex configuration, Coalescent affiliation, Monadic perspective, Delimitive extension, Normal essence" + "one who does/is made to do" + "place where one keeps or shelters" 14:49 < Amnadya> so that is what bananas is 14:49 < ctefaho> plus content-case 14:49 < ctefaho> that's bananas man zo'o 14:50 < Amnadya> right on 14:50 < ctefaho> and I still have no idea what that is supposed to mean 14:51 < Amnadya> bit nonsensical 14:51 < ctefaho> does it mean anything to you? 14:51 < Amnadya> not really 14:51 < ctefaho> la'a 14:51 < ctefaho> English gloss would be?;) 14:52 < Amnadya> too painful 14:53 < zipcpi> mensi: doi durka42 I finally discovered why I accidentally add comments to the main thread for a word rather than the definition-specific thread. When you click "view comments for definition", then "post new", it posts to the main thread, not the definition-specific thread 14:53 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.durka42.gy. di'a cusku da 14:53 < Amnadya> one who is made to desire but which is also a place where one keeps desire 14:53 < Amnadya> it is gibberish 14:54 < zipcpi> Deep 14:54 * zipcpi rings a zen gong 14:55 < ctefaho> je'e 14:55 < ctefaho> btw 14:55 < ctefaho> you do remind me of someone 14:56 < ctefaho> who also prefers this teaching style 14:56 < Amnadya> really 14:56 < ctefaho> selpai 14:56 < ctefaho> (pun on senpai that no one seems to get) 14:56 < durka42> ua 14:56 < mensi> durka42: cu'u la'o gy.zipcpi.gy.: I finally discovered why I accidentally add comments to the main thread for a word 14:56 < mensi> rather than the definition-specific thread. When you click "view comments for definition", then "post new", it posts to 14:56 < mensi> the main thread, not the definition-specific thread | 2015-07-09T21:53:23.152Z 14:56 < Amnadya> i was just about to mention it 14:56 < ctefaho> at least iirc from long time ago 14:57 <@xalbo> senpi pajni 14:57 < ctefaho> xalbo: exactly 14:57 < ctefaho> when I he teached me my first jbobau 14:58 < Amnadya> I'm sorta him. But on vacation. 14:58 < ctefaho> hasn't been around much lately unfortunately:| 14:58 < ctefaho> lol man you should play along;) 14:58 < Amnadya> i meant im trying to be like him 14:58 < ctefaho> now everyone will know 14:59 < zipcpi> xu la selpa'i cu selpre so'i da 14:59 < Amnadya> everyone already does know 14:59 < ctefaho> uh huh 14:59 * ctefaho blinks 14:59 < zipcpi> .y. se selpre 14:59 < zipcpi> xu la selpa'i cu se selpre so'i da 15:00 < ctefaho> I really am tired 15:00 < ctefaho> ... 15:00 < Amnadya> ithkuil melted you brain 15:00 < ctefaho> guess I read you too literal 15:00 < ctefaho> also brb 15:02 < Amnadya> How did you know who i was zoho 15:03 < ctefaho> instinct zo'o 15:03 < ctefaho> ...and my brain has been kinda melted the last 2 weeks 15:04 < ctefaho> language wise 15:05 < ctefaho> adding Ithkuil to the mix doesn't help;) 15:05 < Amnadya> you made a good effort today 15:05 < ctefaho> I do see its utility though 15:06 < ctefaho> cramming so much semantic potential into so little space 15:06 < ctefaho> quite the opposite of jbo's formal rafsi system... 15:09 < ctefaho> but the implementation, my god 15:09 <@xalbo> It's almost as though every Lojban predicate needed everything specified in full. {lu'omi'o ca'acaca'obu'uve'uze'avi'eta'enaicasnu} 15:09 < ctefaho> 45 consonants and 13 vowels 15:09 < ctefaho> at that point you can just go over and speak full IPA 15:09 < Amnadya> but that lojban phrase still contains a lot less nuance 15:09 < ctefaho> xalbo: .u'i 15:10 < Amnadya> german has more vowels 15:11 <@xalbo> True. The question is, is it possible in Ithkuil to express {mi'o casnu}, with its own nuance? 15:11 < ctefaho> swedish probably has more 15:11 < ctefaho> 9 or 18 depending how you count... 15:12 * ctefaho closes the Ithkuil book for now though 15:12 < ctefaho> also I just discovered http://www.lojban.org/files/etymology/langstat.99 today 15:12 * xalbo wonders how many consonants and vowels English hath. 15:12 < ctefaho> I don't think I have seen that referenced anywhere 15:13 < ctefaho> i.e. how the gismu scores came to be 15:13 < zipcpi> English has about 13 vowels 15:14 < zipcpi> Pronunciation-wise 15:14 < zipcpi> Not counting dipthongs 15:14 < zipcpi> diphthongs 15:14 < ctefaho> (found it when poking around in gimyzba) 15:16 * nuzba @s_i_d_a: carvi sance .i lo karce zi pagre .i ba smaji sicpi #lojban #haiku [http://bit.ly/1KZiwjN] 15:18 < ctefaho> "Given that the Lojban gismu list is baselined, these numbers are primarily for academic interest." *evil laughter* 15:35 < zipcpi> ku'i pe'i ka'eku lei meti cmana fe'edei'a je fe'eco'u da 15:35 < zipcpi> .y. 15:36 < zipcpi> la'a na drani 16:20 * nuzba @uitki: Welcome!/en - http://mw.lojban.org/papri/Welcome!/en by Wuzzy - Cleaned up list of dictionaries, added la vlasisku [http://bit.ly/1Csyjqb] 16:28 < ctefaho> co'o rodo 19:56 * nuzba @ligervision: @Fnxtx @evanmcm New thread: #lojban #Singularity When the Transhumanist succeeds in becoming God-like, his prison is complete. I can prove. [http://bit.ly/1HgzTbg] 22:55 < gleki> well, {kinda} is {su'ova'e} 23:01 < gleki> alta: li mu cu mo 23:01 < mensi> ([FA {li <mu BOI> LOhO}] [cu {mo VAU}]) 23:02 < gleki> alta: lo mlatu li pinxe 23:02 < mensi> ([FA {lo <mlatu NF> KU}] [li {pinxe NF} VAU]) 23:33 < gleki> http://mw.lojban.org/papri/cu_li --- Day changed Fri Jul 10 2015 00:44 < gleki> Amnadya: mi mo'u jmina zo'ei lo sance http://mw.lojban.org/extensions/ilmentufa/cilretci/ http://mw.lojban.org/extensions/ilmentufa/cilretci/data.js 00:47 < gleki> en: online 00:47 < mensi> 3 da se tolcri: jbovlaste, jibgaucpe, xa'i 00:47 < gleki> u'i 00:47 < gleki> en: jibgaucpe 00:47 < mensi> jibgaucpe [< jibri gasnu cpedu ≈ Job do request] = x1 applies for job x2 to x3 in form x4. |>>> The x3 is likely the 00:47 < mensi> future employer. The x4 can be used to say "apply online" or "apply by mail". See also jibri, gasnu, cpedu. |>>> 00:47 < mensi> djeikyb 00:47 < gleki> en: xa'i 00:47 < mensi> xa'i = [UI3] discursive: imagining/roleplaying - not imagining/"out of character / in real life" |>>> Denotes imaginary/ 00:47 < mensi> roleplaying actions and events. Serves much the same purpose as *star-quotes* or the /me function [= xa'i mi] would in 00:47 < mensi> online communities. Cf. da'i, xanri, xarpei |>>> spheniscine 01:09 < Amnadya> ua doi la gleki 01:10 < Amnadya> ihesai 01:11 < tsani> coi 01:11 < gleki> ei zo cilretci se basti zo'oi PooLearning 01:12 < Amnadya> uhi 01:12 < tsani> .i mi gleki lo nu viska lo so'i nintadni ne lo se .irci .i ca bo tcidu lo se .irci pruselsku 01:12 < gleki> ba'a do ba ze'usai tcidu 01:12 < tsani> .u'i .i mi lo pagbu ku po'o cu tcidu 01:13 < tsani> .i xu ma nuzba lo jbogu'e 01:13 < gleki> la cilretci zei'a mabla vau za'adai 01:14 < tsani> .i zo cilretci ki'a zo zei'a ki'a 01:14 < tsani> .i lo jbobau ku ki'a zo'o 01:14 < gleki> zo zei'a co'e lu fi'o zenba li'u 01:14 < tsani> ua 01:14 < gleki> jbo: zei'a 01:14 < mensi> zei'a = [TAhE] zo zei'a sumtcita gi'e cmavo ma'oi ta'e .i ca'e lo du'u «zei'a ko'a broda» cu du'u «zi'o zenba lo ni 01:14 < mensi> broda kei ko'a» kei gi'e va'i du'u «lo ni broda cu zilzena ko'a» |>>> srana fa zo zenba joi zo dei'a joi zo zilzena |>>> 01:14 < mensi> spheniscine 01:15 < gleki> en: zei'a 01:15 < mensi> zei'a = [TAhE] tense interval modifier: increasingly...; incrementative. Tagged sumti, if present, indicates amount of 01:15 < mensi> increase (lo te zenba) |>>> Alternate form to avoid confusion: ze'ai. See zenba. Opposite of dei'a |>>> 01:15 < mensi> spheniscine 01:15 < gleki> xm 01:15 < gleki> zo dei'a je'e 01:15 < gleki> i la sutysisku zo'u do djuno i na sanji lo drata nuzba 01:15 < gleki> i ua 01:16 < gleki> i mi pu zi finti la culis ku noi bankle 01:16 < gleki> alta: mi li monsuta 01:16 < mensi> ([FA mi] [li {monsuta NF} VAU]) 01:16 < tsani> ua ie .i mi milxe co zgana lo nu do pu pilno 01:16 < tsani> .i na ku'i smujimpe 01:17 < gleki> zo li ka'e dunli zo cu 01:17 < gleki> i smudunli lu mi cu monsuta li'u 01:17 < tsani> .i xu djica lo nu lo jbobau lo tokpona cu binxo 01:17 < tsani> zo'o 01:17 < gleki> lo tokpisini 01:17 < tsani> ua 01:18 < gleki> i lo jbobau co'a bankle be lo tokpisini 01:19 < tsani> .i pe'i mi ca'o no'e smujimpe zo'e ne zo zei'a 01:19 < gleki> i xu do toktoki lo tokpisini i cu'ei ie mi likliki toktoki 01:19 < tsani> .i sa'e lu zi'o zenba li'u nonselsmu mi 01:20 < gleki> mi zenba lo ka certu la lojban i mi zei'a certu la lojban 01:20 < tsani> .i lo di'u nunpli ku .i'e nai 01:21 < tsani> .i lo nu lo sumtcita cu se vajni pe'a ru'e lo du'u makau sumti pamoi cu na'e racli 01:21 < tsani> .i ki'u ma lo nu pilno lo tanru ja lo lujvo cu banzu 01:21 < tsani> naku 01:22 < gleki> enfoka 01:23 < tsani> vlaste: enfoka 01:23 < vlaste> enfoka = x1 is in the focus of x2; x1 is the point where x2 focuses, converges 01:23 < gleki> i lu mi zei'a certu li'u zo'u zo certu enfoka 01:23 < gleki> i lu mi zenba lo ka certu li'u zo'u zo certu na enfoka 01:24 < gleki> i lu enfoka pe'a li'u la'a i ku'i na djuno lo du'u makau poi valsi cu mapti i zo xusra la'a 01:24 < tsani> .i ko jdice lo du'u makau enfoka vau sa'e 01:24 < gleki> i lu mi zei'a certu li'u zo'u zo certu enfoka 01:24 < tsani> .i ma jdice lo du'u zo certu na enfoka lu mi zenba lo ni certu li'u 01:24 < tsani> .u'i 01:25 < gleki> va'o ku zo certu cu nenri lo na'e ralju bridi i seni'i bo na enfoka 01:25 < tsani> .i va'i lo du'u makau enfoka cu na'e vajni mi 01:25 < tsani> .i da'i cusku lu .i mi certu zenba li'u 01:26 < gleki> lu i mi zenba i fe bo certu la lojban li'u zo'u ro lo re selbri cu enfoka 01:26 < gleki> to ku'i zo ka canci i'enai toi 01:27 < gleki> i mutce clani tanru pilno nelci fa do ibo ja'o 01:27 < tsani> .i pe'i smudu'i fa lu mi ze'ai certu li'u lu mi zenba lo ni certu li'u .e lu mi certu zenba li'u 01:28 < tsani> *selsmudu'i 01:28 < gleki> lo tanru na satci 01:28 < gleki> ni'o ma nuzba API pe JVS doi samcre 01:29 < tsani> no da .u'i .i mi zukcfudu'e vau .oi 01:29 < gleki> oidai 01:30 < gleki> i ku'i mi bredi loka favgau la sutysisku si sutytiksku 01:32 < tsani> .i do pacna lo nu la sutysisku cu pilno API ne JVS ku xu 01:32 < gleki> go'i 01:33 < gleki> i ku'i co'a nunstika be JVS proga 01:34 < tsani> .i lo nu favgau JVS cu xagmau lo nu na favgau JVS .i ku'i xagrai fa lo nu gunka te zu'e lo nu la vlasisku ba za basti JVS 01:39 < gleki> mi nitcu fe po'o API 01:39 < gleki> i mi to'e certu tu'a la vlasisku 01:40 < gleki> i nu'oku la vlasisku zenba lo ni prane 01:40 < gleki> *cu 01:43 < gleki> i sa'e mi nitcu API pe lo JVS datni sorcu ju JVS 02:03 * nuzba @uitki: La Bangu: Lojban Dictionary with Examples - http://mw.lojban.org/papri/La_Bangu:_Lojban_Dictionary_with_Examples by Gleki [http://bit.ly/1MjTPNO] 02:11 * nuzba @uitki: L17-01 - http://mw.lojban.org/papri/L17-01 by Gleki [http://bit.ly/1JTub4Z] 03:57 < zipcpi> Hm I noticed that {gei} is a bit weird... exponent first, significand second, base last 03:58 < zipcpi> Guess they decided that order of magnitude was most important? 03:58 < gleki> en: gei 03:59 < gleki> selmaho:gei 03:59 < zipcpi> gei is VUhU 03:59 < mensi> gei = [VUhU2] trinary mathematical operator: order of magnitude/value/base; [b * (c to the a power)]. |>>> 03:59 < gleki> i know i just want to test mensi's links to cll 03:59 < mensi> officialdata 03:59 < mensi> .i lu gei li'u cmavo zo'oi VUhU2 03:59 < gleki> heh 04:00 < gleki> i'll fix sutysisku first 04:02 < gleki> fixed. will appear after the next update of the manifest file or you can just clear you appcache 04:02 * gleki wondering how many ofsuch fixes are needed 04:03 < gleki> we just need links to cll for every selmaho and every subselmaho http://mw.lojban.org/extensions/ilmentufa/i/en/bangu.js 04:03 < Amnadya> la ctefaho ca certu la itkuhile 04:03 < gleki> je'upei 04:03 < Amnadya> nai zoho 04:03 < Amnadya> i kuhi lonu troci cu tolfaho roda 04:03 < gleki> je'uja'aizo'onaipei 04:04 * gleki that last compund cmavo will go into history 04:05 < Amnadya> la vreji cu rejgau gaje lo nahe cinri gi lo cizra je lo banli 04:06 < zipcpi> {dektenfa} is weirdly defined; I'm not sure what x2 is for? 04:14 < gleki> where is the list of all selmaho and subselmaho? 05:04 < zipcpi> vlaste: bitni {gimka} 05:04 < vlaste> no results. http://vlasisku.alexburka.com/bitni+%7Bgimka%7D 05:04 < zipcpi> vlaste: bitni (gimka) 05:04 < vlaste> zo bitni gimkamsmikezypro no catni gismu .i zo bitni gimkamsmikezypro no cipra gismu 05:04 < zipcpi> I think I like {bitni}; matches {baitni}, and reminds one of {klani} 05:05 < zipcpi> And don't have the {birti/bitri} problem 05:06 < zipcpi> Yeah sounds like "beatnik" but I don't think that's an issue. "beatnik" is hardly gimy'inda :p 05:11 < gleki> a female gismu, i see 05:11 < zipcpi> u'i 05:14 < gleki> en: bitri 05:16 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 05:17 < gleki> lmw is superslow 05:23 < gleki> jb: bitni 05:23 < mensi> bitni = bitni — x1(entity) is x2(number) bits in size 05:23 < mensi> :lo se bitni — bit (data). 05:23 < mensi> :lo vi vreji cu bitni lo kilto — This file is 1000 bit. 05:23 < mensi> :Related words: datni 05:23 < gleki> as you wish 05:24 < zipcpi> ki'e 05:26 < gleki> en: bitni 05:27 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 05:27 < gleki> en: baitni 05:27 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 05:29 < zipcpi> la jbovlaste cu maltarti za'ure'u .i li'a tolnei kanpe lo nu mi jmina ai ji'imuno valsi 05:41 * nuzba @uitki: L17-03 - http://mw.lojban.org/papri/L17-03 by Gleki [http://bit.ly/1HS7H25] 05:41 * nuzba @uitki: L17-04 - http://mw.lojban.org/papri/L17-04 by Gleki [http://bit.ly/1HpcTWy] 05:49 < zipcpi> I'm getting these errors: Couldn't open temporary html file; check your definition for unbalanced dollar signs or other wierd characters. 05:51 < zipcpi> http://jbovlaste.lojban.org/dict/kilto 05:54 < gleki> report to github 08:32 * nuzba @garydale: Fahrenheit 1600: The Amazon servers take even more heating… Alltop is hot for the funny. Cartoon via H... http://bit.ly/1HpVcq1 #lojban [http://bit.ly/1HSqIBq] 08:34 < durka42> ckini la lojban ba'e ma 09:01 < gleki> la'a fi'o xanri ra la lojban cu du la cnino ve tavla 09:01 < gleki> ku noi du la nuspik 09:02 < gleki> Fiction: At the GruntWerx Board of Directors Meeting: http://bit.ly/1IRyUON #lojban 09:02 < gleki> cizra 09:04 < gleki> http://cogas.github.io/hajiloji/article/nunctu/18.html 09:04 < gleki> melbi 09:04 < durka42> ua 09:05 < gleki> mo i djica lo ka fanva fi lo glico ibo xu 09:07 < durka42> mi na se bangu 09:08 < gleki> la gugl cu mo 09:08 < durka42> uat 09:09 < gleki> la gugl transleit cu mo 09:09 < durka42> la'a tolcertu 09:09 < gleki> xu vajni va'o nu do jimpe fi le ralju bangu 09:10 < gleki> i ko larcu je satci be na fanva 09:46 * nuzba @octophore: I reflexively trust Esperanto enthusiasts more than Lojban enthusiasts [http://bit.ly/1HiwMj2] 10:18 * nuzba @Nubobot42: @lucasq_ eu falo em lojban [http://bit.ly/1HSgOlj] 10:21 < gleki> ue 10:21 < gleki> i lo zmiku xu 10:23 * nuzba @Nubobot42: Lojban retweeted you [http://bit.ly/1HShBmg] 10:29 < durka42> y 10:30 < durka42> bu'u lo sotygu'e... 10:39 < zahlman> ra'oi sot rafsi ma 10:39 < mensi> y no da se tolcri 10:39 < gleki> uesai 10:39 < gleki> rafsi:sot 10:39 < mensi> y no da se tolcri 10:39 < gleki> i oi la mensi cu spofu 10:39 < zahlman> ue ii la mensi cu menli 10:40 < zahlman> i spuda lo preti 10:40 < gleki> ra'oi sot rafsi ma 10:40 < mensi> zo so'u se rafsi zo'oi sot 10:40 < gleki> di'a tolspofu 10:40 < zahlman> io 10:42 < zahlman> ni'o mo 10:43 < gleki> la lojban co'a mabla 10:43 < gleki> i mu'a zo li co'a smuni dunli zo cu 10:44 < gleki> alta: lo jbocre li jai fenki 10:44 < mensi> ([FA {lo <jbocre NF> KU}] [li {jai <fenki NF>} VAU]) 10:46 < zahlman> ii 10:46 < zahlman> NF mo 10:48 < gleki> sinxa lo famyma'o pe lo me'oi noun 10:48 < gleki> NF ... ku => noun 10:48 < gleki> NF vau => verb 10:49 < zahlman> si'au na sarcu 11:23 < gleki> sarcu lo nu fanva fi lo glico a lo simsa 11:26 < durka42> ta'i ma cusku lo namcu .u'i 11:32 < gleki> alta: li mu li jai se o'unmo 11:32 < mensi> ([FA {li <mu BOI> LOhO}] [li {jai <se (¹o'unmo NF¹)>} VAU]) 12:03 < sorpaas> Is {la .ibu mo} a valid Lojban cmene? Can I say {la .ibu mo cu mo}? 12:04 < gleki> exp: la ibu mo 12:04 < mensi> ([la {<(¹i bu¹) BOI> mo} KU] VAU) 12:04 < gleki> ^ 12:04 < gleki> it is 12:04 < zahlman> ue 12:04 < gleki> la ... ku 12:04 < zahlman> si uenai mi bebna 12:05 < gleki> one of Lojbanists has the name {la mokau} 12:05 < zahlman> tatpi ji'a oi 12:06 < durka42> lo mo pe ibu 12:07 < sorpaas> But {la .ibu} is not valid. Am I correct? 12:08 < durka42> correct 12:09 < sorpaas> ui 12:18 < noncomcinse> coi 12:37 < gleki> coi 12:58 < sorpaas> exp: la go'i cu mo 12:58 < mensi> ([la go'i KU] [cu {mo VAU}]) 12:59 < sorpaas> exp: la mo cu mo 12:59 < mensi> ([la mo KU] [cu {mo VAU}]) 13:09 * nuzba @la_taun: ti ba'e NA pe'i cinmo condi pe'a pemci! .i zbalermorna .i srilermorna .i .a'iru'e ze'i finti .i mu'o #lojban http://t.co/eN7Mkou63h [http://bit.ly/1KVhzL2] 13:46 < sorpaas> I'm just trying to explore wired things... What place structure does {nei} have? What does it mean if I say {mi tavla nei do}? 13:47 < sorpaas> exp: mi tavla nei do 13:47 < mensi> (mi [CU {tavla nei} {do VAU}]) 13:48 < samcinki> coi la .sorpa'as. 13:52 < sorpaas> coi 13:53 < sorpaas> coi la .samcinki. 13:56 < Ilmen> Hi Sorpaas! 13:57 < Ilmen> ze'a lo barda na vitke 13:58 < Ilmen> Did you get the email I sent you several months ago? 13:59 < Ilmen> As for {nei}, it's a brika'i (pro-bridi ~ pro-sentence), so it hasn't any intrinsic place structure. 14:00 < sorpaas> No, I probably didn't :( 14:00 < Ilmen> {mi tavla nei do} --> {mi tavla tavla nei do} --> {mi tavla tavla tavla nei do} --> ... — This seems to be infinitely recursive. 14:00 < Ilmen> Oh. 14:00 < Ilmen> Let me check. 14:01 < sorpaas> Oh, you mean "zy zei voksna"? Yeah, I found it! 14:01 <@xalbo> Worse, it's {mi tavla ke tavla nei ke'e be fa mi bei do do} etc 14:01 < Ilmen> I've sent it to you on 2015-01-31, and it was titled "zy zei voksna" 14:02 <@xalbo> It's like standing between two mirrors. 14:02 < sorpaas> Thanks for the voksa! 14:02 < durka42> na nei 14:02 < Ilmen> sorpaas: Okay, happy to know you successfully received it. 14:02 < sorpaas> {nei} sounds extremely fun! 14:03 <@xalbo> durka42: Dammit, now I need to reboot the universe again. 14:03 < Ilmen> .u'i 14:04 < Ilmen> This one is worth being added to the IRC pearls 14:06 < Ilmen> ta'o ma nuzba lo jbogu'e 14:07 < Ilmen> en: morgana 14:07 < mensi> morgana = x1 is a mirage/Fata Morgana |>>> selpahi 14:09 < Ilmen> ŭe zantufa zmiboto 14:11 <@xalbo> There is a baseball player whose last name is "Trumbo". Every time I see it, I can't see it as anything other than a slinku'i. 14:11 < Ilmen> ba'a nai zo morgana pagbu su'o selsa'a pe la'au za'o li'u 14:15 < Ilmen> In a similar fashion, everytime I hear the French sentence pattern « Tout le monde n'est pas [predicate] » (= naku ro da [predicate]), I've the persistent feeling of there being a logical scope error .u'i 14:16 < Ilmen> s/tout/tous --mute fenki 14:21 < Ilmen> "Tout le monde n'est pas..." = "Not everybody is..." = naku roda... So it feels like a scope error, although it seems it isn't because it's a set expression 14:23 < Ilmen> en: asfalte 14:23 < mensi> asfalte = x1 is asphalt/bitumen |>>> See tarla |>>> selpahi 14:25 < Ilmen> ji'a pagbu lo selsku pe su'o selsa'a .i cinri koinsidensi .i ka'e jijyni'i lo du'u lo selsa'a cu se ciska pu za lo masti be li so'i 14:34 < Ilmen> .i .oi ti'e vi ba zi di'a tcima glare mutce ze'a lo bramau be lo pa jeftu soi mabla 14:34 < Ilmen> .o'o 14:35 < Ilmen> mu'o mi'e monlogo'ysku 15:21 * nuzba @BWACG: Hey #Lojban speakers, a Lojban StackExchange site has been proposed http://area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture?referrer=onNWtue0kWMRAmkuNTrNeQ2 [http://bit.ly/1HSKJK4] 17:34 < ctefaho> camxes: pujenaicajeba 17:34 < camxes> ([{pu <(¹je nai¹) ca> <je ba>} KU] VAU) 17:34 < ctefaho> cll is really fun sometimes 17:34 < ctefaho> calling that a "word" 17:41 < ctefaho> co'o 18:30 < _mukti_> coi jbopre 18:31 < ComradeBecca> coi la mukti 18:53 < _mukti_> ue coi la .bekas. 19:21 < vultrue> Ìû 19:21 < _mukti_> iu dai 19:22 < _mukti_> cpana 20:13 < zipcpi> Crazy thought: Can brivla be trademarked? 21:28 < ocimar> hi 22:35 < gleki> tcika lo nu smaji --- Day changed Sat Jul 11 2015 00:55 < le_jatna> mi viska le nei, I see myself? 00:56 < gleki> yes 00:57 < gleki> not a usual style for Romance languages and English, but normal for e.g. Slavic languages 00:57 < le_jatna> What about, I see a reflection of myself 00:57 < Tirifto> Hello! 00:58 < gleki> "I see a reflection of myself" - mi viska lo minra be vo'a 00:58 < gleki> i'd use {vo'a} here beause it always refers to the main clause ({mi viska} in this case) 00:58 < gleki> hello, Tirifto 00:58 < Tirifto> Question: When I want to accent a syllabe, do I need to capitalize all its letters (ex. lojBAN), or is the first one enough (ex. lojBan)? 00:59 < Tirifto> (Accent as in put emphasis on it/prolong it) 00:59 < le_jatna> ki'e 00:59 < gleki> usually you capitalize the whole syllable but these days a usual stress mark like e.g. ` is more popular 00:59 < gleki> although in general marking stress in Lojban is a very rare practice 01:00 < Tirifto> I see. Changing stress won't change meaning of the word, right? 01:00 < gleki> wont change, yes 01:00 < gleki> and that's the explanation why it isnt popular 01:01 < Tirifto> I see :P 01:01 < Tirifto> Also, how do I use the ` mark? Do I put it before/after the syllabee, or add it to vowels? 01:01 < gleki> i guess like in English, before the vowel 01:03 < Tirifto> ` is used in English/ 01:03 < Tirifto> ?* 01:04 < gleki> in English the stress mark is used beforre the stressed thing, in some other languages after 01:05 < Tirifto> I see 01:05 < gleki> it's not ` of course. it's something else which i dont have on my ASCII keyboard :/ 01:05 < Tirifto> Ohh 01:06 < gleki> Some Lojbanists just use diacritic marks for stress 01:06 < gleki> which is even more problematic for me in particular 01:06 < Tirifto> This thing? ˈ 01:07 < gleki> http://selpahi.de/ 01:07 < gleki> ^ this is an example of Lojban with stress marked using diacritic marks 01:08 < gleki> melbi (pretty) although not easy to use by everybody 01:08 < Tirifto> I see. What difficulties does it bring to some people? 01:08 < gleki> for me is that i dont have such a keyboard since i work from different devices including smartphones 01:09 < gleki> not everywhere i can install the necessary software 01:09 < Tirifto> Ah 01:10 < Tirifto> So is that method generally discouraged? 01:11 < gleki> with diacritics? no, not discouraged. it's readable, it's easy to be used by computer programs analyzing the structure of Lojban text. It's just that not all can use it. 01:13 < Tirifto> Okay! One more question: Do people sometimes use words from lojban vocabulary as names? 01:13 < Tirifto> (Or pseudonyms) 01:13 < gleki> yes 01:13 < gleki> mi'e la gleki 01:13 < gleki> = I'm Happy 01:13 < gleki> where "Happy" is a personal name 01:14 < gleki> (as opposed to "happy") 01:14 < Tirifto> Does the "la" distinguish it from a non-name? 01:15 < gleki> {la} is a prefix word making names 01:15 < gleki> gleki = to be happy 01:16 < gleki> lo gleki = someone happy (noun) 01:16 < gleki> la gleki = Happy (personal name) 01:16 < Tirifto> Oh I see, I mistook it with lo at first. 01:18 < Tirifto> If one was to have multiple names (ex. name surname), do they put "la" in front of each one (la name la surname), or just in front of the first one (la name surname)? 01:18 < gleki> only one {la} is necessary 01:19 < gleki> each {la} creates one name 01:19 < gleki> My full nickname is {la gleki arxokuna} 01:19 < gleki> "Happy Raccoon" literally although it shouldn't be translated. 01:22 < Tirifto> That is a nice name! https://lut.im/5ZZPhZxi/dQSzPquf 01:23 < Tirifto> Thank you for your help! :) 01:24 < gleki> je'e (= You are welcome!) 01:24 < gleki> oh, where did you find my photo, zo'o (= kidding) 01:25 < Tirifto> That's a secret >.> 01:26 < Tirifto> And one more thing: Is the ' pronounced as the IPA/Esperanto "h"? 01:26 < gleki> Usually yes. 01:27 < Tirifto> Okay, thanks! 01:27 < gleki> It *can* be pronounced as any other frricative phoneme not already used by any other Lojban letters. E.g. voiceless "th" or voiceless "wh" but that's rarely used. 01:29 < gleki> mensi: doi Amnad https://github.com/lagleki/glekitufa/blob/master/cilretci/README.md 01:29 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.Amnad.gy. di'a cusku da 01:33 < gleki> mi de'a jundi 01:33 < gleki> (I pause listening) 02:20 < Amnadya> mi nenri lo sorpeka 02:20 < mensi> Amnadya: cu'u la'o gy.gleki.gy.: https://github.com/lagleki/glekitufa/blob/master/cilretci/README.md | 02:20 < mensi> 2015-07-11T08:29:33.921Z 02:21 < Amnadya> jehekihe la gleki 02:30 < Amnadya> i pu fliba tau lo jai se platu sorpeka kihu lo bebna krinu noi mi nohe fuzme i mi ca nenri lo drata sorpeka noi klama ba tau lo jai se platu sorpeka za lo cacra be li ci 02:31 < Ilmen> .oi dai 02:33 < Ilmen> .i .a'o na ku do denpa tau lo sorpeka bu'u lo glare dukse ze'a lo cacra be li ci 02:36 < Amnadya> na fasnu i kuhi tohe kihu nai mabla ikihubo pu nitcu loka erve lo cnino pikta noi kargu sai je noi mutce kargymau lo pamoi 02:36 < Ilmen> .i mi zo'u ca lo prulamdei lo na se kanpe fasnu cu selja'e lo nu mi nitcu lo ka denpa ze'a lo cacra fau lo nu lo solri cu dirce darxi pe'a je bu'u lo glare mutce je fau lo nu lo se ralte djacu cu cmalu 02:37 < Ilmen> s/cacra/cacra kei to ko smaji doi la fenki toi 02:37 < fenki> Ilmen meant to say: .i mi zo'u ca lo prulamdei lo na se kanpe fasnu cu selja'e lo nu mi nitcu lo ka denpa ze'a lo cacra kei to ko smaji doi la fenki toi fau lo nu lo solri cu dirce darxi pe'a je bu'u lo glare mutce je fau lo nu lo se ralte djacu cu cmalu 02:37 < Ilmen> :/ 02:37 < Amnadya> oidai 02:41 < Ilmen> ta'e spaji mi fa lo nu lo .irci cu .irci fau lo nu zvati lo marce 02:42 < Amnadya> cnino mijiha 02:42 < Ilmen> .i mi frimo'i lonu ca da su'o .irci cu .irci tavla fau lo nu sisku lo nicte marbi gi'e na snada 02:44 < Amnadya> ie mi morji 02:44 < Amnadya> i pu sfehero 02:45 < Ilmen> .i pu stidi lodu mi fonxa tavla lo pendo be vo'a .i ku'i mi jungau lodu mi na se bangu lo mapti .i ku'i ba ku ba'a nai snada loka marbi facki kei xoi lo .irci be lo drata cu sidju 02:45 < Amnadya> ie 02:52 < Ilmen> mi na birti lodu xukau zabna fa lonu ze'e kibro jundi ku'i .i za'a so'i da pe lo cabycedra fau lo nu denpa co se sruri lo drata cu jundi lo samfonxa gi'e na jikca troci vau ja'e lo nu simlu loka pe'a zvati lo drata munje 02:52 < Amnadya> taho au pei cpacu lo selsaha selsku pe la Djemynai i xu noda do pu dunda be ru i mi djuno ia lodu do na kakne loka erve 02:52 < Ilmen> .i ji'a lo nu ka'e ku ca ro da lo drata ka'e foxycpe cu la'a fanza 02:53 < Ilmen> ru'a ru'e 02:53 < Amnadya> ie 02:53 < Amnadya> kahe je puhi kibro dukse 02:53 < Amnadya> fa lo remna 03:01 < Ilmen> mi ca na ka'e terve'u .i ku'i .a'o ba kakne .i lo se finti be la .djem. cu banli dukse lo nu mi na ckire dunda su'o da kei ĭa .i ji'a mi da'i na se pluka lo nu jvarivbi cpacu lo selsa'a selsku .i ji'a xamgu fa lo nu selsku smadi ta'i lo ka tinju'i 03:01 < Ilmen> *dukse fi 03:02 < Amnadya> do nobli zaha 03:02 < Amnadya> mi sinma 03:03 < Ilmen> to zo jvarivbi .a'u pei toi 03:03 < Ilmen> je'e 03:03 < Amnadya> mi jimpe 03:04 < Ilmen> pe'i pei mapti lo si'o me'oi cheat 03:04 < Amnadya> pu na pensi loduxukau xagrai valsi 03:04 < Amnadya> ahu uhi 03:05 < Ilmen> va'o lo nu na mabla dukse zo'u mi jmina .ai 03:05 < Amnadya> javrihi sei pajvo 03:08 < Amnadya> mojriha mi lodu aidji loka jmina zo zvasku noi cohe lu zvati sisku lihu 03:08 < Amnadya> je noi ba fengygau la uuzit 03:09 < Ilmen> .u'i je'e 03:09 < Amnadya> i mi ba pilno sebahi zo mitysisku noi lo tarmi mi fegli 03:10 < Amnadya> i lo smuni cu simsa kuhi 03:11 < Ilmen> jbo: sisku 03:11 < mensi> sisku = x1 zukte da lo nu x1 facki lo du'u ma kau poi me x3 cu ckaji x2 |>>> srana fa zo facki .e zo zvasisku .e zo 03:11 < mensi> mitysisku |>>> Ilmen 03:12 < Amnadya> ua zo zvasisku zasti 03:12 < Amnadya> na morji 03:13 < Ilmen> .u'i se'i lo zgike pe la'au ki'e li'u snipa lo mi besna .i lo galto'a voksa cu simlu lo ka cusku zo ki'a ku mi xoi xajmi 03:13 < Amnadya> uhi mi pu kanpe 03:26 < Ilmen> .ai mi citka ta'o .i co'o di'ai do .i mi gleki lo nu casnu 03:27 < Amnadya> jehe i raho 04:10 < gleki> ue lo jbocre cu tavla zu'ai 04:11 < Amnadya> uhi iepei coha tolcafne 04:11 < mensi> ei mi tugni 04:12 < Amnadya> sihau glitce 04:14 < Amnadya> zo zihe morsi i koi zo zihe jmive 04:14 < Amnadya> i mi zgana su nu pilno 04:15 < gleki> oi do li pilno la tcekitau 04:17 < Amnadya> cohaku soha jbocre cu gohi i ei mi fukpi tarti 04:17 < Amnadya> doi li zei prenu 04:19 < gleki> cohaku so'o jbopre cu pante 04:20 < gleki> i ta'o bu'u la'oi #espanol da pu cusku zo'oi nombre ije bazibo mi pensi de 04:21 < gleki> ne lo lujvo 04:21 < Amnadya> jetahe noroi jbosku 04:22 < Amnadya> ihu 04:36 < Amnadya> ei mi ma te cmene lo malgli pagbu be la itkuhile ma ca lonu tavla fe muha la kixadas i lo na jbopre na jimpe tau zo malgli 04:37 < Amnadya> s/mi ma/mi 04:37 < fenki> Amnadya meant to say: ei mi te cmene lo malgli pagbu be la itkuhile ma ca lonu tavla fe muha la kixadas i lo na jbopre na jimpe tau zo malgli 04:38 < gleki> According to G. Cinque the relative order or Tense, Aspect, Mood rafsi is preserved the same across languages. which means that it's possible to find fixed slots and thus create naturalistic lujvo with meaning predictable from its slot structure. but /(de'a / ka'e)+/ isn't immediatly clear to me. 04:39 < gleki> https://www.dropbox.com/home/loi%20cukta?preview=ENG+Cinque%2CG.+-+Adverbs+and+Functional+Heads.pdf 04:41 < gleki> or https://www.academia.edu/2679553/Again_on_Tense_Aspect_Mood_morpheme_order_and_the_Mirror_Principle_ 04:42 < gleki> by this rule {co'a ca} is to be fixed into {PU co'a ca} and probably {ca co'a ca} 04:55 * mumble_ jorne co'a fa la'o fa menli fa 05:18 * mumble_ jorne co'u fa la'o fa menli fa 05:18 * mumble_ jorne co'a fa la'o fa menli fa 05:40 * nuzba @uitki: La Bangu: Lojban Dictionary with Examples - http://mw.lojban.org/papri/La_Bangu:_Lojban_Dictionary_with_Examples by Gleki - /* Types of places */ [http://bit.ly/1JaFZu7] 06:07 < Amnadya2> zutse je zutse je zutse 06:18 < gleki> tersmus: zutse je zutse je zutse 06:18 < tersmus> <<zutse(_)>{je}<zutse(_)>(_)>{je}<zutse(_)>( ) 06:18 < gleki> i'enai 07:00 < niftg> coi 07:02 < Ilmen> coi 07:03 < niftg> la'a lo cabdei ke jbobau nu vokta'a cu mulno pu 07:04 < gleki> xu pu nu tavla zu'ai 07:07 < Ilmen> je'u nu'o voksnu ca lo cabdei 07:08 < Ilmen> .i ku'i mi za'o jundi .i ku'i la'a mi ba zi klama lo bartu 07:09 < niftg> .ua snuti nonseka 07:09 < Ilmen> ni'o mi pu zi finti lo ti kibypapri — http://mw.lojban.org/papri/UserWiki:Ilmen/la_nu_bevri_do 07:09 < Ilmen> .i'u dai .i mi pu cnegau joi cu dragau vau so'o da 07:10 < gleki> zo'oi UserWiki mabla gi'e glico 07:13 < niftg> ta'o .uinai la zbaga mi noda spuda 07:13 < Ilmen> ŭe ru'e 07:15 < niftg> xu do snada lo nu zvati la zbaga doi za'o jundi 07:15 < Ilmen> mi je'a jundi nau ca'a 07:15 * mumble_ jorne co'u fa la'o fa menli fa 07:15 * mumble_ jorne co'a fa la'o fa menli2 fa 07:16 < Ilmen> cizra 07:16 * mumble_ jorne co'a fa la'o fa menli2 fa 07:16 * mumble_ jorne co'a fa la'o fa vreji fa 07:16 < Ilmen> doi la .niftyg. mi ca lifri lo simsa nabmi 07:17 < Ilmen> .i ku'i mi ka'e nerkla .i va'o bo po'o lo nu pilno zo'oi 84.73.73.130 .enai zo'oi zbaga.ax.lt:64738 07:17 < cliva> niftg: Ilmen: si'au da pe DNS pe la zbaga cu nabmi .i ku'i za'a la jinme bunkere di'a .akti 07:18 < Ilmen> je'e pei la .niftyg. 07:18 < Ilmen> .e'u gau do zo'oi 84.73.73.130 basti zo'oi zgaga.ax.lt 07:18 < niftg> je'e je'e 07:18 < Ilmen> *zbaga 07:19 < Ilmen> je'e la cliva 07:19 * mumble_ jorne co'a fa la'o fa SuperUser fa 07:19 * mumble_ jorne co'u fa la'o fa SuperUser fa 07:19 < Ilmen> ta'o .ai mi me do si si cliva si'a 07:20 < cliva> je'e 07:20 < cliva> .i mi ca lo cabdei na zifre tu'a la mumble .i la'a ca lo bavlamdei ja'a co'e 07:20 < Ilmen> je'e 07:21 * mumble_ jorne co'a fa la'o fa niftyg fa 07:21 < Ilmen> .i ba'a mi zifre lo ka voksnu ca lo bavlamdei ku ji'a 07:22 < Ilmen> s/voksnu/vokta'a 07:22 < fenki> Ilmen meant to say: .i ba'a mi zifre lo ka vokta'a ca lo bavlamdei ku ji'a 07:22 < Ilmen> co'o ta'o 07:22 < cliva> jetcyco'o 07:22 < niftg> je'e co'o 07:23 < niftg> snuti snada co jvokatna 07:23 < cliva> .u'i 07:23 < niftg> .u'u katna gi'e ku'i na gismu ciksi 07:38 < niftg> .uinai lo ni mi pixokauroi catra cu te zmadu lo ckuliki lo fetyja'a manti poi se nalci gi'e cliva troci 07:49 < gleki> cinri fa lo nu zo zu'ai na kakne tu'a lo se xi pa broda 07:54 < gleki> For those of you curious to hear the Ithkuil language sung to music, or who are interested in progressive-rock music in general, I have just published a video on YouTube of the first track from the forthcoming album by Kaduatán (the Ithkuil word for “Wayfarers”), the music project I’ve been working on for the past year and a half. 07:54 < gleki> https://youtu.be/KM-r-4dIMic 07:54 < gleki> </quote> 08:02 < _mukti_> Perhaps conlangers can band together and get some attention for the conlang music boom. 08:02 < mensi> _mukti_: cu'u la'o gy.gleki.gy.: pls disable VisualEditor and TweetAnew | 2015-07-11T12:28:55. 08:02 < mensi> 681Z 08:02 < _mukti_> First a lojban rap album, now ithkuil progressive rock? 08:06 * nuzba @_angelesperanto: Ŝajnas al mi preskaŭ pli facile lerni la latinan ol lojban! Ĉu iu esperantisto instruos lojban al mi per interreto? [http://bit.ly/1Hmv1DK] 08:10 * nuzba @lojban_org: @_angelesperanto jes, bonvolu uzi IRC-kanalon #lojban en freenode: http://mw.lojban.org/extensions/ilmentufa/irci/ [http://bit.ly/1HmvRQU] 08:10 < _mukti_> coi ue la selpa'i 08:11 < selpahi> coi 08:11 < selpahi> .i mi pu zi mo'u xrukla lo zdani 08:11 < janAnken> Saluton? 08:11 < _mukti_> fi'i dai 08:11 < _mukti_> coi la'o zoi jan anken zoi 08:11 < janAnken> Mi ne scipovas ankoraŭ la lojbanan lingvon bedaŭrinde. 08:12 < gleki> Saluton 08:12 < selpahi> Cxu vi volas lerni gxin? 08:12 < janAnken> Jes. 08:12 < selpahi> :) 08:12 < gleki> Cxu vi volas lerni gxin *nun*? 08:12 < selpahi> .u'i 08:12 < janAnken> Okej 08:12 < selpahi> bone 08:12 < janAnken> Ĉu .u'i estas "amuza"? 08:13 < gleki> {u'i} estas interjekcio 08:13 < gleki> ~= "amuze" 08:13 * gleki ni bezonas traduki cnima'o en E-on 08:13 < selpahi> ~= lol / haha 08:14 < janAnken> Kompatinda _mukti_. Li ne povos daŭre babili lojbane nun. 08:14 < gleki> kial? 08:15 < janAnken> Ni estas ĉi-tujmesagile babilante esperante. 08:15 < gleki> ah, kompreneble 08:15 < gleki> _mukti_: cxu vi komprenas nenion? 08:16 < _mukti_> mi parolas iomete 08:16 < gleki> janAnken: kiajn lingvojn vi parolas? 08:17 < janAnken> La hispana kiel denaska lingvo, ankaŭ scipovas tokiponon kaj iom esperanton kaj la anglan. Mi studas nunteme tokiponon, la anglan, esperanton kaj la klasiklatinan. 08:18 < janAnken> Kiaj? 08:18 < gleki> kiujn 08:18 < gleki> sed ni povas paroli en la angla 08:18 < janAnken> Maybe... 08:18 < janAnken> I'm not sure if I will understand you. 08:19 < gleki> i think it's the first or the second time someone tried to learn Lojban from Esperanto. Before that i saw only attempts to do that in Japanese, French, Russian 08:19 < gleki> most textbooks are in English http://mw.lojban.org/papri/Learning/en 08:19 < gleki> We are lazy to translate them into Esperanto 08:19 < janAnken> I was trying to learn this language with an english written book and I couldn't. The spanish translation is horrible. 08:20 < gleki> you may choose another resource from the link above. 08:20 < gleki> another book 08:20 < selpahi> Kaduatán is cool. It sounds like Finnish when pronounced like that :) 08:20 < gleki> janAnken: and then ask questions here 08:20 < janAnken> I have a question. 08:21 < gleki> re'i (=Listening, ready to receive) 08:21 < gleki> janAnken: go on 08:21 < janAnken> What is the exact sound for “.” is it a click, glotal sound stop or simply a no-noise time.? 08:22 < janAnken> I can't express myself correctly. 😞 08:22 < gleki> glottal stop or a short pause 08:22 < gleki> "a no-noise time" as you said, yes. You may also use that 08:22 < janAnken> What is more common? Glotal or pause? 08:22 < gleki> pause 08:23 < janAnken> How are you? 08:24 < gleki> I'm fine :) 08:24 < janAnken> Who* sorry 08:25 < janAnken> Are you the owner of “@lojban_org”? 08:25 < gleki> Ehm, what do you mean? 08:25 < janAnken> Have you answered me? 08:25 < gleki> oh, yes and no. Iy's a shared account. Several people use it. 08:25 < gleki> oh, yes and no. It's a shared account. Several people use it. 08:25 < xajmi> coi 08:25 < gleki> But that reply was from me, yes. 08:25 < gleki> xoi la xajmi 08:25 < gleki> sa 08:25 < gleki> i coi la xajmi 08:26 < xajmi> coi la gleki 08:26 < xajmi> coi rodo 08:26 < janAnken> I come in a moment. 08:26 < _mukti_> la xajmi .e la gleki cu zvati .i ku'i la xalbo cu zvati ma 08:26 < janAnken> Okey 08:27 < gleki> za'a la'oi VisualEditor mo'u cliva 08:27 < _mukti_> go'i 08:27 < gleki> i ku'i la'oi TweetAnew na co'e 08:27 < _mukti_> Really, I turned that off, too. 08:28 < janAnken> I was in my smartphone. Could you send me again the links? I have lost them. 08:28 < gleki> i ua i ie 08:28 < gleki> janAnken: mw.lojban.org/papri/Learning 08:28 < _mukti_> I also turned off parsoid. I think that will help with the performance. 08:28 < gleki> _mukti_: so'u da pu pilno i seki'u bo na pu mutce vajni i ku'i ca ti mi bilga lo ka finti lo drata tutci pe la tu'iter 08:28 < xajmi> mi jmisre la jbobau ku 08:29 < _mukti_> .i'e 08:29 < gleki> Actually we need textbooks in Simple English 08:29 < _mukti_> ua nai dai doi la xajmi 08:30 < janAnken> I have to learn the real British English. So :( 08:30 < janAnken> I must. 08:30 < xajmi> mi rapcreze'a 08:30 < xajmi> Good luck with that janAnken 08:30 < janAnken> Thanks. 08:31 < janAnken> I think I have B1 08:31 < janAnken> Do you know many lojban spanish speakers? 08:31 < gleki> lo munje zo'u so'i da se bangu lo glico poi su'eva'e lo ni sampu 08:32 < xajmi> I am a spanish soeaker... but there are not many 08:32 < xajmi> are you? 08:32 < gleki> janAnken: Jorge Llambias from Argentina, Dank Writer (nickname) from somewhere idk 08:32 < janAnken> Yes 08:32 < janAnken> Soy español 08:32 < janAnken> From Spain I mean 08:32 < xajmi> hola janAnken, soy latino 08:33 < janAnken> Hablas lojban? 08:33 < xajmi> estoy aprendiendo 08:33 < janAnken> Is Jorge Lambias Dank Writer? 08:33 < xajmi> es por pura curiosidad... primera vez que veo a otro que habla español aqui 08:33 < _mukti_> coi la durkavore 08:33 < janAnken> It has no sense 08:33 < xajmi> no lo conozco 08:33 < janAnken> JAJAJA 08:34 < durka42> coi 08:34 < janAnken> LA primera vez que entro aqui 08:34 < gleki> janAnken: no, those are two persons 08:34 < xajmi> coi durka42 08:34 < janAnken> La segunda vez que intento empezar a aprender lojban 08:34 < _mukti_> ca lo cabdei ma'a se bangu lo spano .e lo spero 08:34 < gleki> ie 08:34 < durka42> ca lo cabdei mi cliva seka'a lo dotygu'e 08:34 < _mukti_> ua .i'e 08:34 < xajmi> algo que funciona es tratar de traducir cosas que ya sabes 08:35 < janAnken> Dice que hay otros nativo español, Jorge Lambias 08:35 < xajmi> por ejemplo trate de escribir el cuento de la caperucita roja en lojban 08:35 < gleki> xu do aidji lo ka penmi la jboberlinpre doi la durkavore 08:35 < janAnken> Pero hablas ingles tu? 08:35 < durka42> .au 08:35 < _mukti_> http://tiki.lojban.org/tiki/la+cmalu+je+xunre+mapku 08:35 < xajmi> claro... sino no podria entender lojban 08:36 < xajmi> of cpurse 08:36 < xajmi> *course 08:36 < gleki> u'i 08:36 < janAnken> When should I use "le"? 08:37 < gleki> 1. you may always use {lo} and not {le} 08:37 < xajmi> its a plural of "la" 08:37 < _mukti_> ue 08:38 < janAnken> I have just read "le" in several lojban sentences!! 08:38 < gleki> 2. if you really want to use it then it's the same as {lo} but mentioning things in context, e.g. already mentioned, that you have on your mind. 08:38 < gleki> as for plural it's {za'u} in Lojban 08:38 < gleki> {lo mlatu} = cat/cats 08:38 < xajmi> le means "the ones known as..." its a plural "article" 08:38 < gleki> {lo pa mlatu} = cat, one cat 08:38 < gleki> {lo za'u mlatu} = cats (plural) 08:38 < _mukti_> The older dialects of lojban preferred {le}. It's meaning in contemporary lojban is the subject of quite a bit of discussion. 08:39 < _mukti_> coi ui la menli 08:39 < Ilmen> coi 08:39 < janAnken> "lo" is the plural of "la" 08:40 < janAnken> why "lo pa mlatu" is "one cat"? 08:40 < janAnken> la mlatu ??? 08:40 < samcinki> "lo" is not the plural of "la". 08:40 < _mukti_> The most prolific lojbanist these days tend not to use {le} at all. 08:40 < _mukti_> (lojbanists) 08:40 < xajmi> "lo" just takes the "x1" position of any selbri... 08:40 < gleki> {la} creates names 08:40 < gleki> la rozgu = Rose (personal name) 08:40 < gleki> lo rozgu = rose (flower) 08:41 < xajmi> "lo gerku" means "that which is dog/dogs" 08:41 < gleki> 08:40 < janAnken> why "lo pa mlatu" is "one cat"? <-- well, that's how Lojban is made :) 08:41 < _mukti_> {lo pa mlatu} is, more precisely, "that which is a cat and is counted as one, in the way that cats are counted" 08:41 < gleki> xajmi: or simply "dog/dogs" 08:41 < janAnken> lo is used to refer to something which has been said in x1 before, isn't it? 08:42 < gleki> it's the most generic particle, so rather no. 08:42 < gleki> in English or Spanish terms it just makes nouns out of verbs. 08:42 < janAnken> A xajmi que te llamas Jaime no? 08:42 < xajmi> "pa" is the lojban word for one, and "lo" transforms a selbri (verb) into a sumti (noun), so "lo pa mlatu" means "that one cat" or something similar 08:43 < xajmi> No... xajmi means "funny" 08:43 < gleki> sazri = to drive, conducir 08:43 < gleki> lo sazri = driver(s), conductor(es) 08:43 < selpahi> Hello again. Could someone please send me the logs for the last two weeks? My logging facility went down while I was on vacation. 08:43 < xajmi> mi nobre es Darío (dari.o) 08:43 < xajmi> *dari.os 08:44 < xajmi> *nombre 08:44 < selpahi> dari'os / dariios 08:44 < janAnken> Okey, I think I have understand lo 08:45 < janAnken> understood 08:46 < xajmi> not "dari'os"... that puts an unnecessary "h" between "i" and "o"... "dari,os" might be better, but the coma is not exactly accepted in lojban so I use a dot "dari.os" 08:47 < xajmi> Any of you understands lojban very well? 08:47 < xajmi> or are the experts gone? 08:47 < selpahi> The dot makes it two words, the first of which is actually two cmavo 08:47 < selpahi> dari.os = da ri .os. 08:48 < gleki> experts are still here 08:48 < _mukti_> indeed, right now :) 08:48 < xajmi> Oh... thanks... I will use the coma then, because I don't really like the "h" in the middle... sorry about that 08:48 < _mukti_> so'o jbocre cu zvati 08:49 < janAnken> Every word which starts with a vocal has a dot before that letter? 08:49 < xajmi> _mukti_ can you help me with something? 08:49 < selpahi> You can also say {dariios}, which is /da.ri.jos/ in IPA 08:49 < gleki> selpahi: e'u do na ciksi lo cnino javni vau u'i 08:49 < selpahi> If you don't want the h. 08:49 < _mukti_> xajmi: Sure, but I'm not one of the experts. :) 08:49 < gleki> xajmi: just ask questions. someone will answer 08:49 < xajmi> thanks sepahi 08:49 < xajmi> *selpahi 08:50 < durka42> coi ls selpa'i 08:50 < durka42> la* 08:51 < janAnken> Every word which starts with a vocal has a dot before that letter? 08:51 < xajmi> I want to say "grandma leaves using a motor bike" 08:51 < xajmi> I try to say that with "lo rirmanta ku klama fu lo matra relxilma'e ku" 08:51 < xajmi> is it right? 08:52 < selpahi> Yes. (rirmamta) 08:52 < gleki> *rirmamta 08:52 < selpahi> janAnken: Yes. The dot represents a glottal stop. 08:53 < janAnken> a glottal stopa made by the glotis or only a silence time. (being quiet) isn't it? 08:54 < _mukti_> janAnken: In general, yes, but combinations like {ui} and {ii} which start with a consonant sound (help me out, linguists!) don't require the dot 08:54 < xajmi> yes a dot is just a "silent pause" 08:54 < durka42> semivowel* 08:54 < _mukti_> Damn. I typed that then doubted myself. 08:54 < janAnken> u and i consonants?? 08:54 < selpahi> This is the first time I don't feel like reading a two-week log file... 08:55 < gleki> in {ui} {u} is usually pronounced as "w" 08:55 < _mukti_> {ui} sounds like english "we", french "oui", etc. 08:55 < gleki> in {ii} the first {i} is like "y" in Spanish "soy" 08:55 < selpahi> {u} before another vowel is [w] (IPA), and {i} before another vowel is [j] (IPA) 08:56 < xajmi> janAnken and _mukti_ I think that the dot is used to separate words, specially qhen they start with a vowel, because they can sound similar to a diphtong 08:56 < xajmi> *when 08:56 < janAnken> thanks 08:57 < janAnken> How many time someone has to spend to speak lojban? 08:57 < xajmi> ii suena parecido a decir "yi" en español 08:57 < _mukti_> Somewhere mixed into all of that is the idea that, in theory at least, lojban could be written without spaces. 08:57 < xajmi> I think this video might come in handy... 08:58 < xajmi> wait a second 08:58 < gleki> _mukti_: do u have a ready to use tool showing the intensity of this channel in different hours over the last months? 08:58 < _mukti_> Nothing ready to use but it wouldn't be hard to whip something up given the logs. 08:58 < gleki> _mukti_: measured by the number of messages/unique users in a given hour 08:58 < xajmi> https://www.youtube.com/watch?v=d0yGdNEWdn0 08:58 < gleki> hm, logs ... 08:58 < _mukti_> Did someone help selpahi with those, BTW? 08:59 < gleki> _mukti_: i have logs but in certain days there were disrupted. this can damage stats 08:59 < gleki> i ie mi mo'u benji fi la selpa'i 09:00 < gleki> if we combine logs from different people this can decrease deviation 09:01 < gleki> _mukti_: i think i can provide you with the logs but i think it'd be necessary to split the output by weeks so that deviations are easily seen 09:01 < xajmi> why everyone has gone quiet? 09:02 < janAnken> Aprender con hipnosis?? 09:02 < xajmi> no 09:02 < xajmi> u'i u'i u'i 09:02 < gleki> it's weekend here on this planet, so maybe that's why 09:02 < gleki> i just wanted to leave btw 09:02 < xajmi> basicamente tienes que practicar... asi practiques mal 09:03 < xajmi> you want to leave gleki? 09:03 < gleki> janAnken: do you use facebook too?? 09:03 < xajmi> yes 09:03 < gleki> xajmi: i will return, my account will be still connected 09:03 < janAnken> I don't use it too much... 09:04 < xajmi> oh... OK... thanks for the heads up gleki ... 09:04 < xajmi> I don't connect with facefook here 09:05 < xajmi> te interesan los idiomas janAnken? 09:05 < xajmi> aparte de lojban claro... 09:05 < janAnken> Si... ingles, español, tokipona, esperanto, lojban y latin es con los que estoy ahora 09:05 < janAnken> tambien empece japones, frances, portugues y aleman. 09:06 < xajmi> oi'o 09:06 < janAnken> oi'o? 09:07 < xajmi> oi'o = "fuck!!!" 09:07 < janAnken> jajajajaj por que?? 09:07 < xajmi> ouch... cuantos idiomas 09:07 < janAnken> Los de abajo no tengo ni idea y de lojban tampoco 09:07 < Ilmen> Hola 09:07 < janAnken> xajmi tu dia de suerte 09:07 < janAnken> otro que habla español 09:08 < xajmi> hola limen 09:08 < janAnken> Hola Iilmen 09:08 < janAnken> llmen 09:08 < janAnken> IImen 09:08 < xajmi> hablas español? or would you rather english? 09:08 < Ilmen> ¿Estás nuevo por aquí? 09:08 < janAnken> No se si me preguntas a mi o a iimen 09:09 < xajmi> Ilmen 09:09 < xajmi> yo he venido un par de veces... janAnken creo que si es nuevo, y la pregunta era para Ilmen 09:10 < Ilmen> Hablo ambas, luego no me importa 09:11 < xajmi> u'i ... gracioso 09:11 < Ilmen> Cierto, estoy más comfortable con inglés 09:11 < xajmi> no problem... I speako ingleso too... u'i 09:12 < xajmi> (by the way that was a joke) 09:13 < Ilmen> En este canal, se puede utilizar cualquiera lengua, en tanto que se trata de Lojban. 09:13 < xajmi> buen punto... 09:13 < xajmi> para practicar les parece si contamos chistes en lojban? 09:13 < Ilmen> O se puede hablar en Lojban a propósito de qualquiera cosa 09:15 < Ilmen> ju'o ro nu jbobau tavla cu se zanru 09:15 < janAnken> Yo como no entiendo una mierda me voy a poner a leer el famoso libro ese aunque sea en ingles. Ya preguntaré dudas por aquí. 09:15 < Ilmen> (Por cierto, qualquier hablar en Lojban está bienvenido aquí) 09:15 < janAnken> I can't speak due to I don't know Lojban 09:16 < xajmi> janAnken 09:16 < xajmi> espera 09:16 < Ilmen> do se bangu la tokpona ku ji'a xu — Hablas usted Toki Pona también? 09:17 < janAnken> Yes, mi toki e toki pona 09:17 < xajmi> toki pona no hablo 09:17 < Ilmen> je'e — Okay 09:17 < janAnken> sina li toki ala toki e toki ni? 09:17 < xajmi> podriamos traducir lo que estamos diciendo en lojban 09:18 < Ilmen> ma'a ka'e fanva lo se cusku lo jbobau 09:18 < janAnken> Eso estaría bien. 09:18 < Ilmen> da'i zabna 09:18 < selpahi> da'i xamgu 09:18 < Ilmen> (Lo que digo está traducción de lo que diced) 09:18 < Ilmen> *decid 09:19 < Ilmen> ŭe coi la selpa'i fi'i 09:19 < xajmi> da'i xamgu = supongo que esta bien 09:19 < selpahi> lo pu se cusku be mi cu xe fanva lo pu se cusku be do 09:20 < selpahi> coi la .ilmen. 09:20 < selpahi> .i mi mo'u xrukla je ca tatpi 09:22 < Ilmen> Esta demasiado raro que se habla en español por aquí, luego me precipito tan pronto como veo que aguien está hablando español jejeje 09:23 < xajmi> do tatpi ma? 09:23 < Ilmen> lo nu litru ku ru'a 09:23 < selpahi> ie 09:23 < selpahi> .i ze'a lo cacra be li za'u xa cu litru ne'i lo sorpeka 09:24 < xajmi> do tatpi ma? = tu cansado por? (por que estas cansado) 09:24 < Ilmen> .oi dai .i ditybra 09:24 < selpahi> .i ji'a ti'u li 4:30 co'a cikna 09:24 < selpahi> .i mi xa'o jungau do lo du ma kau pu fasnu 09:24 < selpahi> .i lo jai se platu sorpeka cu co'e li'o 09:24 < Ilmen> mi sanji 09:25 < selpahi> .i se jalge lo nu lo nunlitru cu zenba lo ka ditcu 09:25 < Ilmen> .i .au dai dorsipna 09:25 < selpahi> .u'i 09:25 < selpahi> .au .i ku'i lo drata ca cinrymau 09:26 < Ilmen> je'e 09:26 < janAnken> You speak a lot. 09:26 < janAnken> hahaha 09:26 < janAnken> congratulations 09:26 < xajmi> i don't speak that much... just a little 09:26 < Ilmen> jaja sí, como puedes ver, se puede hablar rápido en Lojban 09:27 < Ilmen> Sólo se necesita practicar 09:27 < Ilmen> Como con qualquiera otra lengua 09:28 < janAnken> ya... 09:28 < janAnken> creo que es dificil lo de fa fe fi fo fu y lo de las posiciones 09:29 < _mukti_> es raro que se usen todas las posiciones 09:29 < xajmi> mmm... siempre es mejor tener un diccionario al lado para recordar las posiciones... mira este http://jbovlaste.lojban.org/ o este http://mw.lojban.org/extensions/ilmentufa/i/en/index.html# 09:30 < janAnken> gracias 09:31 < Ilmen> No tan difícil; por ejemplo en la frase inglesa, "I gave you something", solo la pocisión de los argumentos indica sus papeles 09:31 < xajmi> fi'i 09:31 < xajmi> = you are welcomed (de nada) 09:32 < janAnken> ya... 09:32 < janAnken> gracias? 09:32 < _mukti_> fi'i -> bienvenido 09:32 < Ilmen> gracias = ki'e 09:33 < Ilmen> je'e = bien, he comprendido lo que has dicho 09:34 < Ilmen> es: je'e 09:34 < mensi> je'e = [COI] vocativo: recibido (reconocimiento) - reconocimiento negativo; usado para reconocer ofertas y 09:34 < mensi> agradecimientos |>>> xorxes 09:34 < janAnken> la primera posicion va siempre justo delante del selbri? 09:35 < xajmi> janAnken si quieres aprender lojban (o cualquier idioma) tienes que usarlo de alguna forma que te haga recordar las cosas (por ejemplo, si te gusta ver peliculas puedes ver una en el idioma que estas aprendiendo,aunque en lojban no es muy practico) 09:35 < xajmi> site gusta escribir, puedestratar de escribir algo en español y luego tratar de traducirlo a lojban 09:35 < janAnken> Si, eso lo hago con ingles. En los otros idiomas no hay mucha pelicula que ver... latin, esperanto, tokipona... escojo unos idiomas muy raros jajaja 09:35 < Ilmen> janAnken: el ordeno de los argumentos está muy liberado 09:36 < xajmi> coi ctefa'o 09:37 < xajmi> fi'i 09:37 < ctefa`o> coi la xajmi 09:37 < selpahi> coi 09:37 < ctefa`o> ue coi la selpai 09:37 < Ilmen> mi do viska (Yo te veo) = viska do fa mi ("veo tú yo") = do se viska mi (estas visto por mí), etc 09:37 < Ilmen> @ janAnken 09:38 < ctefa`o> .i mi bloti klama 09:38 < janAnken> mi viska do? 09:38 < Ilmen> El ordeno de los argumentes no es rígido 09:39 < Ilmen> "mi viska do" también es correcto 09:39 < Ilmen> "yo veo tú" 09:39 < Ilmen> (te veo) 09:39 < janAnken> vale, el lugar del selbri da igual si estan bien ordenados los sumti 09:39 < xajmi> mi bloti klama = I vehiclel-ing go? 09:40 < Ilmen> "I boat go" 09:40 < Ilmen> (I go via boat) 09:40 < Ilmen> ta'o za'a ca lo cabdei ro ba'u da litru 09:40 < ctefa`o> .i falnu bloti 09:40 < selpahi> ctefalnu bloti 09:40 < ctefa`o> .u'i 09:41 < astroendjisai> Hi. I have a question. Is there a selbri for anything like "x1 convinces x2 to do x3"? 09:42 < ctefa`o> It is... interesting how I have better phone 3G-connection out in the Skärgård than at home 09:42 < Ilmen> I may go for {djigau} (make+want) 09:42 < _mukti_> en: birtygau 09:42 < mensi> bitygau [< birti gasnu ≈ Certain bring about] = g1 convinces/persuades b1 that b2 is true. 09:42 < _mukti_> Not to do something, though... 09:43 < Ilmen> "x1 makes x2 certain that x2 has to do x3", maybe 09:44 < astroendjisai> I suppose I could make it more complicated and say something like "convince that it is a good idea to do" 09:44 < selpahi> jinvygau / kricygau / birtygau / tugnygau 09:46 < Ilmen> astroendjisai: Lojban's vocabulary is still growing, and there's still missing useful words, it's why anybody can create new words and add them to the dictionary 09:46 < Ilmen> The alternative to creating a new word is to use paraphrasis 09:47 < _mukti_> I stumbled on "persuade" a lot during the Diplomacy game. 09:47 < astroendjisai> I'd love to create a new word, but I don't feel I know the language well enough to do so. 09:47 < xajmi> well I gotta go now 09:48 < janAnken> bye 09:48 < _mukti_> co'o di'ai la xajmi 09:48 < xajmi> rodo zdifri 09:49 < xajmi> co'o 09:50 < _mukti_> .i mi di'a jundi ki'u lo nu lo gerki cu djica lo nu barkla 09:50 < Ilmen> X makes Y certain that Y should do Z = X birtygau Y lo du'u Y .iancu Z 09:50 < Ilmen> je'e .i .a'o zanfri 09:50 < Ilmen> astroendjisai ^ 09:50 < astroendjisai> Okay. Let's try a workaround. .i mi tadni la .lojban. .i mi na'e kakne lo su'u lo tadni la .lojban. ku xamgu sidbo 09:51 < gleki> "to convince" in some of its meanings can also be {bapli} 09:51 < astroendjisai> No I did something wrong 09:51 < Ilmen> Here, "lo tadni" fills sidbo's x1 09:51 < Ilmen> and a student is not an idea ;-) 09:52 < gleki> I can't? 09:52 < gleki> maybe I don't think that studying Lojban is a good idea? 09:53 < Ilmen> {mi na jinvi lo du'u da prali fi lo nu tadni lo jbobau} = I don't opine that there's a benefit to studying Lojban 09:53 < astroendjisai> I'm going for "I can't convince anyone that studying lojban is a good idea" but I can see that I've screwed up here. I'm gonna move to the computer and try again. Maybe I'll have better luck when I'm not on my phone. 09:54 < gleki> mi na kakne lo ka gau ce'u da jinvi lo du'u lo nu tadni la lojban cu xamgu sidbo 09:54 < gleki> long 09:54 < Ilmen> {mi na ka'e birtygau fi lo du'u xamgu fa lo nu tadni lo jbobau} = I can't convince that it's good to study Lojba 09:54 < Ilmen> n 09:54 < Ilmen> (you can also use "djigau" instead of "birtygau") 09:54 < Ilmen> (djigau = make want) 09:55 < Ilmen> (djica+gasnu) 09:55 < Ilmen> <Ilmen> {mi na ka'e birtygau fi lo du'u xamgu fa lo nu tadni lo jbobau} = I can't convince that it's good to study Lojban 09:55 < Ilmen> (you can also use "djigau" (make want) instead of "birtygau") 09:56 < astroendjisai> What's the place structure of djigau? 09:56 < Ilmen> na ka'e = cannot; lo du'u + bridi = that + subsentence; xamgu fa lo nu ... = it is good that... 09:57 < Ilmen> djigau / djicygau would mean "x1 makes x2 wants x3 to happens / to be the case" 09:58 < Ilmen> mi djigau do lo nu do barkla = I make you want that you go-out 09:59 < astroendjisai> .i zo'e fliba 09:59 < _mukti_> mi na snada lo ka birtygau da lo du'u xamgu fa lo nu cinri fi la .lojban. 10:00 < Ilmen> (djigau is a regular lujvo, it's basically a shorthand for "gasnu lo nu djica" = make want...) 10:00 < durka42> co'o .i la'a lo za'umoi be zo coi pe mi cu se cusku bu'u lo dotygu'e 10:00 < selpahi> di'ai ! 10:00 < _mukti_> ko snura litru 10:00 < durka42> ki'e 10:01 < Ilmen> zabna doi la .durkas. 10:01 < Ilmen> di'ai ri no'u ko 10:01 < Ilmen> .i mi zo'u .ai de'a jundi 10:01 < Ilmen> co'o 10:01 < _mukti_> coi .ibu 10:01 < _mukti_> co'o 10:01 < Ilmen> co'oi 10:02 < _mukti_> Apparently it's opposites day for me. 10:02 < _mukti_> When I say {coi} I mean {co'o}, {di'a} for {de'a}, etc. 10:04 < _mukti_> mi skudji lo dukti be lo se cusku be mi 10:04 < astroendjisai> So _mukti_ says... "I do not succeed in convincing anyone that it is good to be interested in lojban?" I'm genuinely confused here, how is la .lojban. the x2 of cinri here? 10:04 < _mukti_> Hey, you're getting good already. Yes, I messed up. 10:05 < _mukti_> I meant {cilre}, but on reflection, {tadni} is better here 10:05 < _mukti_> mi na snada lo ka birtygau da lo du'u xamgu fa lo nu tadni la .lojban. 10:06 < astroendjisai> Okay, so that means "I do not succeed in convincing anyone that it is good to study Lojban" 10:07 < _mukti_> Yes. 10:07 < astroendjisai> Yay! Thank you people. Now, why can I not find birtygau in the dictionary but I can find bitygau? 10:09 < gleki> they mean the same 10:09 < gleki> its actually the same word 10:10 < gleki> {bitygau} is shorter so it's in the dictionary 10:12 < astroendjisai> Ah, okay. Cool. .i mi ckire ro do 10:14 < _mukti_> Oh, I didn't even notice that I used the expanded version. 10:14 < _mukti_> I was just thinking {birti} + {gasnu} 10:15 < _mukti_> I knew there was a lujvo, but I forgot it was shorter 10:16 < _mukti_> coi la barda xirpre 10:16 < gleki> now you force me to think of rss-to-twitter script 10:17 < _mukti_> .u'u mi ja bapli tu'a do 10:17 < _mukti_> s/ja/jai 10:17 < fenki> _mukti_ meant to say: .u'u mi jai bapli tu'a do 10:18 < gleki> do ce mi jai bapli tu'a mi 10:18 < gleki> i mi ji'a pu cpedu 10:19 < _mukti_> Oh, that's old {ce} ! 10:19 < _mukti_> Took me a moment to parse. 10:20 < gleki> now it's called "old"... 10:20 < _mukti_> na cafne fa lo nu da viska lo lojbo selcmi 10:21 < gleki> zo jo'u clani 10:21 < _mukti_> Well, I called it that, anyway. Because these days when I see {ce} it tends to be {ce'u} 10:23 < _mukti_> So I am interested to know if disabling parsoid helps with the wiki performance. I suspect you will let me know. :) 10:23 < selpahi> la .xorxes. cu stidi lo du'u zo joi basti zo jo'u 10:23 < gleki> zo joi pu se smuni lo drata vau ienaipei 10:23 < mensi> ei mi tugni 10:24 < _mukti_> xu ca stidi 10:25 < gleki> selpahi: ko na co'u morji fi la mupli doi lo pilno ba la'oi PooLearning 10:25 < gleki> *be 10:25 < selpahi> mi na tolmo'i 10:25 < selpahi> .i ko na co'u morji lo du'u mi pu zi sai xrukla 10:25 < gleki> mi morji 10:25 < gleki> i sa'u 10:25 < gleki> mu'o 10:27 < _mukti_> pe'i zo joi cu mutce lo ka se pilno .i to'e mupli fa mu'a zo ce 10:27 < gleki> if anyone else has objections or suggestions related to ASCII table in Lojban please do. because i want to include parts of it into my dictionarymw.lojban.org/papri/ASCII_table_keys_names 10:27 < gleki> http://mw.lojban.org/papri/ASCII_table_keys_names 10:32 < _mukti_> Also: In (X joi Y cu broda), {broda} is non-distributively predicated of {X joi Y}, right? Whereas with (X jo'u Y cu broda), distributivity is elided. No? 10:33 < _mukti_> So if {joi} and {jo'u} were merged we would need to explicitly invoke {gunma} to get explicit collectivity? 10:34 < _mukti_> Seems like the opposite direction of xorlo ... towards singularizers. 10:34 < _mukti_> Which is why I asked if it's something he proposes currently. 10:34 < gleki> yeah, i suppose {joi} is for masses although CLL uses {joi} in casnu1 10:36 < _mukti_> gleki: That doesn't surprise me since the classic definition of {casnu} calls for a "mass" 10:36 < gleki> in la bangu several days ago i renames this and other similar places to "group" 10:36 < _mukti_> And "classic" is the word I should have used re: {ce}, not "old". 10:36 < gleki> *renamed 10:37 < gleki> so all {ce}-capable places are now called "group" 10:37 * gleki cepable 10:37 < gleki> The place structure of “casnu” is: 10:37 < gleki> casnu: the mass x1 discusses/talks about x2 10:37 < gleki> so the x1 place must be occupied by a mass (for reasons not explained here) 10:37 < gleki> [https://lojban.github.io/cll/14/15/] 10:38 < _mukti_> There's definitely a need for a term which avoids the more entangled terms like "mass" and "set". 10:39 < gleki> {abu ce by casnu} and {abu joi by casnu} have different meanings to me with the second being pragmatical nonsense. although several days ago scientists managed to connect brains of several mice into one entity 10:40 < gleki> well, "group" in the recent terminology is what can go into simxu1 and casnu1 10:40 < _mukti_> {abu joi by casnu} makes sense to me if {casnu} describes a non-distributive relationship 10:40 < gleki> maybe a lojbanic term needed for that 10:41 < _mukti_> If casnu means something like "is a discussion group" 10:41 < gleki> _mukti_: isn't it a mixture of abu and by, whereas what is intended here is {simxu lo ka ce'u ce'u tavla}? 10:41 < _mukti_> (which is how I understand it, pretty much) 10:41 < gleki> hm 10:41 < gleki> maybe then 10:42 < _mukti_> Well, the CLL definition of a mass is inherently problematic with the assertion that the mass has all the properties of every member of the mass, so if we're going to make any sense of it, I think we need to exclude that. 10:42 < gleki> then {abu ce by} and {abu joi by} is a useful distinction 10:43 < gleki> {abu joi by} is more similar to saying {lo re prenu cu casnu} 10:43 < _mukti_> As I recall, CLL is non-committal about what kinds of things can actually be predicated of a set. 10:43 < _mukti_> (aside from membership) 10:43 < gleki> mi e do casnu, mi ce do casnu, mi joi do casnu, lo re prenu cu casnu - 4 cases, all are useful 10:44 < _mukti_> I understand {abu cu by cu casnu} as something like {lo gunma be abu .e by cu casnu} 10:45 < _mukti_> Oops, messed that up 10:45 < _mukti_> I mean {abu joi by cu casnu}, of course 10:45 < gleki> looks like 10:48 < gleki> mailed to mriste about lmw 10:49 < _mukti_> On reflection, I would suppose that {abu} and {by} are quantified, so {pa gunma be .abu .e by} might be more accurate 10:49 < gleki> *bei 10:50 < gleki> oh 10:50 < gleki> sorry 10:50 < gleki> i see 10:52 < _mukti_> Assuming that I'm not interested in the idea that .abu and by might actually comprise more than one {casnu gunma}... 10:52 < selpahi> Quantfied? 10:53 < _mukti_> Can you have an unquantified reference? 10:53 < _mukti_> Assuming, btw, that abu and by are references 10:55 < _mukti_> I guess I should walk back those assumptions. 10:55 < _mukti_> .i mi nelci lo cakla .i ku'i cy boi palci 10:56 < gleki> due to recent rants about LI I'm instead removing letteral anaphora from la bangu 10:57 < _mukti_> I'm conflating anaphora and reference 11:01 * nuzba @oka_iu_tcan: このトピック面白いな…。ロジバンの口語としての姿が書かれている https://groups.google.com/forum/#!topic/lojban-soudan/jvNvLR5d5oM [http://bit.ly/1HuEfe4] 11:01 < Ilmen> "PooLearning" .u'i 11:02 < Ilmen> ta'o doi la selpa'i — http://mw.lojban.org/papri/UserWiki:Ilmen/la_nu_bevri_do 11:02 < gleki> yes, i successfully launched a new online education resource. 11:18 < _mukti_> .i ue mi viska lo cmene be do doi la gleki do'u bu'u lo mutpapri pa la'o zoi NooLearning zoi 11:19 < _mukti_> .i xu do djuno lo du'u su'o prenu pe nyly ba vitke lo jbosla 11:19 < xajmi> excuse me, does this means what i think it means? (warning swearing ahead) 11:19 < xajmi> do e'o ko gletu do 11:20 < _mukti_> s/vitke/vitke fi/ 11:20 < fenki> _mukti_ meant to say: .i xu do djuno lo du'u su'o prenu pe nyly ba vitke fi lo jbosla 11:20 < _mukti_> The first {do} isn't necessary, but otherwise, yes it does suggest a certain action which is difficult to perform nonkansa 11:21 < _mukti_> To be precise, it requests it 11:21 < xajmi> u'i u'i u'i 11:22 < xajmi> hilarious 11:22 < xajmi> thanks 11:22 < _mukti_> "Please copulate with yourself" 11:22 < xajmi> exactly... hilarious 11:22 < xajmi> thanks for clearing that up 11:22 < _mukti_> Now it could be argued -- and hell, I'll do it -- that this is not in an of itself a hostile thing to say, since {gletu} in lojban doesn't inherently imply {mabla} 11:23 < _mukti_> Which is to say, you could be requesting that the person achieve a healthy and enjoyable sexuality 11:25 < _mukti_> And as I think about it more ... I'm not sure it's desirable to assume that theres any reason that x1 and x2 in gletu couldn't co-refer... that may be creeping heteronormativity 11:25 < selpahi> ua je'e la .ilmen. 11:25 < selpahi> .i mi morji 11:25 < selpahi> .i lu bevri lo nu mi'o pensi'u li'u za'o melbi mi 11:33 < Ilmen> melbi mi ji'a 11:35 < Ilmen> doi la selpa'i lo krasi prosa cu co'e zoi ponbau いつかきっと出会う僕らを乗せて ponbau .i valsi panra fa lu bevri mi'o noi pensi'u li'u .i ku'i ĭa lu bevri lo nu mi'o pensi'u li'u smuni mapti zmadu gi'e melbi zmadu 11:36 < selpahi> je'e 11:36 < Ilmen> sa'u lo rarbau poi slabu mi cu jai cafne fai lo ka se nandu tu'a lo simsa be lo'u bevri lo nu le'u 11:36 < Ilmen> .i la'a basti fa lo panra be lu bevri ko'a noi co'e li'u 11:37 < Ilmen> mu'a 11:38 < Ilmen> mu'o mi'e di'a jundi 11:38 < selpahi> je'e 11:46 < Ilmen> mi nelci lo krasi selsa'a .i mi ri so'o roi sanga .i ri fanmo lo .anlime skina be lo nu lo re verba cu litru te zu'e lo nu zvafa'i lo dzena ke vofli daplu noi zo .laputas. cmene 11:49 < Ilmen> be'o noi cmene fa lu la .laputas. noi tsani mradi'u li'u 11:54 < Ilmen> .i .oi vi ku la ka vacri glare dukse zo'u pu ku zvati fa ri .i je ba ku zvati fa ri ja'a ku 11:54 < Ilmen> ti'e 11:55 < selpahi> .u'i 11:55 < selpahi> .oi dai 12:08 < xajmi> co'o 12:44 * mumble_ jorne co'u fa la'o fa menli2 fa 12:57 < ctefa`o> Fun fact: I have suddenly started speaking more "old"/formal swedish 12:57 < ctefa`o> Too much jbo exposure surely 13:25 * mumble_ jorne co'a fa la'o fa menli2 fa 13:40 < astroendjisai> I was just gonna be here to lurk but now I have a question. I'm seeing in a word definition z1, z2, z3, etc. rather than x1, x2, x3, etc. Is there a difference in meaning there? 13:40 < _mukti_> Yes, you're probably looking at a lujvo (compound word) ? 13:40 < astroendjisai> That's correct. 13:41 < _mukti_> So, by convention, in giving the place definitions of the lujvo, the corresponding places in the component words are identified by the initials 13:41 < _mukti_> en: skudji 13:41 < mensi> skudji [< cusku djica ≈ Mention desire] = d1 means/intends to say c2 to c3 via medium c4. |>>> 13:41 < mensi> arj 13:42 < _mukti_> skudji = cusku + djica so that d1 corresponds to the first place in djica, c2 to the second place in cusku and so on 13:42 < _mukti_> does that make sense? 13:42 < astroendjisai> Yes, that makes perfect sense. 13:47 < astroendjisai> So... .i coi .i lo nu mi jundo ro do kei ku terzu'e lo nu mi zvati kei ku .i xu la'edi'u drani 13:47 < astroendjisai> *jundi 13:48 < selpahi> drani 13:48 < astroendjisai> .ui 13:48 < _mukti_> .i'e 13:49 < astroendjisai> And am I correct in guessing I can elide pretty much every terminator there? 13:49 < selpahi> Almost... nu jundi ro do terzu'e is a tanru, so not there. 13:50 < astroendjisai> Okay. So the 'selbri word' doesn't have to be right next to the other one to form the tanru. 13:51 < selpahi> Well, nu jundi ro do kei is a selbri, and terzu'e is a selbri. They are next to each other 13:52 < astroendjisai> Oh oh okay now I get it. And I couldn't just elide the kei? 13:52 < selpahi> No, because it gets auto-inserted as late as possible, which is after {ro do} 13:53 < selpahi> The {terzu'e} can't be part of the abstraction anymore, because it already contains another selbri 13:53 < selpahi> So the abstraction has to end, and that's where {kei} gets auto-inserted. 13:53 < selpahi> Ah, you meant just saying {ku}? 13:54 < astroendjisai> Yeah, that's what I meant. 13:54 < astroendjisai> Only elide the kei, I suppose. That would have been more accurate. 13:54 < selpahi> Yes, you can do that. The {ku} has nothing else to close but the {lo} 13:54 < selpahi> And {kei} is inside it, so it's the same. 13:55 < astroendjisai> Okay, so the most 'trimmed down' version of that sentence would be: .i lo nu mi jundo ro do ku terzu'e lo nu mi zvati 13:55 < astroendjisai> *jundi 13:55 < astroendjisai> Bah I just copied and pasted and forgot about my typo 13:56 < selpahi> Yes 13:56 < astroendjisai> ckire 14:05 < Ilmen> ju'i plise si si coi cilre 14:21 * mumble_ jorne co'u fa la'o fa niftyg fa 14:24 * mumble_ jorne co'u fa la'o fa menli2 fa 14:24 < astroendjisai> .i lu si .i pluka mi 14:25 < Ilmen> zo si xu pluka do 14:25 < Ilmen> .i ma bangu do 14:25 < Ilmen> .a'u 14:26 < astroendjisai> .i la .gliban. bangu mi 14:27 < Ilmen> spagetufa: ro lo ma'a datka cu lalxu limna va 14:27 < spagetufa> ([FA {<ro BOI> <lo (¹NE ma'a¹) datka KU>}] [cu {<lalxu limna> <(¹va KU¹) VAU>}]) 14:27 < Ilmen> .i'u nai cnino zmiku 14:27 < Ilmen> je'e la .astros. 14:28 < selpahi> so'i bultufa 14:29 < Ilmen> bultufa ki'e si ki'a 14:29 < selpahi> bu'a zei gentufa 14:30 < selpahi> gentufa zantufa ilmentufa spagetufa 14:30 < Ilmen> .i ca lo cabdei no nu voksnu cu fasnu .i ku'i ka'e cabna lo bavlamdei 14:30 < Ilmen> je'e 14:30 < selpahi> je'e 14:30 < selpahi> .i mi pu kucli 14:30 < Ilmen> je'e 14:30 < selpahi> je'e 14:30 < Ilmen> .i la'a lo nu mi notci benji fliba cu rinpau 14:31 < selpahi> ki'u ma lo prenu za'o nitcu tau lo notci .i sa'u dikni 14:31 < selpahi> .i na nandu 14:31 < Ilmen> .u'i ĭe 14:33 < _mukti_> .i ta'o mi pu pacna lo nu jbota'a ca lo cerni 14:34 < _mukti_> .i cumki fa tu'a lo bavlamdei 14:34 < Ilmen> ŭa .i ku'i la'a cabna lo nu mi darno lo skami 14:34 < Ilmen> .i ke'u ba'a mi vokta'a zifre ca lo bavlamdei 14:35 < _mukti_> je'e .i a'o mi kansa do 14:36 < Ilmen> je'e .i ja'o .ai mi ranji jundi la zbaga pe la .mambl. 14:37 < selpahi> xu .itku'ile vokta'a 14:37 < Ilmen> nau zo'u tcika lo nu mi sipna zukte 14:37 < selpahi> (to ju'i nai toi) 14:40 < Ilmen> .i mi senpi lo nu .itku'ile vokta'a ca lo tai cabjbi kei .u'i .i ku'i lo nu tcidu fi lo .itku'ile papri cu stodi lo ka pagbu lo mi .aiste .i cumki fa lo nu ca da la terdi cu bevri lo nu mi'o .itku'ile voksnu .u'i ĭa cu'i 14:43 < Ilmen> .i .e'i mi snada lo ka jimpe lo na'e frili javni pe lo nu jungau lo du'u ma kau slaka gi'e se basna 14:43 < selpahi> .u'i 14:43 < selpahi> koi pacna fa mi'o 14:49 < selpahi> ie lo basna lerfu javni cu pluja 14:50 < selpahi> á - ko basna, à - ko na basna 14:51 < Ilmen> je'e .i ku'i lo nu se zunti lo mokca re mei lefpau cu rinka lo nu na'e frili .i .a'o ka'e ciska zo'oi ö́ mu'a zo'o ru'e 14:52 < selpahi> ie 14:54 < selpahi> zoi fa. attàwîl .fa ji'a xamgu mupli 14:55 < selpahi> î = clani .i'y-zei-sance .i ku'i ji'a ki'u tau zoi ta. à .ta lo ro moi slaka cu se basna 14:56 < Ilmen> je'e .i ku'i ma togmo'a sinxa 14:57 < selpahi> lo tonga na srana lo du'u basna ma kau 14:57 < Ilmen> tugni 14:57 < selpahi> .i lo tonga lerfu cu lidne lo valsi 14:57 < Ilmen> ŭa 14:57 < selpahi> no lerfu = farlu tonga 14:58 < Ilmen> je'e 14:58 < selpahi> li'o li'o .i ku'i 14:58 < Ilmen> cuxycau tonga 14:58 < selpahi> sei vajni la .itku'ile cu se stika ja'e lo nu lo tonga cu cenba lo ka se pilno fi ma kau 14:58 < Ilmen> ca lo cabjbi xu 14:59 < selpahi> ba'o je'u 14:59 < selpahi> .i jalge fa lo nu ji'i ro se xusra pe lo cukta cu ca jitfa 14:59 < selpahi> (to poi srana lo tonga toi) 14:59 < Ilmen> ku'i ba'a la .itku'ile julne cu sarxe lo vifne traji klesi 15:00 < selpahi> .i ku'i mi zanru (to ra'u toi) .i mi pu no'e nelci lo laldo tonga bo ciste 15:00 < Ilmen> je'e 15:00 < selpahi> na sarxe 15:00 < selpahi> .i la .kixadas. pu'o ningau 15:00 < selpahi> .i sa'e ... ba jmina lo PDF sfaile 15:01 < Ilmen> ki'e kajde 15:01 < selpahi> sa'u ko tolju'i lo srana be lo tonga 15:01 < selpahi> .i rau drata vajni cu zasti 15:02 < selpahi> .i pu ku lo tonga cu tutci lo no'e cinri be mi 15:02 < selpahi> .i lo cnino cu cinrymau 15:02 < selpahi> .i ku'i pe'i spofu 15:02 < selpahi> .u'i .oi 15:02 < Ilmen> .itlokuti ĭe nai pei ju'o nai 15:02 < selpahi> na srana 15:02 < Ilmen> je'e .i ja'o mi mo'ifli 15:03 < selpahi> wî = ca'e zei .itlokuti 15:03 < selpahi> we = ko'oi 15:03 < selpahi> etc 15:03 < selpahi> .i na dukse lo ka pluja 15:03 < Ilmen> ei = ei (ba'anai ru'e) 15:03 < selpahi> attàwîl 15:03 < selpahi> ie .u'i 15:04 < selpahi> -> ca'e rinsa 15:04 < Ilmen> .a'u je'e 15:04 < selpahi> lo .itlokuti cu jbini lo genrkasu joi lo .y gunma rafsi 15:05 < selpahi> igral = citka, igrawel = ko citka 15:05 < selpahi> .infikse 15:05 < Ilmen> je'e 15:05 < selpahi> fu'au dai zo'o 15:06 < selpahi> .y fu'i dai * 15:08 < selpahi> Ein-n igraswa kei. 15:08 < selpahi> (to ia drani toi9 15:08 < selpahi> toi) * 15:08 < selpahi> -> .ei do citka 15:09 < selpahi> .u'i lo lojbo cu cmalu 15:09 < Ilmen> la .kixadas. da'i jilra zo'o 15:10 < selpahi> Ei igral ku. -> do [zukte lo ka] de'a citka 15:10 < selpahi> ei = de'a 15:10 < selpahi> .i cfipu xu zo'o 15:11 < Ilmen> na'e cfipu .i ba'a lo si'o bilga kei rafsi cu bi'ai .infikse 15:11 < Ilmen> soi jalge fa lo nu na'e .ambigu ru'a 15:11 < selpahi> jetnu 15:12 < selpahi> ei igral == an-nei igral 15:13 < Ilmen> je'u pei la .itku'ile cu claxu tau zo co'u .i da'i cizra fa lo nu ku'i je'a co'e zo de'a 15:13 < selpahi> vs ein-n igral 15:13 < selpahi> an-n vs ei-nn 15:13 < Ilmen> za'a 15:13 < selpahi> ein-n * 15:13 < selpahi> .i ji'a, an-n igral == igral 15:14 < Ilmen> zo'oi - ma sinxa 15:14 < selpahi> .claxu zo co'u poi sumtcita .i na claxu lo .aspekte 15:14 < Ilmen> je'e 15:14 < selpahi> Ü igral -> co'u citka 15:15 < selpahi> n-n zo'u, sinxa lo du'u slaka karsna 15:15 < Ilmen> je'e 15:15 < selpahi> .i re mei fa lo ny zei sance 15:15 < selpahi> .i .ei bacru ro lo re mei 15:16 < Ilmen> xu zo'oi nn cu sinxa su'o drata 15:16 < selpahi> .i do'e lo .adjunte lo re moi be lo ny zei lerfu cu se basna ro roi 15:16 < selpahi> go'e 15:16 < selpahi> .i geminate 15:16 < selpahi> sei lazni 15:16 < Ilmen> je'e 15:17 < selpahi> an-n -> [an.'n] IPA 15:17 < selpahi> ua pei 15:18 < Ilmen> xu lo pa moi ny zei zunsna cu se zifygau pu tau lo re moi 15:18 < selpahi> pe'i na go'i 15:18 < selpahi> .i po'o pagu lo drata slaka 15:18 < selpahi> pagbu* 15:18 < Ilmen> je'e 15:23 < Ilmen> ki'e te gusni be lo .itku'ile tcila 15:23 < selpahi> je'e 15:23 < selpahi> .i .oi mi tatpi 15:23 < Ilmen> .i .ai mi klama lo senva munje .i .a'o lo nicte cu ricfu lo ka stidi lo cfika sidbo do 15:23 < selpahi> ki'e do 15:24 < Ilmen> .i ĭe ko litru surla 15:24 < selpahi> .i .a'o do te dunda lo simsa se vamji 15:25 < Ilmen> je'e .a'o 15:25 < Ilmen> co'o 15:25 < selpahi> co'o 17:52 < zipcpi> oiro'a toi'e ba'a la selpa'i mo'u jmina lo zi'ejvo la jbovlaste 18:06 < astroendjisai> coi .i lo nu jundi ro do ku terzu'e lo nu mi zvati .i mi cilre fu la'edi'u 18:29 < astroendjisai> .i ro do smaji 18:29 < astroendjisai> si tolylau 18:37 < bigcentaur> coi 18:41 < astroendjisai> .ui nai lo rirni mi na nelci barduku 18:41 < mensi> i ji'a mi na nelci mi'e la mensi 19:02 < senpi_gunka> coi 23:56 < gleki> right, NE should be autoinserted --- Day changed Sun Jul 12 2015 01:24 < gleki> spagetufa: ro lo ma'a datka cu lalxu limna va 01:24 < spagetufa> ([FA {<ro BOI> <lo (¹NE ma'a¹) datka KU>}] [cu {<lalxu limna> <(¹va KU¹) VAU>}]) 01:24 < gleki> spagetufa: ro lo ma'a je do datka cu lalxu limna va 01:24 < spagetufa> ([FA {<ro BOI> <lo (¹NE [ma'a je do]¹) datka KU>}] [cu {<lalxu limna> <(¹va KU¹) VAU>}]) 01:24 < gleki> spagetufa: ro lo ma'a je ne do datka cu lalxu limna va 01:24 < spagetufa> SyntaxError: Expected [,], [0123456789], [lL], [mM], [nN] or [rR] but "v" found. 01:24 < gleki> cliva: ^ 01:24 < gleki> alta: lo mi je ne mi mensi cu mo 01:24 < mensi> ([FA {lo <(¹[NE {FIhA mi} GEhU] [je {ne <FIhA mi> GEhU}]¹) (¹mensi NF¹)> KU}] [cu {mo VAU}]) 02:13 < gleki> Altatufa is now able to autoinsert I omitted before selbri 02:13 < gleki> alta: prami mi prami lo mi gerku 02:13 < mensi> ([{FA ZOhE} {CU <prami NF> <FE mi> VAU}] [I {FA ZOhE} {CU <prami NF> <FE (¹lo [{NE <FIhA mi> GEhU} {gerku NF}] KU¹)> VAU}]) 02:27 * ctefa`o cu cilce tolcanci 02:31 < gleki> cizra fa lonu zo tolcanci cu lujvo gi'e valsi lo vajni 02:37 < ctefa`o> gives 6 c, love me some c-s 02:39 < ctefa`o> One thing - what were the reasons for words like "sincrboa" being disallowed? Because the "boa" was hard to distinguish from "boua"? 02:41 < gleki> most likely 02:41 < ctefa`o> boua/boia* 02:46 < gleki> in la alta dialect they are synonyms to {-bo'a} 02:46 < ctefa`o> Ah yes bo'a too. boa vs bo'a boia boua 02:47 < gleki> altatufa is so long now due to all these autorestorations ... 03:55 < gleki> alta: lo cidja cu noi mi djica lo ka cikta ke'a 03:55 < mensi> ([FA {lo <cidja NF> KU}] [cu {DU ZOhE <noi (¹FA mi¹) (¹CU [djica NF] [FE {lo <ka (²[FA ZOhE] [CU {cikta NF} {FE ke'a} VAU]²) KEI> KU}] VAU¹)> KUhO}]) 03:56 < gleki> en: poi'i 03:56 < mensi> poi'i = [NU] x1 is such that poi'i abstraction is true; x1 binds ke'a within the abstraction. |>>> Originally proposed 03:56 < mensi> by And Rosta around 2001. |>>> selpahi 03:56 < gleki> crap 03:57 < gleki> NU can't bind ke'a 04:03 < ctefa`o> I think poi'i/voi shouldn't be considered a NU 04:03 < ctefa`o> More like noi and poi 04:04 < ctefa`o> Same for kai'u 04:04 < gleki> kai'u is definitely in NU since it takes ce'u 04:05 < gleki> as for poi'i it's some kind of shortcut, closer to compound cmavo 04:05 < ctefa`o> ti'e it is considered to expand to a ka 04:05 < gleki> as with xu'u = lo du'u 04:05 < gleki> kai'u isn't precisely expanded into ka due to variable typing. 04:05 < ctefa`o> Well that was what I ti'ed 04:06 < gleki> e.g. {mi} isn't an abstraction, thus {ka} can't be applied to it if you apply strong typing 04:06 < gleki> but you can se {kai'u} 04:06 < ctefa`o> poi'i is definitely not a NU though 04:06 < ctefa`o> Put it in POIhI or something 04:06 < gleki> *but you can use {kai'u} 04:06 < Ilmen> poi'i is a NU. NU is a *syntactic* class 04:06 < gleki> it's NU but some kind of NU10000 04:07 < Ilmen> for semantic distinctions, a number is usually appended to the selmaho name 04:07 < gleki> because no other NU can take ke'a 04:07 < Ilmen> UI1, UI2, UI3... 04:07 < selpahi> https://www.reddit.com/r/Ithkuil/comments/3d05jj/the_locative_case_is_ambiguous/ 04:07 < gleki> ke'a is a NOI thing 04:07 < ctefa`o> Well I don't think NU should be able to bind ki/ke'a 04:07 < gleki> dative is similarly ambiguous 04:07 < gleki> mentioning Russian there is just that JQ misread something somewhere 04:07 < Ilmen> je'e la selpa'i .i .ai ba tcidu .i ta'o .ai mi di'a citka co'o 04:07 < ctefa`o> but then again I have...different semantics than you 04:09 < gleki> it's a no-winner game. if you put it into NU or whether you make a new selmaho you gain nothing but a mild stisfaction that everything parses 04:09 < gleki> https://www.reddit.com/r/Ithkuil/comments/2ann3i/why_the_dative_case_for_to_have_something/ 04:09 < gleki> Secondly, like some Western languages (e.g., Russian), Ithkuil uses a dative construction in lieu of any verb ‘to have’ in reference to possession or attribution. It would therefore be used in translating sentences such as 04:09 < selpahi> mi pu za tcidu 04:10 < gleki> Gleki replied: Russian doesn't use dative for possession. "There is a book to me" is meaningless in Russian. 04:10 < gleki> We say "у меня есть книга". It doesn't have any dative cases at all. 04:10 < selpahi> .i la .kixadas. cu tugni fi lo du'u su'o lo mupli pe la .dativ. cu naldra 04:11 < gleki> i ku'i na mo'u stika le se cukta 04:12 < gleki> i krinu lo nu mi pu noroi jinvi lo du'u ibu me lo bangu poi prali lo prenu fa lo nu baupli 04:13 < selpahi> ga'a do ma genrkasu do'e lo do rusko jufra 04:13 < gleki> i sa'u lo ibu cukta cu simsa lo bauske cukta ma'i mi 04:14 < gleki> selpahi: lo genrkasu zo'u na'inmo i lo sumtcita ba'e e lo genrkasu cu cupra lo te sumti stura 04:14 < gleki> i va'i zoi ry.у.ry. mutce vajni 04:14 < selpahi> ua 04:15 < gleki> у меня есть книга => ~=At I-GENITIVE exists book-NOMINATIVE 04:15 < gleki> At+Genitive => to have 04:16 < gleki> ta'osairu'e 04:16 < gleki> с книгой = with BOOK-INSTRUMENTAL = with a book 04:16 < gleki> i ku'i 04:17 < gleki> с книги = with BOOK-GENITIVE = from a book 04:17 < selpahi> ie mi djuno ru'e 04:17 < gleki> i lo sumtcita cu mutce vajni i lo genrkasu na roroi cupra lo te bridi stura 04:17 < selpahi> https://www.reddit.com/r/Ithkuil/comments/3d06tl/the_way_negation_works_is_either_underspecified/ 04:18 < gleki> i ku'i na mtce srana la itkuiile cukta i sa'u Jy Ky bu pu tavla fi lo na drani pe lo rusko 04:19 < gleki> selpahi: uenai 04:20 < gleki> i o'i lo nu do pante ka'e rinka lo JQ cu catra le bangu 04:21 < selpahi> xu nabmi .i la .itku'ile va'o lo nu na renvi lo tai cmalu nu pante cu na .inda 04:22 < gleki> inda lo ka mi lanli ce'u gi'e pilno lo tcila be ce'u tu'a lo jbobau 04:22 < gleki> alta: lo noi mi citka ke'a cidja 04:22 < mensi> ([FA {lo <(¹noi [{FA mi} {CU <citka NF> <FE ke'a> VAU}] KUhO¹) (¹cidja NF¹)> KU}] [CU {COhE VAU}]) 04:23 < gleki> spagetufa: lo poi'i mi citka ke'a cidja 04:23 < spagetufa> ([FA {lo <(¹poi'i [{FA mi} {CU <citka (²[FE ke'a] VAU²)>}] KEI¹) cidja> KU}] VAU) 04:23 < gleki> spagetufa: lo noi mi citka ke'a cidja 04:23 < spagetufa> ([FA {lo <noi (¹[FA mi] [CU {citka <(²FE ke'a²) VAU>}]¹) KU'O> cidja KU}] VAU) 04:23 < gleki> spagetufa: lo poi'i mi citka ke'a vau cidja 04:23 < spagetufa> ([FA {lo <(¹poi'i [{FA mi} {CU <citka (²[FE ke'a] vau²)>}] KEI¹) cidja> KU}] VAU) 04:24 < gleki> spagetufa: lo poi'i mi citka ke'a vau vau cidja 04:24 < spagetufa> SyntaxError: Expected [,] or end of input but "c" found. 04:25 < gleki> ma xamgu mupli lo'u lo poi'i le'u 04:25 < gleki> i sa'e ma pluka mupli 04:26 < selpahi> "mi kecti lo poi'i lo nu lo remna ta'e ckasu ke'a cu cafne" 04:27 < selpahi> "mi manci lo poi'i no da djuno lo du'u ke'a mipri balcre" 04:27 < gleki> mi kecti zo'e noi lo nu lo remna ta'e ckasu ke'a cu cafne 04:28 < gleki> i pei 04:28 < gleki> di'u 04:28 < selpahi> mi na pilno zo zo'e 04:29 < selpahi> https://www.reddit.com/r/Ithkuil/comments/3d07k4/how_to_specify_destination_another_missing_case/ 04:29 < gleki> i xu no drata se pante 04:29 < selpahi> lo cutci 04:29 < selpahi> .i no drata mi se pante 04:29 < gleki> je'e je'e 04:30 < selpahi> .i ku'i ko'a ta'e poi'i ro da jinvi lo du'u ke'a sipna 04:30 < selpahi> sei mupli 04:30 < selpahi> .i krinu lo nu .ei zo poi'i cmima ma'oi NU 04:30 < gleki> mi pu kucli tu'a lo'u lo poi'i le'u 04:31 < gleki> i ie ie ma'oi nu i sa'u mi pensi tu'a y smuske nu gentufa 04:31 < gleki> *lo 04:31 < gleki> alta: mi kecti lo noi lo nu lo remna ta'e ckasu ke'a cu cafne 04:31 < mensi> ([FA mi] [CU {kecti NF} {FE <lo (¹[noi {<FA (²lo [nu {<FA (³lo [remna NF] KU³)> <CU (³ta'e [ckasu NF]³) (³FE ke'a³) VAU>} KEI] KU²)> <cu (²cafne NF²) VAU>} KUhO] COhE¹) KU>} VAU]) 04:32 < gleki> no'i sa'u mi jinvi lo du'u zo noi ce zo xoi ce zo poi'i cu tcesimsa zu'ai 04:33 < gleki> i ja'o mi na nelci lo nu le gunma cu so'ijecimei 04:33 < mensi> ji'a mi mutce nelci i ie 04:33 < selpahi> zo'o ru'e krinu lo nu rimni 04:33 < gleki> rimni i ku'i ci nai mei 04:33 < gleki> vau je'u 04:34 < gleki> noi:poi::noi'i?:poi'i::?:xoi 04:34 < gleki> i xa ei mei 04:36 < selpahi> go'i .i ku'i zo noi'i mo 04:36 < selpahi> .i na mapti lo drata 04:36 < selpahi> .i zo poi'i no'e srana zo poi 04:37 < gleki> srana zo noi xu 04:38 * gleki ba'a la xorlo bazi co'a se casnu 04:40 < selpahi> na srana 04:40 < selpahi> .i srana zo me'au je zo ckaji .i na srana zo poi ja zo noi 04:41 < gleki> xu zo lo ca'o srana lu zo'e noi li'u 04:41 < selpahi> go'i 04:41 < gleki> zo ckaji je'e 04:47 < gleki> i ba'a milxe nandu fa lo nu gentufa galfi zo poi'i zo ckaji 04:47 < Ilmen> lu ckaji lo ka ce'u goi ke'a zo'u ... li'u 04:48 < Ilmen> %) 04:48 < selpahi> satci 04:48 < selpahi> la .redit. na curmi lo nu mi benji 04:49 < gleki> benji ma 04:49 < selpahi> lo cnino se casnu 04:49 < Ilmen> .ai mi tcidu tau la bredi ditcu 05:08 < ctefa`o> Now that more people are around -- what exactly was the ban on "-boa" "-bae" "-bea" based on? The problem of distinguishing boa from boua/boia/bo'a? 05:08 < ctefa`o> Or something else as well? 05:09 < gleki> it was initially banned in Lojban. from it's beginning. 05:09 < selpahi> When Lojban was made, they decided against keeping the VV from Loglan, which is why cmavo and rafsi have V'V instead. It makes no sense to then allow them in brivla. 05:14 < ctefa`o> So they were initially only allowed in brivla and then not at all? 05:21 < selpahi> "Vowel hiatus is often not stable: it tends to turn into diphthongs. 05:21 < selpahi> Loglan, in fact, uses "ao" for the diphthong written "au" in Lojban; 05:21 < selpahi> Loglan "aa", "ae", "au", "ea", "ee", "eo", "eu", "oa", "oe", "oo", "ou" 05:21 < selpahi> are all instances of hiatus. Adding "'" not only eliminated hiatus, but 05:21 < selpahi> added lots more vowel pairs (meaning more cmavo and rafsi) to play with. 05:21 < selpahi> In addition, eliminating things like "tiu" in favor of "ti'u" reduced 05:21 < selpahi> the chance that the "t" would be palatalized into "tcu". " 05:22 < selpahi> -John Cowan 05:23 < ctefa`o> je'e .a'u ki'e 05:35 < ctefa`o> And for die/[dje] not being ok?;) 05:37 < ctefa`o> Similar reasons I suppose 05:41 < ctefa`o> Guess that easily becomes [dije] 05:45 * mumble jorne co'u fa la'o fa Ilmen fa 05:45 < selpahi> It becomes {dje} 05:46 * mumble jorne co'a fa la'o fa Ilmen fa 05:53 < ctefa`o> jehe 05:55 < ctefa`o> iirc some were discussing that putting a denpa-bu would improve that ".die" but I don't see how 05:58 < selpahi> I don't recall any such thing. 05:59 < selpahi> Maybe you are confusing it with {au ua} needing a glottal stop in the middle. 06:00 < ctefa`o> no it was definitely about putting denpa-bu before CiV/CuV. Maybe I dreamt it 06:01 < ctefa`o> yeah au.ua or we get "[awwa]" 06:02 < ctefa`o> ([awa] li'a) 07:07 < gleki> i can't ever imagine pronouncing {die} as {dje}. probably because i utter it dentally 07:21 * nuzba @oka_iu_tcan: mi finti lo lojbo selsanga pe me'e lu claxu be do crisa li'u #lojban https://soundcloud.com/cogas-iuk/sunday-10-38-pm?utm_source=soundcloud&utm_campaign=share&utm_medium=twitter [http://bit.ly/1M4doMZ] 07:31 < rutytar> coi 07:33 < ctefa`o> gleki: yeah that die is kinda distinct from dje 07:47 < Ilmen> What's wrong with [ʔaʊwa] for {.auŭa}? 07:49 * nuzba @s_i_d_a: はじロジ http://cogas.github.io/hajiloji/ 一通り読んで適当にロジバン作文とかしてみて、足りないところはwavelessonsの翻訳 http://misonikomilojban.blogspot.jp/search/label/lojban%20wavelessons で学習すればとりあへず僕を追ひ越す程度のレベルにはなれるかと。 [http://bit.ly/1M4f1KS] 07:58 < Ilmen> za'a la .itku'ile finti cu spuda tau la selpa'i .i za'a sutra lo ka spuda 08:07 < gleki> my leg got hit... 08:10 < gleki> mi ka lo jamfu be ce'u pu co'a se darxi 08:20 < gleki> wat's the difference between klani and gradu then 08:22 < Ilmen> .i'unai 08:32 < gleki> en: incita 08:32 < mensi> incita = x1 is an adposition linking its complement (argument) x2 with x3 |>>> x2 can sometimes be called 'noun phrase 08:32 < mensi> of adposition'. x3 can e.g. be the current clause (lo bridi), x2 of it. |>>> gleki 08:33 < gleki> what's wrong with its place structure 08:33 < gleki> if it's wrong it should be aligned to the place structure of similar words 08:33 < gleki> en: bridi 08:33 < mensi> bridi = x1 (du'u) is a predicate relationship with relation x2 among arguments (sequence/set) x3. |>>> Also: x3 are 08:33 < mensi> related by relation x2 (= terbri for reordered places). (x3 is a set completely specified); See also sumti, fancu. |>>> 08:33 < mensi> officialdata 08:35 < gleki> incita looks like {se te bridi} 08:37 < selpahi> xu na simsa zo tcita 08:38 < gleki> en: tcita 08:38 < mensi> tcita = x1 is a label/tag of x2 showing information x3. |>>> See also sinxa, cmene, judri. |>>> 08:38 < mensi> officialdata 08:38 < gleki> xm 08:38 < gleki> na simsa 08:39 < gleki> i lo te incita so'eroi bridi 08:39 < gleki> i lo incita cu te jorne lo sumti so'eroiku lo bridi 08:40 < gleki> i ie simsa lu te jorne li'u 08:40 < gleki> i do stidi ma 08:41 < gleki> i vajni ke bauske tcila 08:41 < gleki> fa lo incita 08:46 < ctefa`o> Ilmen: I like [ʔaʊwa] too 08:51 < ctefa`o> [ɂaɪja] as well 08:54 < astroendjisai> Hi. I have a question. How would you ask a question requesting permission to do something? 08:54 < astroendjisai> Like "may I sit here?" 08:54 < gleki> e'a pei 08:54 < gleki> e'a pei mi zutse vi 08:55 < gleki> gloss: e'a pei mi zutse vi 08:55 < mensi> Permission emotion ? me sit here 08:55 < astroendjisai> Cool, thanks. 08:56 < _mukti_> I prefer to place the {vi} in front of the {zutse}, in part so I remember that it's a tense rather than an object 08:56 < _mukti_> Nice glosser, by the way :) 08:57 < gleki> no 08:57 < astroendjisai> That's nifty. Hmm. 08:57 < astroendjisai> gloss: do viska mi 08:57 < clakre> it's offline 08:57 < astroendjisai> Oh 08:57 < gleki> wait a bit 08:57 < clakre> it quit the channel less than a minute ago, it'll be back 08:57 < astroendjisai> Drat it all 08:58 < astroendjisai> Ah 08:58 < ctefa`o> coi clakre 08:58 < gleki> i have no clue what's going with the server right now 08:58 < gleki> but mensi bot doesnt want to join 08:58 < astroendjisai> Yeah. 08:58 < astroendjisai> Weird 08:58 < astroendjisai> gloss: do viska mi 08:58 < mensi> You see me 08:59 < astroendjisai> Nice 09:00 < gleki> gloss: e'a pei mi zutse vi 09:00 < mensi> [permission granted] ? me sit here 09:00 < gleki> maybe this is a bit better 09:01 < gleki> for those who haven't heard please rregister here and upvote 5 questions with rating below 10 area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 09:06 < _mukti_> exp: to'e ki'u nai bo 09:06 < mensi> SyntaxError: Expected [yY] but "o" found. 09:06 < _mukti_> ua nai 09:06 < _mukti_> .i to'e ki'u nai bo ty cu tròci lo nu clìva lo bìrka be la mùdypre kei 09:12 < _mukti_> Curious. It needs some kind of structuring cmavo to parse 09:12 < _mukti_> camxes: .i to'e ki'u nai bo 09:12 < camxes> (i [to'e ki'u nai] bo) 09:12 < _mukti_> camxes: to'e ki'u nai bo 09:12 < camxes> SyntaxError: Expected [,], [gG], [lL], [mM], [nN], [rR] or [vV] but "b" found. 09:12 < _mukti_> camxes: ni'o to'e ki'u nai bo 09:12 < camxes> SyntaxError: Expected [,], [gG], [lL], [mM], [nN], [rR] or [vV] but "b" found. 09:13 < _mukti_> camxes: lu to'e ki'u nai bo li'u 09:13 < camxes> SyntaxError: Expected [uU] or [yY] but "i" found. 09:13 < _mukti_> Or maybe a very specific cmavo. 09:13 < _mukti_> I think this may be a camxes bug 09:14 < _mukti_> Unless there's some kind of rationale for why it works with {.i} and not {ni'o} 09:14 < _mukti_> (or for that matter, why it doesn't work without either) 09:15 < Ilmen> coi 09:16 < _mukti_> coi la .ilmen. 09:16 < _mukti_> I've been working on errata for "Oz". 09:17 < _mukti_> There's a line that seems to have been mangled by search-and-replace that I'm trying to reconstruct. 09:17 < _mukti_> .i je ty. to'e ki'u nai gi lo blànu cu ba'o milxe lo ka kadbì'o gi'a'a mèlbi tàxfu 09:17 < _mukti_> Original: "and although the blue was somewhat faded with many washings, it was still a pretty frock." 09:17 < Ilmen> za'a 09:18 < _mukti_> The part about the washings seems to have been deliberately omitted. 09:18 < _mukti_> (Unless the search and replace swallowed even more text) 09:19 < Ilmen> .a'a pei doi la selpa'i 09:19 < _mukti_> I'm tempted to reconstruct it without the gihek 09:20 < _mukti_> .i je ty to'e ki'u nai lo nu lo blanu cu ba'o milxe lo ka kadbi'o cu melbi taxfu 09:20 < _mukti_> (or {cu ja'a melbi taxfu}) 09:21 < Ilmen> mapti simlu 09:22 < _mukti_> exp: ge nai lo ka stedu noi dy pu viska tu'a ke'a gi lo ka kukte ninmu 09:22 < mensi> ([{ge nai} {lo <(¹ka [CU {stedu VAU}] KEI¹) (¹noi [{dy BOI} {CU <pu viska> <tu'a ke'a LUhU> VAU}] KUhO¹)> KU} gi {lo <ka (¹CU [kukte ninmu] VAU¹) KEI> KU}] VAU) 09:22 < _mukti_> exp: ge nai lo ka stedu noi dy pu viska tu'a ke'a gi ja'ai lo ka kukte ninmu 09:22 < mensi> ([{ge nai} {lo <(¹ka [CU {stedu VAU}] KEI¹) (¹noi [{dy BOI} {CU <pu viska> <tu'a ke'a LUhU> VAU}] KUhO¹)> KU} {gi ja'ai} {lo <ka (¹CU [kukte ninmu] VAU¹) KEI> KU}] VAU) 09:23 < Ilmen> za'a sumti gi'enai mulno jufra 09:25 < _mukti_> ie na mulno jufra 09:25 < _mukti_> la .dòrotis. cu se spàji lo nu la bànli te màkfa cu ckàji ... 09:39 < gleki> since when {to'eki'unaibo} should parse? 09:39 < gleki> fragments are even dying 09:39 < gleki> alta: i to'e ki'u nai bo 09:39 < mensi> (i [{to'e <ki'u nai>} bo] [FA ZOhE] [CU {COhE VAU}]) 09:40 < gleki> should i allow autoinserting I? i think not 09:40 < gleki> even {prami mi prami lo mi gerku} is a super rare example 09:44 < _mukti_> It seems to me that just about anything that follows {.i} should also parse without the {.i} 09:44 < _mukti_> And that just about anything that parses after {.i} should also work after {ni'o} 09:45 < _mukti_> But maybe there are reasons that I'm not aware of that would preclude this 09:45 < gleki> if you mean making i elidable then it makes sense only at the very beginning of the text 09:45 < _mukti_> Yes, that's what I'm talking about 09:45 < gleki> because making {i} elidable in the middle makes sense in a very minor nummber of examples 09:45 < _mukti_> camxes: .i to'e ki'u nai bo mi broda 09:45 < camxes> (i [{to'e ki'u nai} bo] [mi CU] [broda VAU]) 09:45 < _mukti_> camxes: to'e ki'u nai bo mi broda 09:45 < camxes> SyntaxError: Expected [,] but "b" found. 09:46 < _mukti_> camxes: ni'o to'e ki'u nai bo mi broda 09:46 < camxes> SyntaxError: Expected [,] but "b" found. 09:46 < Ilmen> camxes: ni'o je broda 09:46 < camxes> SyntaxError: Expected [,] but "b" found. 09:46 < _mukti_> So, the issue of fragments is a separate one. 09:46 < Ilmen> Yet {ni'o je} appears a fair amount of times in la Oz 09:46 < Ilmen> ba'anai 09:47 < _mukti_> Hmm, I'm not finding that 09:47 < _mukti_> {ni'o je'u} and {ni'o jetnu}, however 09:47 < Ilmen> Maybe I'm misremembering 09:47 < gleki> alta: mi prami 09:47 < mensi> (I [FA mi] [CU {prami NF} VAU]) 09:47 < gleki> as you wish 09:48 < gleki> alta: ki'unai bo mi mo 09:48 < _mukti_> {ni'o to'e ki'u nai}, however does appear 09:48 < mensi> SyntaxError: Expected [yY] but "o" found. 09:48 < gleki> ^ wont parse since {bo} is about the previous clause which can't appear in the beginning of the text 09:53 < _mukti_> I see a discussion in 2006 09:53 < gleki> alta: ni'o je bo mi mo 09:54 < mensi> (ni'o je bo [FA mi] [CU {mo VAU}]) 09:54 < gleki> fixed 09:54 < gleki> text_1 = expr:(I_clause (jek / joik)? (stag? BO_clause)? free* paragraphs / NIhO_clause+ (jek / joik)? (stag? BO_clause)? free* su_clause* paragraphs? / I_clause_elidible paragraphs) {return _node("text_1", expr);} 09:54 < selpahi> I already told you gi'a'a is {gi ja'a}. But the washing part seems to be missing. 09:55 < _mukti_> selpahi: Yes, I included {ja'a} but the sentence still needed more help 09:55 < selpahi> Why? 09:55 < _mukti_> to'e ki'u nai gi lo blànu cu ba'o milxe lo ka kadbì'o gi'a'a mèlbi tàxfu 09:56 < gleki> spagetufa: to'e ki'u nai gi lo blànu cu ba'o milxe lo ka kadbì'o gi ja'a melbi taxfu 09:56 < spagetufa> (CU [{<to'e (¹ki'u nai¹)> gi} {<FA (¹lo blanu KU¹)> <cu (¹[ba'o milxe] [{FE <lo (²ka [CU {kad/bi'o VAU}] KEI²) KU>} VAU]¹)>} gi {CU <(¹ja'a [melbi taxfu]¹) VAU>} VAU]) 09:56 < _mukti_> What to do with the first {gi} ? 09:56 < gleki> ma nabmi vau uanai 09:56 < selpahi> BAI gi ... gi 09:56 < selpahi> {to'e ki'u nai} is BAI 09:57 < _mukti_> I'm not familar with BAI gi ... looking now 09:57 < _mukti_> CLL 23 ua 09:58 < _mukti_> Cool 09:58 < selpahi> Man I'm not sure what to make of JQ's responses. 09:58 < _mukti_> Where? 09:59 < selpahi> On reddit. I've been posting a few problems I've found with Ithkuil 09:59 < selpahi> (ithkuil reddit) 09:59 < gleki> xe'e 09:59 < gleki> i ibu larcu bangu 10:00 < gleki> i xu do djica lo nu co'a prane ja logji bangu 10:00 < selpahi> djica lo nu ka'e cusku lo se skudji 10:00 < gleki> i ma'a pinfu lo rarna bangu i ibu ji'a jai bapli lo nu ma'a na kakne tu'a da 10:02 < gleki> im slowly getting rid of fragments in alta 10:03 < gleki> fragment = expr:(!term !tag quantifier / links / linkargs) {return _node("fragment", expr);} 10:03 < gleki> only 3 cases need to be reimplemented za'adai 10:04 < _mukti_> I think fragments are useful. 10:04 < cliva> exp: mo'e ti mo'e ta roi 10:04 < mensi> ([{<(¹mo'e ti TEhU¹) (¹mo'e ta TEhU¹)> roi} KU] VAU) 10:04 < cliva> ^ ma jai va'o pilno sei malgli 10:05 < _mukti_> Which is to say, I don't see the value in making fragmentary utterances unparseable. 10:05 < gleki> cliva: this happened this-some times 10:05 < gleki> _mukti_: they are parseable 10:05 < gleki> alta: je 10:05 < mensi> (I [FA ZOhE] [CU {COhE VAU}] [je {CU <COhE VAU>} VAU]) 10:06 < gleki> the question is not in this syntactic parsing 10:06 < _mukti_> Oh. Curious! 10:06 < gleki> exp: je 10:06 < mensi> je 10:06 < gleki> which is better? na'i 10:07 < _mukti_> I see. You're accepting them, but interpreting them in an implicit context. Neat. 10:07 < gleki> the question is what is COhE 10:07 < cliva> gleki: ro lo re zo mo'e zei co'e cu pagbu lo mekso poi pagbu lo roi zei klauza 10:07 < gleki> xu naku lo klauza cu du lo bridi 10:07 < cliva> me PEG jargone 10:08 < gleki> i ua 10:08 < _mukti_> Does the implicit "FA ZOhE" mean that co'e is 1-ary? 10:08 < cliva> sa'e nai 10:08 < gleki> _mukti_: lojban grammar says that selbri is >=1-ary 10:09 < selpahi> (Toaq Dzu has nullary predicates) 10:09 < _mukti_> Where's that? I'm not surprised, but curious. 10:12 < gleki> for a moment i becae angry at such a weird application like mahantufa only a second later realizing that it was written by me 10:12 < gleki> *became 10:12 < ctefa`o> "nullary predicate"? 10:12 < ctefa`o> MO 10:13 < _mukti_> I see there's a jbovlaste feature request to support nullary predicates. 10:14 < _mukti_> So I'm wondering if there is a rationale for why there aren't or can't be. Or if it's just an arbitary limitation in mu'a jbovlaste 10:14 < gleki> if there are then what is {lo broda} 10:15 < gleki> and what is bridi head then 10:16 < gleki> but for guaspi it's obvious that 0-ary predicates can exist 10:17 < _mukti_> Supposing that {broda} were nullary, I don't think {lo broda} would have any defined sense. It would be like {lo se blanu}, no ? 10:17 < _mukti_> Or {lo zi'o crino} to use an example from the wiki 10:17 < gleki> {zo'e noi fa ke'a broda} 10:18 < gleki> that's not the proof of course. the proof is that bridi head must exist. 10:19 < gleki> if it exists then ke'a goes to x1 10:20 < _mukti_> Right, so I follow you in saying that {lo broda} wouldn't be valid for nullary {broda}. But I could imagine it working for non-sumti. 10:21 < gleki> if you change PEG and CLL then sure :) 10:21 < gleki> currently it's against them 10:21 < _mukti_> Hmm. I'm not sure this question touches the formal grammar. And as for CLL, I believe you if you are saying that the CLL says that. I just want to see the details. 10:22 < _mukti_> Maybe there's a good reason to disallow nullaries. I'd just like to know what it is. Or alternately, if nullaries are prohibited without a rationale, I'd like to know about that, too. 10:23 < gleki> someone create a repo on github called "Questions to jbocre" and open this issue there 10:23 < gleki> :) 10:26 < ctefa`o> camxes: su zei si 10:26 < camxes> ([{su <zei si>}] VAU) 10:28 < gleki> The general rule, then, is that the selbri may occur anywhere in the bridi as long as the sumti maintain their order. The only exception (and it is an important one) is that if the selbri appears first, the x1 sumti is taken to have been omitted: 10:28 < gleki> [ http://lojban.github.io/cll/9/2/index.html ] 10:28 < gleki> _mukti_: ^ 10:29 < gleki> this "exception" works the same for {LE selbri KU} 10:29 < gleki> where x1 is "omitted" 10:29 < _mukti_> That seems to me more like a description of how word order works than a constraint requiring a sumti. 10:30 < gleki> so you can define nullary predicates, it's just that you can never use them in Lojban 10:30 < _mukti_> I see how it could be read that way, but that's not the way I perceive the intent. 10:31 < gleki> well, then instead of one rule you begi treating LE+selbri and bare selbri differently 10:31 < _mukti_> How so? 10:32 < gleki> 10:20 < _mukti_> Right, so I follow you in saying that {lo broda} wouldn't be valid for nullary {broda}. But I could imagine it working for non-sumti. 10:32 < selpahi> The reason why it works in Toaq Dzu is that there is no {zo'e}, and unfilled places are "not there" (i.e., they are in a sense {zi'o}'d out), so a nullary predicate can happily stand on its own. 10:32 < _mukti_> Say we have a 2-ary predicate, {broda}. In the statement {.i broda}, we assume that there are two magic {zo'e} that are implicitly available. 10:33 < gleki> _mukti_: the grammar knows nothing about x2 here. 10:33 < _mukti_> If {brode} is 1-ary, then {.i brode} supplies only one 10:33 < gleki> only semantics knows 10:33 < _mukti_> And in the case of {.i nilbrode} ? 10:33 < _mukti_> s/nilbrode/zilbrode/ 10:33 < fenki> _mukti_ meant to say: And in the case of {.i zilbrode} ? 10:33 < gleki> in {i zilbrode} at least x1 exists 10:34 < gleki> e.g. {i zilblanu} which means someone is gonna change the dictionary :) 10:34 < _mukti_> selpahi: Yeah, I get that {zo'e} is a mess, and I've followed, with some befuddlement, the attempts to correct it with {zi'o} 10:35 < selpahi> Toaq Dzu predicates are more akin to prolog's, where a different arity means a different predicate, although the differences tend to be smaller than in prolog. 10:35 < _mukti_> .a'u 10:36 < _mukti_> That's a familiar concept to anyone who uses a programming language with method "overloading". 10:38 < gleki> curtis suggested something similar recently. depending on how you fill a sumti you change arity of "this predicate" 10:38 < _mukti_> I guess what I'm saying is I don't see the implicit x1 that is supplied in the passage that gleki quoted as having anything to do with a lower bounds on the number of sumti that a selbri may take. It assumes selbri take at least 1 sumti because they are all defined that way. But we're also given {zi'o} which let's us get to a place where that assumption is not valid. 10:39 < gleki> well, {i zi'o broda} is {i fa zi'o broda}, yes. semantically you get 0-ary predicate 10:39 < _mukti_> The way I read it is "if the selbri appears first, the x1 sumti is taken to have been omitted" ... if the selbri takes such a sumti. 10:40 < _mukti_> Already we have the case where a magic number of {zo'e}s is provided based on the arity of the predicate. If the arity is zero, that's not a special case. No {zo'e} is provided. 10:41 < gleki> "... if the selbri takes such a sumti" <-- where is that said? 10:41 < _mukti_> That's my text, placed outside the quote to mark it as such. 10:42 < gleki> i can only repeat that i see the same rule for sumti 10:44 < _mukti_> I see a necessary restriction for descriptive sumti, in the sense that the way they are defined, they require a place for {ke'a} 10:45 < _mukti_> I don't see that restriction having anything to do with the predicates themselves 10:45 < gleki> i only see one rule, that's it. one rule is easier to remember for newcomers 10:46 < gleki> https://groups.google.com/forum/#!topic/bpfk-list/EV1avCV8_PY 10:50 < _mukti_> There's only one rule that I'm describing: descriptive sumti require a predicate with at least one argument 10:51 < _mukti_> As opposed to what you are describing: All sebri require at least one argument 10:51 < gleki> no, you have two rules 10:51 < gleki> the second rule is about bare bridi 10:51 < _mukti_> (Which incidentally requires a corollary to explain what happens when {zi'o} erases a predicate's last argument) 10:52 < gleki> which would mean that bare bridi and bridi converted to sumti behave differently 10:52 < _mukti_> Explain to me how they behave differently? 10:53 < gleki> you say that in sumti x1 must exist, in bare selbridi they may not 10:53 < gleki> two rules 10:54 < _mukti_> I'm saying that the definition of sumti requires that they exist, and that's not a separate rule. That's part of the definition of descriptive sumti. 10:54 < _mukti_> No separate rule is then required for selbri. 10:55 < gleki> compare {broda ko'e ko'i} and {lo broda be ko'e bei ko'i} 10:55 < _mukti_> Ok... go on ... 10:55 < gleki> the same rule is used for both of them if x1 is assumed to always exist. 10:56 < _mukti_> How is it different without that assumption. 10:56 < _mukti_> ? 10:56 < gleki> the rule in short is "bridi head can't be empty, it's prefilled with at least {fa zo'e}." 10:56 < _mukti_> I think you're innovating. 10:56 < gleki> rephrasing since CLL explanations arent convincing you 10:57 < _mukti_> I don't see a reason why the bridi head can't be empty. Either logically, or in tradition. 10:57 < gleki> i have only two points mentioned above 10:57 < _mukti_> I think you are reading a restriction into the CLL which is not necessary. 10:57 < selpahi> tersmus: broda fa mi gi'e brode fa do 10:57 < tersmus> (broda(mi) /\ brode(do)) 10:58 < gleki> im just perceving two parts of CLL as one rule. this allows LE-sumti and selbri be perceived equally. 10:58 < selpahi> No shared head^ 10:58 < gleki> selpahi: srana ma 10:58 < selpahi> lo bridi stedu 10:58 < gleki> pe'i na srana lo pu zi nu casnu 10:58 < selpahi> je'e 10:59 < gleki> i ni'i ma pa da poi stedu cu se ponse re de 10:59 < gleki> i re stedu vau i'a 10:59 < gleki> en: uilkinse 10:59 < mensi> uilkinse = x1 is semantically regularized in x2 (property of x1) |>>> Can be applied to languages, linguistic constructs. 10:59 < mensi> See also kalsa, gansu, smuni, prane |>>> gleki 11:00 < gleki> di'u ta'osai 11:00 < gleki> i oise'i 11:00 < gleki> i zo ganzu li'a 11:00 < _mukti_> That's interesting. So how would {broda fa mi} be conciled with the doctrine of "no empty head" ? 11:01 < gleki> 10:46 < gleki> https://groups.google.com/forum/#!topic/bpfk-list/EV1avCV8_PY 11:01 < _mukti_> Ah. So you want to hoist FA-sumti into the head. 11:01 < gleki> which is again confirmed by {be fa} usage in another part of CLL. 11:02 < gleki> ofc. you may rewrite CLL applying these new rules, just dont forget of domino effect. otherwise you'll increase the number of onsconsistencies in CLL 11:02 < _mukti_> (or maybe just fa-sumti) 11:03 < _mukti_> Definitely not interested in unilaterally rewriting anything. But I do challenge whether the interpretation you suppose in this case is necessary and useful. 11:06 < _mukti_> And now I think I understand the point with the tersmus parse. Martin apparently does not subscribe to the interpretation that a fa-term is hoisted into the head. Because if he did, then the {fa mi} in the first bridi tail would be forced into the x1 position for the second bridi tail. 11:06 < gleki> i think you want to inject semantics into the grammar. if we were in ##English channel i wouldn't say anything but i thought the main advantage of Lojban was separating syntax and grammar in a speakable language 11:07 < gleki> _mukti_: no, i agree with Martin's parse 11:07 < gleki> because FAMs and ZAMs are different things 11:07 < _mukti_> Well, that is clear. But you see how it's relevant now? 11:08 < gleki> no, i dont see this example relevant 11:08 < _mukti_> I don't think this is about semantics. If there is a prohibition on 0-ary predicates, that's a syntactic concern. Or am I high? 11:08 < gleki> i meant that one needs to look into the dictionary to see if x1 can be filled with {zo'e} 11:09 < _mukti_> How is that different from needing the dictionary to see if x2 can be filled with it? 11:09 < gleki> grammar doesnt know anything about x2, it's not its concern 11:10 < gleki> it only says that infinite number of xN is possible 11:10 < _mukti_> In {lo broda be koa} both x1 and x2 are invoked 11:10 < _mukti_> s/koa/ko'a/ 11:10 < fenki> _mukti_ meant to say: In {lo broda be ko'a} both x1 and x2 are invoked 11:11 < gleki> no'i if you want to know how i parse selpa'is example. {fa zo'e cu broda fa mi gi'e brode fa do} = {fa zo'e broda gi'e brode i mi broda i do brode} 11:11 < gleki> _mukti_: in {lo broda be ko'a} x2 is explicitely filled according to sequential filling of places. 11:11 < _mukti_> Ok, but I think you're introducing arbitrary restrictions in the service of building a parser. And I think parser building is a worthwhile pursuit, as you know. But the tail should not wag the dog. 11:12 < _mukti_> Right, and the sequential first argument is used in the definition of the description. 11:12 < gleki> it's the nature of lojban. you use {lo broda}, you prefill broda1 11:13 < _mukti_> {zo'e noi ke'a broda} ... no controversy there 11:15 < rutytar> sorry to interrupt, but does anyone feel like proofreading a translation i did? 11:15 < gleki> yes, i think you want 0-ary predicates in the dictionary. but im not against that. then x1 will be always =zi'o. 11:15 < gleki> rutytar: go on 11:16 < gleki> _mukti_: i think x1 will have a special variable type: zi'o. that's how it can be reconciled with the rest of the language 11:17 < _mukti_> .u'i and here I was thinking it was you who submitted the feature request to add nullary predicates to jbovlaste 11:17 < _mukti_> https://github.com/lojban/jbovlaste/issues/126 11:17 < _mukti_> rutytar: I'd like to see your translation 11:18 < rutytar> http://pastebin.com/raw.php?i=R1h9U17e 11:18 < Ilmen> I guess that zo'e are injected only when the bridi is complete and that some sumti slot haven't been filled 11:19 < Ilmen> So that in the case of {citka gi'e klama} the zo'e isn't necessarily shared 11:20 < Ilmen> actually I wonder how ambiguous is this. In the case of gihek, are ellipsis slot always not-shared by default? 11:21 < Ilmen> If it's left to context whether they're shared or not, it would be a logical ambiguity 11:21 < _mukti_> rutytar: I think I understand. You may want {ja'e lo nu} each time {ja'e} appears 11:22 < _mukti_> or maybe just the first time 11:22 < _mukti_> No, the second, too. 11:22 < Ilmen> tersmus: citka gi'e klama 11:22 < tersmus> (citka( ) /\ klama( )) 11:23 < gleki> i think in {mo gi'e mo} the head is shared syntactically. but i dont care whether they are shared semantically 11:23 < rutytar> oh yeah 11:23 < _mukti_> rutytar: I always do that! 11:23 < rutytar> is there any other good way to say "because"? 11:24 < _mukti_> {ki'u lo nu} is my go to phrase ... but there are many shades, many ways 11:24 < _mukti_> {ki'u lo nu} ... and the reason is this (event) 11:24 < gleki> ja'e seems to be the most generic (except ni'i which is a separate case) 11:24 < gleki> {i ki'u bo} is another option. it has the same number of syllables if that matters 11:25 < gleki> btw, was it meant to be {ja'e} or {se ja'e}? 11:25 < rutytar> does ja'e have place structure? 11:26 < selpahi> This is the version I had: http://pastebin.com/NkkEBx40 11:26 < Ilmen> Having the gihek-head empty by default (as long as there is no explicit sumti there) enables {sipna gi'e jmive} to be {sipna fa zo'e gi'e jmive fa zo'e}, where context may tell that both zo'e refer to the same thing, or the converse way. 11:26 < rutytar> interesting 11:26 < gleki> what if zo'e=anything 11:26 < gleki> rutytar: if you needed "because" it's {se ja'e} 11:27 < rutytar> doesn't xa'o imply that something started too early? 11:27 < selpahi> Not in the sense that "too early" is a bad thing 11:27 < ctefa`o> doi gleki 11:27 < selpahi> Just before some expected or otherwise relevant point 11:28 < gleki> ctefa`o: mi xu 11:28 < ctefa`o> gleki: "lajiar", is that troublesome for you to say? 11:28 < gleki> ctefa`o: depending on how to utter that 11:28 < rutytar> but i guess it's "already" in the sense of "anyways" 11:28 < ctefa`o> or "lavuar" 11:28 < ctefa`o> Well like jbo 11:29 < ctefa`o> {lajiar} with i=[j] 11:29 < gleki> as "lah-zhyah-r"? no, that's easy. {lah-v-wah-r} is more difficult 11:29 < selpahi> You could replace xa'o with {va'o ro da} "in any case" 11:29 < selpahi> "no matter what" 11:30 < gleki> mi de'a jundi 11:30 < rutytar> ah, "ri" is what i was trying to use "jy" for 11:30 < ctefa`o> i am on phone and it takes eternity to put ipa signs together 11:31 < selpahi> And to imagine that Amnadya had to type Ithkuil on a phone. 11:31 < Ilmen> I find none of [laʒjar] and [lavwar] difficult to utter 11:32 < ctefa`o> Interesting I have a bit of trouble with it 11:32 < rutytar> selpahi: where is the kei on line 2? is it just nested? 11:32 < Ilmen> In French, the word for "to see" is pronounced /vwar/ 11:33 < ctefa`o> How hard is it to distinguish from /var/ and /war/ for you? 11:34 < ctefa`o> to distinguish /vwar/* 11:34 < Ilmen> [var] from [vwar]? 11:35 < _mukti_> exp: ba za lo cacra be li cu'o mi broda 11:35 < mensi> SyntaxError: Expected [,] but "b" found. 11:35 < _mukti_> exp: ba za lo cacra be li su'o mi broda 11:35 < mensi> ([{<ba za> <lo (¹cacra [be {li <su'o BOI> LOhO} BEhO]¹) KU>} mi] [CU {broda VAU}]) 11:36 < selpahi> rutytar: The kei is elided, it's before the {na} 11:36 < Ilmen> Well I find them easy to distinguish, but indeed that highly depends on one's linguistic background. I'm sure I'd have a hard time distinguishing sound pairs that are easy to distinguish for people with another linguistic background than mine. 11:36 < Ilmen> @ ctefaho 11:36 < _mukti_> That seems like an uncontroversial rule to merge into camxes. 11:37 < _mukti_> (ba za lo broda) instead of (ba za) (lo broda), I mean 11:37 < selpahi> Ithkuil has h vs hh, that's difficult for me. 11:37 < rutytar> gleki: how is the place structure of ja'e determined? from the gismu it's derived from? 11:37 < selpahi> {ja'e} is a preposition. 11:38 < rutytar> but why is it a preposition for what causes rather than what is caused? 11:38 < selpahi> {ja'e X} means "resulting in X" 11:38 < ctefa`o> Ilmen: /vwar/ from /var/ and /war/ 11:39 < _mukti_> {jalge} is a result 11:39 < selpahi> Because {jalge}'s x1 is a result. 11:39 < Ilmen> {ja'e} is a oddball, it has it's meaning reversed as compared with other causals (ki'u/ni'i/ri'a/mu'i...) 11:40 < Ilmen> You can use {seja'e} for tagging the cause instead of the consequence 11:40 < Ilmen> (se + ja'e) 11:40 < rutytar> if {ja'e X} means "resulting in X", wouldn't {xunre ja'o lo blanu} be "blue because red"? 11:40 < rutytar> "red results in blue" 11:41 < Ilmen> ctefa'o: I can pronounce and distinguish all the three easily 11:41 < ctefa`o> Ilmen: good 11:41 < selpahi> "it being red results in a blue event" 11:41 < selpahi> Me too 11:42 < rutytar> do i don't want {se ja'e} 11:42 < rutytar> maybe gleki was misinterpreting my intended meaning 11:42 < selpahi> You can use {ki'u X} for "because X" 11:43 < selpahi> If the {se} confuses you with {ja'e}. 11:43 < rutytar> it's supposed to be {because of X, there is Y}, so i used {X ja'e Y} 11:43 < _mukti_> What's the original text? 11:43 < rutytar> isn't that correct? 11:43 < rutytar> _mukti_: http://wiki.lesswrong.com/wiki/Litany_of_Gendlin 11:44 < selpahi> {ki'u X, Y} 11:44 < _mukti_> Would I be wrongheaded in considering {ni'i} for this? 11:45 < selpahi> No. 11:45 < _mukti_> Since these are "reasons" in the sense of what persuades or justifies, but logically inexorable effects... 11:45 < rutytar> well there aren't any axioms, so i don't think it's a logical conclusion 11:45 < rutytar> something like "the care wouldn't start because it's out of gas" isn't a logical statement 11:46 < rutytar> but it isn't a persuasion or justification either 11:47 < _mukti_> I fumble around with the causals, and usually default to {ki'u} (though {ja'e} would be a much more prudent default) 11:51 < rutytar> how would i say "ri ka'e se rinka" without the se? where would the ka'e go? 11:52 < rutytar> could i say "rinka ri ka'e"? 11:52 < selpahi> ka'e rinka ri 11:53 < selpahi> You *can* say {rinka ri ka'e}, but I don't recommend it unless you're writing a poem or are being colloquial. 11:53 < selpahi> The thing is that if you follow it up with another sumti, then it gets swallowed by {ka'e}, which is also a preposition 11:53 < rutytar> i see 11:54 < rutytar> well i am sort of writing a poem, but ka'e rinka ri actually fits better 11:54 < selpahi> If you need the {ri} to come first (hypothetically) then {fe ri ka'e rinka} is an option that doesn't involve {se} 11:55 < _mukti_> {selri'a} is another option 11:55 < selpahi> (this happens to be extremely rarely done in practice, don't ask me why) 11:56 < rutytar> for this specific project i don't want to use any lujvo 11:56 < _mukti_> je'e 11:56 < rutytar> or rafsi, rather 11:58 < rutytar> rafsi is an interesting component of the language, it reminds me of japanese where children's books have katakana written above the kanji 11:58 < rutytar> there's an advanced and simple version of the alphabet 11:58 < _mukti_> furigana 11:58 < rutytar> ah cool 11:59 < _mukti_> That's a neat parallel. I like it. 11:59 < selpahi> Lujvo are not always just a collapsed version of two gismu. Lujvo (and zi'evla) are needed to make novel concepts that gismu alone can't give us. 11:59 < rutytar> because of that, it seems to me that a feature of gismu should be lack of specificity 12:00 < selpahi> Avoiding lujvo entirely means giving up meaning. 12:02 < selpahi> Lujvo are sometimes advertized as "advanced" brivla, because they are made up of other gismu, and making them is "better left to experienced users", so beginners are scared for no reason. 12:02 < selpahi> There is no difference between looking up a lujvo and looking up a gismu. Or looking up an English word and using a lujvo or a gismu you find. 12:03 < _mukti_> I think part of the difficulty with lujvo is not having a reliable list of the most frequent and accepted ones. 12:03 < selpahi> You don't need to know about their components. 12:03 < rutytar> i thought that was mainly because there isn't enough room to make rafsi for every gismu, and they didn't want people going around using them up 12:03 < _mukti_> Part of the difficulty regarding beginners, I mean. 12:03 < rutytar> part of the problem with analyzing lojban usage is that most of lojban speech is people talking about lojban 12:04 < _mukti_> Part of the failure of the infamous Dictionary project is that Lojbanistan has never canonized lujvo. 12:04 < clakre> I thought part of the reason not to have beginners make lujvo was the slinku'i problem 12:04 < clakre> or does that only occur with zi'evla? 12:05 < selpahi> Some people are strongly anti-lujvo (e.g. gleki), some are strongly anti-non-jvajvo, and some don't care at all. There is little unity. 12:05 < _mukti_> rutytar: On the other hand, we usually do that in English. 12:05 < selpahi> Making lujvo is different from using already existing ones 12:05 < selpahi> A beginner can do the latter. 12:06 < clakre> right, but you said above that lujvo were advertized as "advanced" and added that _making_ them was "better left to experienced users", which I think is true 12:06 < rutytar> http://pastebin.com/raw.php?i=Q0gRdv3k any problems with this one? 12:06 < clakre> if peolpe are interpreting that is _using_ them being advanced, that's unfortunate 12:06 < selpahi> I myself am not the biggest fan of lujvo, but only because I dislike the random rafsi. I like my free form lujvo just fine, and treat them like any old zi'evla or gismu. 12:06 < _mukti_> Also as a beginner, I know I was spooked -- and still am somewhat discomfited -- by the jvajvo/non-jvajvo split. Not knowing how much meaning I can infer from rafsi when encountering an unknown lujvo. 12:06 < selpahi> Right, clakre. 12:07 < selpahi> rutytar: You need {lo} before {nu} 12:07 < selpahi> {nu broda kei} is a selbri 12:08 < selpahi> {lifri} -> {se lifri} 12:08 < selpahi> And {ja'e} in the last line needs to be {ki'u}/{ni'i}/{se ja'e}. Also the scope is wrong, but it's too advanced I'm afraid 12:08 < rutytar> okay. cool 12:09 < selpahi> That's why I use {.i ni'i bo} 12:09 < selpahi> used * 12:09 < rutytar> oh! i get the ja'e thing now 12:10 < rutytar> could you give me a synopsis of the scope thing? or a link? 12:10 < selpahi> 'fraid not. 12:10 < selpahi> (about the link) 12:11 < selpahi> {ka'e stali ja'e lo nu renvi} has {ka'e} have scope over {ja'e}, so it means "staying with the result of surviving is possible" 12:11 < selpahi> Or better with the correct version with ni'i: 12:12 < selpahi> {ka'e stali ni'i lo nu renvi} means "staying because your already surviving is possible" 12:12 < selpahi> While {ni'i lo nu renvi cu ka'e stali} means "Because you're already surviving, you can stay" 12:12 < selpahi> (also stali lo jetnu is malgli for "stand the truth") 12:13 < ctefa`o> stali bu'u lo jetnu sa 12:13 < _mukti_> There seems to be a convention (unwritten? buried in some mailing list thread?) whereby the presence of certain rafsi (inversion, negation, -gau, etc) marks a jvajvo. But there may well be exceptions. 12:13 < _mukti_> Is that formalized somewhere? 12:13 < rutytar> why is stali malgli but not rivbi? 12:14 < selpahi> stali means to stay. 12:14 < rutytar> but doesn't rivbi mean to avoid? 12:14 < selpahi> Yes 12:14 < rutytar> aren't they the same level of metaphor? 12:14 < rutytar> truth isn't an event or an object 12:15 < selpahi> The point is that {stali} does not mean what English "stand something" means 12:15 < rutytar> x1 abides with x2 12:16 < selpahi> mi stali lo zdani - I stayed home. 12:16 < selpahi> This has nothing to do with "I can't stand loud music" 12:17 < rutytar> which is the meaning: to not escape, to be present alongside 12:17 < rutytar> if the definition were "x1 remains at x2" i would agree, but it's "x1 remains with x2" 12:18 < _mukti_> I read "stand" in the original as meaning "bear"... closer to {bevri} than {stali} in metaphor 12:18 < rutytar> the seem to me to imply that it's passive, that it's a matter of not avoiding rather than searching 12:19 < rutytar> i'm unconvinced :p 12:21 < selpahi> I think the gimste definition means "stay at place x2". Personally I use {stali} as "x1 keeps doing/being x2 (ka)", and {zvasta} for "x1 stays at place x2" 12:22 < rutytar> so wouldn't "x1 keeps doing/being the truth" be accurate? 12:22 < _mukti_> sanli2 might be worth considering... 12:22 < _mukti_> The idea of the truth as what you're standing upon. Though it departs from the original. 12:23 < _mukti_> So, yeah, forget that. 12:23 < rutytar> i thought it was more like the weight you're holding up 12:25 < rutytar> or being uncomfortable with something happening 12:25 < rutytar> which again, i think abide is perfect for "can stand" 12:26 < ctefa`o> Is it documented anywhere but the CLL as to why some valid consonant clusters are allowed as initials while others are not? 12:26 < _mukti_> I see the important contrast in those last two lines between {ka'e} and {ca'a} 12:27 < _mukti_> stand and endure seem to be used synonymously with that contrast 12:27 < selpahi> ie 12:27 < rutytar> i don't think so. it's an example of english being tangley 12:28 < _mukti_> ctefa'o: There's a formal PEG morphology. Or you're looking for the rationale? 12:28 < rutytar> for the sake of poetry.. they bring up different ideas 12:28 < rutytar> which is part of why i like stali, because we use renvi later on 12:29 < rutytar> {.i ki'u lo prenu ka'e stali lo jetnu kei lo nu ca'o renvi ri} does this have scope issues? 12:29 < ctefa`o> yes the rationalr 12:29 < ctefa`o> Why skari and not ksari 12:29 < rutytar> the scope issue is with ka'e keeping going after the kei, right? 12:30 < selpahi> That sentence is all jumbled up 12:30 < _mukti_> The details have been debated a lot on the mailing lists, and in the old BPFK bulletin board, etc. But I don't think it's been set down in a coherent piece. So many projects! 12:31 < selpahi> lo prenu ka'e stali lo jetnu kei ki'u lo nu ca'o renvi ri 12:31 < selpahi> This is grammatically correct, but has the wrong meaning 12:31 < selpahi> {ki'u lo prenu} means "because of people" 12:31 < rutytar> doesn't kei end ki'u? 12:32 < selpahi> No. {kei} ends {nu}, but there is no {nu} 12:32 < selpahi> You need {ki'u lo nu .... kei} if you want to put an event after {ki'u} 12:33 < rutytar> oh. i see 12:33 < rutytar> so ki'u is essentially just backwards ja'e? 12:34 < selpahi> Sort of, ignoring the differences between {krinu} and {jalge} 12:37 < rutytar> mi xebni lo bakri i ki'u bo ri punji lo purmo lo xance — I hate chalk. It leaves powder on hands. 12:37 < rutytar> does that mean i could just say {lo prenu ka'e stali lo jetnu .i ki'u lo nu ca'o renvi ri}? 12:38 < selpahi> {.i ki'u bo lo go'i ca'o renvi ra} 12:40 < rutytar> could i do it without go'i? 12:40 < selpahi> {.i ki'u bo ra ca'o renvi ri} 12:41 < rutytar> oh right, i have to call lo prenu and lo jetnu 12:42 < _mukti_> i je dy pu rànda ca lo fànmo gi'e fàrlu lo lòldi vi lo cmàlu ki lo kòrbi be lo makpàpi fòldi 12:42 < _mukti_> Is this sticky {ki} ? 12:42 < selpahi> No. 12:42 < rutytar> http://pastebin.com/raw.php?i=qrLMV8iQ 12:42 < selpahi> It's frame of reference {ki} 12:42 < selpahi> I don't use sticky {ki} 12:43 < selpahi> {ki lo korbi} sets the point of reference for the tenses, here {vi}. 12:43 < _mukti_> Yes, I was surprised 12:43 < rutytar> is that one good? 12:43 < selpahi> A small distance from the edge 12:43 < _mukti_> Ah, so it prefixes 12:43 < selpahi> {ki} is a preposition 12:43 < _mukti_> vlaste: ki (class) 12:43 < vlaste> ki (class) = KI 12:45 < _mukti_> Is there an equivalent in cekitaujau? 12:46 < selpahi> Maybe {ke'a}. 12:46 < rutytar> i'm going to do some art stuff to it, so i want to make sure i don't need to edit it any more 12:46 < selpahi> I'm almost the only one who uses {ki} that way 12:46 < _mukti_> Yeah, I hadn't encountered it yet. And it seems like you've moved on! 12:47 < _mukti_> It's an interesting particle, though. 12:47 < selpahi> It's necessary 12:47 < _mukti_> I see that 12:47 < selpahi> rutytar: The rinka sentence confuses me 12:47 < rutytar> how so? 12:47 < selpahi> "it's true so it can be caused" ? 12:48 < selpahi> Is that intended? 12:48 < rutytar> "because it's true, it's there to be interacted with" 12:48 < selpahi> rinka means cause. 12:48 < rutytar> i guess it's a bit liberal, but it's kind of a weird sentence in english as well 12:49 < selpahi> "interact" is difficult. 12:49 < rutytar> can be effected was more what i was goign for 12:49 < rutytar> it seems malgli to discredit a translation because the place structure seems strange 12:50 < selpahi> It's not about place structure, but about rinka being the wrong word 12:51 < selpahi> It's not about causing the truth, since the truth is already there. Interact here means affect or interact 12:51 < selpahi> But jikca doesn't work 12:51 < selpahi> I'd make a new word. 12:51 < selpahi> Because the lujvo would be too long 12:52 < rutytar> what about "se zukte"? 12:52 < selpahi> affect or change or be changed by * 12:52 < rutytar> the english is "it is what is there to be interacted with" 12:53 < rutytar> seems basically to be "what there is to work with" 12:53 < selpahi> I read it as "what there is to face" 12:54 < rutytar> deal with, then 12:54 < rutytar> kurji? 12:54 < rutytar> no, not really 12:55 < rutytar> kakne 12:55 < selpahi> No. 12:55 < selpahi> A new brivla is needed. 12:55 < selpahi> This isn't the first time I'm struggling with "interact" 12:56 < selpahi> It's a kind of ckini 12:56 < selpahi> But there is no good seltau 12:58 < rutytar> nu doesn't work then, it should be ka 12:59 < rutytar> ri jetnu ja'e ka ka'e — ri 12:59 < rutytar> wouldn't that be better? 12:59 < _mukti_> {jikca}, but not social 12:59 < rutytar> ri jetnu ja'e ka ka'e kakne ri 12:59 < rutytar> ^ does that work grammatically? 12:59 < _mukti_> jikca ro'a cu'i 13:07 < selpahi> You want to follow {ja'e} with a sumti, so use {lo} 13:08 < selpahi> {nu} is better than {ka} 13:08 < selpahi> {ja'e lo ka} or {ja'e lo nu} 13:08 < rutytar> okay 13:08 < selpahi> {kakne ri} is weird. 13:08 < rutytar> how about {.i ri jetnu ja'e ro lo nu ka'e kakne ri} 13:08 < rutytar> ignoring kakne for now 13:08 < selpahi> The truth isn't an ability 13:09 < selpahi> I'm not sure what the point of the {ro} is 13:10 < rutytar> the english is "And because it's true, it is what is there to be interacted with." 13:11 < rutytar> "it is what is there" seems to imply "all" to me 13:11 < selpahi> Okay, but then the {ro} is in the wrong place. And maybe you want {po'o} 13:11 < rutytar> where would you put it? 13:11 < selpahi> Although generally one thing existing doesn't mean other things don't exist 13:12 < selpahi> "it is all there is" is {po'o} 13:12 < selpahi> "only the truth is there" 13:12 < selpahi> That's what your "all" means 13:13 < selpahi> {ro} doesn't work without heavy refactoring 13:13 < rutytar> but ro is in the wrong place… where is the right place? 13:13 < selpahi> {ro X} means "each X" 13:13 < selpahi> The right place is in a differently worded sentence. 13:14 < selpahi> In your sentence you need {po'o} 13:14 < rutytar> :( 13:14 < _mukti_> Still thinking about "interact". {jai galfi} ? 13:14 < _mukti_> Or since its "interacted with", se galfi 13:14 < selpahi> I think interact is something like "being in a cause-and-effect relationship", so rinka/ni'anfu ckini or something. 13:15 < _mukti_> rinka simxu 13:15 < selpahi> In my version I used {kau} instead of {ro} or {po'o} 13:17 * nuzba @hubskihose: ZA'O by Djemynai, a Lojban rap album #language #music http://dlvr.it/BVkvLH [http://bit.ly/1L9fBFk] 13:19 < _mukti_> ui 13:21 < _mukti_> So I see you used {jikca} in your version. Do you think jikca should be more broad than the "social" gloss would imply? 13:21 < rutytar> so {ki'u lo nu ri jetnu kei ri kau se jikca} - could this be worded as {{ki'u lo nu ri jetnu kei kau jikca ri}? 13:21 < selpahi> No, jikca is wrong. 13:22 < selpahi> It's a place holder 13:22 < rutytar> but grammatically, are the equivalent? 13:22 < rutytar> *they 13:22 < selpahi> I was talking to mukti 13:22 < rutytar> ah 13:22 < selpahi> They aren't equivalent. 13:23 < selpahi> The {kau} has to follow {ri} 13:23 < selpahi> And {nu jetnu kei jikca} is a tanru 13:23 < selpahi> Add {cu} 13:23 < selpahi> {ki'u lo nu ri jetnu cu jikca ri kau} 13:23 < selpahi> This is equivalent 13:24 < rutytar> okay 13:24 < rutytar> thanks 13:24 < rutytar> i'm going to stick with kakne for now 13:24 < rutytar> that's how i interpret it 13:24 < selpahi> In the last line, the {ri} should be {ra} I think 13:25 < selpahi> In my opinion {kakne} makes no sense, but it's up to you :) 13:26 < rutytar> {kakne ri} = "ri can be done" is what i'm thinking 13:26 < rutytar> i guess truth can't really be done 13:27 < rutytar> pilno? 13:27 < rutytar> it's too vague a concept 13:28 < selpahi> If you don't want lujvo or zi'evla, then a tanru like {stika ckini} is probably your best bet 13:28 < _mukti_> ni'o mi va'o lo nu tanxe lo tabjme cu djica lo ka jinvi lo du'u tanxe lo tabjme .i je bo mi va'o lo nu na ku tanxe lo tabjme cu djica lo ka jinvi lo du'u na ku tanxe lo tabjme .i je bo .ai mi na krici lo na se djica be mi 13:28 < selpahi> Replace stika with something from galfi/rinka/cnegau/ni'anfu if you like 13:30 < selpahi> That last sentence needs to be treated with c are. 13:30 < selpahi> care* 13:30 < _mukti_> Yes, I was sensing a disturbance in the Scope 13:30 < selpahi> An example where {jai} makes all the difference 13:31 < selpahi> {krici lo na se djica} vs {krici lo poi'i mi na djica lo nu mi krici ke'a} 13:31 < selpahi> This is an ambiguity that English has but Lojban doesn't 13:31 < _mukti_> {poi'i} ... where does he get all these wonderful toys 13:32 < rutytar> i've got to go now. thanks everyone 13:32 < selpahi> Take care 13:32 < _mukti_> co'o la .rutytar. 13:32 < rutytar> i just started school so my time i can devote to lojban is thin, but i'll try anyway 13:32 < rutytar> co'o 13:33 < selpahi> mi xagji 13:34 < _mukti_> .e'u ko citka 13:35 * mumble jorne co'u fa la'o fa Ilmen fa 13:35 * mumble jorne co'a fa la'o fa Ilmen fa 13:36 < _mukti_> ua la .cogas. cu sanga 13:36 < _mukti_> https://twitter.com/oka_iu_tcan/status/620236273416978432 13:49 < selpahi> mi pu za tinju'i 13:49 < selpahi> .i ui jbozgike 13:49 < selpahi> .i xu do pu'i tinju'i lo .itku'ile zgike noi cnino 13:50 < selpahi> (https://www.reddit.com/r/Ithkuil/comments/3cw5xt/introducing_kaduat%C3%A1n_featuring_david_j_peterson/) 13:51 < selpahi> mi nelci .i ie mi nelci 13:54 < _mukti_> ie mi pu tinju'i .i ku'i mi na se bangu lo .iku'ile 13:54 < Ilmen> .a'u 13:54 < Ilmen> jbo: ni'anfu 13:54 < mensi> ni'anfu = su'o pagbu be x1 cu rinka su'o pagbu be x2 |>>> ahernai 13:54 < Ilmen> wa 13:54 < _mukti_> .i ku'i mi gleki lo nu tinju'i lo cnino runbau zgike 13:54 < Ilmen> simsa zo rinpau pe mi 13:55 < astroendjisai> .i xu la .mensi. xamgu tarti 13:56 < selpahi> la mensi cu xamgu tarti 13:56 < astroendjisai> xamgu 14:08 < astroendjisai> lujvo: zukte bitygau 14:08 < mensi> zuk[2949], zu'er[4438], zukty[5079] 14:08 < astroendjisai> Just curious 14:09 < selpahi> looks like a bug 14:09 < selpahi> gleki: ^ 14:10 < astroendjisai> bitygau is already a lujvo, should that matter? 14:10 < selpahi> Even if one isn't supposed to enter lujvo there, it still shouldn't output something like that 14:11 < astroendjisai> Ah 14:11 < astroendjisai> So is it not possible to create a lujvo from a tanru already containing a lujvo? 14:11 < selpahi> It's possible 14:11 < astroendjisai> Ah. 14:12 < astroendjisai> Since the other day I've been trying to think up some way to say "1 convinces 2 to do action 3" 14:12 < selpahi> But it can also be thought of as making an entirely new lujvo with those same three rafsi, rather than from a lujvo and a gismu 14:12 < astroendjisai> Ah 14:12 < selpahi> afk, eating. Maybe someone else can help you 14:15 < astroendjisai> I'm not in any hurry, that same day we found a workaround by just using 'convince it is true that it is good to do [action]. Example: mi bitygau da lo du'u xamgu fa lo nu tadni la .lojban. (not that that ever happens) 14:15 < Ilmen> mo'u zgana la'oi Kaduatán 14:15 < Ilmen> .i pluka 14:29 * mumble jorne co'u fa la'o fa Ilmen fa 14:30 < ctefa`o> co'o rodo 14:31 < Ilmen> co'o cliva 17:25 < zipcpi> vlaste: cmevo (gimka) 17:25 < vlaste> zo cmevo gimkamsmikezypro no catni gismu .i zo cmevo gimkamsmikezypro no cipra gismu 17:28 < zipcpi> http://mw.lojban.org/papri/brivycme 17:29 < zipcpi> e'a ro prenu poi aidji cu jmina 17:58 < zipcpi> Someone needs to add "la vrici"; I still don't know what it means 18:02 < zipcpi> k: jbo'irci 18:02 < mensi> (CU [Z:jbo'irci VAU]) 18:06 < zipcpi> k: jbouitki 18:06 < mensi> (CU [Z:jbouitki VAU]) 19:12 < deltab> how is 'brivycme' formed? 19:15 < zipcpi> brivo + cmene 19:16 < zipcpi> brivo is an experimental gismu; perhaps {brivlacme} should be a synonym 19:18 < deltab> ki'e 19:21 < zipcpi> k: mikranti 19:21 < mensi> (CU [L:mikranti VAU]) 19:32 < astroendjisai> I'm wanting to try to translate something (short prose or something) as an exercise in figuring out how Lojban works. I'm going this route since I can't convince anyone to learn it with me so I can speak it with them. Any suggestions for what to try and translate? 19:32 < zipcpi> Short stories? Something you might be interested in 19:33 < zipcpi> Perhaps even instructional texts; Lojban does have "too many" translated fiction 19:33 < astroendjisai> True. 19:33 < astroendjisai> Heck, I could even just straight up try to write something. 19:33 < zipcpi> Mhm 19:33 < astroendjisai> I have a mad plan. AHAHAHAHAHA 19:34 < astroendjisai> Write a daily journal but in Lojban. And perhaps Japanese since that's the language I'm actually supposed to be learning. 19:35 < zipcpi> :p 19:35 < Tachyon`> anyone translated the last question into lojban? 19:36 < zipcpi> Oh that one 19:38 < astroendjisai> Actually I won't do that for privacy issues since I'll end up wanting to get someone to check it for me. Hmm. 19:39 < clakre> near as I can tell, the most popular use for lojban is to discuss changes to lojban in lojban 19:39 < astroendjisai> I thought it would be fun to use it as sort of a code language with my friends but nobody'll go for it. 19:41 < astroendjisai> I guess I can use it to keep notes or something for myself that only I'll understand. At the very least I'll develop some sort of pidgin-Lojban. 22:15 < zipcpi> k: mikca'o 22:15 < mensi> (CU [L:mikca'o VAU]) 22:33 < zipcpi> k: poltepruxi 22:33 < mensi> (CU [Z:poltepruxi VAU]) 22:33 < zipcpi> k: poltevanci 22:33 < mensi> (CU [Z:poltevanci VAU]) 22:34 < zipcpi> k: polteru'i 22:34 < mensi> (CU [Z:polteru'i VAU]) 22:44 < zipcpi> k: uitkipedi 22:44 < mensi> (CU [Z:uitkipedi VAU]) 23:04 < gleki> im not against lujvo. im more against short rafsi 23:05 < gleki> for "I stay in the building" i just say it literally {mi stali bu'u lo dinju} 23:36 < gleki> https://en.wikipedia.org/wiki/Hornbostel%E2%80%93Sachs 23:36 < gleki> ^ that's nice --- Day changed Mon Jul 13 2015 00:05 < gleki> mensi: doi selpa lo anfora zo'u da'i ma'a vimcu lo lerfu anfora i je ku'i co'aku ma xe fanva zoi gy. I use WWW .gy. 00:05 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.selpa.gy. di'a cusku da 00:12 < zipcpi> vlaste: jinda (gimka) 00:12 < vlaste> zo jinda gimkamsmikezypro no catni gismu .i zo jinda gimkamsmikezypro no cipra gismu 00:14 < durka42> coi 00:14 < gleki> coi jinda i coi durka 00:14 < durka42> mi zvati la .tubingen. 00:15 < gleki> xu la iutubingen 00:15 < zipcpi> tu~bi~nge~n .y. 00:15 < zipcpi> ki'a 00:19 < zipcpi> http://mw.lojban.org/papri/brivycme doi durka do'u xu do stidi lo ka'e se jmina 00:20 < zipcpi> .ije xu da mabla 00:22 < zipcpi> mi no'e birti tu'ai <la terjansu> li'u ki'u lo nu mi na kelci 00:22 < gleki> xu la nu jansu 00:22 < zipcpi> ju'ocu'i 00:28 < gleki> only now i realized that djemynai is DJ-emynai 00:55 < durka42> zipcpi: when we were playing we called it {la kibjasyselkei} 00:55 < durka42> or {la jasyselkei} for the non-web version I guess 00:55 < zipcpi> ua 00:56 < zipcpi> I think I'll list both {la jasyselkei} and {la terjansu} as variants 00:58 < Amnadya> I called it {la nu jansu} 00:59 < zipcpi> I somewhat dislike NU in names because sometimes it needs to be closed, which ruins the "name"-ness IMO 00:59 < zipcpi> Unless it is going to actually have sumti 00:59 < selpahi> . 00:59 < mensi> selpahi: cu'u la'o gy.gleki.gy.: lo anfora zo'u da'i ma'a vimcu lo lerfu anfora i je ku'i co'aku ma xe fanva zoi gy. I 00:59 < mensi> use WWW .gy. | 2015-07-13T07:05:55.882Z 01:00 < gleki> xu lu mi pilno la'e li vy bu vy bu vy bu li'u 01:06 < Amnadya> .i re da ralju cumki .i pa mai sa'u co'u ta'e pilno lo lerfu tu'a lo zasni sa'e nai ru'e .anfora (to mu'a lo mlatu cu klama .i my. xagji sei zasni .anfora toi) je co'a pilno fi ra'u tu'a lo na'e zasni sa'e nai ru'e .anfora (to mu'a CLL ja BPFK ja NSA sei cmene simsa toi) (to jalge fa lo nu jdika lo ka cafne toi) 01:07 < Amnadya> .i ja re mai 01:07 < Amnadya> pavgau lo lerfu joi lo namcu 01:07 < Amnadya> ja'e lo nu lo lerfu cu namcu 01:07 < gleki> je'e 01:08 < Amnadya> .i lo re moi danfu cu jai se sarcu lo nu co'a pilno lo simsa be zo li ja zo me'o 01:08 < Amnadya> .i da'i finti lo cnino poi cupra lo cmene 01:08 < gleki> i ku'i ma ca ba'e ti cu snura traji i pe'i lo remoi 01:08 < Amnadya> .i ji'a sarcu fa lo nu finti lo cnino je xamgu .anfora bo sumka'i 01:09 < Amnadya> kei va'o lo nu lo lerfu co'u se pilno 01:09 < gleki> lo nu go'i na nabmi mi 01:09 < Amnadya> je'e 01:09 < gleki> i ku'i lo mupli pe zo'oi WWW cu jai nabmi 01:10 < Amnadya> en:la'ai 01:10 < mensi> la'ai = [LOhU] otherwise lojbanic name, ending in a vowel; multiple names delimited by pauses. 01:10 < Amnadya> en:li'ai 01:10 < mensi> li'ai = [LI] unevaluated mekso as name. 01:10 < gleki> ma'oi li vau o'u 01:10 < gleki> i sa'e na o'u cnino selmaho 01:10 < gleki> jbo: li'ai 01:10 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 01:10 < gleki> je'e 01:11 < gleki> en: research 01:11 < mensi> 3 da se tolcri: boltsemaku, lenguua, skestu 01:11 < Amnadya> .i va'o lo nu fenki ka'e gau basti zo li'ai fa zo lai ju'i nai lobypli 01:11 < gleki> en: inquir 01:11 < mensi> 2 da se tolcri: nabmi, nadycai 01:11 < gleki> en: investiga 01:11 < mensi> 2 da se tolcri: cipcta, xigzo 01:11 < zipcpi> mi za'o nelci ru'e le lerfu anfora ki'u lo nu va'o lo nu pilno zo goi kei ka'e pilno lo lerfu anfora noi zmadu lo ka morji se kakne 01:12 < zipcpi> .y. noi zmadu lo ze'oi ko'a anfora 01:12 < Amnadya> mi ta'e no roi pilno zo goi 01:12 < zipcpi> mi so'oroi pilno .i plixau va'o lo ve ciksi 01:13 < gleki> mi pu kucli tu'a lo snura tadji 01:13 < gleki> i le bangu cu cenba i mi nitcu lo stodi 01:13 < Amnadya> la'a su'o snura cu zasti .i ku'i cumki fa lo nu fegli 01:14 < gleki> mi pu vimcu lu abu li'u pe la alsi lo mi vlaste i ku'i zo'oi WWW ca'o renvi 01:14 < Amnadya> .i lo'u la'e me'o le'u mu'a jai snura je ku'i cu fegli mi 01:14 < gleki> sa'ai alis le'ai 01:15 < gleki> i la vy bu vy bu vy bu co'e 01:15 < zipcpi> .y. di'u lerfu anfora 01:16 < gleki> exp: i la vy bu vy bu vy bu co'e 01:16 < mensi> (i [la {<(¹[vy bu] [vy bu] [vy bu]¹) BOI> co'e} KU] VAU) 01:16 < zipcpi> la co'e pe vy.bu vy.bu vy.bu 01:16 < gleki> i ku'i cmene 01:16 < zipcpi> la mi dirba xirma 01:16 < zipcpi> le gerna cu simsa 01:16 < gleki> i na dunli lu la co'e ku pe vybu vy bu vy bu li'u 01:16 < zipcpi> exp: la co'e ku pe vybu vy bu vy bu 01:16 < mensi> ([{la co'e ku} {pe <(¹[vy bu] [vy bu] [vy bu]¹) BOI> GEhU}] VAU) 01:17 < gleki> exp: lo co'e pe vy bu 01:17 < mensi> ([lo {co'e <pe (¹[vy bu] BOI¹) GEhU>} KU] VAU) 01:17 < zipcpi> .y. mi claxygau zo ku 01:17 < gleki> ie i zo ku vajni 01:18 < zipcpi> va'o lonu zo ku cu claxu kei le mezopemei cu pagbu le cmene 01:18 < zipcpi> .y. mezopemoi 01:19 < gleki> ie ie 01:19 < gleki> i ku'i na danfu tua lo lerfu cmene 01:19 < zipcpi> zo li'ai ka'e .iku'i ka'e fegli 01:21 < gleki> ma'i mi ca'a fegli 01:21 < gleki> i ku'i ca ba'e ti le nabmi cu zasti i mu'a zo'oi 111 ka'e cmene i ta'i ma pilno lo tai cmene 01:22 < zipcpi> pe'i nitcu zo li'ai ja lo'u la'e li le'u 01:22 < gleki> exp: la'e me'o 111 cu bebna 01:22 < mensi> ([la'e {me'o <(¹pa [pa pa]¹) BOI> LOhO} LUhU] [cu {bebna VAU}]) 01:23 < gleki> fegli 01:23 < gleki> i fegban 01:23 < zipcpi> u'i 01:23 < gleki> y zo'oi 911 cmene 01:24 < gleki> i je'u mi puza kucli 01:24 < gleki> https://groups.google.com/forum/#!topic/lojban-beginners/jeEZY2nYku4 01:25 < zipcpi> https://groups.google.com/forum/#!topic/lojban-beginners/jeEZY2nYku4 xy~'y zo aksese 01:25 < gleki> mo 01:25 < zipcpi> mi zasni finti zo aksese tau lo cmene po'u zo miky'aksese 01:26 < gleki> i xu do na kakne lo ka tcidu 01:26 < gleki> zoi gy.Therefore one should write {.i ko klacpe la'e me'o sopapa}..gy. jai se stidi la gejyspa 01:26 < zipcpi> ku'i "The missing concept includes opening a file, accessing a webpage, opening a 01:26 < zipcpi> cupboard, opening a book, or accessing a memory location. Opening a cupboard 01:26 < zipcpi> is "kargau", but that's what you do to the door; I'm talking about what you do 01:26 < zipcpi> to the cupboard. Any suggestions?" 01:26 < gleki> zo kargau jai se stidi mi 01:27 < gleki> lo nu kargau lo canko cu simsa lo nu kargau lo cukta 01:27 < zipcpi> kargau is what you do to the door; there needs to be a general word for "access" 01:27 < gleki> kalri = to be accessible 01:28 < zipcpi> Hmm... not sure about that 01:28 < gleki> it's what the definition of ganlo says 01:28 < gleki> although the two words are poorly defined as noticed by guskant 01:28 < zipcpi> x1 is defined as a portal, unless we just want to ignore that 01:29 < gleki> en: ganlo 01:29 < mensi> ganlo = x1 (portal/passage/entrance-way) is closed/shut/not open, preventing passage/access to x2 by x3 (something being 01:29 < mensi> blocked). |>>> As a doorway, but also perhaps a semi-permeable membrane. x3 is blocked by x1. x2 is inaccessible. See 01:29 < mensi> also kalri, pagre, canko, vorme, zunti. |>>> officialdata 01:29 < gleki> x2 is inaccessible 01:29 < gleki> ganlo = to'e kalri 01:29 < zipcpi> zilkarli 01:29 < zipcpi> x1 is accessible to x2 01:29 < zipcpi> *zilkalri 01:29 < gleki> jb: ganlo 01:29 < mensi> ganlo = ganlo — portal x1(entity) is closed, preventing access to place x2(entity) for object x3(entity) 01:29 < mensi> :lo ganlo — closed. 01:29 < mensi> :lo vorme cu ganlo lo nenri lo fange — The door is closed not letting the strangers in. 01:29 < mensi> :Related words: kalri, pagre, canko, vorme, zunti 01:30 < gleki> ^ that definition used guskan'ts comments 01:30 < Amnadya> lo vorme je nai lo vrogai cu ganlo 01:30 < gleki> ie 01:30 < gleki> i pe'i lo vrogai na ka'e ganlo 01:31 < Amnadya> .i pe'i pei lo cukta cu simsa lo vorme 01:31 < zipcpi> The thing is x2 is important here, not x1. Like when accessing a webpage, it might be somewhat stretching to put something in x1 01:31 < gleki> do jinvi ma 01:31 < Amnadya> pe'i no'e simsa .i na pagre lo cukta ja lo nenri be ri 01:31 < gleki> lo cukta cu ganlo lo se cukta mi 01:32 < Amnadya> .i lo cukta gacri cu gacri lo se cukta 01:32 < Amnadya> .i srana lo si'o polje 01:32 < zipcpi> I translated Microsoft Access as {la miky'aksese}, very malgli but I just made it up because all I needed was a name, and even Chinese just calls it "Microsoft Access" 01:32 < gleki> lo cukta cu ganlo ri'a tu'a lo gacri 01:33 < gleki> i la mikrysofto kalri 01:33 < zipcpi> And now we want something that {aksese} might mean, but we might want to change it to a different form. 01:35 < Amnadya> lo cukta gacri je lo cukta papri pe ca lo nu "ganlo" cu simsa lo vrogai je nai lo vorme .i lo cukta cu gunma lo gacri joi lo papri .i ma vorme 01:36 < Amnadya> .i pagre ma 01:36 < Amnadya> .i ca lo nu ganlo sei naldra valsi kei lo no gusni cu muvdu fi lo papri fe lo kanla be lo zgana 01:37 < Amnadya> s/lo no/no 01:37 < fenki> Amnadya meant to say: .i ca lo nu ganlo sei naldra valsi kei no gusni cu muvdu fi lo papri fe lo kanla be lo zgana 01:37 < gleki> ma'i do ma ka'e mupli lo ka ganlo 01:37 < Amnadya> lo canko je lo vorme je lo kevna je lo simsa 01:37 < Amnadya> .i srana lo si'o na ka'e pagre 01:37 < Amnadya> .i da zunti ja fanta lo nu pagre 01:38 < Amnadya> .i je da ka'e se vimcu 01:38 < Amnadya> jbo:ganlo 01:38 < mensi> ganlo = lo nu pagre x1 boi (to noi vorme toi) x2 fa x3 fo zi'o cu se zunti zi'o zi'o 01:38 < Amnadya> .i ma pe lo cukta cu vorme 01:40 < gleki> lo cukta cu vorme pe'i 01:40 < gleki> i sa'u lo cukta ka'e cenba lo ka ganlo 01:41 < tsani> .i ku'i naku lo ka vorme cu ka ka'e cenba lo ka ganlo 01:42 < gleki> uanai 01:42 < gleki> i je'upei 01:42 < tsani> je'u 01:42 < tsani> .i lo ka vorme cu srana lo ka ka'e zunti lo ka pagre 01:42 < gleki> i xu naku lo nu lo vrogai cu muvdu cka'e rinka lo nu lo vorme cu cenba lo ka ganlo 01:43 < tsani> .i la .amnadi'a cu drasku pe'i 01:43 < gleki> *ka'e 01:43 < tsani> vlaste: vorme 01:43 < vlaste> vorme = x1 is a doorway/gateway/access way between x2 and x3 of structure x4. 01:44 < gleki> i ja'o ma poi srana lo vorme cu cenba lo ka ganlo 01:44 < tsani> .i ie lo vorme cu cenba lo ka ganlo 01:44 < gleki> i li'a na vrogai 01:44 < Amnadya> .i lo cukta cu vorme ma ma ma 01:44 < tsani> .i ku'i mi si'a jinvi lo du'u lo cukta na vorme 01:45 < tsani> (to ta'o lo kibjudri batke (to me'oi link toi )cu vorme xu toi) 01:45 < gleki> lo cukta cu vorme lo tcidu lo se cukta lo muvdu pelji 01:46 < gleki> i lo kibro tutci zo'u oi aunai casnu 01:46 < gleki> valdi = x1 is a software poorly localized to Lojban by Lojbanist x2 01:47 < gleki> la vi valdi = brivycme, some software near you that is poorly localized 01:58 < zipcpi> vlaste: jonse (gimka) 01:58 < vlaste> zo jonse gimkamsmikezypro no catni gismu .i zo jonse gimkamsmikezypro no cipra gismu 02:07 < zipcpi> http://jbovlaste.lojban.org/dict/jonse 02:39 * nuzba @okau_junpe: .ui zabna La kanto nomiĝas "somero sen vi" En Loĵbano. Låten heter "sommar utan dig". På lojban. https://twitter.com/oka_iu_tcan/status/620229220187291648 [http://bit.ly/1TwytBa] 03:04 < ctefa`o> coi 03:06 < gleki> coi 03:07 < ctefa`o> cio gleki 04:20 < Amnadya> I made a little Ithkuil learning collection: http://selpahi.de/ithkuil/learning/collection.html 04:20 < Amnadya> Some links and some quizzes. 04:48 * nuzba @garydale: The unexpected benefits of climate change: Sure, half of Western Canada was on fire, polar bears were ... http://bit.ly/1Hq2afr #lojban [http://bit.ly/1fCmrrB] 04:59 < gleki> da'i do zbasu lo cukta pe ibu i ku'i ba'a JQ na zanru 05:00 < gleki> i ji'a lo uilkinse liste lo itkuiile me'oi root cu mo i xu da ca'o favgau 05:00 < Amnadya> mi djuno no da lo nu go'e 05:00 < zipcpi> xu lo me'oi root cu rafsi 05:00 < gleki> su'ova'ei 05:01 < gleki> i frica lo prefiksi e lo subyfiski e lo infiksi 05:01 < zipcpi> pe'i ei zo rafsi ja lo drata valsi ka'e se pilno fi tau lo drata bangu 05:02 < gleki> y lo so'u bangu cu kakne tu'a lo jetnu rafsi 05:03 < zipcpi> zo rafsido'a zo'o 05:04 < zipcpi> mu'a zo ki'orbi zo'u: zo'oi -kirbi- cu mo 05:04 < zipcpi> http://jbovlaste.lojban.org/dict/ki'orbi 05:04 < gleki> en: ki'orbi 05:05 < mensi> ki'orbi = x1 is 1024 (2^10) of x2 in dimension/aspect x3 (default units) |>>> Informal combining form: kirbi-. See 05:05 < mensi> kirbibitni, kirbibaitni, kilto. // Series: ki'orbi me'orbi gi'orbi te'orbi pe'orbi xe'orbi ze'orbi go'orbi |>>> 05:05 < mensi> spheniscine 05:06 < gleki> the most pressing fix for alta now is {cu broda noi} support 05:06 < zipcpi> zo selma'o ji'a jai nabmi fai lo simsa 05:07 < zipcpi> xu lo na'e cmavo ka'e se selma'o 05:08 < zipcpi> pe'i zo vlagenkle ja lo simsa 05:10 < Amnadya> is zbaga down? 05:10 < gleki> na ka'e doi la zipcpi 05:10 < gleki> i zo vlakle e'u 05:12 < zipcpi> vlasmukle ji vlagenkle 05:57 < niftg> la mensi .e la makfa mensi ca zvati vau za'a 05:58 < gleki> la makfa mensi cu mo 05:59 < niftg> sampu xe fanva zo'oi magical-imouto 06:00 < niftg> (sa'e se cmene lo go'i) 06:26 < Amnadya> mi piltcu lo do cunso valsi tutci doi la cliva 06:27 < Amnadya> (to #pajvo toi) 06:45 < gleki> en: piltcu 06:45 < mensi> [< pilka nitcu ≈ Crust need] lo nu mulno sisku zo'u: y no da se tolcri 06:45 < mensi> piltcu[5878], pi'atcu[6367], pilnitcu[7937], pilkytcu[8008], pi'arnitcu[9526], ... 07:27 < ctefa`o> coi'oi Amnadya 07:47 < Amnadya> Attál. 07:47 < Amnadya> Glawulač żi'alukta ki? 07:49 < niftg> Atta/l .i mi nu'o kakne co jimpe fi lo jufra smuni co itku'ile ku ta'i lo sutra 07:50 < niftg> ithkuil: Atta/l 07:50 < ithkuil> error 07:50 < Amnadya> En-n üareč. 07:50 < niftg> ithkuil: Attál 07:50 < ithkuil> Attál: P1S1/STA-tt-OBL-UNI/CSL/M/DEL/NRM-EXS-FML/UNFRAMED 07:51 < Amnadya> lu rinsa li'u 07:51 < niftg> xy'y .i nitcu lo glos zei nuncki 07:51 < niftg> lu rinsa li'u zo'u mi pu'i ba'o co'a jimpe .i po'o cipra 07:55 < Amnadya> Glawulač żi'alukta ki? -- xu do ca lo cabdei cu banzu kanro lifri 07:55 < Amnadya> En-n üareč. -- a'o na mabla 07:56 < Amnadya> Exhawúl ki? -- xu do jimpe 07:56 < niftg> zoi .itk. Glawulač żi'alukta ki? .itk. zo'u mi puziku jimpe lo du'u lo ka kanro jo'u lo djedi jo'u mi te tavla ku po'o .i caku mi mulno jibni jimpe 07:56 < Amnadya> ua 07:56 < Amnadya> .i'e 07:57 < niftg> nu'o certu tu'a la'oi Adjuncts 07:57 < niftg> la'oi PRA zo'u so'u va'e vau po'o 07:58 < Amnadya> lo .adjunte na jai frili fai ro da .i ku'i la .modalitis. cu jai frili 07:58 < Amnadya> .i ie PRA zo'u nandu sai 08:06 < samcinki> ŭa .itku'ile casnu coi doi nei 08:07 < gleki> CIA ca se nandu fa lo ka tcidu jimpe lo smuni be di'u 08:08 < niftg> sa'e mi nu'o tavla fo lo itku'ile ja'e lo nu nu'o itku'ile casnu 08:09 < niftg> (to'e do'e zo'oi Atta/l) 08:10 < niftg> .ija ta'u casnu lo itku'ile ca'a 08:11 < Amnadya> mi sruma lo du no nu tavysi'u bau lo .itku'ile pu'i fasnu 08:11 < Amnadya> .i da'i lo jbopre ba pa moi 08:12 < Ilmen> Attàwîl 08:12 < Amnadya> lo ka jbopre cu ka certu lo ka facki fi lo cfila 08:12 < gleki> alta: mi ca tavla noi mutce cizra 08:12 < mensi> (I [FA ZOhE] [CU {FASNU <FA (¹LO [NU {<FA mi> <CU (²ca [tavla NF]²) VAU>} KEI] KU¹) (¹noi [{FA ZOhE} {CU <mutce (²cizra NF²)> VAU}] KUhO¹)> VAU}]) 08:13 < Amnadya> .i la'a ku'i se jalge lo nu so'i prenu cu na'e se pluka lo ka jikca lo jbopre 08:13 < Amnadya> .u'i zo'o 08:13 < Ilmen> .u'i 08:14 < Ilmen> xu lo cnivla cu zasti lo .itku'ile 08:14 < Amnadya> ru'e 08:14 < Amnadya> .i lo rafsufikse je lo .y me'oi bias cu zasti 08:15 < Ilmen> mu'a pei .i ma da'i mapti zo ŭi mu'a 08:15 < gleki> ca ti mi bilga lo ka zukte lo cipra be lo nu xukau la altatufa ku pe ca co'a spofu ki'u lo nu lu i broda noi mo li'u co'a gendra 08:16 < Amnadya> aqtal -- gleki, .i ku'i zo ui zo'u na djuno .i ka'e na zasti 08:16 < Ilmen> alta: lo nu banfi carvi noi fasnu cu cizra 08:16 < mensi> (I [FA {lo <nu (¹[FA ZOhE] [CU {FASNU <FA (²LO [NU {<FA ZOhE> <CU (³banfi [carvi NF]³) VAU>} KEI] KU²) (²noi [{FA ZOhE} {CU <fasnu NF> VAU}] KUhO²)> VAU}]¹) KEI> KU}] [cu {cizra NF} VAU]) 08:17 < gleki> pu je ca gendra 08:17 < Ilmen> je'e la .amnadiĭa 08:17 < Amnadya> vi ji'a carvi fa lo banfi je lo mlatu je lo drata 08:18 < Ilmen> .i ta'o do ma jinvi lo .itku'ile selsa'a 08:18 < gleki> ti'a bu'u lo glico fe ji'a lo gerku cu carvi 08:18 < gleki> sa 08:18 < gleki> i ti'e bu'u lo glico fe ji'a lo gerku cu carvi 08:18 < gleki> sa 08:18 < gleki> i ti'e bu'u lo glico fa ji'a lo gerku cu carvi 08:18 < gleki> i oi se'i mi tatpi 08:18 < gleki> i de'a jundi 08:18 < Amnadya> .i clani sai .i ku'i mi nelci 08:19 < Amnadya> .i ca so'o da ze'u sanga na ku 08:19 < Ilmen> ma'i lo .itku'ile lo bi valsi selsa'a cu da'i banzu lo nu clani mutce 08:19 < Ilmen> .i ja'o banli 08:20 < Ilmen> .i ju'o lo barda pagbu cu voksa claxu .i ku'i lo gunma be ro lo voksa pagbu cu barda banzu ju'o 08:20 < Amnadya> ie 08:21 < Ilmen> * lo ro 08:22 < Amnadya> za'a zoi ty. r .ty simsa zo ui 08:22 < Amnadya> .i mulno jufra fa zoi ty. T. .ty 08:22 < Amnadya> .u'i 08:22 < Amnadya> .yyy 08:23 < Amnadya> s/T/R. 08:23 < fenki> Amnadya meant to say: .i mulno jufra fa zoi ty. R.. .ty 08:23 < Amnadya> .. 08:23 < Amnadya> ithkuil: r 08:23 < ithkuil> error 08:23 < Amnadya> na naldra doi do! 08:23 < Amnadya> .u'i 08:24 < Amnadya> ithkuil: kal'r 08:24 < ithkuil> kal’r: P1S1/STA-k-OBL-l’r-EXS-IFL/UNFRAMED 08:24 < Amnadya> uat 08:24 < Amnadya> .oi se'i 08:24 < Amnadya> ithkuil: kala'r 08:24 < ithkuil> kala’r: P1S1/STA-k-OBL-UNI/CSL/M/DEL/NRM-EXS-GRT-IFL/UNFRAMED 08:24 < Amnadya> GRT = ui 08:24 < Ilmen> lo .evjekti zunsna cu simlu loka na'e jai cafne 08:25 < Amnadya> zanfu'a 08:26 < Amnadya> ithkuil: an-nar 08:26 < ithkuil> an-nar: Verbal adjunct 08:26 < Amnadya> ... 08:26 < Amnadya> .i'a .i'a 08:27 < Ilmen> zo'oi T mo .i xu co'e zo'oi þ 08:27 < Amnadya> lerbasysre .i zo'oi R 08:27 < Amnadya> .i sinxa no da .i sa'u ga'e zei lerfu 08:27 < Amnadya> .i lo pa moi lerfu pe lo jufra cu ga'e zei co'e ta'e 08:28 < Amnadya> "R." .i mulno jufra 08:28 < Ilmen> je'e 08:29 < Ilmen> zo'oi . cizra .u'i .i sa'u na ba'a ka'e sinxa su'o prosodi 08:29 < Amnadya> ___'ř / ř / an-nař -> .e'o 08:29 < Amnadya> .i lo mokca lerfu cu sinxa lo fanmo be lo jufra 08:29 < Amnadya> bau so'e bangu 08:30 < Ilmen> ku'i ta'e panra lo nu lo prosodi sinxa cu fanmo lo se bacru jufra 08:31 < Amnadya> ie ie 08:32 < Amnadya> ___'ç / ç / an-naç -> ue 08:32 < Ilmen> je'e 08:33 < Amnadya> Ikala'ç ku -> ue do [zukte lo ka] klama 08:33 < Ilmen> cizra fa lo nu lo xromana .itku'ile cu selpau zo'oi dh .i paunai ki'u ma na co'e zo'oi ð 08:33 < Amnadya> an-naç ikal ku -> ue do [zukte lo ka] klama 08:33 < Amnadya> .i dunli 08:33 < Amnadya> .i na djuno 08:38 < Amnadya> uebryspoja 08:38 < Ilmen> xu lo galxe ganlo zunsna cu na ka'e linlami si li'erla'i lo moklu ganlo zunsna soi jalge fa lonu zo'oi p’ na ka'e .ambigu 08:40 < Amnadya> xu lo moklu ganlo zunsna cu nazbi zunsna 08:40 < Amnadya> .i ia lo tai girzu na zasti 08:40 < Ilmen> je'e 08:40 < Ilmen> .i mi de'a jundi 08:48 < cliva> Amnadya: zbagazasni.ax.lt 08:48 < cliva> (to lo laldo me DNS ri'a lo na jai se djuno cu jai nabmi toi) 08:48 < Amnadya> ua ki'e 08:58 < fushgrhfhxdz> y 09:02 < fushgrhfhxdz> xu mi jorne 09:02 < Ilmen> coi 09:03 < cliva> jorne 09:03 < cliva> .i xu zu'i za'o na .akti 09:03 < fushgrhfhxdz> ma me zuhi 09:04 < cliva> la parti servere 09:04 * mumble jorne co'a fa la'o fa Ilmen fa 09:04 < fushgrhfhxdz> oi 09:05 < Ilmen> za'a mi ka'e nerkla 09:05 < fushgrhfhxdz> i la irsis na zvati 09:05 < cliva> ba'a ka'e kompili 09:06 * mumble jorne co'a fa la'o fa gunro fa 09:07 < fushgrhfhxdz> lo du be lo du coha jorne 09:17 < sorpaas> A small question: how do I express "every day" and "every week" in Lojban? 09:19 < cliva> common are: ca ro djedi/jeftu; pa/su'o roi ro djedi/jeftu 09:20 < cliva> though these aren't exactly correct because of overlapping 09:22 < cliva> detydei and detyjeftu maybe 09:27 < sorpaas> Do you mind explaining what means "overlapping" in this case? {ca ro djedi} seems to mean "during all days"? What if I want to express "once a day"? 09:30 < cliva> pa roi ro (lo) detydei, probably 09:30 < cliva> A djedi is just a 24-hour period, not necessarily from 0:00 to 24:00 09:32 < sorpaas> I think I get what you mean. ki'e! 09:40 < Ilmen> ca ro djedi = at each day 09:41 < Ilmen> ca ro djedi pa roi (ku) = at each day one time 09:41 < Ilmen> (one time a day) 09:41 < Ilmen> @ sorpaas 09:41 < gleki> pa roi ro djedi 09:42 < Ilmen> I'm currently checking what's the role of roi's tagged sumti 09:43 < Ilmen> roi: "Quantified tense. Converts a number to an objectively quantified tense interval modifier, indicating that the event takes place n times. When tagging a sumti, the sumti specifies the interval in which the repetitions occur. " 09:43 < Ilmen> So {pa roi ro djedi} seems fine too 09:43 < Ilmen> je'e pei la'oi sorpaas 09:53 <@xalbo> Again, {ro djedi} is weird. There's a {djedi} from midnight to midnight, and another one from 0:01 to 0:01 the next day, and another one from 0:02-0:02, and so on. {ro djedi} would seem to include all of those. 09:54 <@xalbo> This very moment is a part of an infinite number of overlapping djedi. 09:55 <@xalbo> I guess if something happens *exactly* once ever 24 hours, then for any given 24 hour djedi it would happen exactly once during that period. But it still feels wrong. 09:58 < Ilmen> en: djefau 09:58 < mensi> [< djedi fasnu ≈ Full day event] lo nu mulno sisku zo'u: y no da se tolcri 09:58 < mensi> djefau[5847], deirfau[6936], djefasnu[7917], djedyfau[7977], deirfasnu[9006], ... 09:58 < Ilmen> jbo: djefau 09:58 < mensi> djefau [< djedi fasnu ≈ Djedi fasnu*] = pa da poi ctemidju fi x2 zo'u x1 se krasi da gi'e se fanmo lo poi'i ge ke'a 09:58 < mensi> ctemidju x2 gi'e balvi da gi no ctemidju cu purci ke'a gi'e balvi da |>>> srana fa zo djedi .e zo ctemidju |>>> 09:58 < mensi> Ilmen 09:59 < Ilmen> za'a I did made a word for "day / period between two consecutive midnights" 10:00 < Ilmen> {djefau} 10:02 < Ilmen> za'a ji'a mi pu lazni loka finti lo rarbau velski 10:02 < gleki> 20 more questions with a score of 10 or more 10:02 < gleki> probably because today i posted to G+ and pinned the area51 post to the top 10:11 < gleki> in altatufa i now have even more difficult tasks, namely autorestore ke'a and ce'u. i m not sure i can do that in pure peg 10:13 * nuzba @Dank_scriptor: bredi lo nu cunso ciska .i ui #lojban [http://bit.ly/1K1Qshc] 10:23 < gleki> anyone can express loeb and moeb functions in Lojban? 10:49 * mumble jorne co'u fa la'o fa Ilmen fa 10:50 * mumble jorne co'u fa la'o fa gunro fa 10:50 * mumble jorne co'a fa la'o fa gunro fa 11:09 < gleki> if loeb in programming can make sense e.g. in creating spreadsheet-like apps then in Lojban i have no clue what to aim at. 11:33 * mumble jorne co'a fa la'o fa Ilmen fa 13:09 * mumble jorne co'u fa la'o fa gunro fa 13:14 * mumble jorne co'a fa la'o fa eqil fa 13:14 * mumble jorne co'u fa la'o fa eqil fa 13:14 * mumble jorne co'a fa la'o fa Amnadya fa 13:16 < omni___> coi 13:17 < bigcentaur> cei 13:18 < omni___> cai 13:18 < bigcentaur> .i .u'i lo mi degji cu sakli 13:19 < omni___> je'e 13:24 < omni___> doi la cliva ja la akmnlrse .i ma ralju cmene .i .ei mi ma sivni tavla 13:26 < durka42> coi 13:27 < omni___> coi 13:29 < durka42> mi ja'a ru'e ru'e jorne lo kibro 13:31 < bigcentaur> .i ri'a ma 13:31 < durka42> lo nu mi zvati lo me'oi Jugendherberge 13:32 < durka42> lo .iugendxerberge 13:32 < bigcentaur> .ua 13:32 < durka42> oi valslinku'i 13:32 < omni___> -dx- ! 13:32 < durka42> lo .iurgendxerberge 13:32 < omni___> coi la durkavore .i ui do mo'u litru 13:32 < durka42> xu tolcurmi 13:32 < durka42> mi ca'o litru 13:33 < omni___> go'i .i dy. fa'u xy. voksa fa'u na'e voksa zunsna 13:33 < durka42> mi mo'u vofli 13:33 < omni___> ie 13:33 < durka42> lo .iu nai .iurgenderberge 13:34 < durka42> .iu cu'i * 13:34 < omni___> ua ba'e no zabna jbovla be ru cu zasti 13:35 < omni___> (to zoi gy. youth .gy toi) 13:35 < durka42> lo kibro te jorne cu masno jecu se ditcu lo cacra be li mo'a pimu 13:36 < bigcentaur> (sometime, can i get someone to help me understand how zabna / mabla are used?) 13:36 < durka42> lo nalmakcu .u'i 13:36 < omni___> lo cifnu ku ji'a nalmakcu 13:36 < omni___> .i da'i no'e makcu 13:37 < omni___> (to sa'e zoi gy. adolescent .gy toi) 13:37 < omni___> .i ka ca'o binxo lo ka makcu 13:41 < omni___> zo makcubi xu 13:41 < omni___> mi na xebni 13:41 < omni___> .i xu do na nelci doi la mensi 13:41 < mensi> ji'a mi mutce nelci i ie 13:41 < omni___> ie 13:43 < omni___> xu su da pante 13:43 < omni___> 3 13:43 < omni___> 2 13:43 < omni___> 1 13:43 < omni___> za'a na go'i 13:43 < omni___> zo'o 13:44 < omni___> .i ta'o mi pu zi te sidbo fi lo cnino .i lu .u'y zei sance li'u je lo simsa xoi cafne cu se cusku 13:44 < omni___> .i mi pu pensi zo .asna fa'u zo .esna fa'u zo .isna fa'u zo .osna fa'u zo .usna 13:45 < omni___> .i ku'i zo .asna zasti 13:46 < omni___> .i lo zunsna zo'u ma ni mabla fa mu'a zo sancebe noi sinxa lo by zei zunsna 13:46 < omni___> .i la'a ru'e la zipcpi cu se cinri 13:50 < omni___> http://jbovlaste.lojban.org/dict/makcubi 14:01 < cliva> spagetufa: iugynderberge 14:01 < spagetufa> (CU [iugynderberge VAU]) 14:01 < omni___> .u'i 14:02 < omni___> makcubi xotli 14:05 < omni___> lo sumstura be zo xotli cu mabla 14:05 < omni___> .i pa mai lo stuzi na vajni .i re mai ma xabju 14:06 < omni___> .i ja ma te friti 14:06 < omni___> .i koi panra ru'e zo gusta 14:11 < dutchie> en:skina 14:11 < mensi> skina = x1 is a cinema/movie/film about x2 [plot/theme/subject/activity], filmmaker x3, for audience x4. 14:11 < dutchie> both the place of viewing a film and the film itself? 14:12 < dutchie> hmm, from looking at other languages, that's probably just wrong 14:13 <@xalbo> Yeah, I suspect that "cinema" is a mnemonic, but not a good gloss. 14:13 < omni___> x1 is a film, not a place. 14:16 < dutchie> can i edit the definition? 14:16 < dutchie> i don't know the protocol around that sort of thing 14:17 < omni___> You shouldn't add officialdata 14:17 < omni___> edit* 14:19 < omni___> http://mw.lojban.org/papri/BPFK_Section:_gismu_Issues 14:19 < omni___> This is a good place to add problematic definitions 14:19 < omni___> Then the BPFK will take care of it. 14:20 < dutchie> ki'e 14:29 < dutchie> mo'u jmina 14:30 * mumble jorne co'u fa la'o fa Ilmen fa 14:31 * mumble jorne co'a fa la'o fa Ilmen fa 14:48 * mumble jorne co'u fa la'o fa Amnadya fa 14:51 * mumble jorne co'u fa la'o fa Ilmen fa 14:52 * mumble jorne co'a fa la'o fa Ilmen fa 15:03 * mumble jorne co'u fa la'o fa Ilmen fa 16:13 < zipcpi> uese'i mi mabla lo ka malpli lo cipra rafsi 16:13 < zipcpi> u'ese'i 16:21 < zipcpi> http://badconlangingideas.tumblr.com/post/123211547822/404 // va'i, finti lo jbobau, jecu jetmlugau lo du'u ri ja'a rarbau 16:27 < zipcpi> coi la bardykentauru 16:28 < bigcentaur> coi la zipcpi 16:29 < bigcentaur> .i do mo 16:32 < zipcpi> mi kanro ki'e 16:36 < bigcentaur> je'e 16:39 < bigcentaur> .i bu'u la'oi montreal cu glare mutce 16:40 < zipcpi> mi zvati la .kuualalumpur. 16:41 < bigcentaur> .ua 16:41 < bigcentaur> .i xu ji'a glare .u'i 16:42 < zipcpi> ne'a ca jacke'o li rexa 16:44 < zipcpi> k: faxrenai 16:44 < mensi> (CU [FA:fa L:xrenai] VAU) 16:44 < bigcentaur> .i do xabju je'i vitke 16:45 < zipcpi> xabju 16:45 < zipcpi> k: falrenai 16:45 < mensi> (CU [Z:falrenai VAU]) 16:46 < bigcentaur> .i .ua mi pu noroi klama la .kuualalumpur. 16:47 < zipcpi> ne'i le mejgu'e (to va'i la'oi Malaysia toi) 16:50 < bigcentaur> .i mi no'e djuno be fi la'oi Malaysia mutce 17:03 < astroendjisai> Hi. Question. Would "le famyma'o be selsa'a" be a good translation for "The Terminator Song"? 17:04 < astroendjisai> Because I have an...idea that I'd like to implement. 17:08 < zipcpi> I don't think {be} is right 17:08 < zipcpi> It could be left out altogether, or replaced with {bo} if you need it for meter 17:09 < astroendjisai> Okay. This is the title, it doesn't actually appear in the song. Thanks 17:09 < astroendjisai> Oh yes now I see what I did wrong 17:41 < astroendjisai> gloss: .i mi klama do le du'u zdani fa lo du'u mi tavla le du'u mikce mi vau kei ku ko'a goi le zdani ku fi'o cusku fe'u lu ko'a xamgu vau li'u vau kei ku vau kei ku vau 17:41 < mensi> . me come(s) you the bridi abstract nest (1:) a(n) bridi abstract me talk the bridi abstract doctor me ] end abstraction, it 1 pro sumti assign the nest, selbri to modal say end modal selbri 17:41 < mensi> < it 1 nice ] > ] end abstraction, ] end abstraction, ] 17:42 < astroendjisai> Now I'm just hoping that isn't a load of nonsense. 17:44 < astroendjisai> This isn't as good of an idea as I thought. Hmm. 17:45 < zipcpi> u'i 17:46 < astroendjisai> The idea was to get a ridiculously long string of famyma'o at the end and make that the chorus of the song 17:46 < zipcpi> Hm kinda like a grammatical old woman that swallowed the fly or something 17:47 < zipcpi> Start with a simple sentence 17:47 < zipcpi> Then gradually complicate it 17:47 < astroendjisai> Oh hey that'll work 17:49 < astroendjisai> I'm trying to end up saying "I come to you from the house about which me and my doctor spoke about saying that it was good" 17:56 < zipcpi> k: vultru'e 17:56 < mensi> (CU [Z:vultru'e VAU]) 18:02 < vultrue> huh? 18:03 < zipcpi> Just testing :p 18:03 < zipcpi> Now and then I like to quiz myself on the morphology 18:12 < phma> coi 18:16 < zipcpi> coi 18:18 < phma> I have an idea: at jbonunsla, post the letters B, D, N, and V on the walls at the appropriate positions and talk in geographic directions. 18:19 < zipcpi> Hehe 18:19 < zipcpi> Too bad I'll never attend one 18:20 < phma> I've been reading about the Yimidhirr and other peoples who talk like that. 18:20 < zipcpi> Oh, you mean as opposed to left-right etc. 18:21 < phma> yes. {manti be'a lo do jamfu} and the like. 18:21 < zipcpi> If I ever become an eccentric billionaire, all my buildings will have compass roses on the ceilings 18:22 < zipcpi> It's also interesting that in video-game worlds, characters talk about compass directions a lot more than in real life :p 18:23 < phma> Why won't you attend one? 18:23 < zipcpi> Distance 18:23 < phma> Dismoklu 18:23 < zipcpi> u'i 18:24 < phma> How far are you from the nearest Lojbanist? 18:24 < zipcpi> ra'oi -dis- cu rafsi zo dinso noi cipra gismu 18:25 < phma> cpitance 18:25 < zipcpi> mi na sanji lo drata jbopre poi meljo 18:25 < phma> bindo xu? i sralo xu? 18:26 < zipcpi> na sajni ji'a 18:26 < zipcpi> mi paroi stuvi'e le sralygu'e 18:27 < zipcpi> melbi zanfri 18:33 < mi_se_pilno> coi ro do 18:33 < phma> coi la se pilno 18:34 < mi_se_pilno> do mo 18:34 < mi_se_pilno> or coi la phama 18:34 < phma> mi vreta gi'e tumymre gi'e sampla li'o 18:37 < mi_se_pilno> .i xu do sampla 18:39 < phma> go'i 18:40 < mi_se_pilno> .i xu do nelci 'it'/ how do you say it? 18:40 < phma> in this case, "la'edi'u" 18:40 < phma> go'i 18:41 < mi_se_pilno> oh yeah! need to remember that 18:43 < mi_se_pilno> .i do ma sampla 18:43 < ldlework> https://www.youtube.com/watch?v=kI1xuE6OUuU 18:45 < phma> mi sampla zoi .gy. Bezitopo .gy lo nu tumymre 18:48 < mi_se_pilno> .i xu nu tumymre cu simlu la gugyl maps 18:50 < phma> ckini .iku'i na simlu 18:50 < zipcpi> http://mw.lojban.org/papri/brivycme 18:50 < zipcpi> la guglycartu ? 18:52 < mi_se_pilno> mo .iku'i 18:53 < mi_se_pilno> go'i la zipcpi 18:54 < zipcpi> mi finti lo vreji befi lo brivycme 18:54 < zipcpi> ki'oi la jbouitki 18:54 < phma> la bercribe cu targri 18:55 < zipcpi> ua moi'ibu'o 19:01 < mi_se_pilno> you guys lost me at ckini .iku'i. .i ro do ma tavla 19:02 < tsani> .i coi la se pilno .i do se pilno fi ma 19:05 < mi_se_pilno> .i nu mi se pilno cu pluka ro do 19:24 < bigcentaur> is there a good brivla for something like "x1 swears that x2, invoking x3"? 19:25 < noncomcinse> coi lo jundi jbopre do'u 19:25 < zipcpi> There is {nupre} but I'm not sure there is a specific word for "swear" in that sense 19:26 < noncomcinse> What sense? 19:26 < zipcpi> The sense of invoking a deity or a fate 19:26 < bigcentaur> yeah "i swear by the beard of zeus" 19:26 < bigcentaur> or "by grabthar's hammer" 19:26 < bigcentaur> etc 19:26 < zipcpi> {xusra} is also related 19:26 < bigcentaur> yeah 19:26 < bigcentaur> it's related 19:27 < bigcentaur> maybe {xusra} with a good {fi'o ...} 19:27 < zipcpi> Problem is I don't think there's a good fi'o 19:27 < noncomcinse> "I swear to God, I'll kill you." - .i mi nupre loka do mrobi'o kei la .cev. 19:27 < zipcpi> du'u ai, I think 19:28 < zipcpi> Not ka; {ka} requires a ce'u 19:28 < noncomcinse> Right. 19:28 < noncomcinse> {nu}, then. 19:28 < noncomcinse> du'u is a bridi, nupre2 is an event/state 19:29 < noncomcinse> xu malgli 19:29 < zipcpi> "aicai do mrobi'o xoi mi nupre ke'a la cevni" 19:29 < zipcpi> Dunno 19:29 < noncomcinse> The ke'a in xoi becomes a lonu, right? 19:30 < zipcpi> Basically yes 19:30 < noncomcinse> .i'e 19:30 < noncomcinse> I should do flash cards for the UI I don't know 19:32 < noncomcinse> en: a'u 19:32 < mensi> a'u = [UI1] attitudinal: interest - disinterest - repulsion. |>>> See also cinri, selcni. |>>> 19:32 < mensi> officialdata 19:33 < noncomcinse> .i mi na djuno lodu makau smuni kei la'e lo'u<< .a'a .a'e .a'i / .e'a .e'e .e'i / .i'a .i'o .i'u / .u'a .u'o / .o'a .o'e .o'i .o'u / io >>le'u 19:33 < noncomcinse> .i na djuno fi lo paxa cinmo valsi 19:35 * nuzba @bgcarlisle: Little bit of irony: {jundi} was the Lojban gismu that I always had the hardest time remembering when doing the flashcards. [http://bit.ly/1HYU57f] 19:37 < noncomcinse> I should do gismu flashcards, too. 19:50 < zipcpi> http://jbovlaste.lojban.org/dict/dimnu'e *shrugs* 19:55 < bigcentaur> whoa 19:55 < bigcentaur> you found it 19:55 < zipcpi> No I made it 19:55 < bigcentaur> oh 19:55 < bigcentaur> who 19:55 < bigcentaur> whoa 19:55 < bigcentaur> you made it! 19:56 < zipcpi> I think my {nupre} example is generally sufficient, but it might be useful to have a short word for "He takes an oath" 19:56 < zipcpi> Also it makes the semantic meaning of "promising to a bad fate" a bit clearer 20:01 < bigcentaur> i like it 20:04 < zipcpi> Hm think x3 is just a fate; deities should be tu'a-raised. Because you might also want to tu'a-raise a treasured possession, like the hairs on your head or something :p 20:05 < zipcpi> "mi dimnu'e lo du'u la'e di'u jetnu kei lo ka lindi ce'u" = "I swear that is the truth, or may lightning strike me down" :p 20:12 < clakre> deities should definitely always be tu'a-raised; they get quite cross if they're not. 20:12 < zipcpi> Haha 20:12 < noncomcinse> .i lo xeckule zerle'a lo mi naljmive krirmsa tricu mi 20:13 < zipcpi> I think you left out a {cu} 20:14 < noncomcinse> lo'ai xeckule sa'ai xeckule cu le'ai 20:15 < noncomcinse> .i mi ji'a xruti zerle'a baku 20:17 * nuzba @la_taun: ge'uru'e milxe cando pixra je ciska .i .y. #lojban http://t.co/8189jWlLrq [http://bit.ly/1CB6HiN] 20:18 < noncomcinse> TIL ma'oi GOI has a famyma'o 20:18 < zipcpi> lo'u ge'uru'e le'u ki'a 20:18 < noncomcinse> la'a srera 20:19 < zipcpi> u'i 20:19 < noncomcinse> .i ku'i nalvai lonu mi cilre fi zo goi 20:19 < noncomcinse> si zo ge'u 20:19 < zipcpi> ie 20:27 < noncomcinse> .i ca mo fa lo jbobau 20:30 < zipcpi> xu sanga 20:49 < noncomcinse> ja'a sanga 20:59 * nuzba @slaimsan: lo mi bindo pendo be lo xindo xendo cu ca ca'o cando #lojban [http://bit.ly/1CB9Pes] 21:08 < zipcpi> u'i 21:45 < zipcpi> {cei} might not actually be very useful except for use with {me'oi}/{mu'oi}, because it's hard to define a place structure that way 21:46 < zipcpi> I think {lo ka mi tavla do ce'u ce'u kei goi me'ei broda} may be generally more practical. 23:02 < tsani> .i .u'i lo pu selsku cu jai mukti lo nu mi co'a tinju'i la'e lu lo jbobau cu mo li'u 23:04 < zipcpi> mi puzi co'a sanji zo vandani u'isai 23:05 < gleki> so it's generally that i need to connect to this postgres db to start developing an edidint ginterface. anyone has a copy of JVS db? 23:05 < zipcpi> lu mulno caksova li'u pe'a 23:06 < zipcpi> srana fa zo'oi 糟糕 23:06 < zipcpi> lu festi kenka li'u pe'a 23:07 < zipcpi> xu lo pe'a mulno caksova cu te zbasu lo pe'a festi kenka iau u'i 23:07 < gleki> zo caksova cu morsi vau xunai 23:07 < zipcpi> .y. mi ca'a pilno .iku'i le terbri... 23:08 < zipcpi> ka'eku zo djaso'a xagmau? 23:09 < gleki> jb: sovda 23:09 < mensi> sovda = sovda — x1(entity) is an egg or pollen from organism x2(entity) 23:09 < mensi> :lo sovda — egg. 23:09 < mensi> :mi pu lebna so'o lo sovda be lo jipci — I took several hen eggs. 23:09 < mensi> :Comment: sovda is pollen, tsiju is spore or embryo-containing egg. Hen eggs used as food are usually unfertilized and 23:09 < mensi> thus are sovda. 23:09 < mensi> :Related words: ganti, gutra, mamta, patfu, rorci, lanbi, tarbi 23:09 < zipcpi> je'eru'e 23:10 < zipcpi> xu lo pe'a mulno sovda cu te zbasu lo pe'a festi kenka --- Day changed Tue Jul 14 2015 00:19 < gleki> alta: mi broda do vau noi mo 00:19 < mensi> (I [FA ZOhE] [CU {FASNU <FA (¹LO [NU {<FA mi> <CU (²broda NF²) (²FE do²) vau>} KEI] KU¹) (¹noi [{FA ZOhE} {CU <mo VAU>}] KUhO¹)> VAU}]) 00:19 < gleki> alta: mi broda noi mo 00:19 < mensi> (I [FA ZOhE] [CU {FASNU <FA (¹LO [NU {<FA mi> <CU (²broda NF²) VAU>} KEI] KU¹) (¹noi [{FA ZOhE} {CU <mo VAU>}] KUhO¹)> VAU}]) 00:20 < gleki> i also changed prenex rule 00:21 < gleki> subsentence = expr:(prenex* bridi) 00:27 < durka42> alta: i 00:27 < mensi> (i [FA ZOhE] [CU {COhE VAU}]) 00:28 < gleki> alta: 00:28 < mensi> (I [FA ZOhE] [CU {COhE VAU}]) 00:29 < gleki> i think what's missing here is the prenex 00:29 < gleki> i lo nu xo'e zo'e cu co'e do'e zo'u fa zo'e cu co'e 00:31 < durka42> alta: mi 00:31 < mensi> (I [FA mi] [CU {COhE VAU}]) 00:31 < durka42> alta: noi 00:31 < mensi> (I [FA ZOhE] [CU {ME ZOhE <noi (¹FA ZOhE¹) (¹CU [COhE VAU]¹)> KUhO}]) 00:31 < durka42> alta: noi noi 00:31 < mensi> (I [FA ZOhE] [CU {ME ZOhE <noi (¹FA ZOhE¹) (¹CU [ME ZOhE {noi <FA ZOhE> <CU (²COhE VAU²)>} KUhO]¹)> KUhO}]) 00:31 < durka42> alta: noi vau noi 00:31 < mensi> (I [FA ZOhE] [CU {ME ZOhE <noi (¹FA ZOhE¹) (¹CU [FASNU {FA <LO (²NU [{FA ZOhE} {CU <COhE vau>}] KEI²) KU> <noi (²[FA ZOhE] [CU {COhE VAU}]²) KUhO>} VAU]¹)> KUhO}]) 00:32 < durka42> cunso .u'i 00:37 < gleki> do kusru 00:40 < ComradeBecca> ma kusru ma 00:41 < gleki> ei ma'a kurji lo gentufa 00:41 < gleki> i ku'i la durka cu jai bapli lo nu lo gentufa cu dukse gunka 00:45 < gleki> durka42: xu do kakne lo ka jmina fi la nuzba fe RSS pe le uitki 00:46 < durka42> cumki 00:46 < durka42> ma judri 00:51 < gleki> http://mw.lojban.org/api.php?hidebots=1&days=7&limit=50&action=feedrecentchanges&feedformat=atom 00:53 < durka42> je'e 01:12 < gleki> LIVLABOT: a simple version of la sutysisku searching for Ithkuil rafsi made available http://mw.lojban.org/extensions/ilmentufa/i/ithkuil/index.html#sisku/banana 02:48 * nuzba @kotatsu_mi: ロジバン - Wikipedia https://ja.wikipedia.org/wiki/%E3%83%AD%E3%82%B8%E3%83%90%E3%83%B3 [http://bit.ly/1Rw36ce] 03:02 * mumble jorne co'a fa la'o fa Ilmen fa 03:05 < gleki> https://github.com/pegjs/pegjs/commit/da57118a43a904f753d44d407994cf0b36358adc 03:31 * mumble jorne co'u fa la'o fa Ilmen fa 03:36 < gleki> the master version of pegjs didnt want to work for me though 03:45 < zipcpi> exp: broda cei me'au ka mi brode do ce'u ce'u 03:45 < mensi> SyntaxError: Expected [yY] but "e" found. 03:45 < zipcpi> exp: me'au ka mi brode do ce'u ce'u 03:45 < mensi> SyntaxError: Expected [yY] but "e" found. 03:45 < gleki> me'au eats only sumti 03:45 < zipcpi> Right 03:45 < zipcpi> exp: broda cei me'au lo ka mi brode do ce'u ce'u 03:45 < mensi> (CU [broda {cei <me'au (¹lo [ka {mi <CU (²brode [do ce'u ce'u] VAU²)>} KEI] KU¹) MEhU>}] VAU) 03:46 < zipcpi> Just wondering what's the best way to define a word with a place structure 03:46 < zipcpi> Within a text 03:47 < zipcpi> Really perhaps {cei me'au lo ka} should be one word. 03:49 < zipcpi> Maybe it could be in SEI 03:49 < zipcpi> SEI broda mi brode do ce'u ce'u 03:50 < zipcpi> exp: sei broda mi brode do ce'u ce'u 03:50 < mensi> ([sei broda SEhU] [mi {CU <brode (¹do ce'u ce'u¹) VAU>}]) 03:50 < zipcpi> exp: sei broda ko'a ko'e ce'ai mi brode do ce'u ce'u 03:50 < mensi> ([sei broda SEhU] [{ko'a ko'e} ce'ai] [mi {CU <brode (¹do ce'u ce'u¹) VAU>}]) 03:51 < zipcpi> Oops forgot to replace the ce'us :p 03:57 < zipcpi> exp: sei broda tu'e brode .i brodi tu'u 03:57 < mensi> ([sei broda SEhU] [tu'e {<CU (¹brode VAU¹)> <i (¹CU [brodi VAU]¹)>} tu'u]) 04:15 < gleki> translation challenge: She sent her son for a doctor. 04:15 < zipcpi> "sent" 04:15 < zipcpi> Hm... 04:15 < zipcpi> What sense is used here... 04:16 < gleki> it's more about "for" 04:16 < zipcpi> Yeah but I assume it's not "send to a doctor"? 04:16 < gleki> en: dispatch 04:16 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 04:16 < zipcpi> send for: order or instruct (someone) to come to one; summon. 04:16 < gleki> To dispatch an agent or messenger to convey a message, or to do an errand. 04:17 * mumble jorne co'a fa la'o fa Ilmen fa 04:17 < zipcpi> I think "summon" might be the desired sense 04:17 < zipcpi> Hm {klami'e} might be too specific though 04:17 < zipcpi> There's also {klacpe} but I'm not sure that's right either 04:18 < gleki> dirce zo'o 04:21 < zipcpi> Who do I submit the request to use the bo'V set to? I want it for "x1~x5" 04:22 < gleki> hm, framenet says "for" is non-core "purpose" case 04:22 < zipcpi> I think the meaning is something like "She asks for a doctor for her son" 04:23 < zipcpi> asks/summons 04:23 < gleki> i thought she literally made him go 04:24 < zipcpi> That would be "sent her son to a doctor"? 04:24 < gleki> no, "to a doctor for what?" 04:24 < zipcpi> "send for" is idiomatic 04:25 < gleki> https://framenet2.icsi.berkeley.edu/fnReports/data/lu/lu4491.xml?mode=annotation 04:26 < gleki> dont think so. i think it's "purpose" FE with raising 04:27 < zipcpi> Yeah, "send ko'a to ko'e for ko'i (abstraction)" means one thing, but "send ko'a for ko'i (noun)" means another; I think this is one of many, many ways that English doesn't match Lojban in regularity :p 04:28 < gleki> i transalted as {te zu'e tu'a} 04:29 < zipcpi> I understand what you're trying to say; I just think the English actually means something else, due to a malrarbau idiom :p 04:31 < gleki> no idea 04:31 < zipcpi> uesai The grammarians haven't touched your "google" sentence yet 04:31 < gleki> havent found yet probably 04:31 < zipcpi> Ironically it may be your Lojban sentence that's considered informal, since I literally just rewrote the definition of {gugle} today, while Curtis' original def was downvoted. 04:32 < zipcpi> Well, yesterday actually 04:39 < gleki> en: gugle 04:39 < mensi> gugle = x1 searches for x2 using search engine x3 |>>> See sisku, guglsisku |>>> 04:39 < mensi> spheniscine 04:40 < gleki> for me it's just "x1 googles x2" 04:40 < zipcpi> Well, te gugle = la gugle is somewhat implied :p 04:41 < zipcpi> Maybe it's better to have a word for "search engine" in general though 04:41 < zipcpi> That isn't so... trademark-clashing 04:42 < zipcpi> mi gugle da la .bing. iau zo'o 04:42 < zipcpi> https://www.youtube.com/watch?v=tYVCk10AzS0 04:46 < zipcpi> ko gugle fi la .bing. 04:47 < fushgrhfhxdz> pehipei zoi zoi bing zoi jbovla 04:47 < fushgrhfhxdz> zo bin dramau 04:48 < zipcpi> xy'y mi na nelci 04:48 < mensi> ji'a mi mutce nelci i ie 04:50 < zipcpi> su'oroi nitcu lo nu ficygau zo'oi /n/ zo'oi /ŋ/ .iku'i ro re xai cu se sinxa me'o ny 04:50 < fushgrhfhxdz> ikuhi meho gy na mapti 04:52 < omni___> coi 04:53 < omni___> .i mi .omni lo ka rinsa kei lo ka .irci lo me dei moi je na me mi 04:53 < gleki> i mi'a casnu lo jufra pe la mupli 04:54 < zipcpi> na'o ku me'o ny cu to'ai ba'urtadji zo'oi /ŋ/ .ijo lo nu le go'i cu li'erla'i me'o gy ja me'o ky ja me'o xy 04:54 < omni___> lo sance be zoi zoi gy. bing .gy noi glico na se pagbu lo gy zei sance 04:54 < omni___> .i ku'i .a'u cu'i 04:55 < omni___> s/zoi zoi/zoi 04:55 < fenki> omni___ meant to say: lo sance be zoi gy. bing .gy noi glico na se pagbu lo gy zei sance 04:55 < Ilmen> za'a ca ti la cnino tankoi cu traji zvajbi la .pluton. 04:55 < Ilmen> s/za'a/ti'e 04:55 < fenki> Ilmen meant to say: ti'e ca ti la cnino tankoi cu traji zvajbi la .pluton. 04:55 < zipcpi> a'u 04:55 < omni___> .ei la fenki cu cusku lu ko'a pu skudji li'u je nai lo glico 04:56 < Ilmen> mi catlu lo vifne pixra pe la .pluton. u'e 04:56 < gleki> ei la su'o da cu cikre la fenki 04:56 < zipcpi> http://www.bbc.com/news/science-environment-33524589 04:56 < zipcpi> u'e 04:56 < gleki> i ai mi ji'a catlu 04:57 < zipcpi> ca co'a djuno tu'a le skari 04:58 < Ilmen> ĭe 04:59 < gleki> i oinai la pluton cu simsa lo bicyladru 04:59 < gleki> i kukte ba'a 04:59 < gleki> sa'ai bisyladru le'ai 04:59 < zipcpi> u'i 05:00 < omni___> doi la zipcpi .i xu do sanji tu'a lo mi sidbo pe lu .a'y zei sance li'u 05:01 < zipcpi> mi pu pensi lo du ka'e makau lo bifce se ladru 05:01 < zipcpi> doi omni .i la'a naku 05:01 < omni___> .i mu'i lo nu cafne joi lo nu clani cu stidi lo du'u zo .asna fa'u zo .esna fa'u zo .isna fa'u zo .osna fa'u zo .usna ka'e basti .i ku'i ui nai zo .asna xa'o zasti 05:01 < omni___> .i lo ka .esna cu ka .e'y zei sance 05:01 < omni___> .i plixau lo bangu casnu 05:02 < zipcpi> je'e 05:02 < omni___> .i mi pu kanpe fi li su'o fe lo nu do na to'e se cinri 05:02 < gleki> zo asna zasti i ku'i lo nu pilno enai tu'a lo sepli remna cu rinka 05:02 < gleki> exp: mi enai tu'a mi cu mo 05:02 < mensi> ([mi {e nai} {tu'a mi LUhU}] [cu {mo VAU}]) 05:04 < gleki> i ua zo vanlamigu pa da co'a se tunba 05:04 < zipcpi> ma tunba 05:04 < gleki> lo clani valsi 05:04 < zipcpi> http://jbovlaste.lojban.org/dict/mardurudra 05:05 < zipcpi> simsa lo itku'ile iau pe'i 05:05 < omni___> .i ku'i na srana 05:05 < durka42> xu la .lyndyn. ta'e se tcima lo marduradra 05:05 < zipcpi> u'i 05:06 < omni___> mi senpi 05:12 < omni___> .i mi se mansa lo se cuxna be mi 05:12 < omni___> .i pu ze'u troci lo ka jdice 05:14 < omni___> .i za'a doi la gleki do finti la .itku'ile sutysisku .i .ai mi cupra lo gredile pe lo valstani 05:15 < omni___> .i ku'i barda .i .au lo drata .itku'ilpre mi sidju 05:16 < gleki> ie i mi na ba sidju 05:16 < omni___> .u'i 05:16 < gleki> i ta'o la mupli 05:20 < gleki> lai cnino korbi ca cliva la pluton to ju'oi http://whereispluto.info/en toi 05:21 < durka42> ua .i xu lo se kacma cu tsuku lo terdi 05:22 < zipcpi> je'u // http://www.bbc.com/news/science-environment-33524589 05:22 < durka42> ki'e 05:23 < Ilmen> .u'a dai 05:24 < zipcpi> Hey durka I'm taking bo'V; all five of them 05:24 < durka42> lol what for 05:24 < zipcpi> For "x1 - x5" 05:24 < durka42> ta'o lo cecmu cu cuxna 05:25 < durka42> zipcpi: https://xkcd.com/927/ 05:25 < zipcpi> Well I think it's kinda long past due lol 05:25 < durka42> don't we have enough notations for those 05:25 < durka42> I thought we were waiting for something really important to use the bo'V series 05:25 < zipcpi> The problem is that they are all based on lerfu-anaphora 05:25 < durka42> (just playing devil's advocate) 05:26 < zipcpi> Which has the {boi} problem, and which selpa'i wants to replace (I'm kinda skeptical of an adequate replacement though; at the very least I like them for {goi}) 05:27 < gleki> for x1 - x5 we have FA bu 05:27 < durka42> heh, it dosn't have the {boi} problem with xornunspo 05:27 < zipcpi> Yes that is the set I like the best, but ... 05:28 < zipcpi> FA bu does have the boi problem :p 05:28 < zipcpi> Even though I prefer it to xy xi pa 05:28 < durka42> the {boi} thing is rarely a problem 05:28 < zipcpi> Except in definitions lol 05:28 < durka42> only when you're defining one broda in terms of another and really "forwarding" all the xN in order 05:29 < gleki> i wanted full period to be expressed via {i bu} from {.i} but ... that already has another meaning which is a nice illustration of this {bu} problem 05:29 < zipcpi> Yes... which is even more complicated with the CKTJ reassignment of {zo'u} to {.u} :p 05:31 < zipcpi> What's more I just defined {cei'e} (selma'o SEI) to make it easier to define new words (probably involving zai'e) within a text 05:31 < gleki> maybe {... bu} should be explicitly made lo sumka'i but not letterals[3~[3~[3~ 05:31 < zipcpi> So yeah, gonna need the x1-x5 things for that 05:32 < durka42> you can use {zo i bu} 05:32 < durka42> camxes: zo.ibu 05:32 < camxes> ([{<zo i> bu} BOI] VAU) 05:32 < omni___> Five cmavo just so you can talk more easily about Lojban grammar. People aren't crazy when they complain that Lojban is only used to talk about Lojban... 05:32 < zipcpi> u'i 05:32 < durka42> ie 05:33 < durka42> ta'o I fell off my xkcd translations during kibjasyselkei and I have to get back on that 05:33 < durka42> .o'anaise'i 05:35 < zipcpi> OK it doesn't like me adding "x3" as a keyword for some reason, even though it accepted "x1" and "x2" 05:35 < omni___> pe'i {bu} is not a solution for anything permanent, like the names of keys or specific characters of an alphabet. {bu} is ad-hoc, and ambiguous between referring to a letter or a thing starting with that letter. I much prefer making brivla. 05:35 < zipcpi> I think me'o refers to the letters or letter-strings 05:36 < omni___> {bu} is not compositional, so it's bad Lojban 05:36 < zipcpi> Or {li'ei} post xornunspo :p 05:36 < omni___> cmavo compounds should not have irregular meanings 05:36 < omni___> nor should brivla + cmavo 05:36 < zipcpi> The problem is how do you spell out, say, a URL? 05:36 < omni___> broda bu is super irregular 05:36 < zipcpi> Not sure brivla helps with that 05:37 < omni___> denpa bu is such an ugly term. 05:37 < b_jonas> ienai doi omni___. I think {bu} is a special case, because {bu} compounds are similar to lujvo, they're words that should go to a dictionary 05:38 < b_jonas> omni___: {bu} compounds and {zei} compounds shall both get permanent meanings, and not reused in incompatible ways 05:39 < omni___> Sure, that's an option. 05:39 < gleki> {denpa bu} is fine provided it never refers to {lo denpa} 05:39 < durka42> exactly, that's the problem ^ 05:39 < gleki> i already removed all letteral anaphora from la bangu 05:39 < durka42> then {mlatu bu} can never refer to {lo mlatu} 05:40 < durka42> only ^_^ 05:40 < gleki> similarly {a bu} is "a" as an object (e.g. letter as text) 05:41 < omni___> {.abu} can refer to Alice, and also to the letter A, or anything else. 05:41 < omni___> So can any BY. 05:41 < gleki> y mi stidi lo nu co'u co'e 05:41 < durka42> the letter is {me'o .abu} 05:41 < durka42> but yes 05:42 < ctefa`o> Hey guys. Related to my question yesterday, how many find it hard to distinguish [baw] vs [baʊ]? 05:43 < durka42> in english, would that be the sound difference between "caw" (the sound a crow makes) and "cow" (the animal)? 05:43 < zipcpi> No 05:43 < omni___> Not *that* easy, but doable. However the difference quickly disappears in fast speech, especially if another [w] follows 05:43 < ctefa`o> Not sure about english counterparts 05:43 < durka42> hmm 05:43 < durka42> if you can transliterate into english or spanish, I can tell you :) 05:44 < omni___> Distinguishing /bawa/ from /bauwa/ is hard in fast speech. 05:45 < durka42> {ba ua} / {bau ua} ? 05:45 < zipcpi> {bau ua} technically needs a pause 05:45 < durka42> that's why I wrote it without one 05:45 < durka42> I was asking if it's the same equivalence 05:45 < durka42> I should just stay out of IPA discussions 05:47 < ctefa`o> doi gleki your opinion? 05:47 < durka42> <gleki> in la alta you can use ' to avoid the problem 05:47 < durka42> zo'osai 05:49 * nuzba @slaimsan: lo tutra xatra cu sutra matra lo gutra catra #lojban [http://bit.ly/1O3yrNx] 05:50 < gleki> my opinion that {v} should be banned :D 05:50 < gleki> too close to [w] 05:51 < zipcpi> u'i 05:51 < durka42> .u'i 05:51 < ctefa`o> wu'i 05:58 < omni___> Here is the prototype of the ithkuil dictionary spreadsheet. Any complaints before I fill in the rest? 05:58 < omni___> https://docs.google.com/spreadsheets/d/1UFwAZz-PpCZ6t5jU0wCd5TR2LgskBNInlEpKiktYCoQ/edit?usp=sharing 06:00 < gleki> + most obvious example 06:01 < gleki> for roots each P/S combination should have its own explanation 06:01 < gleki> +examples too 06:01 * mumble jorne co'u fa la'o fa Ilmen fa 06:01 < gleki> roots should also link to series they belong tooo (i.e. P/S patterning) 06:01 < gleki> *belong to 06:01 * mumble jorne co'a fa la'o fa Ilmen fa 06:01 < omni___> mi xa'o gasnu 06:01 < omni___> .i ko catlu 06:03 < gleki> mi na zgana 06:03 < gleki> i do cusku zo'ei ma 06:03 < gleki> i lo mupli xu 06:03 < gleki> i xu lo me'oi root 06:03 < omni___> "roots should also link to series they belong tooo (i.e. P/S patterning) " 06:04 < gleki> mi na zgana lo me'oi root 06:06 < gleki> i ca ti ki'e 06:07 < omni___> je'e .i ku'i lo mupli zo'u ka'e jmina ca lo drata 06:07 < omni___> .i mi na .aidji 06:07 < gleki> lo mupli pe lo me'oi root zo'u ca ti i'a 06:07 < omni___> .i dukse lo ka fanza fa lo ka finti 06:08 < zipcpi> exp: tu'e broda .i brode tu'u sei brodi 06:08 < gleki> i ku'i lo mupli pe lo me'oi case zo'u e'u lo formative cu jorne lo me'oi case ji'a 06:08 < mensi> (tu'e [{CU <broda VAU>} {i <CU (¹brode VAU¹)>}] tu'u [sei brodi SEhU]) 06:08 < b_jonas> omni___: "{.abu} can refer to Alice, and also to the letter A, or anything else." -- no no no now you're mixing it up, that's just as true for {ly.}. {me'o .a bu} refers to the letter A, {.a bu} alone as a sumti is a pro-sumti with no fixed meaning and almost certainly doesn't refer to the letter A itself, 06:09 < gleki> actually it was discussed just recently 06:09 < b_jonas> and {li .a bu} is a also variable that has no fixed meanign and could refer to a different thing each time, only it's used in mekso context. 06:09 < b_jonas> omni___: but the same is true for {me'o ly.} and {ly.} respectively 06:09 < gleki> https://groups.google.com/d/topic/lojban/F_bOJaX-SuA/discussion 06:09 < gleki> .title 06:09 < fenki> [ Google Groups ] 06:10 < b_jonas> omni___: now if you say that {me'o .a bu} is a bad name for the letter "A" and we should use {a'y} instead, that's a position I can sort of understand though I don't really agree, and even then, I still think {bu} is useful for at least some things like this. 06:10 < b_jonas> because even if the letter "A" is so common or {.a bu} in particular is a bad name, {bu} has lots of other good uses 06:19 < zipcpi> I prefer {pi bu} for dot, but that might be mabla ma'i lo'e ropno :p 06:21 < b_jonas> what? {pi bu}? why not just {pi} 06:21 < b_jonas> isn't {pi} the dot? 06:22 < zipcpi> pi is not BY, it's PA 06:22 < zipcpi> Oh 06:23 < zipcpi> Within {me'o} 06:23 < zipcpi> Well that kinda depends on whether you're talking post xornunsep or not, but... *shrugs* 06:23 < b_jonas> zipcpi: yes, the grammar differs a bit 06:24 < zipcpi> Still though I feel kinda weird going {me'o gy obu obu gy ly ebu *pi* cy obu my} 06:24 < b_jonas> but if you're reading out something, {pi} could be fine for a dot, the only problem is that some people may think it's a comma 06:24 < zipcpi> Yeah 06:24 < b_jonas> which is why it may be better to use {denpa bu} if you want dot ambiguously 06:25 < zipcpi> The problem is I think of {denpa bu} as space, but, definitions are definitions... 06:25 < zipcpi> Think space is typically {kunti bu}? Don't think that's defined though 06:28 < zipcpi> But yeah I'm not fond of what xornunsep does to LI...LOhO. I invented {li'ei} just to protest it 06:28 < zipcpi> Though it has four votes now o.o 06:29 < gleki> i proposed several words for dot. but two jbopre were against 06:29 < gleki> so i changed back to {denpa bu} 06:29 < zipcpi> Was {pibu} one of them? 06:30 < gleki> no 06:30 < gleki> now try you :) 06:30 < zipcpi> Haha 06:30 < gleki> https://groups.google.com/d/topic/lojban/hEEu_l9TxJw/discussion 06:31 < zipcpi> .latex. bu zo'o 06:32 < zipcpi> lu sumti bu li'u noi krasi li'ai py xy.bu py 06:32 < zipcpi> .y. se krasi 06:33 < zipcpi> merko bu zo'o 06:37 <@xalbo> zipcpi: The problem with {pibu} is that a lot of cultures use ⟨.⟩ as a thousands separator, and ⟨,⟩ as the decimal separator. 06:38 < zipcpi> I know... that's why I said it might be mabla fi lo'e ropno :p 06:38 < zipcpi> But come on... denpa bu... x.x 06:39 < gleki> {pi bu} is fine. it's just that its spelling varies across localizations 06:55 < zipcpi> One of my little zbusufukai eccentricities is using the tilde rather than the hyphen for {bi'i} in English texts 06:55 < zipcpi> Because mabla 06:56 < zipcpi> mabla lo ka smuvrici 07:00 < gleki> anyone can provide me with the postgres db of JVS? 07:05 * cliva ku ji'a se cinri tau le sorcu 07:34 < zipcpi> Guess gleki really didn't like {gugle fi la .bing.} :p 07:36 < zipcpi> I mean, maybe a trademark is hard to argue for gimy'inda. It's already used a lot though lol 07:37 < omni___> mi se mansa zo gugylsisku 07:45 < gleki> not really. i cant be serious with this proposal 07:45 < gleki> it's an ad hoc gismu 07:45 < zipcpi> Yeah 07:45 < gleki> en: guglbon 07:46 < gleki> vlaste: guglbon 07:46 < vlaste> guglbon = Google Wave 07:46 < mensi> guglbon = Google Wave |>>> sarefo 07:46 < gleki> how many people remember? 07:47 < omni___> Google Bone 07:47 < zipcpi> Which is also why I didn't take Curtis' suggestion for the rafsi -gul- 07:47 < omni___> coi la niftagi 07:47 < zipcpi> Even though I'm pretty bad with abusing experimental rafsi :p 07:47 < niftg> coi omni jo'u me ro do 07:48 < omni___> I wonder what it would be like if selckiku was here. 07:48 < zipcpi> I still don't know what his glotu or whatever mean 07:48 < omni___> Then you could make experimental rafsi together 07:49 < omni___> selckiku would approve of mardurudra! 07:49 < zipcpi> u'i 07:50 < zipcpi> mi morji tau la guglyboxna .iku'i mi noroi pilno 07:50 < zipcpi> zo guglyzambo'a u'i 07:51 < omni___> zamboni 07:51 < gleki> la gugle boxna cu mabla 07:51 < omni___> la .gugyl. boxna cu mabla 07:55 < omni___> en:menstruate 07:56 < omni___> za'a zo gutrypilcri 07:57 < gleki> LIVLABOT: "lujvo:" command now outputs only lujvo but not cmevla 07:57 < gleki> lujvo: klama gasnu 07:57 < mensi> klagau[5847], klagasnu[7917], klamygau[7977], klamygasnu[10047] 07:57 < mensi> klagau [< klama gasnu ≈ Come(s) do] = x1 brings x2 to x3 07:58 < omni___> xamgu 07:59 < gleki> na djuno 08:00 < gleki> i mi su'oroi nitcu lo jvocmevla 08:00 < zipcpi> jvocmevo: klama gasnu 08:00 < zipcpi> da'i 08:34 < niftg> .ua la zantufa ji'a ca'o zvati 08:45 < zipcpi> zantufa: do mo 08:45 < zantufa> (do CU [mo VAU]) 08:48 < omni___> zantufa: .abu za'u re'u broda 08:48 < zantufa> ([{a bu} BOI] CU [{za'u re'u} broda] VAU) 08:48 < zipcpi> exp: me'olo'o 08:48 < mensi> SyntaxError: Expected [yY] but "e" found. 08:48 < zipcpi> Huh maybe that's what Curtis' {zai'e} was for 08:49 < zipcpi> Then again though, I don't see why that has to break 08:50 < zipcpi> Also no idea what his selma'o is for lol 08:50 < zipcpi> XOhEhOhE 08:51 < niftg> zantufa: jo'au zantufa xipire .i domo 08:51 < zantufa> ([jo'au {zantufa <xi (¹pi re¹) BOI>} DOhU] [i {do CU <mo VAU>}]) 08:52 < niftg> zantufa: jo'au zantufa xipire .i fajefedomo 08:52 < zantufa> ([jo'au {zantufa <xi (¹pi re¹) BOI>} DOhU] [i {<fa (¹je fe¹)> do} CU {mo VAU}]) 08:52 < niftg> .uo 08:52 < ldlework> coi ny 08:53 < niftg> zantufa: jo'au zantufa xipipabi .i fajefedomo 08:53 < zantufa> ([jo'au {zantufa <xi (¹pi pa bi¹) BOI>} DOhU] [i {<fa (¹je fe¹)> do} CU {mo VAU}]) 08:53 < niftg> .uonai 08:53 < niftg> coi cadgu'a 08:53 < zipcpi> e'u ko cipygau tu'a lu <jo'aunai> li'u 08:53 < niftg> .u'i zo'o dai 08:54 < zipcpi> ku'i mi ja'a jmina ri la jbovlaste 08:54 < zipcpi> http://jbovlaste.lojban.org/dict/jo'aunai 08:54 < niftg> .u'e 08:55 < niftg> jbo:jo'aunai 08:55 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 08:55 < niftg> en:jo'aunai 08:55 < mensi> jo'aunai = [COI*3] denial of dialect 08:55 < niftg> je'e .i jo'au noda 08:55 < zipcpi> u'i 08:55 < niftg> sa'e jo'au papo'oda 08:56 < zipcpi> jo'au na'e co'e 08:56 < zipcpi> ei mi cliva co'o 08:57 < niftg> co'oi 08:59 < niftg> .ai lunbe lumci vo'a gi'e surgau lo vo'a skami co'o 08:59 < gleki> zantufa: mi broda noi mo 08:59 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 08:59 < gleki> ue 09:00 < niftg> gau ko xagze'a stika 09:05 < gleki> exp: mi broda noi 09:05 < mensi> SyntaxError: Expected ['h] but "i" found. 09:13 * xalbo suspects it would be possible to make the error messages from mensi less useful, but isn't sure how. 09:15 < gleki> less or more? 09:15 < gleki> if more then i linked today to a new tracing option in peg.js 09:15 < gleki> but i couldnt run the master version of it 09:15 < clakre> less, it's a joke 09:15 < clakre> that's the jocular way in english to say "the messages aren't very useful, in fact they could hardly be less useful" 09:16 < gleki> translation challenge: "to pluck a fruit (~=to tear away)" 09:16 < gleki> vimcu? 09:16 < gleki> lebna? 09:16 < gleki> definitely not degycpu 09:17 < e`ogan> Shouldn't it depend on what author means though? 09:18 < gleki> I plucked a fruit from the tree.Eve plucked fruit from the tree of knowledge. 09:20 < gleki> alta: mi klama noi cizra noi se kanpe 09:20 < mensi> (I [FA ZOhE] [CU {FASNU <FA (¹LO [NU {<FA mi> <CU (²klama NF²) VAU>} KEI] KU¹) (¹noi [{FA ZOhE} {CU <FASNU (²FA [LO {NU <(³FA ZOhE³) (³CU [cizra NF] VAU³)> KEI} KU] [noi {<FA ZOhE> <CU (³se 09:20 < mensi> [kanpe NF]³) VAU>} KUhO]²) VAU>}] KUhO¹)> VAU}]) 09:22 < gleki> double NOI is impossible then 09:22 <@xalbo> Maybe something with {lacpu}. {lebna} is definitely useful. Depends on whether you want to emphasize it no longer being attached, or gaining it by removal, or something. 09:22 < gleki> i wonder how soi was meant to work if it operates on main level. 09:22 <@xalbo> Oh, duh: {crepu}. 09:23 < gleki> xalbo: degycpu and lacpu doesnt express the notion of tearing away. rather {snada lo ka lacpu}? 09:23 <@xalbo> {crepu} is the answer for most purposes, I think. 09:23 < gleki> ki'e 09:25 < gleki> that's definitely something backlinks in gimste missed 09:26 < gleki> there is also 09:26 < gleki> en: ampute 09:26 < mensi> ampute = x1 amputates, surgically removes x2 from x3 leaving remainder x4 |>>> See also vimcu, katna |>>> 09:26 < mensi> gleki 09:26 < gleki> but that's cheating 09:27 < omni___> xu do .ampute lo plise lo tricu 09:27 < gleki> ie i sepi'o lo dakfu 09:27 < gleki> i xu do na morji doi la fanza najnimre 09:27 < gleki> sa 09:27 < gleki> i xu do na vedli doi la fanza najnimre 09:27 < gleki> i ro ma'a pu se ampute 09:28 < gleki> exp: mi klama soi mo soi mo 09:28 < gleki> soi mo 09:28 < gleki> doi la mensi 09:28 < mensi> (mi [CU {klama <soi (¹CU [mo {soi <CU (²mo VAU²)> SEhU} VAU]¹) SEhU> VAU}]) 09:29 < gleki> omni___: xu do tugni lo cipra gerna 09:29 < omni___> gi'i 09:29 < omni___> go'i 09:30 < gleki> i ba'a zo soi damba lo sko'opu gi'enai damba lo nu lo bridi cu nenri 09:30 < gleki> vau iepei 09:30 < mensi> mi tugni i ie mi tugni 09:30 < gleki> jb: crepu 09:30 < mensi> crepu = crepu — x1(entity) harvests x2(property of x3) from x3(entity) 09:30 < mensi> :lo crepu — harvester. lo se crepu — harvested crop. lo te crepu — crop field. 09:30 < mensi> :ra crepu lo vanjba lo purdi — She is harvesting grapes from the garden. 09:30 < mensi> :Related words: lebna, vimcu, latna, critu, sombo, jmaji 09:30 < gleki> *katna 09:47 < gleki> https://github.com/guskant/gerna_cipra/blob/master/minajimpe.sh 09:59 * nuzba @graywyvern: ba'o klama le plini nomoi ka'ucu'i #pluto #lojban [http://bit.ly/1UYYElE] 09:59 < gleki> xu no moi 10:00 < gleki> i pe'i na go'i 10:00 < gleki> i la tcerera cu plini ji'a 10:00 < gleki> vau mu'a 10:00 < gleki> jb: plini 10:00 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 10:00 < gleki> je'e 10:00 < gleki> en: plini 10:00 < mensi> plini = x1 is a planet revolving around x2 with planetary characteristics x3, orbital parameters x4. |>>> See also lunra, 10:00 < mensi> mluni, terdi, solri, munje. |>>> officialdata 10:01 < gleki> i think we should add "a planet/dwarf planet" 10:12 < gleki> alta: noi 10:12 < mensi> (I [FA ZOhE] [CU {FASNU <FA (¹LO [NU {<FA ZOhE> <CU (²COhE VAU²)>} KEI] KU¹) (¹noi [{FA ZOhE} {CU <COhE VAU>}] KUhO¹)> VAU}]) 10:12 < gleki> the complexity increases 10:13 < gleki> soon a single "alta: le" will output the whole explanation of xorlo + all kurtyvla explained as a bonus. 10:28 <@xalbo> soon, "alta: le" will explain the current state and history of {le}, in automatically generated Ithkuil, and will also emit an autodidact to teach Ithkuil, no matter what languages you already know. 10:31 < gleki> hm, in pictures then 10:31 < gleki> btw can we both have {astro} and {astra}? 10:32 < b_jonas> gleki: no 10:32 < b_jonas> oh wait 10:32 < b_jonas> sure, you can have both 10:32 < b_jonas> those aren't gismu 10:32 < b_jonas> but it could be confusing of course 10:32 < gleki> berbere, berberi ... 10:35 < omni___> xalbo: .u'i 10:36 <@xalbo> berbera'a'a'a'a'a 10:39 < gleki> en: astro 10:39 < mensi> astro = x1 is a celestial body, heavenly body |>>> plini describes planets (lo planeta), dwarf planets (lo jdikyplini), 10:39 < mensi> asteroids (lo cmaplini). See also kensa, tarci, solri. |>>> gleki 10:41 < b_jonas> i fo'i jubme gi'e ka'e zu'e galfi lo ka fo'i galtu .i bu'u lo bemro so'i xu briju cu stuzi su'o fo'i 10:41 <@xalbo> astro = x1 is a rhotic space dog of breed x2. See also skubi. 10:45 < gleki> i think next i will concentrate on Tatoeba. issues with altatufa dont seem to be very interesting to me 10:48 * mumble jorne co'u fa la'o fa Ilmen fa 12:15 < gleki> xalbo: it was asked here about "once a week". in zipcip's datetime system it'd be {de'i ro li jy} with possibly {xo'e} at the end 12:17 <@xalbo> I like {di'i/di'inai lo jeftu}, but I think others strongly disagree with that for {di'i} as sumtcita. 12:17 <@xalbo> ({di'i lo jeftu} being "regularly, with intervals of one week", and {di'inai lo jeftu} being "irregularly, but roughly once a week") 12:19 < gleki> how it can solve the problem of intersecting intervals 12:21 <@xalbo> Under my theory, {di'i} as sumtcita takes as its argument a single event (or something with spacetime duration), and uses that as the interval. So, since all weeks are of the same length, it doesn't matter which week-long-thing we use as {lo jeftu}. 12:23 <@xalbo> {lo jeftu [be li pa [be'o]]} is any week-long thing, so {di'i lo jeftu [be li pa [be'o]]} is "at regular intervals, each of them being the same duration in time as {lo jeftu [be li pa [be'o]]}". 12:23 < gleki> no i mean if a 7 day interval intersects with another one at ppoin 3.5 days then even when taken together they will form regular intervals of 3.5 days 12:25 < gleki> each of them forms intervals of 7 days but the sum function can turn that into 3.5 days and so on 12:25 <@xalbo> I don't understand. My {di'i} more or less throws out its argument, or rather, it evaluates it only for its duration. 12:25 <@xalbo> Where is this sum function being added? 12:26 < gleki> two sets of 7 day intervals intersect in sum forming a 3.5 interval since the second set of 7 day intervals is shifted 3.5 days compared to the first one. 12:27 <@xalbo> But we're not intersecting them. We're only evaluation them for their length, and they have the same length. 12:28 <@xalbo> {di'i lo nu la .linkn. jmive}: at regular intervals, each of which is the same length as the amount of time that Lincoln lived. 12:28 < gleki> yes, they all have the same length but nothing is said about their absolute coordinates. we know period but not phase. 12:28 <@xalbo> Yes. That's all {di'i} does is give you a period. 12:29 < gleki> it's jeftu2 that gives you period 12:29 <@xalbo> Oh, are you trying to say something like "once per calendar week"/"once in each period from Sunday-Saturday" (or wherever your week starts/ends)? 12:32 < gleki> one {jeftu li pa}period can start on monday 00:01am, another on wednesday 12:01pm. if they are sine functions then we have a perfect regular (di'i) pattern that however has the total period of 3.5 12:33 * gleki de'a jundi 12:34 <@xalbo> That's like saying that two dogs can kill each other, so {lo gerku} means the bloody remains of two dogs. 12:39 <@xalbo> I really think we're failing to understand each other. My point is entirely that {lo jeftu be li pa} is one thing that lasts one week. {broda di'i ko'a} says "how long is {ko'a}? Well, broda happens once per that amount of time." 12:41 <@xalbo> {ko'a} could be {lo jeftu be li pa} or it could be {lo temci be lo nu do jbena bei lo nu lo remna cu zvati le lunra}, or it could be the length of a Presidential term, or anything else that has a spacetime duration. 12:41 < b_jonas> xalbo: sure. so {ze'a lo pa jeftu} means it lasts for one week long, and {ca ro jeftu} means once every week? 12:43 < b_jonas> xalbo: or {ca ro jeftu fi'o se rapli li pa} if you want to be really clear 12:43 <@xalbo> {ca ro jeftu} can be problematic, because there we're taking the weeklong things and treating them as locations in time, instead of durations. 12:43 < b_jonas> = on every week, repeated once 12:43 < b_jonas> xalbo: meh, {ca ro ko'a fi'o se rapli li pa} and you started that 12:43 <@xalbo> {fi'o se rapli li pa} is pretty much {pa roi}, I think. 12:44 < b_jonas> if you don't want to use {lo pa jeftu}, fine, whatever 12:44 < b_jonas> xalbo: yes, it's the same 12:44 < b_jonas> xalbo: I just don't like {roi} and {re'u} 12:44 < b_jonas> so I use {fi'o se rapli} and {fi'o te krefu} instead 12:45 < b_jonas> feel free to say {pa roi} if you wish 12:46 <@xalbo> My problem with {ca ro jeftu} is one of pragmatism intersecting with literalism. That is, for things I do once a week, then unless I do it at the exact same moment each week, it's not true literally true. 12:48 <@xalbo> This week I did it at 8 am Monday. The next week, at 8:30 Monday morning. So there are infinitely many week-long intervals during which I didn't do it. (There's one starting at 8:01, there's another starting at 8:01:02, there's another starting at 8:01:02.334543, etc). 12:48 <@xalbo> My pragmatist says "So? It's close enough, just deal with it." 12:48 < b_jonas> xalbo: dunno, feel free to use something else then. but do you at least agree that {ze'a lo jeftu be li pa} means the event lasts for a week long? 12:49 <@xalbo> Yes. That I'll absolutely agree with. 12:52 < b_jonas> xalbo: um, what if you try to figure out a way to say “calendar week” ({detcmenyciste jeftu} or something)? would that help so then you could say {ca ro detcmenyciste jeftu fi'o se rapli li pa}? 12:52 < b_jonas> as in, once every calendar week, which means a week long interval starting at some fixed day of the week and at a fixed time of the week 12:53 < b_jonas> eg. Monday midnight in some cultures 12:53 <@xalbo> Yeah, that might be what I'd want. 12:58 * mumble jorne co'u fa la'o fa Ilmen fa 13:19 < b_jonas> xalbo: ok, figure out something better than {detcmenyciste} then because that's horrible 13:20 <@xalbo> veldetri? 13:20 <@xalbo> A veldetri is a system by which numbers are assigned to days. Doesn't seem terribly wrong. 13:24 < b_jonas> xalbo: yes, that could work 13:25 < b_jonas> for calendar at least 13:25 < b_jonas> but I'm not sure if {veldetri jeftu} is a good way to say calendar week 13:25 < b_jonas> it might be 14:08 < Ilmen> coi 14:08 < omni___> coi 14:09 < Ilmen> Do you think that {na ka'e broda} ≠ {ka'e nai broda}? My recollection seems to imply so. 14:11 < omni___> Absolutely. 14:11 <@xalbo> My gut reaction is to simultaneously think they sound really similar, and that there's probably some subtlety I'm overlooking. 14:11 < omni___> I say this all the time 14:11 < omni___> I always tell people that "ka'e nai" is always wrong 14:12 <@xalbo> Oh, now I remember: *{ka'e nai} is a parsefail under the official grammar. 14:13 < omni___> {ka'e nai} is self-contradicting. 14:13 <@xalbo> How so? 14:14 < omni___> Like any {TAG nai}, {ka'e nai broda} expands as {lo su'u broda cu fasnu gi'e na cumki}. 14:14 < omni___> Compare with {ca nai broda} -> {lo su'u broda cu fasnu gi'e na cabna} 14:14 < Ilmen> Should I deduce that {ka'e broda} = {lo su'u broda cu fasnu gi'e cumki}? 14:15 <@xalbo> omni___: So are you saying that {ka'e} expands to include a {fasnu}‽ 14:16 < Ilmen> Or is it just that {TAG nai} expands differently to a regular TAG? 14:16 <@xalbo> (Viddy well, my little droogies: this is why unifying all your tags together can be taken too far.) 14:16 < Ilmen> s/to/from 14:16 < fenki> Ilmen meant to say: Or is it just that {TAG nai} expands differently from a regular TAG? 14:17 < Ilmen> So {TAG nai} would always be incidental to the main claim. ie pei 14:17 < mensi> ba'e mi na tugni 14:19 < omni___> Weird... of course {ka'e broda} does not assert {broda}. 14:19 < Ilmen> Whereas {TAG ja'ai} may change the main bridi's realisness, depending on the predicate underlying this tag. 14:19 < Ilmen> (If I follow correctly) 14:21 < omni___> Hm... 14:21 < omni___> {broda fi'o na se cumki} 14:21 < Ilmen> ka'e broda ---> broda xoi cumki ---> lo nu broda cu cumki || ka'enai broda ---> broda sei cumki ---> broda .i lo nu go'i cu cumki 14:22 < omni___> Clearly CAhA doesn't work like that 14:22 < Ilmen> .i cizra ku'i 14:22 < omni___> Doesn't broda xoi cumki assert broda? 14:22 < Ilmen> *broda sei na cumki 14:23 <@xalbo> I treat {ko'a ka'e broda} as {ko'a se jinzi lo ka kakne lo ka broda}, roughly. 14:23 < omni___> x1 centered tags? 14:24 < omni___> To me it's just {lo nu broda cu cumki} 14:24 <@xalbo> Not really. It's harder to paraphrase {ko'a ko'e ka'e broda}, but it should have the same relationship. The property of being able to be in a broda relationship is inherent in the tuple. 14:24 < Ilmen> I want xoi to be able to change the bridi's realisness, so that xoi can be used to describe/replace any of the current tag word 14:25 < Ilmen> So that ka'e = xoi cumki (vau) 14:25 < Ilmen> If one wants to make an incidental claim, there's always {sei}, for example 14:25 <@xalbo> Actually, for me it's more "{ko'a ko'e broda} doesn't fail a compile time type check. {ko'a} and {ko'e} are the sorts of things that can {broda} each other." 14:26 < Ilmen> .i'e nai pei 14:26 <@xalbo> la mensi ku la'a na tugni 14:26 <@xalbo> si zanru 14:26 < omni___> Ilmen: Well I understand why you might want that. I've always treated {xoi} as bridi-{je}. 14:27 < omni___> {broda xoi purci} -> broda happens and is in the past 14:28 < omni___> In your model, don't we lose the broda assertion? 14:28 < Ilmen> It would be weird to claim it's in the past of something that doesn't happen 14:28 < Ilmen> pei 14:28 < Ilmen> (in the currently considered world) 14:29 < omni___> broda xoi ke'a purci tu'a dei 14:29 < omni___> Is it really weird? {lo nu broda cu purci lo nu brode} makes sense for hypothetical events 14:30 < Ilmen> Yes, but then we're talking about an hypothetical world (even if that's not explicit) 14:30 < Ilmen> I guess Ithkuil handles better hypothetical situations marking 14:30 < omni___> But do you want to have to interpret the sentence before deciding if the bridi is asserted? Shouldn't it be algorithmic? 14:32 <@xalbo> Not really. I mean, I think the point is that the assertedness comes from the {purci} in that case. 14:33 < omni___> Don't we agree that {purci} and similar words don't automatically assert their abstractions? 14:33 < Ilmen> In the case of {lo nu broda cu purci lo nu brode} is true only if it is actually the case (in the currently considered world) that the events of brodaing are in the past of those of brodeing 14:34 < omni___> {lo nu do na jimpe cu mukti lo nu do retsku kei ko}. Asserted or not? 14:34 < omni___> xalbo's favorite. 14:34 < Ilmen> omni__: I'd say so; you assert that an event happened before another one, so it entails that they happened 14:35 < Ilmen> at least the speaker claims so 14:35 <@xalbo> xalbo almost always adds a {da'i} to that, for exactly that reason. 14:35 < omni___> I assert that one is before the other, but they might not happen 14:35 < Ilmen> Please show me an example, .e'o 14:35 < omni___> That's why {pu broda} is not {lo nu broda cu purci} but {lo nu broda cu fasnu gi'e purci} 14:37 < omni___> {purci} only talks about the order of two events, hypothetical or real. 14:37 < Ilmen> What do you mean with "they might not happen"? Happen in which world? 14:39 < Ilmen> Does {lo pavyseljirna pu pinxe bu'u lo rirxe} entails that in our perceivable physical world there was in the past an Unicorn drinking at a river? 14:39 < Ilmen> Couldn't it just be that it happens in an imaginary/hypotherical world? 14:39 < omni___> It doesn't matter. It's relative to the domain of discourse. 14:40 < Ilmen> Why does it doesn't matter for fasnu but matters for purci? 14:40 < omni___> Can you rephrase? 14:41 < Ilmen> Maybe I'm too tired for understanding that now 14:41 < Ilmen> .u'u 14:41 < Ilmen> s/doesn't/Ø 14:41 < fenki> Ilmen meant to say: Why does it Ø matter for fasnu but matters for purci? 14:42 < omni___> The point of {fasnu} is to assert that an event takes place. It is not the point of {purci} to assert that something takes place, but to specify the relative order of two events. 14:42 < Ilmen> !q 14:42 < Ilmen> si 14:42 < omni___> Whether those two events are hypothetical or actually occur(ed) is not important to {purci} 14:43 < omni___> The general rule in Lojban is that subordinate bridi are not by default asserted. 14:43 < Ilmen> doens't temporal order implies time and thus implies a world, even if hypothetical? ju'inai mi'e tatpi .u'i 14:44 < omni___> "My birth is before my marriage, that's for sure, although I might never marry" 14:44 < Ilmen> I would say that whether subordinate clause are asserted wholly depends on the predicate they're passed to 14:44 < Ilmen> Otherwise, {lo nu carvi cu fasnu} wouldn't assert "carvi" 14:45 < omni___> It takes specific brivla, as I said, like {fasnu} (or {jetnu} etc) 14:45 < omni___> They exist precisely to assert those subordinate clauses 14:47 < Ilmen> How to say {In our real world, clouds gather before rain, but in imaginary world X, rain precedes cloud gathering."? 14:48 < Ilmen> %) 14:50 < omni___> fatci fa lo nu lo dilnu cu jmaji pu lo nuncarvi .i ku'i xanri fa lo nu lo nuncarvi cu purci lo nu lo dilnu cu jmaji 14:52 < Ilmen> It doesn't keep the "X" part though (world naming). So "fatci" is definitely bound to our own real world? 14:52 < Ilmen> Well I have to meditate about all this. 14:53 < omni___> Referring to specific imaginary worlds isn't something I'm used to doing on a daily basis :) 14:53 < omni___> Except maybe for my novels 14:54 < omni___> I thought {zasti} had a fitting place, but it seems not 14:54 < Ilmen> Can be usefule though. "In the world of Harry Potter, people can fly riding brooms" 14:54 < omni___> Yes of course 14:54 < omni___> I didn't mean to say that it's rare to talk about imaginary worlds 14:55 < omni___> But it's always "the world of X", "the world where Y happens" 14:55 < omni___> ta'o do po'o nai tatpi 14:56 < Ilmen> I've long wondered whether {fasnu} doesn't lack an essential world slot. (Possibly filled by default with the real world, if deemed necessary) 14:56 < Ilmen> je'e 14:56 < Ilmen> Thank you for casting the light of a different opinion on this matter. 14:57 < Ilmen> ki'e jinvi cusku 14:57 < omni___> (mu'ei exists, but a brivla would be better) 14:57 < omni___> mi na djuno lo du'u xu kau zo munje drani 15:00 < omni___> ta'o nai lu ka'e nai li'u zo'u ki'u ma no da jungau mi lo du'u mi srera doi poi'i ke'a djuno lo du'u mi ca tavla ke'a 15:01 < omni___> .i .ai mi di'a gunka lo ka mulgau lo .itku'ile vlaste noi barda .oi .i ro .itku'ile gismu cu cupra pabi valstani .i je mi macnu ciska gi'e macnu fukpypu'i 15:01 < Ilmen> ni'o ba'a .a'u sai ca lo bavlamdei lo cnino pixra be la .pluton. cu gubni 15:01 < omni___> .a'u dai 15:01 < omni___> .i la .pluton. do cinri ma 15:01 < Ilmen> .i .oi dai di'ai gunka 15:02 < Ilmen> .i sa'u darno joi fange munje 15:02 < Ilmen> .i se manci 15:02 < omni___> do ma trajynei lo ma'a plini 15:03 < omni___> .i mi go'i la .iupiter. ze'u pu 15:03 < Ilmen> mi na mutce lo ka certu lo ka trajynei cuxna 15:03 < omni___> .u'i .i'a 15:03 < Ilmen> .i ba'a ma'a tugni fi tau la terdi 15:03 < Ilmen> :) 15:04 < omni___> do'a nai la terdi ma'a zdani 15:05 < Ilmen> ĭe ni'o .ai di'a ku la terdi cu bevri lo nu mi sipna co'o 15:05 < omni___> je'e .aidji be lo ka lo nu ce sipna cu se bevri la terdi 15:05 < omni___> .i .a'o surla ba 15:12 < Lojbeh> #jbosnu 16:18 < zipcpi> exp: goi'e si coi ko'a tu'e broda .i brode tu'u 16:18 < mensi> ([goi'e si] [coi ko'a DOhU] [tu'e {<CU (¹broda VAU¹)> <i (¹CU [brode VAU]¹)>} tu'u]) 16:50 < zipcpi> Hmm... how about this for the lerfu anaphora problem, at least with the {goi} side: 16:51 < zipcpi> {ko'ai} (selma'o LI) pro-sumti variable with number or lerfu-string 16:52 < zipcpi> And (crazy idea) this might even allow us to drop fo'V 16:52 < zipcpi> I'm not sure we'd need all five ko'V either, but meh 17:16 < zipcpi> Hmm... {ko'ai} does also suffer from the {boi} problem though, at least in {abu za'u re'u cusku} 17:16 < zipcpi> Well, it's in LI, so xornunsep "fixes" that. xornunsep is incompatible with selpamnunpav though :p 17:19 < zipcpi> Oh crap that means I can't complain about LI...LOhO anymore if I want {ko'ai} to work. But meh, I dunno what is best here anymore. Too much contention 19:29 < zipcpi> exp: zo ifle cu se smuni lo ka ge ganai me'au ko'a gi me'au ko'e gi ga me'au ko'a gi me'au ko'i 19:29 < mensi> ([zo ifle] [cu {se smuni} {lo <ka (¹CU [ge {CU <ga nai> <CU (²me'au ko'a MEhU²) VAU> gi <CU (²me'au ko'e MEhU²) VAU> VAU} gi {CU <ga (²CU [me'au ko'a MEhU] VAU²) gi (²CU [me'au ko'i MEhU] VA 19:29 < mensi> U²) VAU>} VAU]¹) KEI> KU} VAU]) 23:10 < gleki> mensi: doi xalbo {broda pa roi ro jeftu} is indeed a fix. no matter what the phase is it will still be once a week (unlike paroi lo jeftu). but for {di'i lo jeftu} im not sure. 23:10 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.xalbo.gy. di'a cusku da 23:11 < gleki> for me {ka'e} expands not to fasnu but to fanbu so {ka'enai} is fine 23:54 < zahlman> en: fanbu 23:54 < mensi> fanbu'u [< falnu bukpu ≈ Sail cloth] = b1 is canvas of material b2. 23:55 < zahlman> y 23:57 < gleki> http://mw.lojban.org/papri/ELG._Subjunctives,_imaginary_situations 23:57 < gleki> zahlman: ^ --- Day changed Wed Jul 15 2015 00:09 < zipcpi> Are there actually any lujvo with -dum-? 00:10 < zipcpi> I'm just wondering if it'd be mabla if I made {dumsku} have the same structure as {cusku}, but with x2 being a du'u 00:10 < zipcpi> I'm somewhat tired of wondering if {la .tom. cu cusku lo jetnu} is a type error 00:10 < zipcpi> And having to use {lu'e} is just ugly there 00:18 < b_jonas> zipcpi: the radical alternative I was considering is to just determine that {xusra} has the same x1 and x3 and x4 places as {cusku} 00:18 < b_jonas> (but it's probably better to make up some new word for that) 00:18 < zipcpi> Ah 00:19 < zipcpi> Hm don't think that'll break anything; adding places isn't that problematic 00:19 < zipcpi> "via expressive medium" might be though lol 00:20 < zipcpi> Or we could just coin {skuxu'a} or something 00:20 < b_jonas> zipcpi: I don't know what would be the best solution, but in any case, when you try to figure out how you want to use these words, please consider how to say not only “states/claims” but also “asks the question” and “requests/commands” 00:20 < durka42> coi .i ma cabdeispo 00:21 < zipcpi> Oh right 00:21 < zipcpi> All that illocutionary force thing again 00:21 < zipcpi> coi durka 00:21 < durka42> mi bapli lo bilma se cusku 00:21 < durka42> ill-locution zo'o 00:21 < zipcpi> lol 00:22 < gleki> there is smusku 00:23 < zipcpi> Right; still don't know how to define those things in Lojban lol 00:24 < gleki> .dict illocution 00:24 < fenki> illocution — noun: 1. (linguistics) The aim of a speaker in making an utterance as opposed to the meaning of the terms used 00:24 < zipcpi> Something with te zukte? 00:25 < durka42> jbo:ca'e 00:25 < mensi> ca'e = [UI2] tcita le se cusku le du'u la'e ri jetnu seja'e le nu ri se cusku 00:25 < durka42> that wasn't helpful .u'i 00:25 < durka42> jbo:ru'a 00:25 < mensi> ru'a = [UI2] tcita le se cusku le du'u la'e ri se sruma 00:26 < zipcpi> Oh by the way I made a replacement for that {ju'oi}={ju'ocu'i} thing. It's {cau'i} 00:26 < gleki> Illocution refers to what in linguistics is usually termed types of speech acts, i.e., the general purpose of a statement such as whether it is an assertion, a command, a declaratory pronouncement, a question, a warning, etc. This is a category which is not generally marked within Western languages in any consistent grammatical sense, the nearest equivalent grammatical category usually being Mood. As was see 00:26 < gleki> n above in Section 5.1, Mood functions in a much narrower grammatical range than in Western languages. When the moods of Western language actually relate to types of speech acts, the equivalent function in Ithkuil is shown by the category of Illocution. 00:26 < durka42> smusku ~ sei mi do lo smuni cu jungau 00:27 < zipcpi> ... darn suddenly Ithkuil comes along. And I still haven't got around to getting that book 00:27 < durka42> ithkuil: smusku 00:27 < ithkuil> smusku: P1S1/STA-sm-IND-COH/CSL/M/DEL/NRM-AMG-IFL/UNFRAMED 00:27 < gleki> actually i did something to illocution section https://docs.google.com/document/d/1SI7G0C-MoJIZ51N8jhL_7rLd6mIHA4Waq3F7w159RRs/edit 00:28 < zipcpi> I have no idea what all that is T.T 00:28 < gleki> selmaho:xu 00:28 < mensi> .i lu xu li'u cmavo zo'oi UI6 00:28 < mensi> http://lojban.github.io/cll/13/13/ 00:28 < mensi> cmavo: xu'au 00:28 < gleki> selmaho:ui6 00:28 < mensi> cmavo: ge'e, ki'a, ko'oi, xu 00:28 < zipcpi> Hm that Essence thing reminds me of veridicality 00:29 < gleki> so ithkuil illocution is xu,ca'e,za'a,ju'a,ko'oi <-- select some of them 00:29 < gleki> also o'i, au 00:30 < zipcpi> It's funny though that maybe Curtis' {cau'o'e} might come in handy; maybe shortened to {cau'e} 00:30 < gleki> hm, then what is irrealis 00:30 < durka42> zo da'i 00:31 < b_jonas> um, stupid question 00:31 < b_jonas> en: seldaumupli 00:31 < mensi> seldaumupli [< se darlu mupli ≈ 2nd conversion argue example] = m1=d2 is a counterexample of common property(s) m2 of 00:31 < mensi> set m3, given by d1 against stand/argument d3. |>>> jongausib 00:32 < b_jonas> ^ how the heck is that lujvo composed? the parts don't seem to imply "counter" anywhere 00:32 < b_jonas> seldarlu means a theory to support, not a theory to disprove 00:32 < b_jonas> (that's the definition from jbovlaste) 00:33 < durka42> confusing 00:33 < durka42> it's a counterexample of the te darlu, meaning it's actually a positive example of the se darlu 00:33 < durka42> I think 00:34 < durka42> iepei (to doinai la mensi toi) 00:34 < mensi> ei mi tugni 00:35 < b_jonas> I'd be totally confused by that word if I read it in a text, rather than met it in a dictionary while searching for some other words 00:35 < durka42> well, that's what dictionaries are for 00:35 < durka42> but yeah, "counterexample" is probably the wrong glossword 00:35 < gleki> mupli lo dukti 00:36 < durka42> that would be a better translation of "counterexample" 00:36 < durka42> dutmupli 00:36 < durka42> what's a better gloss for seldaumupli 00:36 < durka42> .ai galfi 00:37 < gleki> i cant see any justification for {mupli lo se darlu} 00:37 < durka42> "supporting example" i'epei 00:37 < gleki> is it that someone wanted to relex english? 00:37 < durka42> zo seldaumupli na veljvo lu mupli lo se darlu li'u pe'i 00:37 < durka42> y 00:37 < durka42> zo seldaumupli na lujvo fu lu mupli lo se darlu li'u pe'i 00:38 < zipcpi> I think dukti is better 00:38 < durka42> lujvo fu mu'a lu mupli gi'e se pilno lo se darlu li'u 00:38 < zipcpi> Or natfe even 00:38 < zipcpi> Yeah maybe natfe 00:38 < durka42> nafmupli is defined by noralujv 00:38 < durka42> ok so I'll fix it and steal the gloss from seldaumupli 00:39 < zipcpi> Yay 00:39 < b_jonas> je'e 00:39 < durka42> nafmupli is glossed as "exception" though 00:39 < durka42> that's the same isn't it 00:39 < zipcpi> I think so? 00:40 < durka42> x1 is a counterexample to property x2, normally exemplified by set x3 00:40 < durka42> i'epei 00:40 < gleki> exception can be from a rule 00:40 < gleki> so nafmupli is not an exception 00:40 < durka42> counterexample of a property 00:40 < durka42> exception to a rule 00:40 < durka42> same thing, no? 00:41 < gleki> that needs {javni} 00:41 < gleki> example may not imply forming a rule. 00:42 < durka42> I don't understand the objection, but I'll leave out "exception" I guess 00:42 < gleki> e.g. wetry to form a theory, i provide examples, you provide counterexamples so we cannot form a rule. then we agree and form a rule including all those examples. suddenly in 10 years someone comes and provides a new examples. it will be counterexamples but it'll now be an exception to the rule 00:43 < gleki> so those are different perspectives. 00:43 < durka42> but even the first theory we were trying to form was a rule 00:43 < durka42> it just wasn't a good rule .u'i 00:43 < gleki> yes, but before the rule was formed there were no exceptions to it 00:43 < gleki> since the rule didnt exist 00:44 < gleki> we just listed excamples and counterexamples 00:45 < durka42> ah now I understand the intent behind {seldaumupli}, .u'i 00:45 < gleki> btw, what happened to {kalri},. did y'all finally kill it? i dont {lr} in it. 00:46 < durka42> fabu nafmupli febu boi fibu .ije fobu pilno la'edi'u lo ka darlu fi fubu 00:46 < gleki> *dont like 00:48 < durka42> mi jmina lo jbovelcki zo seldaumupli .ije canau .ei jdice co cikre lo glico 00:49 < durka42> sa'u .ai mi jmina lo klesi lo pavmeixelfanva 00:51 < zipcpi> (CU [nu {CU <(¹ju'e gi¹) broda (¹[gi brode] [gi brodi] [gi brodo] [gi brodu]¹) GIhI> VAU} KEI] VAU) 00:51 < zipcpi> u'i doi zantufa 00:52 < b_jonas> I would guess {seldaumupli} is example of those silly lujvo mistakes that have found themselves to a dictionary, like {selxraci'a skina} which is very likely a typo for {xraselci'a skina} especially as {xraselci'a} is already in the dictionary 00:52 < durka42> http://jbovlaste.lojban.org/natlang/en/counterexample 00:52 < zipcpi> She repurposed {gi'i}; heh 00:53 < durka42> huh, she made n-ary gi? 00:53 < zipcpi> Yep 00:53 < durka42> I thought that had really bad ambiguity problems 00:53 < zipcpi> I think she fixed it by repurposing {gi'i} 00:53 < zipcpi> Hence the GIhI 00:53 < durka42> what's the latest URL for zantufa? 00:53 < durka42> ua 00:53 < durka42> zantufa: coi 00:53 < gleki> zantufa: mi prami do 00:54 < durka42> na ircysampre 00:54 < zipcpi> http://guskant.github.io/gerna_cipra/zantufa-0.2.html 00:54 < durka42> zantufa1: coi 00:54 < akmnlrse> zantufa: cipra 00:54 < zantufa1> (CU [cipra VAU]) 00:55 < gleki> zantufa1: noi 00:55 < zipcpi> Hm selpa'i-jecu works, but selpa'i-guje doesn't work 00:55 < akmnlrse> spuda tu'a zo zantufa lu'u po'o 00:55 < gleki> zantufa: noi 00:55 < zantufa1> minajimpe: .i kanpe tu'a ['h] .i ku'i "i" cu se tolcri 00:55 < akmnlrse> fa ji'a lo fadni gerna na zanru la gudje 00:56 < durka42> zipcpi: yeah it seems {gu} was never implemented 00:56 < durka42> probably an oversight 00:56 < durka42> oh, zantufa doesn't have {gaje} either 00:56 < durka42> strange 00:56 < akmnlrse> lo'u gi je le'u basti 00:56 < akmnlrse> soi cizra 00:56 < durka42> zantufa: ge broda gi ga brode gi brodi gi brodo 00:56 < zantufa1> (CU [ge broda {gi <ga brode (¹[gi brodi] [gi brodo]¹) GIhI>} GIhI] VAU) 00:56 < durka42> zantufa: ge broda gi ga brode gi brodi gi'i gi brodo 00:56 < zantufa1> (CU [ge broda {<gi (¹ga brode [gi brodi] gi'i¹)> <gi brodo>} GIhI] VAU) 00:56 < zipcpi> By the way I also found a second purpose for {bo'V}. Namely an easier way to make n-ary ka 00:57 < zipcpi> Currently {ka'ei} 00:57 < durka42> nope, can't do it 00:57 < durka42> you already stole bo'V for FAbu 00:57 < durka42> it's taken 00:57 < durka42> see the problem? xD 00:57 < zipcpi> It's both 00:57 < akmnlrse> li'ai fabu ba'e du lo me zo ce'u moi 00:57 < zipcpi> Essentially the FAbu purpose is a hidden ka'ei 00:58 < durka42> wtf is ka'ei 00:58 < durka42> that's a krtis-level definition 00:58 < zipcpi> It's {ka}, but with bo'V instead of ce'u 00:58 < durka42> damn 00:58 < durka42> why didn't you write anything related to that 00:58 < zipcpi> I did 00:58 < durka42> abstractor: predicate relationship abstractor; x1 is the relationship expressed by [bridi], binds to bo'a-set 00:58 < durka42> please 00:59 < zipcpi> Notes 00:59 < zipcpi> I put a lot of things in notes x.x 00:59 < durka42> the notes are acceptable, but the definition should also be not just a string of random words 00:59 < zipcpi> u'i 00:59 < durka42> en:ka 00:59 < mensi> ka = [NU] abstractor: property/quality abstractor (-ness); x1 is quality/property exhibited by [bridi]. 00:59 < durka42> well that one is no good either 01:00 < gleki> btw, i removed {i broda cu [me zo'e] noi mo} from altatufa since it conflicted with {i broda noi ke'a mo} 01:00 < durka42> I mean your definition would be better as "ka'ei: [NU] see notes" 01:00 < zipcpi> u'i 01:00 < zipcpi> I dunno; is the {ka} one any better? 01:01 < zipcpi> And that doesn't even have notes 01:01 < gleki> jb: ka 01:01 < mensi> ka = ka [clause prefix] — turn clause into infinitive 01:01 < mensi> :mi djica lo ka sipna — I want to sleep. 01:01 < mensi> :la Alis cu nitcu lo ka te dunda lo birje — Alice needs to be given beer. 01:01 < mensi> :ta'i ku do ba cirko lo ka sinma ce'u — This way you will lose the respect. 01:01 < mensi> :lo pa tadji be lo ka se pendo cu ka pendo — The only way to have a friend is to be a friend. 01:01 < mensi> :Comment: x1 (property) is the property of the clause that follows. Has at least one ce'u inside referring to a noun of 01:01 < mensi> the outer clause. ka without a ce'u inside is the same as ka ce'u. Used to express infinitives. 01:01 < durka42> no, it's not any better 01:01 < durka42> but that's original ma'oste 01:01 < zipcpi> lol 01:01 < durka42> which is like the benchmark for terrible definitions 01:01 < durka42> so... anything we write should be better 01:01 < zipcpi> OK I gotta go though 01:01 < akmnlrse> .ei lo za'o lifri be lo purci be CLL cu salci 01:01 < zipcpi> See ya 01:01 < zipcpi> I'll see about fixing it when I get back 01:02 < durka42> I'll try to think of a better one 01:02 < durka42> co'o uai 01:02 < akmnlrse> uai bi ka zai setso 01:04 < akmnlrse> di'e zo'u .ei fa ji'a na'e la zantufa cu se pagbu lo simsa 01:04 < akmnlrse> zantufa: gi ba bo broda gi brode 01:04 < zantufa1> (CU [{gi ba bo} broda {gi brode} GIhI] VAU) 01:05 < akmnlrse> .i li'a selsmudu'i lo'u ba gi le'u .i ku'i da'i curmi lo simsa be di'e 01:05 < akmnlrse> zantufa: ba gi ba bo broda gi brode 01:05 < zantufa1> (CU [ba {gi ba bo} broda {gi brode} GIhI] VAU) 01:06 < akmnlrse> zantufa: ba gi ba bo broda gi brode gi'i je mo 01:06 < zantufa1> (CU [ba {<gi ba bo> broda <gi brode> gi'i} {je mo}] VAU) 01:06 < gleki> ei lebna so'i da la zantufa 01:06 < akmnlrse> yyy ue 01:06 < akmnlrse> .i cupra lo ba'e .gu'ek. 01:06 < akmnlrse> zantufa: ba gi ba bo cu broda gi brode gi'i je mo 01:06 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 01:06 < akmnlrse> zantufa: ba gi ba bo broda gi brode gi'i je cu mo 01:06 < zantufa1> (CU [{ba <gi ba bo> broda <gi brode> gi'i} VAU] [je cu {mo VAU} VAU]) 01:06 < akmnlrse> zantufa: ba gi ba bo broda gi brode zo'e gi'i je cu mo 01:06 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 01:07 < akmnlrse> zantufa: ba ku gi ba bo broda gi brode zo'e gi'i je cu mo 01:07 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 01:07 < akmnlrse> zantufa: gi je broda gi brode zo'e gi'i je cu mo 01:07 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 01:08 < akmnlrse> zantufa: gi je broda gi brode zo'e gi'i 01:08 < zantufa1> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 01:08 < akmnlrse> zantufa: ge broda gi brode zo'e gi'i 01:08 < zantufa1> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 01:08 < akmnlrse> xu spofu 01:11 < gleki> spagetufa: lo mi je do mo cu mo 01:11 < spagetufa> ([FA {lo <NE (¹mi je do¹)> mo KU}] [cu {mo VAU}]) 01:12 < akmnlrse> di'u zo'u mi za'o na cikre 01:13 < gleki> mo co'u morji lo du'u makau nabmi ibo u'i 01:13 < akmnlrse> spagetufa: lo mi je ne do mo cu mo 01:13 < spagetufa> SyntaxError: Expected [,] but "m" found. 01:14 < gleki> spagetufa: ro lo ma'a je ne do datka cu lalxu limna va 01:14 < spagetufa> SyntaxError: Expected [,], [0123456789], [lL], [mM], [nN] or [rR] but "v" found. 01:14 < gleki> ie 01:14 < gleki> alta: ro lo ma'a je ne do datka cu lalxu limna va 01:14 < mensi> (I [FA {ro BOI} {lo <(¹NE [FIhA ma'a] GEhU¹) (¹je [ne {FIhA do} GEhU]¹)> <datka NF>} KU] [cu {lalxu <limna NF>} {va ZOhE} VAU]) 01:14 < akmnlrse> ja'o .ei mi lebna 01:14 < gleki> cinri i mi spaji mi u'i 01:15 < akmnlrse> zantufa: nu mo .i mo 01:15 < zantufa1> ([CU {nu <CU (¹mo VAU¹)> KEI} VAU] [i {CU <mo VAU>}]) 01:15 < gleki> i lebna e'asai i ku'i la alta ka'e tcefrica 01:15 < akmnlrse> zantufa: nu mo .i bo mo 01:15 < zantufa1> (CU [nu {<CU (¹mo VAU¹)> <i bo (¹CU [mo VAU]¹)>} KEI] VAU) 01:15 < gleki> uasai 01:15 < gleki> i pe'i nasai drani 01:15 < akmnlrse> te zukte ku'i 01:15 < gleki> i zo i jufra ke ralju jai sidbo 01:16 < gleki> i zo i muvgau lo tcidu lo nomoi senta 01:16 < gleki> i ku'i ma'i la zantufa co'u muvgau 01:17 < akmnlrse> simsa tu'a lo .gek. lo ka na kompatibli 01:17 < durka42> uo zo seldaumupli je zo nafmupli cu se jbovelcki 01:27 < gleki> i wrote a rhymed translation of one important text 01:31 < akmnlrse> alta: mobabocumo 01:31 < mensi> SyntaxError: Expected ['h] or ['hiI] but "b" found. 01:31 < gleki> do pu kanpe ma 01:31 < akmnlrse> kanpe no da .i kucli lo du'u xu kau cumki 01:32 < gleki> zantufa: mo babocumo 01:32 < zantufa1> minajimpe: .i kanpe tu'a [0123456789] .i ku'i "b" cu se tolcri 01:32 < gleki> i ua 01:32 < gleki> i lo ciste pe la selpa'i 01:32 < gleki> i je'e 01:32 < gleki> i la jecus 01:32 < gleki> i na djuno i mi na pilno la jecus 01:33 < gleki> i ba'a frili fa lonu jmina 01:33 < akmnlrse> da'i lo'u babocu le'u curmi lo nu pilno lo'u gibabao le'u lo ralju jai se kanpe 01:33 < gleki> alta: mi babo do klama 01:33 < mensi> (I [FA {mi <ba bo do>}] [CU {klama NF} VAU]) 01:34 < akmnlrse> s/babao/babo 01:34 < fenki> akmnlrse meant to say: da'i lo'u babocu le'u curmi lo nu pilno lo'u gibabo le'u lo ralju jai se kanpe 01:34 < gleki> lo sumtcita ce'o zo bo cu terjonma'o vau aidai 01:34 < gleki> anyway here is the poem http://pastebin.com/raw.php?i=snMdD0MJ 01:35 < ctefa`o> coi 01:36 < akmnlrse> melbi 01:36 < gleki> i ku'i se kuspe nabmi 01:36 < akmnlrse> .i zo bi'a co'e li'o .i'e 01:37 < akmnlrse> coi .nictyfam. 01:37 < gleki> zo bi'a pu se finti mivau noi krinu lo nu se permite fi lo ka pilno vau zo'o 01:43 < akmnlrse> spagetufa: mobakubabobroda 01:43 < spagetufa> ([CU {mo <(¹ba ku¹) VAU>}] JU'E ba bo [CU {bròda VAU}] VAU) 01:45 < gleki> oi 01:45 < gleki> i nandu 01:46 < akmnlrse> la'oi abs_tag_term cu nitcu tu'a lo nu lutkaxede lo du'u zo bo na se lidne 01:46 < gleki> zo lutkaxede mo 01:46 < akmnlrse> me'oi lookahead sei mabla 01:47 < gleki> oi ei finti lo jbovla 01:47 < akmnlrse> bavycipra ja serlidcipa ja mo kau 01:47 < akmnlrse> s/dc/nc/ 01:47 < fenki> akmnlrse meant to say: bavycipra ja serlincipa ja mo kau 01:47 < gleki> na djuno 01:48 < gleki> i e'uzo'o zo murmurai prane dunli 01:48 < gleki> i e'uzo'o zo murmura i prane dunli 01:48 < akmnlrse> je'u 02:21 < akmnlrse> alta: lo pu ku je ne ti mo 02:21 < mensi> (I [{FA ZOhE} {CU <FASNU (¹FA [LO {NU <(²FA [lo {pu COhE} ku]²) (²CU [COhE VAU]²) (²je [CU {COhE VAU}] VAU²)> KEI} KU] [ne {FIhA ti} GEhU]¹) VAU>}] [I {FA ZOhE} {CU <mo VAU>}]) 02:23 < akmnlrse> alta: lotige'ujepukumo 02:23 < mensi> (I [FA {lo <(¹[NE {FIhA ti} ge'u] [je {NE <pu ku> GEhU}]¹) mo> KU}] [CU {COhE VAU}]) 02:29 < durka42> zo fi'a xu 02:29 < akmnlrse> je'u la'a zo do'e matmau 02:30 < gleki> li'a ma'a ka'e na tugni la altatufai ku'i ma drani se cikre stura 02:30 < akmnlrse> mi na pante .i sa'u .ei jdice lo du'u ma kau jinga 02:31 < durka42> mi jinga 02:31 < durka42> uo 02:31 < akmnlrse> le zei le 02:31 < gleki> zo do'e zo'u ei finti lo cnino javni to lo ni javni cu zenba toi 02:32 < durka42> sa'u zo fi'a jai preti .i lu lo ne ti mo li'u na jai preti 02:32 < durka42> y ja'a preti .i ku'i drata preti .u'i 02:32 < gleki> zo do'e na srana ma'oi fa 02:32 < akmnlrse> xa'o ku ji'i ci'i da noi srana lo te sumti nu facku cu javni 02:33 < akmnlrse> s/cku/cki/ 02:33 < fenki> akmnlrse meant to say: xa'o ku ji'i ci'i da noi srana lo te sumti nu facki cu javni 02:33 < durka42> zo do'e ka'e basti ma'oi fa 02:33 < gleki> je'upei 02:34 < durka42> mi tugni do'e do 02:34 < durka42> mi tugni fi'o se tugni do 02:34 < durka42> mi tugni fe do 02:34 < akmnlrse> lo se jinvi cu su'ai frica 02:34 < gleki> zo su'ai morsi ti'e 02:35 < gleki> alta: mi ne do cu mo 02:35 < mensi> (I [FA {mi <ne (¹DOhE do¹) GEhU>}] [cu {mo VAU}]) 02:35 < akmnlrse> zo su'ai se gu morsi gi za'o se nelci mi 02:35 < durka42> zo su'ai morsi .i ko ze'u jmive doi zo su'ai 02:36 < gleki> en: su'ai 02:36 < mensi> su'ai = [PA4] digit/number: precise to within the stated sigfigs (significant figures/digits); approximately, measured 02:36 < mensi> to be approximately, with some error/rounding |>>> In the su'e and su'o family. See also: su'au. |>>> 02:36 < mensi> krtisfranks 02:36 < akmnlrse> oi 02:36 < durka42> oi 02:36 < gleki> zekri 02:36 < gleki> en: su'a'ai 02:36 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 02:36 < gleki> xm 02:36 < akmnlrse> zo kai'a zo ju'oi zo zo'au zo la'ai li'o li'o li'o 02:36 < akmnlrse> .i uu doi morsi 02:38 < gleki> http://jbovlaste.lojban.org/dict/su'a'ai 02:38 < akmnlrse> xa'o ni'adzau 02:38 < akmnlrse> fa ma 02:38 < gleki> mi 02:39 < akmnlrse> ua 02:39 < gleki> i la kurtis pu lebna i seki'ubo mi finti lo cnino 02:39 < gleki> i mi pu milxe na'e nelci si si nelci be naku zo su'ai i ki'u bo lo gerna cu cnino 02:40 < akmnlrse> ka'e xusra lo dunli pe zo zu'ai 02:42 < gleki> zosu'ai vimcu lo cmima be lo te bridi i mabla simsa zo zi'o 02:42 < gleki> y i na go'i 02:43 < gleki> i zo su'ai cmavo zo se gi'e ku'i tarti tai zo zi'o i je lo nu ba'e go'i cu nabmi 02:45 < durka42> na nabmi .i ko pensi zo jai noi naru'e cmavo ma'oi se .i tarti lo dukte be zo zi'o 02:46 < durka42> su'a ma'oi se zo'u galfi lo te bridi 02:48 < gleki> oi na djuno i do'o jdice 02:48 < gleki> i sa'u zo zu'ai milxe nolmalgli ma'i mi gi'e ku'i na se cfila i seki'ubo mi pilnp 02:49 < gleki> *normalgli 02:49 < durka42> en:zu'ai 02:49 < mensi> zu'ai = [KOhA7] shows mutual activity between this place and the first place of the current bridi; members participating 02:49 < mensi> in the activity are put into the first place (that e.g. can be formed by connecting sumti with ce or 02:49 < mensi> jo'u) 02:49 < durka42> simsa lo tcepru me zo soi 02:51 < gleki> to u'i lo normalgli na bi'ai normale toi 02:52 < gleki> zo zu'ai simsa zo simxu i ku'i zo soi na go'i 02:52 < durka42> do pu cusku zo nolmagli ga'idaipei .u'i 02:52 < gleki> pu srera 02:56 < gleki> i ju'oru'e le uitki cu zenba lo ka stodi 02:59 < durka42> lo se uitki ji lo selfu cu stodi 03:00 < gleki> lo selfu 03:00 < durka42> je'e ie 03:15 < zipcpi> But yeah I kinda like how {bo'V} turned out. They're essentially variables that automatically assign themselves to {ce'u} through an implied {ce'ai} prenex whenever {ka'ei} is used. Definitions of brivla are just {ka'ei}-based. 03:15 < zipcpi> fi'i la .fancylojban. 03:17 < zipcpi> There's a reason I put them in KOhA8 with {ce'u} 03:30 < zipcpi> exp: ko'a jo'u ko'e jo'unai ko'i 03:30 < mensi> ([ko'a {jo'u ko'e} {jo'u nai} ko'i] VAU) 03:30 < zipcpi> off: ko'a jo'u ko'e jo'unai ko'i 03:30 < mensi> ([ko'a {jo'u ko'e} {jo'u nai} ko'i] VAU) 04:30 < Spheniscine_> zantufa: gaje gajanai me'au ko'a gi me'au ko'e gi'i gi gaja me'au ko'a gi me'au ko'i 04:30 < zantufa1> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 04:30 < Spheniscine_> zantufa: gaje gajanai me'au ko'a gi me'au ko'e gi gaja me'au ko'a gi me'au ko'i 04:30 < zantufa1> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 04:31 < Spheniscine_> ba'a nandu 04:31 < Spheniscine_> zantufa: gajanai me'au ko'a gi me'au ko'e 04:31 < zantufa1> minajimpe: .i kanpe tu'a [oO] .i ku'i "e" cu se tolcri 04:31 < Spheniscine_> zantufa: ganai me'au ko'a gi me'au ko'e 04:31 < zantufa1> (CU [{ga nai} {me'au ko'a MEhU} {gi <me'au ko'e MEhU>} GIhI] VAU) 04:31 < Spheniscine_> ua 04:32 < Spheniscine_> zantufa: ge ganai me'au ko'a gi me'au ko'e gi'i gi ga me'au ko'a gi me'au ko'i 04:32 < zantufa1> (CU [ge {<ga nai> <me'au ko'a MEhU> <gi (¹me'au ko'e MEhU¹)> gi'i} {gi <ga (¹me'au ko'a MEhU¹) (¹gi [me'au ko'i MEhU]¹) GIhI>} GIhI] VAU) 04:32 < Spheniscine_> zantufa: ge ganai me'au ko'a gi me'au ko'e gi ga me'au ko'a gi me'au ko'i 04:32 < zantufa1> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 04:34 < Spheniscine_> So la zantufa implemented JE cu and repurposed gi'i, but did not implement GAJE and GUJE 04:35 < Spheniscine_> coi la jenca 04:53 * nuzba @graywyvern: zo drani .enai zo draci cu drani #lojban [http://bit.ly/1Hu7TTE] 05:17 < Spheniscine_> Wait maybe zantufa took a different approach... 05:17 < Spheniscine_> zantufa: jegi broda gi brode 05:17 < zantufa1> (CU [{je gi} broda {gi brode} GIhI] VAU) 05:17 < akmnlrse> zantufa: gi je broda gi brode 05:17 < zantufa1> (CU [{gi je} broda {gi brode} GIhI] VAU) 05:18 < akmnlrse> re versiio cu se zanru 05:18 < Spheniscine_> Ah 05:18 < Spheniscine_> How does it distinguish between tanru and bridi forethought though? 05:19 < Spheniscine_> zantufa: gije cu broda gi cu brode 05:19 < zantufa1> ([gi je] [cu {broda VAU}] [gi {cu <brode VAU>}] GIhI) 05:19 < akmnlrse> na djuno .i mi pu troci lo ka gau ce gentufa tai tau lo na'e tanru fortote 05:19 < akmnlrse> gi'e fliba 05:19 < akmnlrse> sisi kei gi'e fliba 05:19 < akmnlrse> .i ue 05:20 < akmnlrse> zantufa: gi je broda gi brode zo'e gi'i je cu mo 05:20 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 05:20 < durka42> oi la zantufa cu pilno lo mipri ke terjonma'o ciste 05:20 < akmnlrse> zantufa: gi je broda gi brode zo'e gi'i 05:20 < zantufa1> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 05:21 < akmnlrse> zantufa: gi je cu broda gi brode zo'e gi'i 05:21 < zantufa1> ([gi je] [cu {broda VAU}] [gi {CU <brode (¹zo'e VAU¹)>}] gi'i) 05:21 < akmnlrse> zantufa: gi je cu broda gi brode zo'e gi'i je cu mo 05:21 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 05:21 < akmnlrse> zantufa: gi je cu broda gi brode zo'e gi'i gi'e mo 05:21 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 05:21 < Spheniscine_> moi'iru'e da ve ciksi 05:21 < durka42> learning by brute force .u'i 05:22 < Spheniscine_> http://mw.lojban.org/papri/zantufa_jonma%27o#lo_li.27erla.27i_jonma.27o 05:22 < Spheniscine_> ku'i le kibystu cu spofu ga'a mi 05:23 < durka42> za'a 05:23 < Spheniscine_> oi la jbouitki 05:23 < Spheniscine_> oi la vrici 05:23 < Spheniscine_> mu'i ma mu'i ma mu'i ma 05:28 < gleki> mo 05:28 < gleki> i le uitki na spofu 05:28 < zipcpi> ca na spofu 05:30 < gleki> http://mw.lojban.org/extensions/ilmentufa/i/en/index.html#sisku/commerce 05:30 < gleki> u'i 05:31 < zipcpi> u'i 05:31 < zipcpi> zo merce ki'a 05:33 < gleki> la'a simsa la vi valdi 05:33 < durka42> comymerce* 05:34 < gleki> under mensi's lujvo katna algoritma {commerce} is acceptable 05:34 < durka42> in other words the lujvo katna algorithm is too loose 05:36 < zipcpi> Meh, let's just go Ithkuil; everything is phonemic 05:36 < zipcpi> If you cough, you're likely insulting someone's mother 05:36 < durka42> ithkuil: lujykatny'algoritma 05:36 < ithkuil> error 05:36 < gleki> mensi simply isnt too harsh when people forget {y} 05:37 < gleki> People, I need help from someone who understands Lojban and another constructed language. There is a relay,a broken phone game. I just translated one text from Interlingue to Lojban and now someone else should translate it further to another conlang. 05:38 < zipcpi> mi na se runbau lo drata 05:38 < durka42> .ai mi bazu se spero 05:38 < gleki> https://docs.google.com/document/d/1iDee0cS55ogA2QSxaNE_nFnFSaYtBMltpK4q6R8n2qU/edit#heading=h.oeaisag0j10f 05:43 < zipcpi> zantufa: lo nu broda ibo brode 05:43 < zantufa1> ([lo {nu <(¹CU [broda VAU]¹) (¹i bo [CU {brode VAU}]¹)> KEI} KU] VAU) 05:43 < zipcpi> makfa 05:43 < gleki> mi na nelci 05:43 < mensi> e'u do vrude pajni gi'e nai ze'i co'a cinmo lo ka nelci 05:44 < zipcpi> .iku'i janli pe'a la'e lo'u <.ibo> le'u pe la gleki 05:45 < durka42> oi 05:45 < durka42> mabla mabla 05:45 < durka42> now she's gone too far 05:45 < zipcpi> ei mi finti la maltufa iau zo'o 05:46 < zipcpi> ro da me ma'oi UI 05:46 < durka42> lo lo'u zei gentufa 05:46 < zipcpi> u'i 05:46 < durka42> zantufa: lo nu broda .ije brode 05:46 < zantufa1> ([lo {nu <(¹CU [broda VAU]¹) (¹i je [CU {brode VAU}]¹)> KEI} KU] VAU) 05:47 < durka42> if {.i} doesn't start a new sentence... 05:47 < durka42> oi oi oi 05:47 < zipcpi> Hm yeah I see a problem 05:47 < durka42> it basically eliminates elidible terminators 05:47 < zipcpi> I don't really use .ibo, but .ije might be problematic 05:48 < durka42> zantufa: lo nu broda .i brode 05:48 < zantufa1> ([{lo <nu (¹CU [broda VAU]¹) KEI> KU} VAU] [i {CU <brode VAU>}]) 05:48 < durka42> so it breaks compositionality as well 05:49 < zipcpi> I kinda like the {gi'i} thing though. Even though it does break some old constructs that rely on it being binary 05:49 < durka42> ie 05:50 < gleki> what is this gi'i 05:50 < durka42> terminator for n-ary gi 05:51 < zipcpi> https://groups.google.com/forum/#!topic/lojban/NRUIud_OCj8 05:51 < gleki> i hope someone can translate this text e.g. to Esperanto or toki pona 05:52 < zipcpi> e'u do pante ki'oi ta 05:52 < gleki> en: ki'oi 05:52 < zipcpi> fi'o se kibzva 05:53 < durka42> je'e 05:53 < gleki> actually yes, zihe is now in ja/joi in altatufa 05:57 < gleki> i also hope that SI was implemented fine by her 06:15 < zipcpi> cau'a co'e do'e xo'e zo'e ge'e cau'i 06:19 < gleki> http://mw.lojban.org/papri/on_Lojbanic_Style 06:19 < gleki> lost forever 06:19 < gleki> web archive doesnt have it 06:19 < zipcpi> .y. seja'e ma 06:19 < zipcpi> ua la'oi geocities 06:20 < zipcpi> ze'e cirko 06:21 < zipcpi> oi la'oi robots.txt 06:21 < gleki> ie ie 06:21 < gleki> i la'oi Athens mo 06:22 < gleki> and that's the reason i dont like new lojban websites 06:22 < zipcpi> ju'oinaisai 06:22 < gleki> i spent too much time trying to restore such links 06:22 < gleki> when i was the only editor of lmw 06:31 < zipcpi> I really don't know what {jo'a} is for. {na'i} = "the statement is not framed properly to have a proper truth value"; {jo'a} = "it *is* framed properly"? 06:32 < gleki> yes 06:32 < zipcpi> But doesn't actually say anything about the truth value 06:32 < gleki> "your e.g. question makes perfect sense, this questions has right to be asked" 06:32 * zipcpi nods 06:32 < gleki> jb: jo'a 06:32 < zipcpi> Then it's probably used wrongly here then: https://soundcloud.com/cogas-iuk/sunday-10-38-pm 06:33 < zipcpi> Probably he means {je'u} 06:33 < mensi> jo'a = jo'a [interjection] — the clause can make sense in meaning (metalinguistic 06:33 < mensi> affirmer) 06:33 < mensi> :lu xu le prenu noi sanli ne'a lo va dinju cu ninmu li'u lu jo'a li'u — 'Is the person standing near that building a 06:33 < mensi> woman?' 'The sentence is formed correctly (there is a person standing near the building, the question of whether it's a 06:33 < mensi> woman is applicable).' 06:33 < mensi> :Comment: na'i is metalinguistic negation. 06:33 < mensi> :Related words: drani 06:34 < gleki> ah, I_clause in altatufa and camxe.js in general is ugly 06:35 < gleki> text_1 til statement_3 need to be rewritten 06:36 < zipcpi> Hm... so {jo'a .iku'i mi na djuno} = "That's a good question, but I don't know"? 06:38 < gleki> where is this quote from? 06:38 < zipcpi> I just made it up 06:39 < zipcpi> Hm is the sour grapes thing meant to be a sort of "everyone translate and compare styles" thing? 06:39 < gleki> already in La Bangu :) 06:39 < zipcpi> You added it? 06:39 < gleki> yes, but not only that. there was already North Wind and the sun 06:39 < gleki> https://docs.google.com/document/d/13rEkPCDO3EX0JH5gle8tm68SfLVFBIDhilOgwNOXRHk/edit 06:40 < zipcpi> zo cu'uxna 06:41 < zipcpi> .y. mi se cfipu 06:42 < zipcpi> bo'a sanji lo nu bo'a ka'e ckini ro bo'i fi bo'o ku gi'e ca bo zukte lo ka ckini bo'e noi me bo'i ku'o bo'o 06:42 < zipcpi> oiro'e 06:44 < zipcpi> Hm I think this is just Ilmen trying to add an x4 to cuxna, I suppose 06:44 < zipcpi> x4 being what x1 chooses to do with x2 06:45 < zipcpi> coi la .ilmen. 06:45 < zipcpi> mi casnu zo cu'uxna pe do 06:45 < Ilmen> coi ru'e 06:45 < zipcpi> <zipcpi> Hm I think this is just Ilmen trying to add an x4 to cuxna, I suppose 06:45 < zipcpi> <zipcpi> x4 being what x1 chooses to do with x2 06:46 < Ilmen> xu da pe zo cu'uxna na jai se jimpe 06:46 < zipcpi> mi pu na jimpe .iku'i pe'i co'a jimpe 06:47 < Ilmen> sa'u zo cuxna cizra mi .i na'e satci gi'e smuni to'e klina vau soi simlu 06:47 < Ilmen> jbo: cuxna 06:47 < gleki> la vrici 06:47 < gleki> i mi ca troci lo ka kompili la altatufa i ku'i co'e 06:47 < Ilmen> mi zmanei zo cu'uxna 06:48 < zipcpi> Oh by the way I think you'll like {ka'ei} 06:49 < Ilmen> Xorxes' def of {cuxna} is {x1 jdice lo du'u x2 traji lo ka se nelci ja se zanru x1 gi'a mapti kei x3}, it seems 06:49 < zipcpi> And also what I chose to assign to the coveted {bo'V} set 06:49 < Ilmen> cu'uxna is {x1 sanji lo nu x1 ka'e ckini ro me x3 fi x4 ku gi'e ca bo zukte lo ka ckini x2 noi me x3 ku'o x4} 06:50 < Ilmen> With cu'uxna, you can indicate explicitly what is the choice's inherent purpose 06:50 < Ilmen> you can't really do that with cuxna, as far as I can tell 06:50 < zipcpi> Yeah I understand now 06:50 < Ilmen> {cuxna tezu'e} just don't work 06:50 < zipcpi> cu'uxna4 is what you choose to do with the thing 06:50 < zipcpi> Mm 06:52 < durka42> ba'o pante 06:52 < Ilmen> mi cuxna lo xunre lo tanxe te zu'e lonu.... --- here the fact that I've chosen the red box (to be in some unspecified relation with it) is for purpose/motivation... 06:52 < gleki> i use property in cuxna2 06:53 < gleki> but i suggest that you complain stronger so that i can get rid of this c/x/fricative intensive gismu 06:53 < Ilmen> Like {mi cuxna lo ka xunre ku lo so tanxe}? 06:53 < zipcpi> Can I just add a // definiton to cu'uxna? 06:53 < durka42> a what definition 06:53 < Ilmen> ki'a 06:53 < durka42> a commented out definition? xD 06:54 < zipcpi> Just a "second definition" that may be less precise but doesn't jump all over the terbri 06:54 < durka42> en:cu'uxna 06:54 < mensi`i> lo nu mulno sisku zo'u: y no da se tolcri 06:54 < zipcpi> plipe pe'a le terbri ciste 06:54 < mensi> cu'uxna = x1 is aware that x1 can be in relation x4 with all of x3 (plural) and choose to be in relation x4 with x2 06:54 < mensi> (which is among x3) 06:54 < Ilmen> Why not adding a note? 06:54 < zipcpi> OK 06:54 < durka42> wat 06:54 < durka42> .u'i 06:54 < zipcpi> " $x_1$ chooses $x_2$ out of set $x_3$ to do/have relationship $x_4$ (binary ka)" 06:55 < durka42> I think I remember proposing a definition quite similar to that 06:56 < gleki> mi cuxna lo ka xunre lo ka karce 06:57 < gleki> *vau 06:57 < Ilmen> For many gismu it feels like they have irrelevant sumti slots (e.g. material slots etc), but there are a few case where it feels like there's something missing, and that's the case with cuxna 06:57 < zipcpi> OK I edited the notes 06:58 < gleki> Ilmen: can u translate a short text from Lojban to Esperanto? 06:58 < Ilmen> gleki: it then becomes difficult to say things like "Please take the pen you have chosen" 06:59 < Ilmen> gleki: Have you a particular text in mind? 06:59 < zipcpi> There's a translation relay 06:59 < gleki> Ilmen: it can be raised to {lo penbi} too 06:59 < gleki> Ilmen: https://docs.google.com/document/d/1iDee0cS55ogA2QSxaNE_nFnFSaYtBMltpK4q6R8n2qU/edit 07:00 < Ilmen> .ai catlu 07:00 < zipcpi> Crap, then we run into specificity again 07:01 < zipcpi> lo ka me la .samtin. 07:02 < Ilmen> gleki: Do we translate the original, or the last translation result, with potential information loss/alteration? 07:02 < Ilmen> If it's the latter, it would be funny to finish by translating the last translation back to the original language 07:02 < Ilmen> %) 07:02 < durka42> I think that's the idea 07:02 < durka42> like the game "telephone" 07:02 < Ilmen> je'e 07:03 < gleki> Ilmen: we translate from Lojban 07:03 < gleki> Ilmen: this broken phone game started with Ido probably 07:03 < Ilmen> Who made the Lojban translation? 07:04 < zipcpi> So should I also translate Sour Grapes, and cram in as many weird experimental cmavo as I can zo'o 07:04 < Ilmen> la slari kurtyma'o 07:04 < zipcpi> u'i 07:05 < gleki> Ilmen: me 07:05 < Ilmen> je'e 07:05 < gleki> and english translation too since interlingue is ... kinda superlost and superunderformalized language. 07:05 < gleki> to'e-U 07:06 < durka42> en:slari 07:06 < mensi> slari = x1 is sour/tart to observer x2. 07:06 < durka42> ua 07:06 < zipcpi> ni'o cei'e ikla goi'e ko'ai no'e'uno'ai tu'e xe'enai pa lorxu cu noi'e xagji 07:06 < durka42> oi .u'i 07:07 < gleki> i wanted to translate loeb function to Lojban but got lost at the initial stage. how to express "function(arg1,arg2){return arg1+arg2;} ? 07:08 < durka42> lo ka makau sumji ce'u ce'u 07:09 < gleki> no, function itself! 07:09 < gleki> fancu4? 07:09 < Ilmen> en: funksiio 07:09 < durka42> ka creates functions :) 07:09 < Ilmen> en: funtiio 07:09 < akmnlrse> gleki: mi smadi tu'a lu 07:10 < akmnlrse> lo ka lob cu ka fy ry. ce'ai ge fy. boi ry. dunli lo ka pormei ma kau gi ro da poi se pormoi fi fy. zo'u lo pormoi be da bei ry. cu ka ce'u ry. me'au lo pormoi be da bei fy 07:10 < akmnlrse> li'u noi je'u ku'i detklarativa 07:10 < Ilmen> "lo sumji be X bei Y" 07:10 < zipcpi> I think that's just me'ei sumji 07:11 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 07:11 < mensi> funtiio = x1 (plural of ordered pairs) is an exhaustive set (possibly infinite) of inputs/outputs defining a unary 07:11 < mensi> function which yields result x3 when given input x2; x1(x2) = x3 |>>> "f(x) = z" == "fy funtiio xy fi zy". "lo ro se 07:11 < mensi> zilbri be lo ka ce'u sumji ce'u li pa" is a way to extract the input/output set of the relation "f(x) = x−1", so that 07:11 < mensi> you can use it as funtiio's x1. Using a la-name for referring to a function is also an option, for example "la .fankcyn. 07:11 < mensi> cu funtiio li re li ci". See also fancu, fancuka. |>>> Ilmen 07:11 < zipcpi> Then you run the function by taking {lo poi'i ke'a sumji ko'a ko'e} 07:11 < durka42> 3 lojbanists, 4 solutions :) 07:11 < Ilmen> As for the Lojban→Esperanto translation, I'll see if I can do that tomorrow 07:12 < zipcpi> Ugh... I really need to test out the crazy fancylojban stuff I added. Maybe translate a rulebook or a program 07:12 < durka42> vlaste: cu'uxna (n) 07:12 < vlaste> cu'uxna (notes) = To summarize, "x1 chooses x2 out of set x3 to do/have relationship x4 (binary ka)". See also {cuxna} 07:12 < durka42> zipcpi: yes please that would make you at least better than krtisfranks, even if I still disapprove of wasting the cmavo space 07:15 < zipcpi> Hmm... maybe something like Euclid's algorithm for greatest common factor 07:17 < durka42> but challenge yourself to actually use the frameworks you've created, instead of creating any new cmavo along the way :) 07:17 < Ilmen> co'o 07:17 < durka42> that way you can find out if it's sufficient or not 07:18 < zipcpi> Well I'd probably need a few brivla at least 07:18 < durka42> ia nai 07:18 < zipcpi> {ifle} already exists, so that's great. Dunno about "assignment"; I'm not sure how {goi} would work 07:20 < durka42> but that's what I'm saying, if you use the translation as a way to fish for new words to invent, you won't get around to testing the words you already made :p 07:20 < latro`a> try to be careful to distingusih the *algorithm* from the *function*, if you can, when dealing with something like the euclidean algorithm 07:21 < zipcpi> But I didn't create any brivla for programs; only some assignment cmavo 07:21 < latro`a> I understand; that framework would be hard to embed 07:21 < gleki> exp: tu'e i mi mo i do mo tu'e i mo tu'u i mo tu'u 07:21 < mensi> SyntaxError: Expected [yY] but "u" found. 07:22 < latro`a> but here's the catch: "k is the greatest common factor of m and n" is already a function 07:23 < gleki> exp: tu'e mi mo i mi mo tu'u 07:23 < mensi> (tu'e [{mi <CU (¹mo VAU¹)>} {i <mi (¹CU [mo VAU]¹)>}] tu'u) 07:23 < zipcpi> The main thing is probably style 07:23 < latro`a> so you'd wind up saying something like "the greatest common factor function is equal to the following recursively defined function ..." 07:23 < zipcpi> {ty ca'egoi by}, or {ko goingau ty fi by 07:23 < zipcpi> } 07:23 < durka42> latro`a: if two functions calculate the same thing from the same inputs, are they the same function? 07:24 < latro`a> as functions, yes 07:24 < latro`a> that's extensionality 07:24 < latro`a> yet they have more structure that we care about, which is why the distinction is important 07:24 < durka42> sure 07:25 < latro`a> there is the related problem, common in graph theory: find *a* solution to this problem 07:25 < latro`a> (for instance, find *a* minimal path from A to B) 07:25 < latro`a> here you can have two algorithms both of which find members of the relation but which define different functions 07:27 < latro`a> hm, I wonder how that would work with ma kau 07:27 < latro`a> you have some predicate {broda} which is satisfied by more than one object, how do we think about {lo ka ma kau broda}? 07:28 < latro`a> does it return all of them? 07:29 < zipcpi> ca'e lo ka {gcd} fa ce'u goi ko'a ce'u goi ko'e cu ka'ei {whileloop} fa lo du'u bo'e na du li no kei lo nu ju'egi ko'a ca'egoi bo'e gi ... OK that didn't go too well 07:30 < zipcpi> Mostly the problem is that it's not like defining a Lojban word in x1~x5 style; the variables themselves are reassigned 07:30 < latro`a> if you write it in tail recursive form it should be simple 07:30 < zipcpi> Oh right 07:31 < latro`a> well, straightforward 07:31 < latro`a> I shouldn't say simple 07:31 < latro`a> it'll look complicated, but it will not actually be complicated 07:33 < durka42> heh 07:33 < latro`a> this also gives a nice way to make up words for common programming idioms 07:34 < latro`a> like f(x)=h(x) if P(x),f(g(x)) otherwise 07:34 < zipcpi> We have {ifle} 07:34 < latro`a> (which, incidentally, is the form of the euclidean algorithm) 07:34 < gleki> can anyone write this all down? 07:35 < zipcpi> gleki: I'm gonna create a page to store my translation attempt 07:35 < zipcpi> Of course now we want a brivla for "Greatest common divisor" 07:35 < zipcpi> After all we are defining *something* :p 07:35 < latro`a> nah, you can define it in terms of the programming idiom 07:35 < latro`a> :p 07:36 < zipcpi> Also the tu'e...tu'u problem that I raised in the mriste is kinda biting me in the butt 07:37 < zipcpi> Cause how else would you include a long algorithm within {ifle} 07:37 < zipcpi> Or whatever brivla for while-loop 07:37 < latro`a> x1 is the output upon application to x2 of the recursive function f, where f(x)=x3(x) if x4(x), else x5(x) 07:38 < zipcpi> Am I just supposed to assume that tu'e...tu'u works the way I want it to to make my life easier 07:38 < latro`a> now x1 is GCD, x2 is the pair of numbers in question, x3 gets the nonzero entry from a pair, x4 checks if a pair has a zero element, and x5 takes the remainder 07:39 < latro`a> *shrug* I never really understood it, frankly 07:39 < latro`a> the syntax is fine 07:39 < latro`a> but the syntax doesn't inform the semantics very well 07:39 < latro`a> unlike most of jbo 07:39 < zipcpi> What, tu'e...tu'u? 07:39 < latro`a> yeah 07:39 < latro`a> at least that's been my experience 07:40 < zipcpi> Well, join the fray: https://groups.google.com/forum/#!topic/lojban/NRUIud_OCj8 07:41 < latro`a> nah, I don't understand it enough to try 07:41 < zipcpi> u'i 07:41 < latro`a> and I've tried several times to understand it 07:41 < latro`a> my inability to understand it is part of why I like this simple-but-verbose style of using functions all over 07:42 < zipcpi> The main thing is that {ifle} takes three arguments, and is very familiar as an "if-else" clause 07:42 < zipcpi> But often you want to include more than one statement within the IF or ELSE block 07:42 < latro`a> sure, haskeller's call the equivalent if' 07:42 < latro`a> *haskellers 07:43 < latro`a> this seems like it would be helped by ju'ei 07:43 < zipcpi> Ah yes, ju'ei 07:43 < zipcpi> I actually mentioned it in that thread 07:43 < latro`a> not that I ever spelled out exactly how ju'ei works in weird scope situations 07:44 < latro`a> I can tell you how it works in NU and NOI 07:44 < latro`a> which is where I wanted it 07:44 < latro`a> but tu'e tu'u? not a chance 07:45 < zipcpi> Right, the problem I see with ju'ei is that it basically means that we have be very aware of how "deep" we're in, and use the correct number of terminators for everything. 07:45 <@xalbo> zipcpi: That's just embedding another {lo nu ifle} as one of the ifle2/ifle3 07:45 < mensi> xalbo: cu'u la'o gy.gleki.gy.: {broda pa roi ro jeftu} is indeed a fix. no matter what the phase is it will still be 07:45 < mensi> once a week (unlike paroi lo jeftu). but for {di'i lo jeftu} im not sure. | 2015-07-15T06:10:26. 07:45 < mensi> 330Z 07:45 < zipcpi> It's fine for that Bible example, because the separated bridi are very simple 07:46 < zipcpi> xalbo: Yeah but then we're forced to use {kei ju'e nu} or something 07:46 <@xalbo> Oh, I see. 07:46 < zipcpi> And *also* suffer from the same problem of having to be aware of how deep we're in 07:47 < zipcpi> Not being able to reset anything with a simple {i} or even my {i'au/iau} 07:47 <@xalbo> Functional Lojban is like Lisp, so of course you need to count parens. 07:47 < latro`a> one workaround would be to define the "chunks" and *then* their relation to one another 07:48 < latro`a> ifle fa ko'a lo nu ifle fa ko'e ko'i ko'o kei ko'u 07:48 < latro`a> is not so bad 07:48 < zipcpi> That's not the problem. The problem is if ko'i contains more than one statement 07:49 < zipcpi> Unless we first assign ko'i to something 07:49 < latro`a> that was the point 07:49 < zipcpi> Enter {goi'e} 07:49 < zipcpi> goi'e ko'i tu'e broda .i brode tu'u 07:49 < gleki> eN: goi'e 07:49 < gleki> en: goi'e 07:49 < mensi> goi'e = [COI3] assign tu'e...tu'u group to sumti |>>> goi'i assigns a bridi to a sumti; goi'e assigns an entire tu'e... 07:49 < mensi> tu'u group. May be placed within the group (usually right after tu'e), or right after tu'u. |>>> 07:49 < mensi> spheniscine 07:49 < gleki> in natlangs nesting is solved by declaring new varibles or {la'edi'u}/{go'oi} 07:50 < zipcpi> Actually I changed it so that it works both for bridi and tu'e...tu'u group 07:51 < durka42> I don't see how COI solves that problem anyway 07:51 < durka42> camxes: coi tu'e broda .i brode tu'u 07:51 < camxes> ([coi DOhU] [tu'e {<broda VAU> <i (¹brode VAU¹)>} tu'u]) 07:51 < durka42> it doesn't go "inside" 07:51 < gleki> would it be bad if i parse the sentence {mi klama na gi'a mo} as {mi klama na zo'e VAU gi'a mo}? 07:51 < zipcpi> No, it's not supposed to go inside; all it does is associate the tu'e-clause to ko'i 07:51 < durka42> um yes that's bad .u'i 07:52 < durka42> {nagi'a} is one connective 07:52 < gleki> isnt it the same as negating {mi klama} and then adding {gi'a}? 07:52 < zipcpi> .i coi ko'i tu'e broda .i brode tu'u .i brodi 07:52 < zipcpi> exp: .i coi ko'i tu'e broda .i brode tu'u .i brodi 07:52 < mensi> (i [coi ko'i DOhU] [tu'e {<CU (¹broda VAU¹)> <i (¹CU [brode VAU]¹)>} tu'u] [i {CU <brodi VAU>}]) 07:52 < durka42> well actually 07:52 < durka42> I guess it is equivalent in that case yeah 07:53 < zipcpi> Works, ga'a mi... 07:53 < latro`a> gleki: in isolation yes, in >2 tails no, afaict 07:53 < gleki> latro`a: je'e 07:53 < zipcpi> Except yes I do admit that COI is a weird choice and accepts things I don't want it to accept 07:53 < latro`a> gi'e ... na gi'a -> ~(p ^ q) v r 07:53 < latro`a> which is not (p ^ ~q) v r 07:54 < zipcpi> CUhE was my preferred choice but doesn't work properly with tu'e...tu'u. Unless we add a {zo'u}, which breaks the symmetry with {cei'e} 07:54 < latro`a> it can also eat a sumti 07:54 < latro`a> which you don't want 07:54 < zipcpi> I want it to eat a sumti. But not a selbri 07:55 < latro`a> but tu'e tu'u is not a sumti 07:55 < latro`a> hence the problem 07:55 < zipcpi> goi'e ko'i tu'e broda .i brode tu'u 07:55 < latro`a> mm 07:55 < durka42> use LAhE for that 07:56 < zipcpi> Well part of these fancylojban stuff is to avoid having to rely on vague-scope co'ephora like {la'e di'e} 07:58 < durka42> no I mean put it in ma'oi LAhE to eat sumti and only sumti 07:58 < zipcpi> exp: la'e ko'i tu'e broda .i brode tu'u 07:58 < mensi> SyntaxError: Expected [yY] but "u" found. 07:58 < zipcpi> fliba 07:59 < durka42> exp: la'e ko'i .i tu'e broda .i brode tu'u 07:59 < mensi> ([{la'e ko'i LUhU} VAU] [i {tu'e <(¹CU [broda VAU]¹) (¹i [CU {brode VAU}]¹)> tu'u}]) 07:59 < zipcpi> exp: cu'e ko'i zo'u tu'e broda .i brode tu'u 07:59 < mensi> ([{cu'e ko'i} zo'u] [tu'e {<CU (¹broda VAU¹)> <i (¹CU [brode VAU]¹)>} tu'u]) 08:00 < zipcpi> Thing is, I want it to work as both {goi'e ko'i broda}, and as {goi'e ko'i tu'e broda .i brode} 08:01 < zipcpi> And I'd much prefer not having to add {.i} or {zo'u} to those constructs to make it work, so as to match {cei'e} 08:02 < zipcpi> Cause {cei'e ilka bo'a broda} and {cei'e ilka tu'e bo'a broda .i bo'e brode} work 08:02 < zipcpi> Ugh leaving out the {tu'u}s 08:02 < zipcpi> cei'e is in SEI 08:02 < durka42> heh well they are elidible I guess 08:03 < durka42> hmm 08:03 <@xalbo> My experience with functional languages was years ago, but isn't part of the point that you pretty much don't need multiple statements in a row? I mean, you need tail recursion and calls to other functions, and that's about it. 08:03 < durka42> if old-SOI were still around... :p 08:03 < zipcpi> u'i 08:03 < zipcpi> GOIhE, same grammar as old-SOI! 08:04 < zipcpi> except no one cares about old-SOI anymore! 08:04 < durka42> xalbo: generally there is still some kind of (do ...) syntax to sequence things 08:04 < durka42> well not the same grammar 08:05 < zipcpi> off: soi ko'a broda 08:05 < mensi> ([soi ko'a SEhU] [broda VAU]) 08:05 < zipcpi> exp: soi ko'a broda 08:05 < mensi> ([soi {ko'a <CU (¹broda VAU¹)>} SEhU] VAU) 08:05 < zipcpi> off: soi ko'a tu'e broda .i brode tu'u 08:05 < mensi> ([soi ko'a SEhU] [tu'e {<broda VAU> <i (¹brode VAU¹)>} tu'u]) 08:05 < zipcpi> Hey it works 08:05 < zipcpi> off: soi ko'a mi broda 08:05 < mensi> ([soi ko'a mi SEhU] [broda VAU]) 08:06 < zipcpi> Oh 08:06 < zipcpi> Yeah that's bad 08:06 < zipcpi> One and only one 08:06 < zipcpi> Bad soi. Bad greedi soi 08:06 < zipcpi> Go home 08:07 < latro`a> xalbo: you don't NEED to, in the sense that if you have purely functional code in multiple statements, you can perform substitution to make it one statement 08:08 < latro`a> the only exception is certain recursive paradigms which can't be directly translated into a form where the function doesn't have a name for itself 08:08 < zipcpi> I've never used a functional language for programming, so I'm just working all these out as I go along 08:08 < latro`a> but in practice it is far more common to split into multiple statements 08:08 < durka42> eventually someone will write a Y combinator in lojban 08:09 < latro`a> hell, haskell's do notation is actually a way of turning something which "shouldn't" be able to split into multiple statements into multiple statements anyway 08:09 < durka42> true 08:09 < durka42> .ei mi cliva 08:09 <@xalbo> I'm trying to write GCD in funtional Lojban, and coming against the fact that I don't remember the GCD algorithm well at all. .oi 08:10 < durka42> .u'i 08:10 < zipcpi> co'o la durka 08:10 < zipcpi> xalbo: https://en.wikipedia.org/wiki/Euclidean_algorithm 08:10 < zipcpi> That's the same thing I was working off from 08:10 < latro`a> take (m,n), n<=m; if n divides m then the result is n; otherwise repeat with (n,m%n) 08:10 < zipcpi> Somewhat similar is the definition for {refkusi} 08:11 <@xalbo> Thanks. 08:11 < latro`a> so given the 5 place brivla I made up earlier, it works out provided you can define (m,n)->(n,m%n) 08:11 < zipcpi> The JVS definition I entered is basically translated off of some fancylojban that tsani wrote. His version used {ce'ai} assignments, and didn't use {ifle} 08:13 < zipcpi> {bo'V}, {ka'ei} and {cei'e} should be sufficient to simplify most uses of {ce'ai} (although part of the trade off is that it only works for one level, so {ce'ai} still has some use. 08:13 < zipcpi> ) 08:15 * nuzba @kaleidic: OH: "I'm looking forward to strong AI so I can hear masterpiece lojban rap" [http://bit.ly/1K7rsoR] 08:15 < latro`a> I take a weird sort of pride in seeing my cmavo used in complicated ways, and even replaced with more powerful cmavo 08:15 < latro`a> :) 08:17 <@xalbo> http://pastebin.com/npeLKcWa 08:17 <@xalbo> GCD in functional lojban, with no experimental cmavo at all. 08:18 < zipcpi> Cool 08:18 <@xalbo> Quite possibly some bugs in it, if I have arguments backwards. 08:18 < zipcpi> Hm 08:18 < latro`a> nice 08:18 < latro`a> (of course I think these days it would've been idiomatic to use {ce'ai}, but I see the motivation) 08:19 < latro`a> you might need xor 08:19 <@xalbo> {ga} is "if", {gi} is "else}, {giga} is "else if" 08:19 < latro`a> oh, then don't you want {ga nai} 08:19 < latro`a> ? 08:20 < latro`a> and {gi ga nai}? 08:20 <@xalbo> Oh, i guess it's not "if", really. 08:21 < latro`a> and there is a bug, you want li no in dilcu4 08:21 < latro`a> er nvm 08:21 < latro`a> missed the fo'a in that clause 08:21 < zipcpi> veldilcu: x1 = x2 % x3 08:22 < latro`a> is it the quotient at the end? 08:22 < zipcpi> Yes 08:22 < latro`a> I guess so 08:22 < latro`a> yeah, had to check 08:22 < latro`a> I tend to stop one step later 08:22 < latro`a> when one of the entries is already zero 08:22 < latro`a> err, when one of them is already 1, rather 08:22 <@xalbo> I should have done that; that's what the wikipedia article did. But I basically copied what you said in the channel just now. 08:23 < latro`a> fair point 08:24 < latro`a> I just got turned around by that clause being all dilcu 08:24 <@xalbo> I don't think I did want {ga nai}. "Either (ko'a>ko'a and fo'a is the GCD of ko'e, ko'a) OR ..." 08:24 < latro`a> rather than "if ... return n" 08:24 < latro`a> ah 08:24 < latro`a> so it's not if 08:24 < latro`a> in the way you wrote it 08:24 <@xalbo> Yeah, it actually isn't if. 08:25 < latro`a> you've basically canceled the negations that would've arisen in the "if" formulation 08:26 < zipcpi> {ifle} might be safer, except you have to wrap each clause in an abstraction, then enter {ju'ei} vs {tu'e tu'u} :p 08:26 <@xalbo> There's no return in my model, it's merely conditions under which x1 is the GCD of x2 and x3. The ugly part is that I have to refer to the name {djisidi} in my "definition" of it. 08:26 < latro`a> that's inevitable without a Y combinator 08:26 <@xalbo> (whence the y combinator, but my brain fails at that) 08:27 < latro`a> the Y combinator is actually a problem for certain things; it creates type problems that you may not want to try to deal with 08:27 <@xalbo> {ifle} is just entirely the wrong thing for this, I think. I mean, it's all about {nu} abstractions happening, instead of {du'u} being true. 08:27 <@xalbo> And, I think I just showed that for this purpose, it's not needed. :) 08:27 < zipcpi> I used {ifle} for {du'u} as well 08:27 < zipcpi> Check out my def for {refkusi} 08:27 < latro`a> the english definition of ifle is a bit off 08:28 < latro`a> I know what you intend, it's more or less determined by the arg type 08:28 < latro`a> but your english implies that the arg has to be nu 08:28 < zipcpi> Mhm 08:28 < zipcpi> I did write a Lojban definition for ifle. 100% cmavo 08:29 < zipcpi> ganai me'au bo'a gi me'au bo'e .ije ga me'au bo'a gi me'au bo'i 08:29 < zipcpi> It's almost singable lol 08:30 <@xalbo> Who took the bo'V for what now? 08:31 < zipcpi> I took bo'V for x1~x5, which also autoassigns to ce'u with {ka'ei} 08:32 < latro`a> oh jeez 08:32 < latro`a> ka and ce'ai had a baby 08:32 < zipcpi> Exactly :p 08:32 < latro`a> the cmavo are literally multiplying 08:32 < zipcpi> u'i 08:33 < latro`a> explains why pilji and pinji sound so similar 08:33 < latro`a> (lame joke) 08:33 <@xalbo> u'i 08:35 * nuzba @palkeo: @janJulo mi sama. taso tenpo ni la mi wile kama sona e toki Lojban. mi pilin e ni : toki Epelanto li sama toki Español en toki Français. [http://bit.ly/1O6xeFb] 08:38 < tsani> vlaste: ka'ei 08:38 < vlaste> ka'ei = abstractor: predicate abstractor. x1 is the predicate expressed by [bridi], using bo'a, bo'e, etc for variables. 08:38 < tsani> ah 08:39 < tsani> So many new cmavo these days. 08:39 < zipcpi> u'i 08:39 < zipcpi> I've been pretty naughty with JVS T.T 08:40 < zipcpi> Almost as infamous as Curtis lol 08:40 < tsani> Yep 08:40 < tsani> And every 2nd post on the mailing list is you .u'i 08:40 < zipcpi> u'i 08:41 < zipcpi> And like, I keep replying to my own posts. Soon it'll be discovered that all jbopre are really just me zo'osai 08:42 <@xalbo> You silly jackdaw! 08:42 < latro`a> something about a bard? 08:46 <@xalbo> Nevermind. 08:50 < zipcpi> On {ju'ei}, I have written: experimental-grammar solution #1: {ju'ei}; essentially acts like a tight-scope {.i} that doesn't close sub-clauses. {lo nu broda ju'ei brode}. Has the advantage of not requiring forethought, but has the disadvantage of not automatically closing any bridi tails; if one of your abstraction-bridi happens to contain a lot of nested NU or POI, they must be closed manually. 08:51 < zipcpi> experimental-grammar solution #2: Allow {tu'e...tu'u} to act as a single bridi in NU clauses: {lo nu tu'e broda .i brode tu'u}. Does require forethought, but gives a new "level zero" for {.i} to automatically reset to until closed by {tu'u}. 08:54 < palkeo> hi ! I'm interested in lojban, and I had a question for people who are fluent : what about the saphir-worth hypothesis ? Do you think your way of thinking changed, or that you can think differently in lojban ? 08:55 < palkeo> toki ! sina mute pi toki pona kepeken e toki Lojban li pilin e seme pi sona pi Saphir-Worth ? :) 08:55 < zipcpi> Well, for me, word-choice is the most interesting aspect; breaking down semantic concepts and metaphors and thinking about what I really want to say. Sometimes I even find occassion to add some color to a sentence that would be hard to back-translate to English. 08:56 < zipcpi> One of my favorite translations is this: http://tatoeba.org/eng/sentences/show/4225938 08:56 < palkeo> I learned toki pona, so I have already experienced that way of breaking semantic concepts, BTW :p 08:57 < zipcpi> Ah... Well Lojban does have a different way of doing it 08:58 < palkeo> yeah, it seems so :) I like the unambiguous grammar, and yeah, the emotion words seems really useful 08:58 < gleki> palkeo: some people think that Sapir Whorf hypothesis is an incorrect question. Anyway it was Loglan, not Lojban that should have tested this. But the best proof/disproof of SWH is Guugu Yimitthir language 08:59 < zipcpi> Toki Pona, from what I hear, relies on metaphor to build new concepts, while in Lojban we tend to try to be more specific as to what each word means 08:59 < palkeo> gleki: can you develop on that ? 08:59 < gleki> btw anyone knows what NAI is doing here: 08:59 < gleki> text = expr:(intro_null NAI_clause* 08:59 < gleki> ... 09:00 < gleki> palkeo: just google that language 09:00 < gleki> That language, not necessarily Lojban says a lot (or nothing) on SWH 09:01 < akmnlrse> gleki: sfosili lo nu pu lo nu pavmeigau ma'oi nai jo'u ma'oi cai cu tadji lo nu curmi lo nu spuda lo zo pei preti 09:01 < akmnlrse> s/sfosili/ve vimcu ja mo kau 09:01 < zipcpi> But one nice thing about Lojban; its vocab basically acts a lot like objects in object oriented programming. If I need a new concept? I just define one like I would new Object :p 09:01 < fenki> akmnlrse meant to say: gleki: ve vimcu ja mo kau lo nu pu lo nu pavmeigau ma'oi nai jo'u ma'oi cai cu tadji lo nu curmi lo nu spuda lo zo pei preti 09:01 < zipcpi> At least for me anyway 09:02 < gleki> akmnlrse: simsa la'oi fragment vau iepei 09:02 < mensi> ba'e mi na tugni 09:02 < gleki> exp: nai 09:02 < mensi> nai 09:02 < akmnlrse> je'u 09:02 < gleki> text = expr:(intro_null NAI_clause* text_part_2 (!text_1 joik_jek)? text_1? faho_clause EOF?) {return _node_nonempty("text", expr);} 09:03 < gleki> zo'oi joik_jek mo 09:03 < akmnlrse> ru'a si'a nu curmi lo nu spuda lo preti 09:03 < akmnlrse> exp: .i je 09:03 < mensi> (i je) 09:04 < akmnlrse> off: .i je 09:04 < mensi> (i je) 09:04 < akmnlrse> cizra 09:05 < zipcpi> It's for je'i / ji 09:05 < zipcpi> I think 09:05 < akmnlrse> gerna: .i je 09:05 < mensi> (0[i je])0 09:06 < zipcpi> {protagonista}? u'i 09:07 < gleki> i just removed both NAI and this joik_jek from alta and ... nothing happened 09:07 < zipcpi> Wait I think I hear something exploding 09:07 < zipcpi> zo'o 09:08 < gleki> new horizons? 09:08 < gleki> butterfly effect detected 09:09 < zipcpi> It's all your fault 09:09 < zipcpi> do po'o malfu'e 09:09 < palkeo> you are doing a parser ? 09:10 < palkeo> zipcpi: yeah, a little like in toki pona, but in toki pona you have to define everything, which is quite… extreme 09:10 < gleki> playing with existing ones 09:10 < gleki> toki: Lojban 09:10 < mensi> lojban = toki Losupan. en la toki pona. taso la toki pona ala 09:10 < palkeo> gleki: thanks for the name of that language, I read an article about it that was really interestand, and I didn't knew about it :) 09:11 < gleki> palkeo: if that story doesnt convince you then *for you* SWH is wrong. 09:11 < zipcpi> "The language is notable for its use of pure geographic directions (north, south, east, west) rather than egocentric directions (left, right, forward, backward),[5] though such "purity" is disputed." 09:11 < zipcpi> Hmm... there's also the Piraha language 09:12 < gleki> some language dont have words for "head" 09:12 < gleki> *languages 09:12 < zipcpi> Which had only two numbers. {so'u} and {so'i} 09:12 < palkeo> gleki: it's interesting, but I don't have an opinion about it for now :) 09:12 < zipcpi> Yep, not even {pa} 09:13 <@xalbo> exp: nai 09:13 < mensi> nai 09:13 < zipcpi> (Translation: only two numbers, "few" and "many". They didn't even have a word for "one") 09:13 < gleki> basic picture textbooks using gestures "I","you" dont work for Guugu Yimitthir. 09:13 <@xalbo> gleki: I suspect it's there so that an answer to {uipei} can be a bare {nai}. 09:13 < gleki> xalbo: yes, akmnrlse already answered 09:13 < palkeo> is there anything interesting related to language processing that was made for lojban ? apart from parsers 09:13 <@xalbo> Ok. didn't see it in backscroll. 09:14 <@xalbo> palkeo: There have been a few random text generators. 09:14 < gleki> a thesis on using Lojban in natural language processing 09:14 < gleki> mentioned somewhere here: mw.lojban.org/papri/Lojban_timeline 09:16 < palkeo> nice 09:17 < palkeo> (it's funny how the smiley may be doing the same job in english as the words to express emotion in lojban… there was a lack of them so smileys were introduced ?) 09:17 < gleki> also Lojban is one of few (may be the only) language where "variable types" have been specified for the whole core vocabulary (mw.lojban.org/papri/La_Bangu:_Lojban_Dictionary_with_Examples) 09:17 < gleki> yes, i suppose smileys appeared since internet allowed for a richer set of typographic symbols 09:18 < gleki> again something related to SWH 09:18 < gleki> typography forced people not to express interjections 09:18 < gleki> now this barrier is somewhat lowered 09:19 < palkeo> yeah 09:20 < palkeo> I didn't understand your variable types message, what is it ? 09:20 < zipcpi> Well you know what I had the darndest time with? "Hm". So many meanings; so many intonations 09:20 < palkeo> you can, like, define something, and then refer specifically to it using a specific word like « this », or something ? 09:21 < zipcpi> Hm? Who are you talking to? 09:21 < palkeo> sorry, internet problem… 09:21 < palkeo> I was contesting to gleki 09:22 < palkeo> about "variable types" 09:22 < gleki> palkeo: about smileys: http://nedroid.com/comics/2014-08-28-beartato-newphone.png 09:23 < palkeo> :) 09:23 < gleki> palkeo: just open this link: mw.lojban.org/papri/La_Bangu:_Lojban_Dictionary_with_Examples Such things as "clause", "entity" are not hard to understand. im not aware of any other similar projects. 09:24 < ldlework> mmm camera juice 09:24 < palkeo> gleki: oh, ok 09:24 < gleki> SWWHish comic actually 09:24 < gleki> SWHish comic actually 09:25 < gleki> palkeo: after you read you ofc. can return and ask more questions so that I improve that page 09:27 < palkeo> ok :) 09:28 < zipcpi> But yeah, the Piraha language: predictably, they sucked at math 09:28 < zipcpi> Though I think now they know about numbers, due to cultural influence 09:30 < zipcpi> Well it says "Being concerned that, because of this cultural gap, they were being cheated in trade, the Pirahã people asked Daniel Everett, a linguist who was working with them, to teach them basic numeracy skills. After eight months of enthusiastic but fruitless daily study with Everett, the Pirahã concluded that they were incapable of learning the material and discontinued the lessons. Not a single Pirahã had learned to count up to ten o 09:30 < zipcpi> r even to add 1 + 1." 09:30 < zipcpi> Dunno how they are now 09:30 < gleki> zipcpi: and they use Spanish for that 09:30 < gleki> zipcpi: just like Guugu Yimitthir use English to express "left/right" 09:30 < zipcpi> They now use Spanish numbers? 09:31 < gleki> are there any hints that numbers are appearing in their language? 09:32 < gleki> even if some particular language is changing that way the general trend are borrowings 09:32 < gleki> E.g. 30% of Quechua words are ... Spanish words. 09:32 < gleki> and Quechua is in no way a rare language 09:33 < zipcpi> I have no idea. Can't find anything that suggests it 09:37 < gleki> the only point guskant may have is that {i} not always resets the 0 level. since {tu'e} overrides that. 09:37 < zipcpi> lu...li'u 09:37 < zipcpi> to...toi 09:38 < zipcpi> And I'm not even sure {iau} works either. It's not meant to be more powerful than {i} 09:41 < zipcpi> In fact {iau} specifically says "I pretend to be .i, then back track, removing the last {vau}" 09:42 < zipcpi> Also the attitudinals thing is just deep magic; ca'e anything that attaches to {iau} attaches to the entire sentence, even though it does not close the sentence 09:44 < gleki> alta: coi 09:44 < mensi> (I [coi DO DOhU] [FA ZOhE] [CU {COhE VAU}]) 09:46 < zipcpi> Grammatically, {iau} acts as {cu'au xi da'a} 09:46 < gleki> off: coi tu'e mi mo tu'u i tu'e u'isai mo 09:46 < mensi> ([coi DOhU] [tu'e {<mi CU> <mo VAU>} tu'u] [i {tu'e <u'i sai>} {mo VAU} TUhU]) 09:46 < zipcpi> Attitudinals however, just attach to the entire sentence as if it is {cu'au xi ro} 09:46 < gleki> alta: coi tu'e mi mo tu'u i tu'e u'isai 09:46 < mensi> ([I {coi DO DOhU} {tu'e <I (¹FA mi¹) (¹CU [mo VAU]¹)> tu'u}] [i {tu'e <u'i sai>} {I <FA ZOhE> <CU (¹COhE VAU¹)>} TUhU]) 09:46 < gleki> now i wonder if what alta outputs is desirable 09:49 < zipcpi> alta: ki'ai 09:49 < mensi> (I [ki'ai DO DOhU] [FA ZOhE] [CU {COhE VAU}]) 09:49 < zipcpi> ki'ai do 09:49 < zipcpi> sa'ei do 09:49 < zipcpi> ci'oi do 09:50 < zipcpi> Though I don't know what bare COI2 is good for anyway lol 09:50 < zipcpi> And bare COI3 is probably even less useful 09:51 < gleki> selmaho:coi3 09:51 < mensi> cmavo: goi'e, jo'au 09:51 < gleki> en: goi'e 09:51 < mensi> goi'e = [COI3] assign tu'e...tu'u group to sumti |>>> goi'i assigns a bridi to a sumti; goi'e assigns an entire tu'e... 09:51 < mensi> tu'u group. May be placed within the group (usually right after tu'e), or right after tu'u. |>>> 09:51 < mensi> spheniscine 09:52 < gleki> O_0 09:52 < durka42> jo'au do'u = ko smadi lo du'u makau gentufa lo se cusku be mi 09:52 < zipcpi> ka'e ku tu'ai jo'au la tokpona li'u zo'o 09:53 < zipcpi> jo'au la itku'ile 09:53 < zipcpi> jo'au lo u zei bangu 09:53 < zipcpi> uuuuuuuuuuuuuu 09:53 < durka42> ithkuil: itku'ile 09:53 < ithkuil> itku’ile: P1S1/DYN-tk-DFF-UNI/CSL/M/DEL/NRM-RPS-IFL/UNFRAMED 09:56 < zipcpi> va'o tu'ai jo'au u zei bangu li'u no da ka'e se pilno fi lo ka cnegau le ve tavla 09:57 < zipcpi> uuuuuuuu 09:57 < zipcpi> na'i na'i na'i na'i na'i 09:58 < zipcpi> xo'ezo'e cau'a co'e do'e ku ge'ecau'i 10:01 < zipcpi> exp: ja'a ja na 10:01 < mensi> SyntaxError: Expected [yY] but "a" found. 10:01 < zipcpi> exp: pu ja ca 10:01 < mensi> ([{pu <ja ca>} KU] VAU) 10:03 < zipcpi> ua mi pu jifsku .i mi ja'a se runbau lo drata .i ri me la u zei bangu 10:03 < zipcpi> .iku'i la'a na se pilno fi tu'a le spofu fonxa kelci'e 10:11 < VENOS> Hey 10:18 < Ilmen> coi 10:19 < ldlework> a'oi 11:07 * nuzba @oka_iu_tcan: @hueruhijiki なきにしもあらずやな http://mw.lojban.org/papri/Hebrew_orthography [http://bit.ly/1M6b8UB] 14:25 < ctefaho> co'o rodo 14:50 < Amnadya> mensi: doi ctefaho http://selpahi.weebly.com/lojban/na-as-tag-re-visited 14:50 < mensi> Amnadya: mi ba benji di'u ba lo nu la'o gy.ctefaho.gy. di'a cusku da 14:51 < Amnadya> mensi: doi zipcpi http://selpahi.weebly.com/lojban/na-as-tag-re-visited 14:51 < mensi> Amnadya: mi ba benji di'u ba lo nu la'o gy.zipcpi.gy. di'a cusku da 14:51 < Amnadya> za'a smaji cacra li so'o 14:54 <@xalbo> Huh, that's interesting. 14:56 <@xalbo> So is the idea something a bit like {kau} in its focus? {na mi citka lo plise} and {mi citka na lo plise} would be truth conditionally equivalent, but with differing emphases? 14:57 < Amnadya> Possibly. 15:01 < Ilmen> Wouldn't {na mi je do} be { na [ mi je do ] }? 15:03 <@xalbo> Hmm, probably. Really, I'm not sure how to fix that either way. 15:03 < ctefaho> Amnadya: So na as a prefix huh 15:04 < Amnadya> Wild pokemon appearing left and right. 15:04 < Amnadya> But that's a good point, Ilmen. I expected something like that... 15:04 < Amnadya> (needing a terminator) 15:04 <@xalbo> The problem is that in CLL Lojban there are distinct connectives for termsets (pe'ejeks} and sumti {eks}. If you combine them both, I'm not sure how you fix that. 15:04 < Ilmen> How about it being a NAhE? 15:05 < Ilmen> sei pensi 15:05 < Amnadya> NAhE[BO] ko'a needs a terminator, too 15:05 < Ilmen> jetnu 15:05 <@xalbo> I mean, the same thing happens with {du'o mi .e do} vs {du'o mi pe'eje do}. 15:05 < Amnadya> But this is why I submit my stuff for peer review here first :) 15:06 < Amnadya> (as opposed to having it forever in a mriste archive) 15:06 < ctefaho> Coincidently I already have a sumti-prefix idea in mind 15:06 < Amnadya> mensi: do bebna 15:06 < mensi> xu do ka'e smusku la'e di'u se pi'o lo drata valsi 15:07 < Amnadya> go'i .i lo ni do snada lo te finti cu cmalu 15:07 < ctefaho> Except it is built upon the corpse of the old quantification and scoping system 15:07 < Ilmen> la mensi cu gasnu mo'oi rinka be lo nu do jinvi lo duu vo'a bebna 15:07 < ctefaho> not gonna like that part I expect 15:07 < Amnadya> na benji lo mi notci 15:07 < Ilmen> je'e 15:08 < Amnadya> I see two options: 1) NA is automatically short-scope. 2) A terminator. 15:08 < ctefaho> I have tinkered so much I don't know where to start 15:09 < ctefaho> should probably come back when all is laid out. 15:09 < Ilmen> First of all, how is called your language? zo'o ru'e 15:09 <@xalbo> Again, it really has nothing to do with NA; the same issue affects any tag with connectives. 15:09 < Amnadya> Of course, but with tags it's intended. 15:09 < Amnadya> This NA wouldn't be a normal tag 15:10 < Ilmen> I also wonder where to add FA 15:10 < ctefaho> how about making a new negator 15:10 < Ilmen> fi na mi je do? na fi mi je do? 15:10 < Amnadya> It's like short-scope NAHE+BO 15:10 < ctefaho> not part of NA 15:10 < ctefaho> let's call it no'oi'oi 15:11 < Amnadya> NA has always been its own selma'o anyway 15:11 < ctefaho> so have FAhA 15:11 < ctefaho> /PU 15:12 < Amnadya> :/ 15:12 <@xalbo> no'uei 15:12 < ctefaho> don't expect many to want them to expand differently than a BAI 15:12 * xalbo runs yuei. 15:12 < ctefaho> a sane BAI* 15:12 < Ilmen> .ai mi cilce tolcanci 15:13 < Amnadya> mi pu kanpe lo nu do xa'o sipna 15:13 < Amnadya> .i la'a la samcinki cu badzga lo se .irci 15:14 < Ilmen> .u'i do kanpe certu 15:14 < Amnadya> .u'i 15:15 < ctefaho> as for my sudden return I was talking with mensi about gismu etymology when I suddenly got a message 15:15 < Ilmen> ǐe la mensi na bi'ai gubni ke notci jungau 15:15 < Ilmen> ba'a nai ru'e 15:15 < Amnadya> mi na pu djuno 15:16 < Ilmen> .ai cipra 15:16 < Ilmen> mensi: doi menli ko prije lo ka si'abzu 15:16 < mensi> Ilmen: mi ba benji di'u ba lo nu la'o gy.menli.gy. di'a cusku da 15:17 < ctefaho> ɂI mi Cìkre lo Gìsmu noị ki Spòfu 15:17 < Amnadya> ma gismu gi'e spofu 15:17 < menli> Yeah, it gave me the message privately 15:17 < menli> when I talked to it privately 15:17 < Amnadya> je'e skepre 15:18 < ctefaho> you want the list? 15:18 < Amnadya> ma mupli 15:18 < Amnadya> gi'e pa mei 15:18 < Amnadya> #trappedinlogic 15:18 < Amnadya> .i sa'e ko cusku pa mupli 15:19 < ctefaho> http://pastebin.com/aa3ju2Wm 15:19 < samcinki> zo stika spofu lo ka mi jimpe tau ce 15:19 < clakre> mi pu zi cfasanji lo du'u la .Amnadya. du la .selpahi. 15:19 < ctefaho> +1 on np amount forgot to increment after adding kanpe 15:21 < ctefaho> repeat after me: [jetɁnu] is bad 15:21 < Amnadya> ma cusku lo'u jet.nu le'u 15:22 < ctefaho> upload me a sample on soundcloud where you say [jetnu] and nothing else in between t and n 15:22 < ctefaho> and only [t] and [n] no cheating 15:23 < ctefaho> clakre can back me on this I believe 15:23 < menli> That's pretty easy: don't release the [t] and begin nazalizing 15:23 < clakre> no, I never heard what you heard, ctefaho 15:23 < ctefaho> uhm 15:24 < Amnadya> The [t] is released nasaly 15:24 < ctefaho> I thought you conceded about "a small denpa-bu in there"? 15:24 < clakre> I only made statements predicated on "*if* there is a denpa-bu in there" 15:24 < clakre> yes 15:24 < clakre> and I said that 15:24 < menli> co'o ro do 15:24 < clakre> if there was a small one in there, then the gramatically significant denpa-bu would be a larger one 15:24 < clakre> that was my position. 15:25 < ctefaho> "released nasally"? 15:26 < Amnadya> Yes 15:26 < ctefaho> and that means? nasalization? 15:26 < Amnadya> The "explosion" you feel is in your nose, not your glottis. 15:27 < Amnadya> [tⁿn] IPA 15:28 < Amnadya> There is no glottal stop. You can feel your throat with your hand 15:29 < ctefaho> ⁿ is a symbol for what? 15:29 < ctefaho> can't find 15:29 < Amnadya> nasalized release 15:29 < ctefaho> ah 15:29 < ctefaho> found it 15:30 < ctefaho> and jetnu with ⁿ ? 15:30 < ctefaho> ... 15:30 < ctefaho> without 15:30 < Amnadya> Also possible 15:31 < ctefaho> how exactly? 15:31 < ctefaho> (low on battery may dissapear) 15:34 < Amnadya> It's "normal" to use a nasal release. You could probably insert a voiceless vowel between {t} and {n} 15:34 < Amnadya> But Lojban also allows full buffer vowels 15:34 < Amnadya> Then you can simply release orally 15:35 < ctefaho> I still want my [jetnu] 15:36 < ctefaho> like I have my [jetmu] 15:36 < Amnadya> Do you mean to say that you don't want [tⁿ] as an allophone of /t/ in Lojban? 15:37 < ctefaho> I want you to tell me you can do /tn/ 15:37 < ctefaho> because I can in no way say jetnu like I can say jetmu 15:37 < Amnadya> I can do it without release. 15:38 < ctefaho> I would like to hear that 15:38 < Amnadya> It's common in Thai or Korean 15:39 < ctefaho> the only way I can simulate it is to very quickly catch up with the nu 15:39 < ctefaho> but still a pause in-between 15:40 < ctefaho> In any cause I need to go now 15:40 < ctefaho> sipna-time 15:40 < Amnadya> Me too. 15:40 < ctefaho> co'o then 15:42 * nuzba @ligervision: @Fnxtx @evanmcm what is lojban for "artist" or "love"? I'm interested cuz Wittgenstein worked on an exact phil lang & decided it's all games [http://bit.ly/1DhN7CY] 15:43 < Amnadya> Good old Ludwig 15:47 < Amnadya> co'o 16:21 < zipcpi> {tu'e lo mabla cu fasnu .i lo zabna cu fasnu tu'u ca lo nu mi stuvi'e le sralygu'e} - how to fix for standard Lojban? 16:21 < mensi> zipcpi: cu'u la'o gy.Amnadya.gy.: http://selpahi.weebly.com/lojban/na-as-tag-re-visited | 2015-07-15T21:51:05. 16:21 < mensi> 672Z 16:21 < zipcpi> Oh haha my {ja'a ja na} thing. Well, that's kinda why I added {cau'a} and {cau'i} 16:23 < zipcpi> {cau'i} particularly, would be like *{ja'ai ja cu'i ja nai}, except you can't connect NAI/CAI... 16:27 < zipcpi> But yeah exercising fancylojban would be interesting, since that's supposed to be Lojban's strength. Formal algorithms like computer programs, and maybe less formal algorithms like cookbooks or game rules 16:30 < janKulisa> .i coi rodo mi'e .kristan. 16:31 < zipcpi> Huh the NAJA thing is interesting. Specifically it's been noted that using NAJA is a bit unnerving, like "blablabla NOT! or blablabla" 16:32 < zipcpi> With "NOT!" being the slang usage, where someone makes a statement, then negates it 16:32 < zipcpi> With the logical gardenpathing being intentional 16:33 < zipcpi> But with NAJA, we're basically asking Lojbanists to use the logical gardenpath 16:33 < zipcpi> For something other than a humorous usage 16:34 < janKulisa> but tokiponans are used to that. 16:36 < janKulisa> conditional phrases are at the beginning of the sentence, and ends with "la" 16:36 < zipcpi> (by the way I think the CLL or maybe the wave lessons translated slang NOT as like {mi broda do *nakusai!*}, but that breaks with the right-scoping {na} proposal. With {i'au/iau} though; {mi broda do iau naisai!} might work ({mi broda do vau naisai} in standard grammar) 16:37 < janKulisa> so it sounds like a normal sentence at first, and then is conditional... 16:37 < zipcpi> Exactly... 16:37 < janKulisa> what's "i'au"? 16:38 < zipcpi> Reset bridi level to zero, so auto inserts terminators to return to the top level sentence. Also, deep magic, as in any attitudinals attached to {i'au} get attached to the entire sentence 16:39 < zipcpi> So I expect NAI/CAI to also work 16:39 < janKulisa> "iau" isn't a legal combination, is it? 16:39 < zipcpi> It is, at least now. Dunno about before 16:40 < zipcpi> Basically iau is treated as semivowel i + au dipthong 16:40 < zipcpi> ia isn't considered a diphthong anymore under this model 16:40 < zipcpi> It's semivowel i + vowel a 16:40 < janKulisa> whatever... 16:41 * zipcpi shrugs 16:42 < zipcpi> There was a phonotactic reform a while back, which obsoleted *{versio} in favor of {versiio} 16:42 < zipcpi> And, a bit more controversially, also affects cmevla 16:42 < janKulisa> two vowels in one syllable (regardless of whether you call one a semivowel) is a diphthong in my book. 16:44 < zipcpi> The point is that i and u are semivowels (/j/ and /w/) when they start a syllable, and are treated differently by the reformed morphology 16:44 < janKulisa> sounds like there's too much reforming going on. when are y'all gonna get your act together? 16:45 < zipcpi> Yeah well, I get that a lot; but I'm not responsible for this reform 16:46 < zipcpi> I was taking a hiatus whan that reform happened 16:47 < zipcpi> And... well, I don't remember how the old morphology worked, but I remember there were some confusing/uncertain bits on what should be a valid zi'evla or not, and treating iV and uV as diphthongs was part of the problem 16:47 < zipcpi> Thus pe'i it's a needed reform 16:47 < janKulisa> well, it's gotta stabilise sometime, if it's gonna make it in the long run. 16:48 < zipcpi> I think morphology is stabilized now... a'o 16:48 < janKulisa> can't keep changing it like that forever! 16:49 < zipcpi> Except maybe *additions* that won't break old stuff, like *{muvdyga} 16:51 < zipcpi> I think it's like, take *{zdurio} 16:51 < zipcpi> If {io} is a diphthong, it "should" be a lujvo... zdu/rio 16:52 < zipcpi> Except rio is not a rafsi 16:52 < zipcpi> And we're not sure we want to make it a legal rafsi 16:53 < zipcpi> What's more, glides after consonants are known to be targets of "assimilation" in fast speech 16:53 < zipcpi> Like *{versio} might become {verco} 16:54 < zipcpi> So that might not be desirable for the same reason we don't mix voiced obstruents with unvoiced obstruents, like *{pofgau} 16:54 < janKulisa> sounds likely. 16:57 < zipcpi> I think there needs to be a page on the morphological reform; what it changes, why it was done 16:57 < janKulisa> well, yeah 16:57 < zipcpi> That way I can just point everyone who asks about it to it 16:58 < zipcpi> I'd write it myself but I wasn't involved in it so I fear I won't have a complete understanding of the justification 16:58 < janKulisa> if you're gonna change things around, at least have a page so we can figure out what's going on! 16:58 < zipcpi> :p 17:00 < janKulisa> trying to learn a language that keeps changing is frustrating! 17:07 < zipcpi> Yeah well, I sincerely apologize. I have pretty mabla tinkering tendencies myself. Heck that's even why I formed the joke group, {la bauspo fazykamni} 17:08 < zipcpi> http://mw.lojban.org/papri/bauspo_fazykamni/en 17:09 < janKulisa> afk 17:09 < zipcpi> je'e 17:11 < zipcpi> But yeah it's funny, being one that took a hiatus, you'd think I'd be against these reforms. But I took to it like a fish to water... thing is I understand the new morphology well enough for general usage. I don't understand the old morphology 17:11 < zipcpi> I just entered stuff into JVS and see what sticks 17:12 < zipcpi> Mainly because stuff like, is {versio} a lujvo or a zi'evla 17:12 < zipcpi> I don't even know how to begin untangling that 17:13 < zipcpi> vlaste: zivlo (gimka) 17:13 < vlaste> zo zivlo gimkamsmikezypro zo zivle noi catni gismu .i zo zivlo gimkamsmikezypro no cipra gismu 17:15 < zipcpi> vlaste: zevlo (gimka) 17:15 < vlaste> zo zevlo gimkamsmikezypro no catni gismu .i zo zevlo gimkamsmikezypro no cipra gismu 17:25 < janKulisa> .i mi bilga lenu jukpa 17:29 < zipcpi> {lo} is prefered there... {mi bilga le nu jukpa} implies there is a particular "cooking" event that you are obligated to, not just "I have to cook" 17:29 < janKulisa> there is 17:30 < janKulisa> it's this particular dinner. 17:30 < zipcpi> Well, I just find {le} a bit weird for "irrealis" uses 17:30 < zipcpi> Like I'd use {ca le cabdei} or {ca le nu jukpa}, but I'd avoid {mi djica le nu jukpa} 17:31 < janKulisa> but it's consistently used in L4B, so I'm in the habit. 17:31 < zipcpi> Yeah 17:31 < zipcpi> Well part of it is what is known as "xorlo". Yep, yet another reform, but quite a popular one 17:32 < zipcpi> And doesn't change grammar 17:32 < zipcpi> Basically {lo} no longer represents {da poi}, but is a completely generic "selbri to sumti" converter 17:33 < zipcpi> Thus can stand in for both realis and irrealis objects 17:34 < janKulisa> I read about xorlo, 17:34 < zipcpi> But part of it is that you'd see {lo NU} a lot more than {le NU} these days 17:34 < janKulisa> but habit is habit :-) 17:34 < zipcpi> :p 17:36 < janKulisa> tokipona reform seems to mostly consist in adding new words; not much grammatical stuff. 17:36 < zipcpi> The thing is though, Tokipona doesn't care about syntactic ambiguity :p 17:36 < janKulisa> if we get too many new words, though, we'll lose the whole point... 17:37 < zipcpi> Hm 17:37 < janKulisa> that it doesn't! 17:37 < janKulisa> much to my frustration sometimes... 17:39 < janKulisa> just today I was having a discussion about whether "mi lukin e ilo open mi" could mean "I'm looking for my keys" 17:39 < janKulisa> trouble is, 17:40 < janKulisa> it already means "I see my keys". 17:40 < janKulisa> and if I see them, I don't need to look for them! 17:40 < janKulisa> ya know? 17:41 < zipcpi> Yeah 17:41 < zipcpi> No word for "search"? 17:42 < janKulisa> not really; jan Kipo suggests "alasa" 17:42 < janKulisa> "to hunt" or "to gather" 17:42 < zipcpi> Interesting 17:43 < janKulisa> and since he's the tp guru, or something... 17:43 < zipcpi> Haha... I don't know tokipona 17:44 < janKulisa> but every time I speak toki pona or Esperanto, it takes all my willpower 17:44 < janKulisa> to NOT start sentences with .i 17:45 < clakre> .i mi lernas la Esperanton, sed mi nelci la lojban 17:45 < janKulisa> it's so addictive :-D 17:46 < clakre> cxu vi auxskultis "za'o"? 17:46 < clakre> (I'm not actually conversational in lojban yet...) 17:46 < janKulisa> clakre: jes 17:48 < zipcpi> Haha I don't know tokipona or esperanto, but Lojban tends to get mixed up with Malay in my mind 17:49 < janKulisa> .i ro lo jan lawa cu nasa! 17:50 < clakre> hehe I don't know the tp 17:50 < janKulisa> "jan li sama kala. kala suli li alasa e kala lili." 17:50 < clakre> so what *does* the toki pona verse mean...? 17:51 < janKulisa> I don't remember all of it at the moment... 17:51 < janKulisa> the just-quoted line means: 17:52 < janKulisa> "people are just like fish. Big fish hunt little fish." 17:53 < clakre> aha, neat. And I knew "lili", so now I know kala and suli 17:53 < janKulisa> pona! 17:54 < janKulisa> if you pm me your email adress I'll try to translate the whole verse for you 17:54 < janKulisa> I promise not to spam :-) 17:55 < janKulisa> and I've already translated the chorus 17:55 < janKulisa> or "hook", as they call it. 17:55 < janKulisa> clakre: ^^ 17:56 < clakre> je'e 17:56 < zipcpi> I'm still wondering what would be a good brivla for "while-loop" 17:56 < janKulisa> I don't know what that is. 17:56 < zipcpi> Also do-while 17:56 < zipcpi> Oh, programming constructs 17:56 < janKulisa> that was my guess. 17:57 < zipcpi> That also might be useful for informal algorithms, like instructional texts 17:57 < janKulisa> I don't do programming. 17:57 < zipcpi> It's kinda on my mind these days. "fancylojban" 17:58 < zipcpi> Like {ifle} 17:59 < zipcpi> Well it's probably doable by defining them in terms of recursion 18:00 < janKulisa> yay for recursion :-P 18:00 < zipcpi> Well, part of it is that if one can learn how programming constructs can be used informally like in cookbooks or game rules, maybe that would make it easier for one to learn programming 18:01 < janKulisa> maybe. 18:03 < zipcpi> {fancylojban} is {fancu} + Lojban. But the similarity to English "fancy" has been noted :p 18:04 < janKulisa> I forget what fancu means... 18:04 < zipcpi> function 18:04 < zipcpi> x1 is a function/single-valued mapping from domain x2 to range x3 defined by expression/rule x4. 18:05 < zipcpi> Hmm... mathmatical function... might need a new word for programming function 18:06 < janKulisa> I don't concern myself with either. 18:07 < zipcpi> I might try translating some game rules one of these days, just to test it out 18:09 * nuzba @fsiefken: @rachidalm thx! thought about something similar today: better reading comprehension with socalled "regular grammar" (esperanto & lojban) [http://bit.ly/1LkhE9v] 18:09 < zipcpi> I think programming functions are just either {ko [brivla] [arguments]}, or {[return value] [brivla] [arguments]}, with {sumji} being an example of the latter 18:10 < zipcpi> With the brivla themselves being defined *within* the program 18:10 < zipcpi> It's why I created {zai'e} and {cei'e} 18:11 < zipcpi> {zai'e} basically gives infinite space for algorithms to define brivla 18:12 < zipcpi> And allows one to use any evocative name for it 18:12 < zipcpi> Redefine a gismu, build a lujvo, whatever 18:13 < janKulisa> volapukaĵo 18:13 < janKulisa> al mi 18:16 < zipcpi> Oh I believe you were asking the other day about irrational emotions? 18:16 < zipcpi> I'm thinking it's {ro'enai} 18:17 < zipcpi> Or {ro'ene'e} possibly, not sure what is meant by "mindless" 18:17 < janKulisa> I thought about that, and wondered if that would do it... 18:17 < janKulisa> "Please connect charger" x.X 18:18 < zipcpi> I do kinda want to promote {ne'e} over {nai} for polar-negation of attitudinals, but it being disyllabic is a problem 18:18 < zipcpi> In CKTJ I experimentally swapped it with {tei} 18:19 < janKulisa> cktj iru feken... 18:19 < zipcpi> I don't know what that means 18:19 < janKulisa> you don 18:20 < janKulisa> 't want to 18:20 < zipcpi> Haha 18:20 < janKulisa> stupid phone 18:20 < janKulisa> .i le mi skami finxa cu bebna 18:20 < zipcpi> fonxa? 18:21 < janKulisa> fonxa 18:21 < zipcpi> Heh... yeah I defined {samfonxa} 18:21 < janKulisa> but I can't type cuz of my stupidphine 18:21 < janKulisa> *stupidphone 18:23 < zipcpi> And yeah CKTJ is probably one of those crazy reforms that you were complaining about, but hey, think of it as being a place for all of us zbusufukai prenu to let off some steam :p 18:23 < janKulisa> gotta go eat dinner 18:23 < zipcpi> OK 18:23 < janKulisa> co'o 18:23 < zipcpi> co'o 18:25 < Envigado> co'o = bye= 18:25 < Envigado> ? 18:25 < zipcpi> Yes 18:26 < Envigado> ah 18:33 < noncomcinse> coi 18:33 < zipcpi> coi 18:35 < noncomcinse> do mo 18:35 < noncomcinse> doi la zipcpi 18:35 < zipcpi> mi kanro .i mi pensi tu'a me zo tu'e je lo drata poi simsa 18:36 < noncomcinse> .i pa da cmavo ma'oi tu'e 18:36 < zipcpi> mi djuno 18:36 < noncomcinse> .i ma drata 18:36 < noncomcinse> gi'e simsa 18:36 < zipcpi> .y. sampla simsa 18:36 < noncomcinse> ua je'e 18:37 < zipcpi> algoritmi 18:37 < zipcpi> https://groups.google.com/forum/#!topic/lojban/NRUIud_OCj8 18:39 < noncomcinse> je'e zy 18:39 < noncomcinse> .i mi pluja cinmo pensi 18:40 < noncomcinse> .i ku'i paunai ma poi cinmo ki na pluja 18:44 < noncomcinse> .i nandu dukse 18:50 < noncomcinse> co'o ru'e .i .au sai ku'i di'a .irci ca lo cabycte 18:51 < zipcpi> ei la dzapre cu ciska lo kibypapri poi ve ciksi tu'a le cnino vlaturge'a .i mi raunaiva'e se slabu le krinu ja le nu casnu 19:23 < zipcpi> zantufa: ca lo nu mi stuvi'e le sralygu'e tu'e lo zabna cu fasnu .i lo mabla cu fasnu tu'u 19:23 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "u" cu se tolcri 19:23 < zipcpi> zantufa: ca lo nu mi stuvi'e le sralygu'e kei tu'e lo zabna cu fasnu .i lo mabla cu fasnu tu'u 19:23 < zantufa1> minajimpe: .i kanpe tu'a [yY] .i ku'i "u" cu se tolcri 19:24 < zipcpi> exp: ca le nu mi stuvi'e le sralygu'e kei ju'egi lo zabna cu fasnu gi lo mabla cu fasnu 19:24 < mensi> ([ca {le <nu (¹mi [CU {stuvi'e <le sralygu'e KU> VAU}]¹) kei> KU}] [CU {ju'e gi} {<lo zabna KU> <cu (¹fasnu VAU¹)>} gi {<lo mabla KU> <cu (¹fasnu VAU¹)>} VAU]) 19:25 < zipcpi> zantufa: gi lo zabna cu fasnu gi lo mabla cu fasnu gi'i ca le nu mi stuvi'e le sralygu'e 19:25 < zantufa1> minajimpe: .i kanpe tu'a [,] .i ku'i "s" cu se tolcri 19:25 < zipcpi> zantufa: gi lo zabna cu fasnu gi lo mabla cu fasnu 19:25 < zantufa1> minajimpe: .i kanpe tu'a [,] .i ku'i "f" cu se tolcri 19:26 < zipcpi> zantufa: ju'egi lo zabna cu fasnu gi lo mabla cu fasnu gi'i ca le nu mi stuvi'e le sralygu'e 19:26 < zantufa1> minajimpe: .i kanpe tu'a [,] .i ku'i "s" cu se tolcri 19:26 < zipcpi> zantufa: ju'egi lo zabna cu fasnu gi lo mabla cu fasnu 19:26 < zantufa1> ([ju'e gi] [{lo zabna KU} cu {fasnu VAU}] [gi {lo mabla KU} cu {fasnu VAU}] GIhI) 19:26 < zipcpi> xy'y 19:27 < zipcpi> exp: tu'e lo zabna cu fasnu .i lo mabla cu fasnu tu'u sei ca lo nu mi stuvi'e le sralygu'e 19:27 < mensi> SyntaxError: Expected [,] but "s" found. 19:27 < zipcpi> exp: tu'e lo zabna cu fasnu .i lo mabla cu fasnu tu'u sei ca lo nu mi stuvi'e le sralygu'e cu co'e 19:27 < mensi> (tu'e [{<lo zabna KU> <cu (¹fasnu VAU¹)>} {i <lo mabla KU> <cu (¹fasnu VAU¹)>}] tu'u [sei {<ca (¹lo [nu {mi <CU (²stuvi'e [le sralygu'e KU] VAU²)>} KEI] KU¹)> cu} co'e SEhU]) 19:28 < zipcpi> Ugh 19:28 < zipcpi> fegli 19:30 < zipcpi> off: tu'e lo zabna cu fasnu .i lo mabla cu fasnu tu'u sei ca lo nu mi stuvi'e le sralygu'e cu co'e 19:30 < mensi> (tu'e [{<(¹lo zabna KU¹) cu> <fasnu VAU>} {i <(¹lo mabla KU¹) cu> <fasnu VAU>}] tu'u [sei {<ca (¹lo [nu {<mi CU> <stuvi'e (²le sralygu'e KU²) VAU>} KEI] KU¹)> cu} co'e SEhU]) 19:56 < noncomcinse> coi 19:56 < noncomcinse> selma'o: xoi 19:56 < mensi> .i lu xoi li'u cmavo zo'oi XOI 19:56 < mensi> cmavo: xoi 19:56 < noncomcinse> famyma'o: xoi 19:56 < noncomcinse> camxes: xoi 19:56 < camxes> SyntaxError: Expected [yY] but "o" found. 19:56 < noncomcinse> camxes: broda xoi brode 19:56 < camxes> SyntaxError: Expected [,] but "b" found. 19:56 < noncomcinse> exp: xoi 19:56 < mensi> SyntaxError: Expected ['h] but "i" found. 19:57 < noncomcinse> sa'ei.xm. 20:02 < noncomcinse> camxes: ia nai bu 20:02 < camxes> (ia [{nai bu} BOI] VAU) 20:02 < noncomcinse> camxes: iabunaibu 20:02 < camxes> ([{<ia bu> <nai bu>} BOI] VAU) 20:02 < noncomcinse> camxes: ia zei nai bu 20:02 < camxes> ([{ia <zei nai> bu} BOI] VAU) 20:03 < noncomcinse> camxes: ia zei nai 20:03 < camxes> ([{ia <zei nai>}] VAU) 20:57 < zipcpi> Is it bad if I think I can find some use for Curtis' {zai'e}? I don't like the selma'o XOhEhOhE though; I won't know how it's meant to be defined 20:57 < zipcpi> Kinda just wish {me'o lo'o} was grammatical 21:00 < zipcpi> Well I probably won't mind it being CVV'VV or even trisyllabic 21:01 < zipcpi> Not gonna be used that often. But I have the idea of a Zendo-style game bouncing around... I heard that is usually done with sentences in Lojban though, not strings 21:02 < zipcpi> But ugh, I tried playing a web Zendo game once. I simply couldn't keep up. 21:02 < zipcpi> Couldn't compete 21:15 < zipcpi> But yeah I've been pretty abusive toward JVS. I kinda treat words like some politicians treat some legislation - "We have to define it in order to figure out what it means and how it should be used" 21:15 < zipcpi> :p 21:17 < zipcpi> exp: na ro da zvati 21:17 < mensi> ([{na KU} {ro BOI} da] [CU {zvati VAU}]) 21:17 < zipcpi> off: na ro da zvati 21:17 < mensi> SyntaxError: Expected [,] but "z" found. 21:17 < zipcpi> cizra 21:19 < noncomcinse> zipcpi: s/na/fi'o xo'i na/ 21:19 < fenki> noncomcinse thinks zipcpi meant to say: off: fi'o xo'i na ro da zvati 21:19 < zipcpi> u'i 21:20 < noncomcinse> doi la fenki coi ri .i ua do xa 21:20 < zipcpi> http://selpahi.weebly.com/lojban/na-as-tag-re-visited 21:20 < noncomcinse> si plixau 21:20 < noncomcinse> .i'e 22:20 < zipcpi> zantufa: ge ganai me'au ko'a gi me'au ko'e gi'i giga me'au ko'a gi me'au ko'i 22:20 < zantufa1> (CU [ge {<ga nai> <me'au ko'a MEhU> <gi (¹me'au ko'e MEhU¹)> gi'i} {gi <ga (¹me'au ko'a MEhU¹) (¹gi [me'au ko'i MEhU]¹) GIhI>} GIhI] VAU) 22:33 < zipcpi> zantufa: ga gi ca gi pu gi ba 22:33 < zipcpi> oi 22:34 < zipcpi> netspliti 22:34 < zipcpi> k: netspliti 22:34 < mensi> (CU [Z:netspliti VAU]) 22:38 * nuzba @la_taun: le cizra flira ku srana lu "gunka" li'u #lojban http://t.co/FudccYHWZL [http://bit.ly/1V5R3Sj] 22:40 < zipcpi> Hm if I ever decide to get a pet parser of my own I might actually call it the maltufa, just because lol 22:41 < zipcpi> Well it might be mabla in the sense that it might actually come with... options 22:42 < zipcpi> * enable CKTJ 22:42 < zipcpi> * enable crazy proposal XYZ 22:42 < isd> .i coi rodo 22:42 < zipcpi> coi 22:43 < isd> .i doi la zipcpi mo 22:43 < zipcpi> silder: Do you think destroying Lojban is a good idea? NO DON'T DO IT <----------------> YES DO IT NOW 22:44 < zipcpi> zo'osai 22:45 < zipcpi> I don't know PEG though 22:45 < zipcpi> mi no'aiva'ei certu la .peg. 22:47 < zipcpi> isd: xu do se cmene lo drata poi mi se slabu 22:52 < zipcpi> exp: ge puku gi ca 22:52 < mensi> ([ge {pu ku} gi {ca KU}] VAU) 22:54 < zipcpi> Ah I know. Everytime you click the {ko bauspo} button, it will swap two randomly selected cmavo! 22:55 < zipcpi> Heck there might even be an option to do it secretly; see if you can figure out which cmavo got swapped! The game fun for the whole family! 22:56 < zipcpi> la tcekitauspo kelci'e 22:56 < isd> zipcpi: Not totally sure what you're asking; "am I new here"? 22:56 < zipcpi> I'm asking if I know you by another name 22:58 < zipcpi> "Do I know you by another name?"... if you're new to Lojban though I can see how that sentence is a little hard to parse; it messes with word order in a way that's strange for English :p 22:59 < zipcpi> true/false: you are-named some-other-thing that I am-familiar-with 23:00 < zipcpi> xu do se-cmene lo-drata poi mi se-slabu 23:06 < isd> ah, yeah I got the literal but interpreted it more as a "your name doesn't look familiar, are you new here", not "it is a different name" 23:07 < isd> presumably lots of people you've not met are named things you are not familiar with 23:07 < isd> A loong time ago I had the nick "wobsite", but that was years ago. I've been idling in here a lot, not speaking up as much as I'd like (often busy with other things) 23:08 < isd> but I haven't gone by anything else for a while. 23:08 < zipcpi> lol Well yes, I did mean it like a "do I know you?" / "are you new here?" sense, but mainly because I'm often blindsided by Lojbanists changing names (seems to be a common IRC problem :p) 23:08 < zipcpi> OK, cool. 23:08 < zipcpi> je'e 23:10 < cliva> xu da cusku zo .optiio 23:10 < cliva> spagetufa: -lerfu .abu za'u re'u cusku 23:10 < spagetufa> (CU [{<(¹[{a bu} za'u] BOI¹) re'u> cùsku} VAU]) 23:10 < cliva> spagetufa: +lerfu .abu za'u re'u cusku 23:10 < spagetufa> ([FA {<a bu> BOI}] [CU {<(¹[za'u BOI] re'u¹) cùsku> VAU}]) 23:10 < zipcpi> u'i 23:11 < zipcpi> doi cliva ko sampla fi la tcekitauspo kelci'e go bo'oiru'e :p 23:11 < gleki> mi pu troci tua lu mi mo babocu mo li'u i ku'i mi fliba 23:12 < zipcpi> implemente le kelci'e 23:12 < cliva> za'a mi za'o na kibdu'a lo jai cabna .i .ai ca go'i 23:13 < zipcpi> uidai'i u'idai'i 23:15 < cliva> mo'u kibdu si kibdu'a 23:16 < zipcpi> da'i a website with a parser, with buttons: "Switch two random cmavo" "Switch two random cmavo (without telling me which)" 23:16 < cliva> .u'i 23:16 < zipcpi> CKTJ, the game you can play at home :p 23:17 < cliva> .i da'i sa'u lo ba'e ro cmavo cu se snigau lo cunso 23:17 < zipcpi> u'i 23:19 < cliva> gleki: si'au di'e jai banzu tu'a la jbabocu 23:19 < cliva> abs_tag_term = start:( 23:19 < cliva> - (!gek tag:tag free:free* !selbri !gek_sentence sumti:( 23:19 < cliva> + (!gek tag:tag free:free* !BO_clause !selbri !gek_sentence sumti:( 23:19 < cliva> [...] 23:20 < zipcpi> la jbabocu ku ki'a 23:21 < cliva> spagetufa: ti broda ba bo cu mo 23:21 < spagetufa> ([FA ti] [{CU <bròda VAU>} JU'E ba bo {cu <mo VAU>} VAU]) 23:21 < zipcpi> ju'e? 23:21 < zipcpi> Oh 23:22 < gleki> je'e la cliva 23:22 < gleki> i ki'e 23:23 < cliva> .ei mi di'a favgau lo gloso co'e ku mu'a 23:23 < cliva> .i la'azni 23:23 < gleki> ue la selpa'i cu finti lo notci poi la altatufa cu pagbu 23:37 < cliva> ze'irsaivi pu je ca je ba ku du'e roi 23:40 < zipcpi> Awr {la'ai} is taken; I wanted it to be to {la'a} what {ba'ai} is to {ba'a} 23:40 < zipcpi> I'm not too fond of {la'ai} though. {la'ai bu-anything} breaks 23:40 < cliva> pu ja'a se pilno tai ku 23:40 < cliva> ti'e 23:41 < zipcpi> ma .i zo la'ai xu 23:41 < cliva> go'i 23:41 < cliva> .i lo cmavo poi se smuni za'u da cu memyze'a soi sutra 23:41 < zipcpi> u'i 23:42 < zipcpi> People actually use {ba'ai}? I thought it was a kurtysmima'o :p 23:43 < cliva> zo ba'ai ja'a cnino mi 23:43 < cliva> yyy 23:43 < cliva> .i na go'oi pilno .i pu cmavo ba'e ma'oi ca'a 23:43 < cliva> je nai ma'oi mai 23:43 < zipcpi> ua 23:44 < zipcpi> By the way I found a possible answer to the {ju'oi} problem. {cau'i} 23:44 < zipcpi> Along with friends {cau'a} and {cau'e} 23:44 < zipcpi> {cau'i} = CAI, {cau'a} = NA, {cau'e} = NAhE 23:45 < cliva> cau'a ku mi nelci 23:45 < zipcpi> u'i 23:46 < zipcpi> Well it's different from {ju'ocu'i}, but would probably mean we'd have to use {ju'ocu'i} less 23:46 < zipcpi> often 23:48 < cliva> su'o se bangu zo'u lu ju'o nai li'u co'a basti lu ju'o cu'i li'u .i tormau do'a nai 23:48 < cliva> (to lo laldo smuni zo'u pilno lu ju'o na li'u ja lo simsa toi) 23:48 < zipcpi> xy'y cinri 23:49 < cliva> si'au zo la poi cmene cu se nelci so'i da 23:49 < zipcpi> la ze'oi la 23:49 < cliva> .i se ju la ckiui'irci cu .ei kajde lo troci be lo ka pilno lo cmene poi se pabgu lo kutybu'i 23:50 < cliva> s/pabgu/pagbu 23:50 < fenki> cliva meant to say: .i se ju la ckiui'irci cu .ei kajde lo troci be lo ka pilno lo cmene poi se pagbu lo kutybu'i 23:53 < zipcpi> I'm pretty bad with inventing cmavo. I treat new words like certain politicians treat certain pieces of legislation *cough*... we have to enter it into JVS to figure out what it means and how it should be used :p 23:54 < cliva> je'u so'o cmavo cu se pilno to'e ki'u nai lo nu ri no roi se jbovlaste 23:54 < zipcpi> And whether it's used often enough to deserve existing at all 23:59 < zipcpi> I'm not too fond of {la'ai}... requires {le'u}... breaks with anything that begins with {bu} or {si}... 23:59 < zipcpi> Wait 23:59 < zipcpi> exp: lo'u si le'u 23:59 < mensi> ([lo'u si le'u] VAU) 23:59 < zipcpi> exp: lo'u bu le'u 23:59 < mensi> ([lo'u bu le'u] VAU) 23:59 < zipcpi> makfa 23:59 < zipcpi> off: lo'u bu le'u 23:59 < mensi> ([lo'u bu le'u] VAU) --- Day changed Thu Jul 16 2015 00:00 < zipcpi> But yeah the le'u is quite a turn-off to me. I'd prefer things like la'oi .ali. 00:01 < zipcpi> It might be desirable to have an indicator of some sort though that says "Pronounce this in the Lojbanic manner", and possibly "This is IPA", but they can probably be typographical rather than real cmavo 00:05 < gleki> altatufa isnt necessarily frowing. i just shrinked terjonma'o section 00:06 < gleki> removing eks, jeks 00:06 < gleki> merging them all into joik_jek 00:06 < gleki> zi'e was also deleted from everywhere except joik_jek line itself 00:07 < gleki> kernelo_lo_terjoma = expr:(JA_clause / JOI_clause / stag BO_clause / A_clause / ZIhE_clause) 00:08 < gleki> this unification ofc. leads to some weird results 00:08 < zipcpi> Re: Lojban use: Of course Lojban is mainly used to talk about Lojban. I mean, when we talk in English, what do we code-switch tho Lojban for? Metalinguistic words like cmavo, bridi etc. 00:08 < gleki> alta: mi e e do 00:08 < mensi> (I [FA {mi <e e do>}] [CU {COhE VAU}]) 00:08 < gleki> no idea what that could mean 00:08 < zipcpi> If we count all that, {cmavo} is probably the most common word :p 00:08 < zipcpi> u'i 00:09 < gleki> but the whole {e / ebo / ebabo} looks weird 00:09 < zipcpi> alta: a e i o u a e i o u i o a e u i o a e u a i e o u a i e o u e o i a u e o i a u 00:09 < mensi> SyntaxError: Expected [,] but "u" found. 00:09 < zipcpi> alta: u u u u u 00:09 < mensi> SyntaxError: Expected [,] but "u" found. 00:10 < gleki> U conlang is always supported 00:10 < zipcpi> u'i 00:10 < zipcpi> alta: u 00:10 < mensi> (I [u {FA ZOhE} {CU <COhE VAU>}]) 00:10 < zipcpi> It probably has the right meaning too :p 00:12 < gleki> tersmus: iju fa zo'e cu co'e 00:12 < tersmus> co'e(c0) 00:23 < zipcpi> https://groups.google.com/forum/#!topic/lojban/B1zZGXJ6ows More illocutionary force 00:26 < zipcpi> Huh if Guskant-gi works out we might not even need ga and gu technically 00:27 < zipcpi> All replaced by JAgi / giJA 00:27 < gleki> new ga and gu or old ones? 00:27 < zipcpi> Selpa'i-ga/gu 00:28 < gleki> they were created because he didnt like xorxe's mad proposals 00:28 < zipcpi> Guskant-gi is mad proposal? 00:28 < zipcpi> I don't know the mad proposals 00:28 < gleki> selpa'i's system is a rework of the original proposal by xorxes back from 1996. 00:29 < zipcpi> gije broda gi brode 00:29 < zipcpi> Is that mad-proposal? 00:29 < gleki> no, jegi broda gi brode 00:29 < zipcpi> Huh... zantufa supports both 00:29 < zipcpi> zantufa: gije broda gi brode 00:29 < zantufa> (CU [{gi je} broda {gi brode} GIhI] VAU) 00:29 < zipcpi> zantufa: jegi broda gi brode 00:29 < zantufa> (CU [{je gi} broda {gi brode} GIhI] VAU) 00:29 < gleki> yes, ilmentufa originally supported all three systems 00:29 < gleki> exp: jegi mo gi mo 00:29 < mensi> (CU [je gi] [CU {mo VAU}] gi [CU {mo VAU}] VAU) 00:30 < gleki> exp: mi mo vau je mo 00:30 < mensi> (mi [CU {mo vau}] [je {CU <mo VAU>} VAU]) 00:30 < gleki> exp: mo gi'e mo 00:30 < mensi> ([CU {mo VAU}] [gi'e {CU <mo VAU>} VAU]) 00:30 < gleki> exp: ga mo gi mo 00:30 < mensi> (CU [ga {CU <mo VAU>} gi {CU <mo VAU>} VAU]) 00:30 < gleki> then they added selpa'is system too 00:30 < zipcpi> Except that selpa'i-gu doesn't work 00:30 < zipcpi> exp: mi guje broda gi brode 00:30 < mensi> SyntaxError: Expected [,] but "b" found. 00:30 < gleki> no idea 00:31 < gleki> i only used JA=A in CC. i ignore the other proposals 00:31 < gleki> because they are changing all the time 00:32 < gleki> btw, today i finally merged NA into tags 00:32 < gleki> except when it has the second meaning when working with connectives 00:33 < zipcpi> Part of the problem with connectives though is the question connectives 00:33 < zipcpi> I don't remember them x.x 00:33 < gleki> they are based on -i 00:34 < gleki> {gi'i} is perfect 00:34 < zipcpi> But the rest... x.x 00:34 < gleki> as i said i used both A and JA and A=JA 00:34 < gleki> in CC 00:34 < gleki> so je'i, ji, je'i=ji 00:34 < zipcpi> Yeah but there's still GA 00:34 < gleki> you mean ge'i 00:35 < zipcpi> Yep... 00:35 < gleki> lai'a is just {sei PAcu'o}? 00:35 < zipcpi> .y. 00:35 < zipcpi> nm then 00:36 < zipcpi> It's death was heralded when I couldn't make it fit {ba'ai} anyway T.T 00:36 < gleki> we need a name for this 4-set problem 00:36 < zipcpi> What set problem? The connectives? 00:37 < gleki> k:zgazganzganaga'aza'a 00:37 < mensi> (CU [Z:zgazganzganaga'aza'a VAU]) 00:37 < gleki> yes, let that be the name for it 00:37 < gleki> https://groups.google.com/forum/#!msg/lojban/6H7JgQOJTzk/R_ZSGGomhp8J 00:38 < zipcpi> Oh 00:38 < gleki> so this lai'a is proliferating of this problem 00:38 < gleki> loglan uses {soi crano} = {sei mi cmila} all the time 00:39 < zipcpi> Nevermind, I downvoted it 00:42 < zipcpi> Sigh... sei 00:47 < zipcpi> exp: broda vau nai 00:47 < mensi> (CU [broda {vau nai}]) 00:47 < zipcpi> off: broda vau nai 00:47 < mensi> SyntaxError: Expected ['h] or [uU] but "i" found. 00:47 < zipcpi> u'i 00:48 < zipcpi> off: broda vaucai 00:48 < mensi> (broda [vau cai]) 01:12 < zipcpi> exp xoi pa cu'o ko'a broda 01:13 < zipcpi> exp: xoi pa cu'o ko'a broda 01:13 < mensi> ([xoi {CU <(¹pa BOI¹) cu'o> <ko'a VAU>} SEhU] [CU {broda VAU}]) 01:13 < zipcpi> xoi pa cu'o vau ko'a broda 01:13 < zipcpi> exp: xoi pa cu'o vau ko'a broda 01:13 < mensi> ([{xoi <CU (¹[pa BOI] cu'o¹) vau> SEhU} ko'a] [CU {broda VAU}]) 01:15 < zipcpi> exp: lo'u blablablacomcomco'e fa'o le'u 01:15 < mensi> ([lo'u {blablablacomcomco'e fa'o} le'u] VAU) 01:15 < zipcpi> exp: lo'u su su su su susasisosu le'u 01:15 < mensi> ([lo'u {su su su su su sa si so su} le'u] VAU) 01:16 < zipcpi> gleki: I think that's the official way to quote people who say problematic magic-words :p 01:16 < zipcpi> Not sure how to quote something with {le'u} though lol 01:18 < zipcpi> Huh 01:18 < zipcpi> http://lojban.github.io/cll/19/9/ 01:19 < zipcpi> 9.5 01:19 < zipcpi> THat's funny, because I thought ri refered to {lo'u le ninmu cu morsi le'u} 01:19 < zipcpi> oi 01:20 < zipcpi> exp: lo'u zo le'u le'u 01:20 < mensi> SyntaxError: Expected [yY] but "e" found. 01:20 < zipcpi> What the heck 01:20 < zipcpi> off: lo'u zo le'u le'u 01:20 < mensi> SyntaxError: Expected [yY] but "e" found. 01:20 < zipcpi> "However, it is possible for a “le'u” to occur within a “lo'u ... le'u” quotation by preceding it with the cmavo “zo”, discussed in Section 10" 01:20 < zipcpi> What is this nonsense 01:22 < zipcpi> ca'e {le'e'u} stands for {le'u} within a lo'u le'u quote, {le'e'e'u} stands for {le'e'u}, so on ad infinitum zo'o 01:24 < zipcpi> cylyly zbusufukai 01:24 < zipcpi> lo jbobau cu zbusufukai 01:24 < zipcpi> ro da zbusufukai 01:24 < zipcpi> ro da ka'e daspo lo jbobau 01:33 < zipcpi> blanu: x1 is blue? Not x1 is blue in aspect x2 to observer x3? :p 01:35 < zipcpi> lol this is the kind of thing that is why I dislike E-prime. Sure maybe English "is" is often used to imply absolutism where it isn't desirable... but I don't believe in getting rid of it altogether 01:36 < zipcpi> It just tweaks with my philosophical sensibilities 01:36 < zipcpi> Theres just no way to say A = A in E-prime 01:40 < ctefa`o> xoi rodo 01:43 < gleki> heh, someone bought lojban.ru domain http://lojban.ru/ 01:43 < gleki> at least it's a lojbanist 01:43 < zipcpi> lol I half-expected "Girls of Lojbanistan" o.o 01:44 < gleki> once it existed. i think we should have bought it 01:44 < zipcpi> What existed 01:45 < zipcpi> A porn site using the name "Lojban"? 01:47 < gleki> no, lojban.ru existed 01:51 < zipcpi> Heh 01:57 < zipcpi> coi selpa'i 01:58 < selpahi> zo'o xu do prami mi 01:58 < zipcpi> ja'a pedypa'i u'i 01:58 < selpahi> .i va'o tu'a na ku zo la jai mapti 01:58 < gleki> xrotu prami 01:59 < zipcpi> zo xropa'i 02:00 < zipcpi> zo xedypa'i zo pedypa'i zo tubypa'i zo lazypa'i zo cinpa'i zo ci'irpa'i 02:01 < zipcpi> mi finti du'e lujvo pe zo prami 02:02 < selpahi> na'i go'i 02:02 < zipcpi> ie du'e na'i 02:02 < zipcpi> dukse fi no da poi vajni 02:05 < zipcpi> "The rightwards scope of {na} is standard Lojban by now. It can hardly be called a proposal anymore." ua i'e 02:07 < gleki> je'e je'e 02:07 < gleki> i ku'i CLL cu xusra lo drata 02:08 < zipcpi> je'e 02:08 < zipcpi> ku'i mi mutce to'e nelci zoi zoi All that glitters is not gold zoi 02:09 < selpahi> Not all that glitters is gold 02:09 < zipcpi> Yeah 02:09 < selpahi> You were asking about {jo'a} the other day. I use it like "[sic]" is used in English. 02:09 < zipcpi> lol 02:09 < gleki> uanai 02:10 < zipcpi> {je'au}? 02:10 < zipcpi> {je'aunai} rather? 02:10 < zipcpi> ... dunno 02:11 < selpahi> "This is meant as written, this is not a typo" 02:11 < zipcpi> Yeah I know what you mean, but I'm not sure that is correct against {na'i} 02:11 < selpahi> It tells the listener that the potentially unusual word or phrase is intentional 02:11 < gleki> lu le'ainai li'u 02:12 < zipcpi> Oh yeah that 02:12 < gleki> i ku'i mi jimpe 02:12 < gleki> i simsa zo'oi sic i ku'i na mulno dunli 02:12 < selpahi> ie 02:12 < gleki> i la'a zo'oi sic ambigu 02:12 < selpahi> pe'i smuvrici 02:13 < zipcpi> I use {je'au} mainly as a sort of informal {sa} / {sa'ai} 02:13 < gleki> en: je'au 02:13 < mensi> je'au = [UI3] discursive: correcting/corrective/correction - inattentive/uncaring/neutral toward the presence of 02:13 < mensi> possible errors - permitting (known/likely/plausible) error/incompleteness/approximation |>>> 02:14 < mensi> krtisfranks 02:14 < zipcpi> But I have no idea what {je'aucu'i} or {je'aunai} mean 02:14 < gleki> sounds like sa'e scale 02:14 < zipcpi> Don't think so... je'au is clearly meant to correct something, pe'i 02:15 < zipcpi> Oh... you mean the scale is *like* sa'e 02:15 < selpahi> Ithkuil has a "corrective" bias, sounds like {je'au} 02:16 < gleki> what is "correcting" if not specifying something more precisely 02:16 < gleki> selpahi: ma nuzba la mupli 02:16 < zipcpi> nebyplovoicpi for heron? Really? 02:17 < selpahi> no da nuzba .i mulgau ba tu'a lo .itku'ile vlaste .i ku'i ji'a mi nitcu lo ka se jibri 02:17 < selpahi> .i lakne ku'i fa lo nu la mupli cu mulno ba za lo jeftu be li su'e ci 02:17 < gleki> ua 02:18 < gleki> B120 #ERROR 02:19 < gleki> i e'u do jmina valsi bu 02:19 < gleki> xm 02:19 < gleki> i e'u do jmina me'o valsi bu 02:19 < selpahi> ua nai 02:19 < gleki> https://docs.google.com/spreadsheets/d/1UFwAZz-PpCZ6t5jU0wCd5TR2LgskBNInlEpKiktYCoQ/edit#gid=0 02:19 < selpahi> ithkuil: falara'x 02:19 < ithkuil> falara’x: P1S1/STA-f-OBL-UNI/CSL/M/DEL/NRM-NA1_1/5-EXS-CRR-IFL/UNFRAMED 02:19 < selpahi> "I meant it's not there" 02:20 < selpahi> je'au na zvati 02:20 < selpahi> me'o valsi bu mo 02:21 < gleki> du me'o kunti bu 02:21 < selpahi> ua mi ca viska 02:21 < selpahi> .i ie cikre 02:23 < gleki> i barda se gunka 02:31 < zipcpi> Hm ... {nakusai} still suffers from the problem of open bridi-tails; if there are any sub-bridi left open, it will only negate one sub-bridi, not the entire sentence 02:32 < zipcpi> I have no more control over {i'au}/{iau} at this point lol 02:32 < zipcpi> It keeps showing up everywhere 02:35 < zipcpi> doi la selpa'i: .i xu do ja lo datpre poi do djuno tauki ka'e ciska lo kibypapri pe la jbouitki je poi ve ciksi tau le cnino vlaturge'a 02:37 < zipcpi> da'i ai mi ciska .iku'i mi na mutce lo ka se slabu le krinu be tau le vlaturge'a 02:37 < selpahi> la cirko ba'o ciska 02:37 < gleki> it's not a sub bridi, it's a bridi joined with another one via a common head 02:37 < zipcpi> ki'oi ma a'u 02:37 < selpahi> lo mriste 02:39 < selpahi> .i ji'a http://mw.lojban.org/papri/informal_description_of_the_PEG_morphology_algorithm laldo 02:40 < zipcpi> What I want is something that says... here's what changed, here's why we think it had to change, etc. 02:41 < zipcpi> Cause I still have some tadni coming to me asking why it changed, and my answer keeps changing each time because they're only guesses 02:41 < durka42> coi 02:41 < zipcpi> coi 02:43 < zipcpi> I took to the new morpohology like a fish to water... the changed happened during my hiatus, but I had trouble figuring out some parts of the old one... like why things like {zdurio} are considered zi'evla rather than lujvo... 02:43 < selpahi> https://groups.google.com/forum/#!msg/lojban-beginners/f-u86pPQ-TQ/BVFmXWTSgSQJ 02:43 < zipcpi> Not to mention the whole issue with assimilation 02:43 < gleki> because rio is not a rafsi 02:45 < zipcpi> lol cirko also called himself {la me zo'e}? 02:45 < ctefa`o> rio would make a nice cmavo though 02:45 < ctefa`o> If you like 02:45 < ctefa`o> Want 100+ new simple cmavo forms 02:45 < ctefa`o> Probably not 02:45 < selpahi> I want fewer cmavo, not more. 02:46 < gleki> rio is already a cmavo in La Alta 02:46 < selpahi> The 10000+ experimental cmavo don't help us at all. 02:46 < ctefa`o> .i'e 02:46 < zipcpi> u'i mi mabla 02:47 < selpahi> I often get asked why there are so few people who speak this language. 02:47 < ctefa`o> TOO NERDY SI SI 02:49 < gleki> for mekso new selmaho might be justifiable. although curtis should think of brivla first 02:50 < zipcpi> selpahi: Hm I like what you sent me but I want the justification for dumping a whole bunch of zi'evla and cmene to the "obsolete" list 02:51 < selpahi> ? 02:51 < zipcpi> It's like, I can see the justification, but I can only guess at what it is 02:51 < selpahi> uanai 02:51 < durka42> you can look at the bpfk list to see the discussion of CGV clusters 02:51 < selpahi> Oh 02:51 < selpahi> The morphology 02:51 < zipcpi> Basically, why is CGV banned 02:51 < durka42> too hard to pronounce, too easy to confuse for CVGV, is the summary 02:52 < durka42> ironically curtis sparked that thread, by defining {jie'e'e} :p 02:52 < durka42> I mean *{jie'e'e} 02:52 < zipcpi> u'i 02:52 < zipcpi> What for I wonder 02:52 < zipcpi> mekso operator: assoviated Legendre polynomial in a with unassociated order b and associated order c 02:52 < durka42> Legendre polynomials, naturally 02:52 < selpahi> http://tiki.lojban.org/tiki/BPFK+Section%3A+Ban+on+consonant-glide-vowel+strings 02:52 < durka42> ua 02:53 < zipcpi> And thus Lojban is destroyed. RIP zo'o 02:54 < durka42> nunkurtspo 02:55 < gleki> 19 more questions with a score of 10 or more 02:55 < gleki> http://area51.stackexchange.com/proposals/83133/jbobau-joi-jboceu-lojban-language-and-culture 03:01 < zipcpi> Hm that vote is... interesting 03:01 < zipcpi> Basically 5 people want CgV in cmevla, 5 people don't want it 03:02 < zipcpi> Didn't count the ones that wanted the more complicated rules 03:02 < durka42> there were too many options 03:02 < durka42> most of the options were a middle ground that would have made the morphology more complex, not less 03:02 < durka42> so we went for the hard line 03:02 < zipcpi> Well, first-past-the-post problem 03:02 < durka42> (though I think I voted for one of the messy ones) 03:03 < zipcpi> u'i 03:03 < durka42> was it first past the post? 03:03 < zipcpi> Well actually they're all the same rule 03:03 < zipcpi> We can just ca'e call it "bad glide combinations" 03:03 < zipcpi> [sibilant or dental C+iV] and [liquid C+uV] 03:03 < durka42> oh yeah I remember someone saying we should discuss voting methods and then guskant said "fuck it here is a google form, we're doing it live" 03:04 < zipcpi> Well, the "winner" is "disallow CgV in cmevla/fu'ivla/ma'ovla", and it has the most votes, but only 45.5% 03:04 < durka42> yeah 03:04 < durka42> plurality! 03:05 < zipcpi> Mhm 03:05 < gleki> su'oi zei zanru 03:05 < durka42> heh, cmavla => cmavo => ma'ovla 03:05 < gleki> vau zo'o 03:06 < zipcpi> Why even have {ma'ovla} 03:06 < zipcpi> Just for poetic assonance I guess? 03:06 < durka42> yeah 03:06 < gleki> en: ma'ovla 03:06 < mensi> [< cmavo valsi ≈ Structure word word] lo nu mulno sisku zo'u: y no da se tolcri 03:06 < mensi> ma'ovla[6367], cmavyvla[7988], ma'orvalsi[9526], cmavyvalsi[10047] 03:06 < zipcpi> na se jbovlaste 03:06 < durka42> I never saw it before that thread, ju'ocu'i 03:06 < durka42> smudu'i zo cmavo 03:07 < zipcpi> logji bangu -> lojban -> lojbo, lojbo bangu -> jbobau 03:07 < zipcpi> cadzu cipni -> dzipo, dzipo cipni -> zipcpi 03:08 < zipcpi> refkusi vlakra 03:08 < gleki> -> jbobo 03:08 < zipcpi> je'u 03:09 < zipcpi> I kinda want rafsi -job- for jbobo, so that if we ever need a zi'evlajvo for jbobo + zi'evla, we have an alternative to -lob- 03:10 < zipcpi> But I dunno lol 03:10 < zipcpi> Probably the ones that invented jbobo would support this idea 03:10 < durka42> waste of space since jbobo is (intentionally) a duplicate 03:10 < zipcpi> :p 03:11 < zipcpi> Problem is that zi'evlajvo does break the doctrine that CCV cmavo is "better" than CVC cmavo 03:11 < zipcpi> Er, rafsi, not cmavo 03:11 < zipcpi> oi 03:19 < selpahi> zi'e vlajvo 03:19 < zipcpi> Ugh 03:19 < zipcpi> zevlyjvo 03:19 < zipcpi> {zevlo} 03:20 < zipcpi> Thus completing the set. {gismu}, {cmavo}, {lujvo}, {zevlo}, {brivo}, {cmevo} 03:21 < zipcpi> Well maybe {zevla}, since it doesn't use the -vo ending 03:21 < zipcpi> vlaste: zevla (gimka) 03:21 < vlaste> zo zevla gimkamsmikezypro no catni gismu .i zo zevla gimkamsmikezypro no cipra gismu 03:23 < zipcpi> vlaste: zivla (gimka) 03:23 < vlaste> zo zivla gimkamsmikezypro zo zivle noi catni gismu .i zo zivla gimkamsmikezypro no cipra gismu 03:23 < zipcpi> Oh right 03:24 < zipcpi> jvozba makes {zifyvlajvo}, but, come on 03:24 < gleki> selpahi: lu na mi e do li'u je'e i ku'i lo gi'e zei jufra cu mo i mu'a lu mi mo na gi'e ja'ai mo li'u mo 03:25 < selpahi> mi na mo gi'e ja'a mo 03:26 < zipcpi> do cau'a mo 03:30 < gleki> exp: ba mi e do mo 03:30 < mensi> ([ba {mi <e do>}] [CU {mo VAU}]) 03:31 < gleki> alta: na mi e do mo 03:31 < mensi> (I [{na ZOhE} {FA <mi (¹e do¹)>}] [CU {mo VAU}]) 03:31 < gleki> ju'inai 03:32 < zipcpi> Yeah jvova'i is a problem sometimes 03:34 < zipcpi> Where a lujvo is used to build as a root to build another lujvo, but oops, tosmabru, so you need to change it. Well the least damage you can do is make it something like {zi'ervlajvo} 03:34 < zipcpi> But nope, jvova'i says that {zifyvlajvo} is "better" 03:34 < gleki> 03:35 < zipcpi> And I don't think we want to make new gismu everytime that problem is encountered lol 03:38 < gleki> alta: na mi e do mo 03:38 < mensi> (I [{na <mi (¹e do¹)>} {FA ZOhE}] [CU {mo VAU}]) 03:38 < gleki> selpahi: xu do tugni 03:39 < gleki> zipcpi: just dont use CV'V rafsi in prefixes 03:39 < zipcpi> So ban {zi'evla}, use only {zifyvla}. I might agree with that but jbovlaste keeps giving me that ominous warning 03:40 < zipcpi> Though I'm not sure if that threat is ever followed up on 03:40 < selpahi> If mr and ml weren't initials, then CCV *and* CV[nmrl] would be safe. 03:40 < gleki> alta: mi pe na e do cu mo 03:40 < mensi> (I [FA {mi <pe (¹[DOhE ZOhE] [na e] [DOhE do]¹) GEhU>}] [cu {mo VAU}]) 03:41 < gleki> nu'o jdice lo du'u xamgu xukau 03:44 < zipcpi> Wuzzy is like the stern librarian of JVS. "Hey the jvova'i isn't optimal." "Hey the morphology is new and weird". " Hey you didn't follow the Dewey Decimal System" 03:44 < zipcpi> :p 03:56 < gleki> lo nu traduki la'e la alis cu jai ri'a tatpi 03:57 < zipcpi> dai 03:59 < zipcpi> {pixydja}? Hey I found a {no'e zi'ejvo} :p 03:59 < zipcpi> -dja is really {cidja}, but {djacu} is a more natural metaphor; however it just happens to fit! 04:12 < zipcpi> k: zi'evlrjvo 04:12 < mensi> (CU [Z:zi'evlrjvo VAU]) 04:12 < zipcpi> Well there's always that, but *shudders at the syllabic consonant* 04:14 < zipcpi> *zi'evlyjvo probably doesn't break anything; but all these unassigned-morphology-proposals might not be compatible with each other 04:14 < zipcpi> Even if they might be compatible with current morphology 04:16 < zipcpi> And, funnily, we might want {lujvla} so that we can talk about compound words in general, like in other languages, or in zi'evla formed from Lojban roots :p 04:17 < zipcpi> And {vlrjv}! Wow 04:18 < zipcpi> So many consonants 04:20 < zipcpi> k: zvaisku 04:20 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "z" found. 04:21 < zipcpi> slinku'i 04:21 < zipcpi> k: zvaiku 04:21 < mensi> (CU [Z:zvaiku VAU]) 04:26 < selpahi> Gah Wuzzy 04:27 < zipcpi> What? 04:27 < zipcpi> What did he do? (Well I know he does many things, but what did he do specifically :p) 04:27 < selpahi> Just all those comments 04:27 < zipcpi> Oh... 04:28 < zipcpi> Yep... told you about the jvova'i librarian thing 04:28 < selpahi> Also he downvotes my words as if that would somehow stop people from using them 04:29 < selpahi> He seems unaware of recent developments (i.e. post-CLL zo'o) 04:29 < zipcpi> :p 04:30 < selpahi> At least he has now defined a pajvo ahaha 04:30 < selpahi> .u'i 04:30 < zipcpi> pajvo? 04:30 < selpahi> zi'ejvo 04:30 < selpahi> selpamnonjvo 04:30 < zipcpi> {zevla} is me 04:31 < selpahi> No, he defined pixdja as drinking water 04:31 < selpahi> pixydja* 04:31 < zipcpi> Oh, {pixydja} 04:31 < zipcpi> It's a no'e zi'ejvo lol 04:31 < zipcpi> Because {cidja} actually makes sense, in a way :p 04:32 < zipcpi> In fact I think that's what Wuzzy thought was your intention all along 04:32 < zipcpi> He put in the etymology 04:32 < zipcpi> Implicit-abstraction lujvo from “nu pinxe kei cidja”. 04:32 < selpahi> Yes 04:32 < selpahi> I don't know. 04:33 < selpahi> I don't care. It's just annoying. 04:33 * zipcpi nods 04:34 < zipcpi> I had to scold him for downvoting all the translations of {ybu} 04:38 < zipcpi> Anyway congratulations on tricking Wuzzy to adopt a zi'ejvo lol 04:38 < selpahi> lol 04:39 < durka42> maybe we should change their vote weight to 0.5 zo'o 04:41 < zipcpi> u'i 04:41 < zipcpi> Hey how to say "adopt" as in "adopt a child"? {runrirbi'o}? 04:42 < selpahi> rirni isn't biological 04:42 < zipcpi> Might be stretching {rutni} though lol 04:42 < durka42> just rirbi'o I guess 04:42 < zipcpi> I know but a mere {rirbi'o} might not be clear 04:42 < selpahi> bi'o is too weak 04:42 < selpahi> says nothing about intention 04:42 < zipcpi> ie 04:42 < durka42> well there's no other way to become a parent of a child who is older than 0, .u'i 04:44 < selpahi> Add it to the nabmyrta'imacusku if you don't find a solution now 04:44 < zipcpi> Well we might just want a word for "adopt"; perhaps several words for different senses 04:44 < durka42> there are already several words for "adopt" in the sense of "choose" 04:44 < durka42> zo plicu'a 04:45 < zipcpi> Ah... well, we need one for "adopting a child/pet" 04:45 < durka42> rirni is a bit weird for pets 04:45 < zipcpi> lazycu'a or something? 04:45 < zipcpi> I know 04:45 < durka42> that would just be kujbi'o 04:45 < durka42> mi plicu'a lo ka lazni vau .u'i 04:46 < zipcpi> https://www.google.com/webhp?sourceid=chrome-instant&rlz=1C1CHFX_enMY573MY574&ion=1&espv=2&ie=UTF-8#q=define%20adopt 04:46 < zipcpi> Defines base sense as "legally take another's child and bring it up as one's own" 04:46 < zipcpi> And defines the rest, like "adopt an idea", as metaphorical senses growing out of it 04:48 < zipcpi> What is "nabmyrta'imacusku"? 04:48 < zipcpi> Is it a mailing list just for undefined concepts? 04:52 < selpahi> It's this https://docs.google.com/spreadsheets/d/1LH_bawafDZBp5h4shGoajnTU7sttOilfpyxqUD4B5BM/edit#gid=0 04:52 < selpahi> Just a list of things we don't know how to say 04:52 < selpahi> can be phrases or words 04:53 < zipcpi> towel? 04:53 < zipcpi> Hm French fries... what is that called in other languages 04:53 < selpahi> pommes frites 04:53 < zipcpi> salami sounds like a zi'evla 04:54 < zipcpi> sarcasm... just make {xo'onku} or something lol 04:55 < zipcpi> generous: zo do'a zo'o 04:56 < zipcpi> "interact with (not the social sense)" Hm my new gismu {jonse} might be related 05:00 < gleki> {co'a rirni} is fine, imo. it doesnt necessarily needs to be "parent" 05:00 < zipcpi> exp: pa bi'i ci moi 05:00 < mensi> (CU [{<pa BOI> <bi'i (¹ci BOI¹)>} moi] VAU) 05:01 < zipcpi> gleki: Fine if you don't need to distinguish between an adoption and the biological sense 05:02 < gleki> zipcpi: for biological sense we have another gismu 05:03 < zipcpi> Do we then, need to say rirni jenai ??? 05:06 < zipcpi> {traide} 05:06 < gleki> co'a is enough imo 05:07 < zipcpi> Context is great and lets us get by with a reduced vocabulary, but sometimes there needs to be a way to distinguish these concepts that we are ramming together 05:08 < zipcpi> The ramming together isn't necessarily a bad thing, cause when we say "parent", it might not be desirable to specifically have to say "adopted parent" 05:08 < zipcpi> But sometimes we need that distinguishment 05:08 < gleki> {mi co'a jai zu'e rirni lo mlatu} {i ue i xu do ji'a mlatu} {i na go'i i mi arxokuna} 05:08 < zipcpi> u'i 05:08 < selpahi> lol 05:09 < zipcpi> mi gerku 05:09 < zipcpi> jenai zipcpi 05:09 < gleki> {rirni} means more "to rear" 05:09 < durka42> you could use flarirni/jberirni 05:10 < zipcpi> Adding that as "danfu" #2 05:10 < gleki> toki pona style is inevitable in every language 05:11 < zipcpi> I don't know what you mean by "toki pona style"? 05:11 < durka42> tanru staile 05:11 < zipcpi> Ah 05:11 < durka42> dotco lujvo staile 05:11 < durka42> zo'o ru'e 05:11 < gleki> not only that 05:11 < gleki> just low precision 05:12 < durka42> well TP has low precision 05:12 < selpahi> Yet other languages have more specific words. 05:12 < durka42> lojban has selectable precision :p 05:12 < gleki> TP is 100 less precise than lojban so what. here is the 100th case when we get this lack of precision. natlangs do then same all the time. we just dont notice that 05:13 < gleki> <quote> to have words for such combinations of properties as 05:13 < gleki> the sun and the water on a swimmer's chest, the vague tremulous rose color we see with our eyes closed, the sensation of being carried along by a river and also by sleep. 05:13 < gleki> </quote> 05:18 < zipcpi> exp: zo kanpylerci (to noi xa'o se jbovlaste toi) 05:18 < mensi> ([{zo kanpylerci} {to <noi (¹CU [xa'o {se jbovlaste}] VAU¹) KUhO> toi}] VAU) 05:23 < gleki> i suggested rafybri so long ago and people are still stuck in anarchic lujvo system 05:23 < zipcpi> I dunno what that is 05:27 < gleki> "merge free and indicators" <-- that's an important task 05:28 < gleki> hm, i can't do that. i need to separate UI, CAI/NAI first 05:28 < gleki> zipcpi: was that split formalized by you? 05:28 < zipcpi> I did write something about it 05:29 < gleki> because some work as independent words, some as modifiers of the first and some attach to the end like e.g. {ro'a} 05:30 < zipcpi> http://mw.lojban.org/papri/jbocre:_Attitudinals_-_time_for_a_split%3F 05:30 < gleki> [ui {(ui-[nai-sai])}-ro'a] 05:31 < gleki> okay, so interjections, interjection modifiers, and .. what shall we call ROhA and friends? 05:31 < gleki> endings? 05:31 < zipcpi> categories 05:32 < zipcpi> Or spheres 05:32 < gleki> iim gonna write new rules for all that then 05:33 < zipcpi> Is the "cu li" dialect inspired by toki pona? :p 05:33 < gleki> yeah, and also by selpa'i who didnt like c- sound. 05:34 < zipcpi> Ahaha 05:34 < gleki> but generally it's inspired by this ugly tcekitau dialect 05:34 < gleki> as opposed to it 05:34 < noncomcinse> coi zo'e do'u 05:34 < gleki> feel fre to add it to BSFK proposals 05:34 < zipcpi> lol 05:37 < gleki> oh my, ROhA are all in UI. who could ve guessed 05:37 < zipcpi> Yes 05:38 < zipcpi> ei ro da me ma'oi ui iau zo'o 05:40 < Envigado> co'o 05:43 < zipcpi> OK... "default", yet another missing word 05:43 < zipcpi> There is {tecycuxskicauzmi} but... come on 05:43 < gleki> i usually used {zmiku} or {na'o} for it 05:43 < gleki> the first is more in GUIs 05:44 < zipcpi> Hm found zmiselcu'a 05:45 < gleki> not necessarily chosen 05:45 < gleki> it can be just {pamoi} value 05:48 < zipcpi> lol I found {li'o ce'ai} to use FA bu in {ka} 06:00 < gleki> so it all means that adding INT, MOD and END (endings) will have to be done manually 06:00 < selpahi> Ilmen uses cuxycau for default 06:00 < gleki> en: cuxycau 06:00 < gleki> jbo: cuxycau 06:02 < zipcpi> Seems like a cimjvo but I don't really care lol 06:02 < selpahi> xu mi na nelci lo cy zei sance ?! 06:03 < gleki> http://selpahi.weebly.com/lojban/parallel-universe-lojban 06:03 < selpahi> ua do tavla fi zo su noi basti 06:03 < selpahi> .i mi ra'u nai sai pante 06:04 < gleki> i babo mi finti la culis 06:05 < selpahi> mango jeli selckiku 06:05 < gleki> "free" is found 267 times in altatufa a'enaisai 06:06 < gleki> interesting what tool guskant is using to develop zantufa 06:07 < gleki> i use ircitufa that has a button to recompile .js files. 06:14 < zipcpi> My current understanding of the grammar of {i'au}/{iau} (and probably the final understanding, a'o) is that it isn't actually a famyma'o. All it does is two things: 1. add all famyma'o to return to the top level of the sentence. Basically, everything {.i} will add except the final {vau}. 2. any UI/NAI/CAI attitudinals that attach to it, scope over the entire sentence 06:14 < zipcpi> {mi broda do iau iau iau iau iau zo'o} would be grammatical under this understanding 06:15 < zipcpi> Because iau isn't a famyma'o itself, only adds them, and it works even when there isn't any famyma'o to add 06:17 < zipcpi> One can also probably sort of abuse it; using it to close NU/NOI construct and just continue on with the sentence, and get by most of the times without knowing about {vau}, {kei}, or {ku'o} at all... until the time they want to only close some clauses, and not all lol 06:30 < zipcpi> ... I just added {.i} (just to hyperlink it to {i}) to JVS, and it called it an "experimental cmavo" lol 06:47 < gleki> looks like i succeded in showing the structure of {ui uinaisairo'a} in altatufa 06:47 < zipcpi> alta: ui uinaisairo'a 06:47 < mensi> ([I {ui <ui (¹nai [sai ro'a]¹)>}] [FA ZOhE] [CU {COhE VAU}]) 06:47 < gleki> no, not here yet 06:47 < zipcpi> Hmm... next do TOIhE lol 06:47 < selpahi> internal grammar of UI 06:48 < selpahi> I should finish my internal grammar of quantifiers 06:48 < zipcpi> Yeah {toi'e} is a kurtyma'o, but it's actually useful, otherwise there's no way to unambiguously scope attitudinals over each other 06:48 < gleki> if/when mensi returns i can show you 06:49 < zipcpi> {i'au} used to be in {toi'e}. That cmavo's been through a lot lol 06:49 < gleki> although i still dont know how to name variable in altatufa 06:49 < zipcpi> used to be in TOIhE 06:50 < gleki> alta: ui ui nai sai ru'e ro'a 06:50 < mensi> ([I {ui <ui (¹nai [sai ru'e]¹)> ro'a}] [FA ZOhE] [CU {COhE VAU}]) 06:50 < zipcpi> Hm, ro'a doesn't attach to the second ui? 06:51 < gleki> i still dont know how to treat TAG nai 06:51 < gleki> it does attach, it's prettifier that is not pretty 06:51 < gleki> INT = expr:(FUhE_clause? (((UI_clause / DAhO_clause) !BU_clause MOD?) (END MOD?)* / free)+ FUhO_clause?){return _node("interjection", expr); } 06:51 < gleki> END = expr:(ROhA_clause !BU_clause) {return _node("interjection ending", expr); } 06:51 < gleki> MOD = expr:((NAI_clause / CAI_clause) !BU_clause MOD*) {return _node("interjection modifier", expr); } 06:51 < zipcpi> alta: banai ko'a broda 06:51 < mensi> SyntaxError: Expected [,] but "b" found. 06:59 < zipcpi> Selpahi: I upvoted {pinxydja} just to counter the Wuzzy's downvotes. I also defined {pixyjau} though, just because. 07:01 < gleki> it's because they were in post_clause, however, currently the idea is that NAI can't work on it's own 07:06 < gleki> so currently NAI can work as an interjection modifiers, in connectives, as vocative modifier (a somewhat separate case) and now for tense modals 07:07 < gleki> altA: banai mi mo 07:07 < gleki> alta: banai mi mo 07:07 < mensi> (I [{ba nai} mi] [CU {mo VAU}]) 07:07 < gleki> alta: basai mi mo 07:07 < mensi> SyntaxError: Expected [yY] but "o" found. 07:07 < gleki> ^ intentionally 07:08 < gleki> similarly {i mlatu sai} makes no sense. maybe it should be autorestored to {mlatu ge'esai} 07:12 < zipcpi> {barda sai} has use though 07:12 < gleki> what does it mean 07:12 < zipcpi> Basically anything that has a scalar interpretation 07:12 < zipcpi> {sai'e barda}, or {mutce lo ka barda} 07:13 < gleki> impossible. then it's not an interjection 07:13 < gleki> en: sai'e 07:13 < mensi> sai'e = [NAhE] scalar intensifier: very... |>>> rei'e:ru'e:milxe ; noi'e:na'oi:cnano ; sai'e:sai:mutce ; 07:13 < mensi> cai'e:cai:tcetce. See http://mw.lojban.org/papri/zipcpi:scalarmodifiers |>>> spheniscine 07:13 < zipcpi> I don't like the {ge'e} interpretatin 07:13 < gleki> sai'e:sai::na'e:nai then 07:14 < zipcpi> Yep... 07:14 < gleki> okay, 07:15 < gleki> only after sumti and after brivla NAI is the same as NAhE, just afterthought 07:15 < gleki> that'd be another issue 07:15 < zipcpi> sumtcita, sumti, and brivla 07:15 < zipcpi> Yeah the attitudinals case is an exception I noted 07:15 < zipcpi> I did add a note to the CKTJ page 07:16 < zipcpi> Wanting to swap {ne'e} with {tei}, then promote it as polar negator for attitudinals instead of {nai} 07:17 < gleki> after sumtcita NAI is sometimes like NA 07:17 < gleki> http://mw.lojban.org/papri/move_NAI_to_CAI#NAI_and_tags 07:17 < gleki> hm, probably the same as na'e ke 07:17 < zipcpi> Well, na is actually very much like na'e; it's just scope 07:18 < zipcpi> In that I think they mean the same thing; it's just the scoping that's different 07:18 < gleki> oh my i just recalled that ZI/VA are modifiers too, modifiers of sumtcita 07:18 < gleki> so many issues 07:19 < zipcpi> Heh... 07:19 < gleki> if bare NAI/CAI are not interjections this is confusing 07:20 < gleki> although inded they are not interjections 07:26 < gleki> then what could {mi jesai do mo} mean 07:26 < gleki> I and especially you? 07:26 < zipcpi> Maybe 07:26 < gleki> i wont allow that for now 07:27 < gleki> i temporaily named NAI/CAI compound {ckiluma'o} 07:28 < zipcpi> Oh I actually did use {zei'asai} / {dei'aru'e} 07:28 < gleki> okay i m reallowing it 07:31 < zipcpi> Hmm... maybe {basai} can mean "a long-ish time later", while {baza} repurposed. #miMUtcelokazbusuFUkai 07:32 < zipcpi> So no more bare bazi/baza/bazu; just baru'e, bana'oi, basai, bacai 07:33 < zipcpi> Well it'd fix your ZI problem at least 07:33 < zipcpi> But... zbusufukai 07:33 < clakre> why would you change that? the zi/za/zu - vi/va/vu parallels would be lost 07:33 < zipcpi> clakre: Just rambling... cause Gleki was talking about problems with how tags should scope over each other 07:34 < zipcpi> I'm not completely serious 07:34 < clakre> u'i 07:35 < zipcpi> Though I seem to be fated for my "not completely serious" proposals to get completely out of my control and suddenly gain popular, if controversial, support 07:35 < zipcpi> So caveat emptor :p 07:36 < clakre> zo'o I think I just don't want you to change the parts I already learned :p 07:36 < zipcpi> u'i 07:42 < gleki> the parts already learnt should never be changed unless they are utter bullshit, which still sometimes happens 07:44 < zipcpi> Though I am actually quite starting to like PU+CAI for that meaning 07:44 < zipcpi> I think I might try it out 07:45 < gleki> is it {pu ze'u} or {puzu} and why 07:46 < zipcpi> puzu, because pu indicates position in respect to time 07:47 < zipcpi> So {cai'e xo'i pu} is probably better interpreted as distance rather than duration 07:48 < gleki> still ZEhA are unfortunate to be a problem 07:48 < zipcpi> I don't think ze'u need modify pu 07:48 < zipcpi> Though ze'e... hm not sure how ze'e works 07:49 < gleki> en: ze'a 07:49 < mensi> ze'a = [ZEhA] time tense interval: a medium length of time. |>>> officialdata 07:49 < gleki> en: za 07:49 < mensi> za = [ZI] time tense distance: medium distance in time. |>>> officialdata 07:49 < selpahi> I can answer your questions about ze'e 07:49 < zipcpi> ze'a dosen't modify pu at all, IMO 07:49 < gleki> so only {za} will remain 07:50 < gleki> and i think i can only accept {zasai}/{zaru'e} 07:52 < gleki> {ze'e} is a perfectly understood cmavo. The point is that it's perfectly understood in two completely different ways by two groups of Lojbanists. 07:52 < zipcpi> So then get rid of ze'i / ze'u as well, and ve'i / ve'u, even vi/ vu. Everything is CAI 07:53 < zipcpi> Only perhaps {ti/ta/tu} remains :p 07:53 <@xalbo> Those are {zo'e pe vaCAIku}. 07:53 < zipcpi> Exactly lol 07:54 < gleki> en: tu 07:54 < mensi> tu = [KOhA6] pro-sumti: that yonder; distant demonstrative it; indicated thing far from speaker&listener. |>>> 07:54 < mensi> officialdata 07:54 < gleki> more than that 07:55 <@xalbo> je'e 07:55 < zipcpi> Oh right, demonstrative 07:55 < gleki> it should be both far from the speaker and the listener 07:55 < zipcpi> Well there's also that wrinkle. But yeah there's also the whole debate about exactly how loosely to interpret "demonstrative" 07:55 < selpahi> de zei monsuta 07:56 < zipcpi> u'i 07:56 <@xalbo> selpahi: Am I to understand with your latest NA idea that {na mi citka lo plise}, {mi na citka lo plise}, and {mi citka na lo plise} would all be truth conditionally the same, and the difference would be a lot like a {kau} ({du'u mi kau citka lo plise} etc)? 07:57 < selpahi> I think that would be useful probably, maybe. 07:57 < selpahi> :) 07:57 < zipcpi> Hm... perhaps it's a bit like what {xu} tags 07:57 < zipcpi> Except that {xu} tags things before it 07:57 <@xalbo> Yes, that's maybe even a better analogy. 07:58 < zipcpi> Being a UI 07:58 < gleki> If Lojban gets 1000 votes in the FB poll, LingQ will add support for the language to their online language learning site. <-- can anyone remember what happened then? they added Lojban. So ... 07:58 <@xalbo> So {na mi citka lo plise} and {mi citka na lo plise} are like the answers to {do xu citka lo plise} and {do citka lo plise ku xu} 07:59 < zipcpi> Yeah I think so 08:00 < zipcpi> Hm... how about... {mi nai citka lo plise}, {mi citka lo plise ku nai} 08:01 < zipcpi> exp: mi nai citka lo plise 08:01 < mensi> ([mi nai] [CU {citka <lo plise KU> VAU}]) 08:01 < zipcpi> off: mi nai citka lo plise 08:01 < mensi> SyntaxError: Expected [,] but "p" found. 08:01 < zipcpi> alta: mi nai citka lo plise 08:01 < mensi> SyntaxError: Expected [,] but "p" found. 08:02 < dutchie> in {mi chitka lo plise ku nai}, it seems to me that {nai} attaches only to {lo plise ku} 08:02 < dutchie> er, citka 08:02 < zipcpi> That's kinda the point 08:02 < dutchie> oh right 08:02 * dutchie looks at backscroll more thoroughly 08:03 < zipcpi> We are discussing this: http://selpahi.weebly.com/lojban/na-as-tag-re-visited#comments 08:08 < zipcpi> Hmm... does it mean though, that {na mi citka lo plise}; {na} only scopes over {mi}? So it'd be different from {naku mi citka lo plise} 08:13 < zipcpi> alta: mi broda 08:13 < mensi> (I [FA mi] [CU {broda NF} VAU]) 08:14 < zipcpi> alta: lo broda cu brode 08:14 < mensi> (I [FA {lo <broda NF> KU}] [cu {brode NF} VAU]) 08:16 < zipcpi> And to play even more with scope + NAI/CAI, there's {sei'o}, which is basically what alta represents by NF 08:18 < zipcpi> Hm, with careful placement, and even {fu'e...fu'o}, it might be possible to do very interesting things with scope 08:19 < zipcpi> And by interesting, I mean, brain-hurting :p 08:41 < zipcpi> Oh wait it was approval voting 08:41 < zipcpi> I think that means you were allowed to pick more than one option? 08:43 < zipcpi> Which is, well, not an ideal voting system (technically none are, because there are a bunch of criterion that most people would say "yes" to if asked if a fair voting system should have that feature, but it is found that it is mathematically impossible to fulfill all of them) 08:43 < zipcpi> But it's better than first-past-the-post 08:57 < gleki> sumti nai are not supported by altatufa atm 08:58 < zipcpi> Hm I think {xo'onku} - x1 is sarcastic in aspect/manner x2 08:59 < zipcpi> -nku comes from {lenku} and {kurki} 08:59 < gleki> only disambiguate between la bangu's def. of {xo'o} and {ckasu} 09:00 < zipcpi> Hmm... 09:00 < zipcpi> Maybe {xo'oksu} is better then? 09:00 < gleki> jb: xo'o 09:00 < mensi> xo'o = xo'o [interjection modifier] — verbal irony, xo'o nai — sincerely 09:00 < mensi> :je'u xo'o do ne se mau djuno — Yes, of course, you know better (sarcastically)! 09:00 < mensi> :xo'o nai sai do pu rau va'e gunka — With full sincerity, I can tell you that you worked enough. 09:00 < mensi> :Comment: For xo'o the intended meaning is the opposite of that expressed by the words used, in xo'o nai it explicitely 09:00 < mensi> marks the intended meaning as matching the words. Sarcasm isn't necessarily implied. 09:01 < gleki> do you mean verbal irony or mocking up? 09:01 < zipcpi> Hm... 09:01 < zipcpi> So that's how you interpret {xo'o}; interesting 09:02 < gleki> parallel to zo'o it's a modifier of interjections 09:02 < gleki> not a separate emotion 09:02 < zipcpi> So xo'o = verbal irony, while sarcasm is the use of verbal irony to {ckasu} 09:02 < gleki> wikipedia has more on verbal irony 09:03 < gleki> i think the second sarcam is better glossed with "to mock", yes 09:03 < gleki> i think the second sarcasm is better glossed with "to mock", yes 09:04 < zipcpi> Is there an attitudinal that specifically means "mock" 09:06 < latro`a> ga'a feels related 09:06 < gleki> ga'i 09:06 < latro`a> er, yes 09:08 < zipcpi> xo'onsku: x1 ironically says li'o... 09:09 < zipcpi> Hm {raxsku} is already defined 09:09 < zipcpi> {raxyckasu} too 09:14 < zipcpi> "generous" is in TMC list 09:14 < zipcpi> "stingy" is mobdu'a 09:17 < zipcpi> Hm how's {dudyxe'o} for "generous"? 09:20 < durka42> mi nelci i ie mi nelci 09:25 < zipcpi> $x_1$ is generous to $x_2$ in giving $x_3$ 09:26 < zipcpi> bo'a xendo bo'e lo nu bo'a dunda bo'e bo'i 09:26 < durka42> x1=d1=xe1 x2=d3=xe2 x3=d2=tu'a_xe1 09:26 < zipcpi> er, bo'a dunda bo'i bo'e 09:26 < durka42> jvomekso 09:27 < durka42> xu ka'e dunda lo temci 09:27 < gleki> alta: ui uinai ro'a nai ro'a dainai dai ui 09:27 < mensi> ([I {ui <(¹[ui nai] [ro'a nai] ro'a¹) (¹dai nai¹) dai> ui}] [FA ZOhE] [CU {COhE VAU}]) 09:27 < gleki> it's more complex than before 09:28 < gleki> ROhA now can string together, they dont modify each other 09:28 < gleki> DAI comes after them 09:28 < durka42> nice 09:28 < durka42> this is an internal grammar for UI? 09:28 < durka42> kinda like the internal grammar selpahi made for PA? 09:28 < zipcpi> Yep, based on some things I wrote a while back 09:29 < gleki> it's basically [(UI [ROhA+]) DAI(DAI(DAI...] 09:29 < gleki> so there are more classes now 09:29 < gleki> DAI and ROhA are different classes 09:29 < durka42> je'e 09:29 < gleki> and ... it starts to look like Ithkuil 09:29 < durka42> dai dai iei nu 09:30 < gleki> of course each of UI, ROA, DAI can be modified using NAI/CAI which dont belong to interjections at all 09:30 < zipcpi> Hm new style; move all the weird jvomekso into notes so that it will be less confusing for nintadni 09:30 < zipcpi> While still providing the often-useful information 09:31 < durka42> yes 09:31 < durka42> yes 09:31 < durka42> mi tugni i ie mi tugni 09:34 < zipcpi> http://jbovlaste.lojban.org/dict/dudyxe'o 09:35 < zipcpi> Oh yes, because of bo'V I no longer feel obligated to write {boi} in the lojban definitions anymore lol 09:36 < durka42> I never did, I figured if $x_2$ is smart enough to replace _ with {xi} then it can handle replacing $ with {boi}, heh 09:36 < gleki> and the problem i saw with DAI is that unlike other parts of interjection grammar they should have been parsed from right to left 09:36 < zipcpi> I could just go all the way and write it with bo'V and then trick JVS with $$, but... I dunno how well that would go with some people :p 09:36 < durka42> JVS would complain :) 09:36 < zipcpi> $$ 09:36 < durka42> you didn't answer my question from before though, "I am generous with my time" does {dunda} still work? 09:36 < durka42> (I lean towards yes) 09:37 < zipcpi> Oh didn't see that... 09:37 < zipcpi> I think it does work. It's might be pe'aru'e but I'm not sure how else to make it work 09:38 < zipcpi> After all you are giving up something 09:38 < zipcpi> For the benefit of someone else 09:39 < durka42> yeah 09:41 < zipcpi> http://jbovlaste.lojban.org/dict/guglyzamgri Example of how I tricked JVS with $$ (the definition I accidentally entered into Loglan instead of Lojban) 09:42 < zipcpi> So yes, I can write pure-bo'a or ce'u definitions if I wanted to. But... *shrugs* 09:43 < zipcpi> You see, like <insert appropriate political target>, all it cares about is the dollar signs :p 09:44 < durka42> can you imagine what Wuzzy would say, plus pe'i it's easier to read the numbers in writing 09:44 < durka42> even if it's worse in speech 09:44 < zipcpi> Inorite lol 09:44 < durka42> ei cliva 09:44 < zipcpi> co'o 10:13 < zipcpi> Oh crap... there might be a use for {pinxe zei cidja} over {pinxe zei djacu} 10:13 < zipcpi> "beverage" 10:14 < zipcpi> There is {selpinxe}, but then problems over ca'a vs ka'e again 10:14 < selpahi> Already thought of that. {djalitki} (for nutritious liquids). 10:14 < zipcpi> Ah 10:42 < Ilmen> en: ka'ei 10:42 < mensi> ka'ei'a = [PA5] digit/number: Conway's look-and-say constant λ ≈ 1.303577269… 10:46 < Ilmen> In this tu'e+NU thread, people seem to use {gi'i} for gek-terminator. But why not vu'o? 10:47 < zipcpi> ... wait what 10:47 < zipcpi> Why did mensi return ka'ei'a 10:48 < zipcpi> Ilmen: I'm not sure; it's Guskant's thing; ask her :p 10:49 < zipcpi> I'm not sure vu'o quite works though? It's not really a famyma'o 10:49 < Ilmen> I already wanted for a while a ka that automatically binds ce'u slots to predefined pronoun sets, however I haven't implemented this idea yet. It seems your ka'ei is in the same vein, am I mistaken? 10:50 < zipcpi> Yes it is 10:50 < Ilmen> I'd have made it a LOhOI though, in order to save one more syllable. 10:50 < zipcpi> Ah 10:50 < Ilmen> (just like xu'u = lodu'u) 10:51 < zipcpi> Part of the problem though is that you can't use {me'ei broda cu ka'ei ...} 10:51 < zipcpi> If it's only defined by a LOhOI 10:52 < Ilmen> Originally I planned to make two or three of these autoassigning ka with the same number of pronoun sets, but kurtis took some of the cmavo I planned to use 10:52 < zipcpi> Oh 10:52 < Ilmen> xD 10:52 < zipcpi> And then I took the coveted bo'V set for that lol 10:53 < zipcpi> I don't think it's a bad use though 10:53 < Ilmen> You can use {mintu ka'ei...} 10:53 < Ilmen> X mintu mi = X mibypre 10:54 < zipcpi> I've seen mibypre, but is {mi} always a prenu? 10:54 < Ilmen> so the choice of ka'ei being a NU or a LOhOI depends on what's the most commonly desired case: sumti or selbri 10:54 < zipcpi> Cause I've been told that {mi} was a suitable pronoun for Magic cards :p 10:54 < Ilmen> ANd I'd expect sumti being by far the most common use case 10:54 < zipcpi> *self-reference pronoun 10:55 < zipcpi> Ah 10:56 < Ilmen> In any case {lo ka'ei} will always be shorter than {lo ka bo'a bo'e bo'i bo'o... ce'ai} 10:56 < zipcpi> And {lo'oi} is just {lo poi'i} / CKTJ {lo voi}. Ah, so much reinventions of wheels lol 10:57 < Ilmen> Yeah. I made lo'oi before poi'i was added to JVS though. 10:57 < Ilmen> But poi'i is more versatile and can be used with gihek as well 10:57 < zipcpi> Mhm 10:59 < Ilmen> I think I've though of one ka assigining everithing to ko'V, another one assigning everything to fo'V, and possibly a third one assigning to a third pronoun set, if deemed useful enough 10:59 < zipcpi> I've seen someone use {li'o ce'ai} then FAbu lol 10:59 < Ilmen> .u'i 11:00 < Ilmen> With FAbu, you sometimes need to use boi, though 11:01 < zipcpi> Ah... I went through a different approach. First I invented {cei'e}, then I realized that it needs something; i.e. things that mean x1 to x5, then I realized that those things are really just reassigned ce'u, so a {ka} variant might be able to "steal" them 11:01 < zipcpi> And yes I noted the boi problem 11:02 < Ilmen> In the logic symbolic language, the variable name can be freely chosen, like ∀k [ broda(k) ] 11:02 < zahlman> coi rodo .i mi puziku glico se cfipu bu'u lo me'oi deli .i (to zarci lo rectu .e lo cirla .e lo salta .e lo se jukpa toi) .i ca lo nu pleji; lo ninmu cu te preti lo du'u xukau lo degtai ke rasyjukpa jipcyre'u cu co'e kei; zoi gy ... come out? gy 11:02 < Ilmen> which is basically {roldza fa lo ka ky ce'ai broda fa ky} 11:03 < zipcpi> Mhm. The thing is that the whole "x1~x5" thing is embedded into Lojban grammar, so I thought, why not use that set of five cmavo we were reserving for that 11:04 < zipcpi> Well, "why not" would be "to wait for something better"... but we can't do that forever :p 11:05 < Ilmen> ce'ai makes thing overly long to my tast though, whence the idea of autoassigning ka words 11:05 * zipcpi nods 11:05 < Ilmen> ce'u can gets ugly in complex nested abstraction with more than one ka level 11:06 < zipcpi> Exactly 11:06 < zipcpi> And cannot be reused unless assigned by goi 11:06 < zipcpi> or ce'ai 11:06 < zipcpi> Also cannot switch places easily 11:06 < zipcpi> So yeah lots of reasons 11:06 < zahlman> .i smuni fa lo ka te vecnu bredi .i ku'i mi pensi u'ise'i lodu'u smuni fa lo ka farlu fi lo tanxe 11:06 < Ilmen> {ce'u xi re} isn't even "the next ce'u of the second outer abstraction", it means "the second ce'u slot of the current ka" 11:07 < zipcpi> Huh, didn't know that 11:07 < Ilmen> The officially proposed solution is {ce'u xi X xi Z} 11:07 < zahlman> .i (to zo'o mi ba'e na pensi lodu'u xukau smuni fa lo ka mitcingle toi) 11:08 < Ilmen> which is uglily ugly 11:08 < zipcpi> fegli fi lo ka fegli 11:08 < zahlman> zo fegli u'i .i cnino xu 11:08 < Ilmen> ta'o je'e la .tsalman. 11:09 < Ilmen> .i cnino ĭe 11:09 < Ilmen> en: fegli 11:09 < mensi> fegli = x1 is ugly to x2 in aspect x3 |>>> See melbi. |>>> Ilmen 11:09 < zahlman> a'i clani lisri 11:12 < zipcpi> The thing that I like about bo'V is that by default, they mean x1~x5, so they can be used right away to render Lojban definitions. Then ka'ei allows you to wrap it all up in an abstraction for fancylojban purposes 11:15 < Ilmen> xN shows open slots in the predicate definition, so {ce'u} is totally suitable to render xN in Lojban definitions 11:16 < Ilmen> the only drawback is that ce'u doesn't show the slot number, so the reader has to count 11:16 < Ilmen> and you can't easily repeat back a previously used slot, as in "x1 talks about x2 to x1" (self-talk) 11:16 < zipcpi> Exactly 11:18 < Ilmen> za'a cilce xarju bu'u lo mi purdi 11:18 < zipcpi> All the same problems that {ce'ai} solves in the general case, but is too wordy for the commonly used case 11:20 < zipcpi> {ce'ai} still has some use; sometimes you might want to use a different set, sometimes you need more than one level... 11:20 < zipcpi> But yeah it's really nice to have a set that can be used by default 11:21 < zahlman> doi la ilmen ue do xabju xu lo cilce 11:21 < Ilmen> xabju fa mi fe lo cmana 11:22 < zahlman> lo nu da'i viska CX bu'u lo cmana u'e 11:25 < selpahi> Sort of relevant to your ce'ai and auto-binding {ka}: http://mw.lojban.org/papri/experimental_anaphorics 11:25 < Ilmen> sa'u me lo mamta xarju jo'u lo ri ci panzi gi'e darca tezu'e lo nu citka lo bilma fusra plise poi cpana lo cnita be lo tricu 11:26 < Ilmen> .i tepcau 11:27 < zipcpi> Oh yeah, how to fix {tu'e lo mabla cu fasnu .i lo zabna cu fasnu tu'u ca le nu mi stuvi'e le sralygu'e}? 11:28 < selpahi> By replacing le with lo. 11:28 * selpahi runs 11:28 < Ilmen> mi de'a jundi 11:29 < selpahi> Well the link I posted contains an option. 11:29 < selpahi> That I am not affiliated with. 11:30 < selpahi> But it's the postnex 11:31 < zahlman> u'e ui lo xarju cu sidju fi lo nu co'e lo na'e fusra 11:32 < zipcpi> {zo'au}? 11:33 < zipcpi> {tu'e lo mabla cu fasnu .i lo zabna cu fasnu tu'u zo'au ca le nu mi stuvi'e le sralygu'e} 11:33 < zipcpi> ? 11:33 < selpahi> I think that should work, since {ca ko'a zo'u tu'e ... tu'u} works 11:34 < gleki> Ilmen: coi do noi a'o na co'u morji fi lo uenzi poi nu'o spero 11:34 < zipcpi> It'll have to be a new selma'o right? 11:36 < gleki> what do we call "tanru unit" in Lojban? 11:36 < gleki> seljvatertau? 11:36 < zipcpi> selbrisle 11:36 < zipcpi> or ziltau 11:36 < gleki> en: ziltau 11:36 < mensi> ziltau [< zi'o tanru ≈ Nonexistent it phrase compound] = t1 is a tanru component. |>>> Cf. tanru, zi'o, seltau, tertau, 11:36 < mensi> selbrisle. |>>> totus 11:36 < gleki> en: tanru 11:36 < mensi> tanru = x1 is a binary metaphor formed with x2 modifying x3, giving meaning x4 in usage/instance x5. |>>> (x2 and x3 are 11:36 < mensi> both text or both si'o concept) See also gismu, smuni. |>>> officialdata 11:36 < gleki> ziltau breaks usual rule of using zil- 11:36 < zipcpi> zi'o tanru bo'a zi'o zi'o zi'o 11:37 < zipcpi> Yeah 11:37 < gleki> en: selbrisle 11:37 < mensi> selbrisle [< se bridi selci ≈ 2nd conversion predicate cell] = x1 is a tanru unit. |>>> selbri selci; ziltau |>>> 11:37 < mensi> gusnikantu 11:38 < gleki> im slowly translating all variable names to lojban 11:38 < gleki> which means soon i'll have hard time syncing from zantufa, spagetufa etc. 11:39 < zipcpi> selpahi: *sigh*... I dunno. If I have to rely on {zo'u} and {zo'au}, then {goi'e} might have to say in COI for now 11:39 < zipcpi> I really want it to be a sumtcita 11:41 < dutchie> u'isai i read that as "i really want to be a sumtcita" 11:41 < zipcpi> Yes, it's my life's dream 11:42 < zipcpi> Can I be ze'e? Or maybe fi'o zipcpi :p 11:43 < zipcpi> Clearly a very useful tag, that needs a BAI 11:43 < dutchie> ie zo'o 11:44 <@xalbo> Go for broke and combine them: be {fi'o ze'e zipcpi} 11:44 < zipcpi> YES 11:44 < gleki> alta: fanai mi nai mo nai 11:44 < mensi> (I [{fa nai} {mi nai}] [CU {mo nai} VAU]) 11:44 < zipcpi> u'i 11:44 < gleki> still doesnt work exactly as NAhE 11:45 < gleki> alta: mi na'e mo nai 11:45 < mensi> (I [FA mi] [CU {na'e <mo nai>} VAU]) 11:45 < gleki> actually {na'e mo nai} should be banned 11:45 < gleki> since it's not known who wins 11:45 < selpahi> na'e [mo nai] 11:45 <@xalbo> mi jai fi'o ze'e zipcpi fe'u co'e 11:46 < selpahi> mi jai fi'o no'a fe'u nei 11:46 < zipcpi> na'e ja'a mo nai ja'ai cau'i 11:46 < zipcpi> FIGHT 11:46 < zahlman> pe'i zo'ei lo'u na'e mo nai le'u ponjo sance 11:47 < zahlman> coi si oi lo cmavo tavla 11:47 <@xalbo> .i na'e nei 11:48 < zipcpi> lol xalbo, you might have just made the shortest Lojicomix :p 11:48 < zipcpi> {na nei}. *BOOM* 11:49 < selpahi> ja'a go'o .i na go'i 11:49 <@xalbo> {na nei} is an old koan, but I'm not convinced it's as bad as people think it is, due to the way Lojban has NA overwrite each other. {na'e nei}, though, may be irredeemable. 11:49 < gleki> {na'e ja'a mo nai ja'ai } <-- this is fine 11:50 < zipcpi> tersmus: na nei 11:50 < tersmus> !nei( ) 11:53 < zipcpi> ETERNAL PENGUINS FOR EVERYONE 11:57 < selpahi> zipcpi: Have you translated The Princess and The Pea yet? 11:57 < zipcpi> Huh? Is it like Sour Grapes where everyone translates it? 11:57 < zipcpi> I need to translate Sour Grapes myself lol 11:57 < selpahi> http://mw.lojban.org/papri/the_Princess_and_the_Pea 11:58 < zipcpi> Ah nice 12:03 < zipcpi> "However, don't change anyone else's translations at all." lol I totally wasn't thinking of doing that <.< >.> 12:03 < selpahi> .u'i 12:04 * xalbo goes off and deletes a few essential terminators to add subtle grammatical errors, then runs away giggling, looking for a puppy to kick. 12:05 < selpahi> .u'i ca lo fanmo co'a cmila fa mi 12:05 < gleki> btw i already broke altatufa 12:05 < gleki> alta: puku mo 12:05 < mensi> SyntaxError: Expected ['h] or ['hiI] but end of input found. 12:05 < selpahi> Expected ['hil] oO 12:06 < gleki> alta: puzo'e mo 12:06 < mensi> (I [pu zo'e] [CU {mo VAU}]) 12:30 < zipcpi> doi gleki do ciskysre 12:30 < zipcpi> lo'u <le loxu pu bikla> le'u 12:49 < bigcentaur> coi 13:10 < zipcpi> mi mo'u fanva la'au loi slari vanjba li'u 13:13 < zipcpi> And... yeah. I didn't use as many experimental constructs as I did in that joke snippet, I still used a lot of them lol 13:13 < zipcpi> *but I still 13:20 < zipcpi> http://mw.lojban.org/papri/Sour_Grapes#loi_slari_vanjba 13:20 < zipcpi> xu mabla lo ka pilno lo malma'o iau u'i 13:25 < zipcpi> And cutting short that {nu} for an anaphora might be cheating lol 13:25 < durka42> I should make my own translation before reading yours, shouldn't i 13:25 < zipcpi> Maybe 13:27 < durka42> I'm doin it 13:31 * nuzba @s_i_d_a: carvi sisti .i zasni tolxanka tsani .i co'a cerni #lojban #haiku [http://bit.ly/1JkZkZx] 13:31 < zipcpi> Er, you might want to do it in a notepad or something 13:31 < zipcpi> I just edited my own translation and I hear the wiki does bad things 13:32 < durka42> yes I am using vim of course 13:32 < durka42> and la vlaste 13:32 < zipcpi> u'i 13:41 < Ilmen> en: lau'u 13:41 < mensi> lau'u = [KOhA2] pronoun: the referent of the last utterance |>>> Equivalent to la'edi'u or lo su'u go'i; compounds 13:41 < mensi> considered common enough in usage statistics to deserve a shortening. |>>> spheniscine 13:42 < zipcpi> la .guskant. cu finti 13:43 < zipcpi> Well, she took the idea from the CKTJ page, which wanted {lau} to be {la'edi'u} 13:43 < Ilmen> mi co'u morji .i du'e cmavo ba'o se finti .i du'e cmavo pu'o go'i 13:43 < zipcpi> So she actually defined a cmavo for it, so that CKTJ will have something to swap with 13:43 < zipcpi> u'i 13:44 < durka42> haha 13:44 < Ilmen> tau zo ki'a na banzu 13:46 < zipcpi> ki'a'a ki'a'e ki'a'au ki'a'ai'a 13:46 < zipcpi> ki'a'o'e'o'e <-- elliptical metalinguistic confusion 13:46 < Ilmen> .u'i 13:48 * nuzba @evegwood: the wiki page for lgbt flags has a one line description of what the flags are in 10 languages, and one of those languages is lojban?? wh [http://bit.ly/1Jl2ixa] 13:49 <@xalbo> Make a complete set of ten ki'a'V'V, where the first V is 1-5 which of the previous five utterances was confusing, and the second is which term in it was confusing. 13:50 < zipcpi> u'i 13:50 <@xalbo> Maybe a third set of 'V corresponding to the ro'V scale for whether it was emotionally confusing, physically, etc. 13:50 < zipcpi> u'i 13:50 <@xalbo> s/ten/25/ 13:50 < fenki> xalbo meant to say: Make a complete set of 25 ki'a'V'V, where the first V is 1-5 which of the previous five utterances was confusing, and the second is which term in it was confusing. 13:51 < durka42> haha wat 13:51 < durka42> https://commons.wikimedia.org/wiki/Sexual_identity_symbols#Flags 13:51 < durka42> (SFW) 13:52 < zipcpi> Awr I was thinking of something much less ambitious. Just define a joke cmavo, {ki'a'au'u'au'i} (modeled after {ci'au'u'au'i}), meaning "metalinguistic confusion caused by too many experimantal cmavo" 13:53 < durka42> .i'e .u'i 13:55 * nuzba @evegwood: lojban more like lojBANNED. because it sucks [http://bit.ly/1gEaZMf] 13:58 < zipcpi> http://jbovlaste.lojban.org/dict/ki'a'au'u'au'i 14:14 < durka42> my translation is very gi-heavy for some reason 14:15 < zipcpi> lol 14:16 < zipcpi> ei mi cliva co'o 14:16 < durka42> co'o 14:26 < durka42> jo'au la .jacus. noi terjonma'o ciste je lo cnino me zo soi je lo ga'ezeilerfu basyvla 14:26 < durka42> .u'i 14:26 < durka42> did I miss anything 14:26 < durka42> do I have to jo'au la .dotsaid. ? 14:29 < jacus> the what you say? Got a couple of things in there I don't recognize. ;-) 14:30 < durka42> coi :) 14:30 < durka42> xu do na terjonma'o ciste 14:32 < jacus> [ struggling to say "I should think not" with the right attitude bau lojbo ] 17:51 < vultrue> Ìû 18:27 < pimlu> coi ma 19:47 < noncomcinse> en: na'oi 19:47 < mensi> na'oi = [CAI] Indicator for moderate or normal attitudinal intensity |>>> Used to indicate the "default" intensity of 19:47 < mensi> attitudinals, the intensity that is implied when for example using a bare "ui" (which is in-between "sai" and "ru'e"). 19:47 < mensi> Useful mostly together with pei. Derived from cnano. |>>> ctefaho 19:48 < noncomcinse> .i pau ma'oi cai zo'u lo'u mo ru'e na'oi sai cai le'u valsi 22:10 < zipcpi> Hmm... I find in practice, I use {lo} to introduce a new object into the discourse; {lo bi'u}, or an informal {da poi} but without all the realis and scope problems of {da} that xorlo is supposed to solve 22:11 < noncomcinse> coi zy. 22:11 < zipcpi> {le} refers to something present to the discourse. Something present in both the speaker's and listener's minds 22:11 < zipcpi> Thus could be anaphora, cataphora, exophora 22:12 < noncomcinse> Exophora? 22:12 < zipcpi> Did you like *the play*? (probably the one that you just watched) 22:13 < zipcpi> It means, referring outside of the text, to something in the subtext 22:13 < noncomcinse> Is it acceptable to use {lo} for that? 22:13 < zipcpi> Yes... {lo} is completely generic 22:14 < zipcpi> I'm just observing how I personally distinguish {lo} from {le} 22:14 < zipcpi> In my own usage 22:14 < zipcpi> {lo}, post-xorlo has no bound semantic though; it's just "selbri-verb to sumti-noun" 22:16 < noncomcinse> Does zo su have an experimental use or grammar? 22:16 < zipcpi> It's {su'o} under cekitaujau dialects 22:16 < zipcpi> http://mw.lojban.org/papri/ce_ki_tau_jau 22:17 < noncomcinse> i 22:17 < noncomcinse> I almost think that's going a bit too far, breaking an i-a-u pattern. 22:17 < noncomcinse> ti ta tu, ri ra ru, si sa su, 22:17 < zipcpi> Oh 22:18 < zipcpi> I was going to say something about {su'o} not being part of such a pattern :p 23:21 < gleki> alta: puku mi mo 23:21 < mensi> (I [{pu ku} {FA mi}] [CU {mo VAU}]) 23:21 < gleki> fixing => breaking => fixing 23:26 < zipcpi> The human condition, folks 23:27 < gleki> im trying to solve the major problem of programming. 23:27 < gleki> for now only for lojban 23:31 < gleki> loglan: speni 23:31 < mensi> speni = D experience/undergo/spend (interval) V. 23:31 < gleki> Ei tu speni lo ka se speni :P 23:40 < zipcpi> exp: ca gi xagji gi co'a sanji lo jisryclu 23:40 < mensi> (CU [ca gi] [CU {xagji VAU}] gi [CU {co'a sanji} {lo jisryclu KU} VAU] VAU) 23:41 < zipcpi> I'm not aware of this use of gi 23:42 < gleki> mw.lojban.org/papri/tags_as_connectives 23:42 < zipcpi> Hm seems useful 23:43 < zipcpi> off: ca gi xagji gi co'a sanji lo jisryclu 23:43 < mensi> ([ca gi] [xagji VAU] gi [{co'a sanji} {lo jisryclu KU} VAU] VAU) 23:44 < gleki> exp: je gi mo gi mo 23:44 < mensi> (CU [je gi] [CU {mo VAU}] gi [CU {mo VAU}] VAU) 23:44 < gleki> alta: jegi mo gimo 23:44 < mensi> SyntaxError: Expected [uU] or [yY] but "i" found. 23:44 < gleki> ofc. 23:45 < zipcpi> lol Mad proposals 23:45 < zipcpi> I actually think it might not be so mad though. I mean it's just like {ju'e gi} 23:45 < zipcpi> Or {joi gi} 23:48 < zipcpi> exp: lo lorxu poi xagji co'a sanji lo gunma be lo vanjba poi sai'e si na'e se jisra be'o poi ga'usai banro ra'i lo skospa 23:48 < mensi> ([lo {lorxu <poi (¹CU [xagji VAU]¹) KUhO>} KU] [CU {co'a sanji} {lo <(¹gunma [be {lo <vanjba (²poi [sai'e si]²) (²CU [na'e {se jisra}] VAU²) KUhO> KU} be'o]¹) (¹poi [CU {<ga'u sai> banro} {r 23:48 < mensi> a'i <lo skospa KU>} VAU] KUhO¹)> KU} VAU]) 23:59 < zipcpi> coi durka42 23:59 < zipcpi> lol that {gi} is interesting --- Day changed Fri Jul 17 2015 00:00 < zipcpi> {ki'u lo nu mi broda kei mi brode} -> {mi ki'ugi broda gi brode} 00:00 < zipcpi> WOW 00:00 < zipcpi> So useful 00:02 < durka42> basically, I hate {kei} 00:03 < zipcpi> And I might start to hate it too :p 00:03 < zipcpi> After all I did use {iau} to close two {kei} 00:07 < zipcpi> Under my idea of how two-letter rafsi work, which I haven't written down yet... sezykufygau -> kufrygase 00:07 < durka42> ua 00:07 < durka42> also, using the tag-connective staile makes the x1s more explicit 00:08 < durka42> (without repeating them) 00:08 < zipcpi> bo'a kufrygase -> bo'a sefsi lo ka ce'u kufryga ce'u -> bo'a sefsi lo ka ce'u gasnu lo ka ce'u kufra 00:08 < durka42> but that {... gi te zu'e gi ...} is a bit ridiculous 00:08 < durka42> yeah I like that 00:08 < zipcpi> Basically, ca'e Lojban is now agglutinative :p 00:09 < durka42> heh it basically already was through the lujvo system 00:09 < durka42> multi-paradigm language zo'oru'e 00:09 < zipcpi> lol 00:10 < durka42> the problem with CV rafsi is it's an even smaller space 00:10 < durka42> -se goes to {sefsi} or {selja'e} or what 00:12 < zipcpi> True... 00:13 < gleki> i hate kei but i use vau :D 00:14 < zipcpi> Maybe sefsi should be -si 00:15 < durka42> simxu is -su? 00:15 < durka42> I'm sure tsani or Ilmen has a putative list of these somewhere :) 00:15 < zipcpi> lol 00:15 < zipcpi> I thought it was la cirko who had one :p 00:15 < zipcpi> I also wanted -la for {klani} 00:16 < zipcpi> So man-hours = {pi'ai gunkyla cacra} 00:16 < zipcpi> Can convert anything into a unit 00:23 < zipcpi> The problem right now with regularized lujvo is that the order of operations aren't regularized 00:24 < zipcpi> Also the JVS chaos problem :p 00:27 < zipcpi> exp: ca gi jangalgau gi tezu'e gi sezykufygau gi cusku 00:27 < mensi> (CU [ca gi] [CU {jangalgau VAU}] gi [CU {<te zu'e> gi} {CU <sezykufygau VAU>} gi {CU <cusku VAU>} VAU] VAU) 00:27 < zipcpi> zantufa: ca gi jangalgau gi tezu'e gi sezykufygau gi cusku 00:27 < zantufa> (CU [ca gi] [jangalgau VAU] [gi {<te zu'e> gi} {sezykufygau VAU} {gi <cusku VAU>} GIhI VAU] GIhI VAU) 00:27 < zipcpi> u'i 00:27 < durka42> it doesn't break 00:28 < zipcpi> Oh right because it's in the second place 00:28 < durka42> well JVS won't let you enter relyle'ujvo right now 00:28 * zipcpi nods 00:29 < durka42> tersmus: ly ca gi jangalgau gi tezu'e gi sezykufygau gi cusku 00:29 < tersmus> EX x1. (x1=. jangalgau(l) /\ (ca)(x1). EX x2. (x2=. sezykufygau(l) /\ (te zu'e)(x2). cusku(l))) 00:29 < durka42> wat .u'i 00:29 < durka42> tersmus: jbo: ly ca gi jangalgau gi tezu'e gi sezykufygau gi cusku 00:29 < tersmus> su'o da zo'u ge fi'o du da zo'u ly jangalgau gi ca da su'o de zo'u ge fi'o du de zo'u ly sezykufygau gi te zu'e de zo'u ly cusku 00:30 < durka42> a wild existential quantifier appeared!! 00:30 < zipcpi> Though perhaps then, relyle'ujvo would have to be treated a bit like {za'ure'u} / {baze'e} cmavo clusters. Not *defined* in terms of any natural language construct; the definition is only meant to tie it with a simple concept found in English (or whatever natural language you can find) 00:30 < zipcpi> Like {certyzegasi} ~ "practice" 00:31 < zipcpi> (which we already have {crezenzu'e} / {crenzuue} for, which incidentally I'm a bit confused as to whether {zukte} should actually work that way 00:32 < durka42> -ze = ? 00:32 < zipcpi> zenba 00:33 < durka42> sefsi lo ka gasnu lo ka zenba lo ka certu? 00:33 < durka42> I don't see where {sefsi} comes from 00:33 < durka42> why not just certyzezu 00:34 < zipcpi> bo'a gasnu lo nu bo'a 00:34 < zipcpi> bo'a gasnu lo nu bo'a zenba 00:34 < durka42> ah 00:34 < durka42> still seems excessive 00:35 < zipcpi> That might be part of the problem. Is {se zukte} always a ka tying it to x1, or it is just {gasnu} with a purpose? 00:35 < durka42> the second one 00:35 < zipcpi> Yeah, hence the need for the sefsi... 00:35 < durka42> yeah I guess 00:35 < durka42> I'm not sure it needs to be so explicit 00:36 < zipcpi> That's what we have zi'evla for 00:37 < durka42> nah 00:37 < durka42> it's impossible to always include all the necessary rafsi in a lujvo 00:37 < durka42> that way leads to German 00:37 < zipcpi> u'i 00:38 < zipcpi> I hear a common kid's game in German is basically la jvozba selkei 00:41 < zipcpi> The thing is that in {certyzega}, -> bo'a gasnu lo nu ba'e*bo'e* zenba lo ka certu bo'i 00:42 < zipcpi> So the -si- would be needed to condense x1 and x2 00:43 < durka42> yeah I get it 00:43 < durka42> jbo:crenzuue 00:43 < mensi> crenzuue = .i lo si'o crenzuue cu si'o ce'u goi x1 zukte lo ka ce'u zenba lo ka ce'u certu ce'uxici goi 00:43 < mensi> x2 00:43 < zipcpi> Crap, ce'u-goi staile 00:44 < durka42> oisai 00:44 < durka42> now I forget what ce'u-subscripts mean 00:44 < zipcpi> Exactly why I took bo'V lol 00:44 < durka42> I think ce'uxici means "jump three levels" 00:44 < durka42> well, two levels, since one-indexing 00:44 < durka42> so it's back up on the level of the first ce'u 00:45 < durka42> so yeah it's correct 00:45 * zipcpi faceplants 00:46 < zipcpi> I think gasnu + sefsi would be a frequent combination 00:47 < durka42> or zukte + sefsi 00:47 * zipcpi nods 00:49 < zipcpi> Also my thinking is that any brivla can be mutated by 'yGA 00:49 < zipcpi> broda'yGA 00:49 < zipcpi> sormei'yGA 00:49 < durka42> that might work... 00:49 < durka42> unless it breaks the morphology somehow 00:50 < zipcpi> Gismu is special in that their final vowel can be elided. Also any lujvo that ends with full-gismu-rafsi 00:50 < durka42> sure 00:51 < durka42> I'm just worried this will introduce new classes of tosmabru and slinku'i and stuff 00:51 < zipcpi> Yeah maybe... I don't think it does though. 'yX already is used for zi'ervlajvo, and combinations of relrafsi aren't valid zi'evla 00:52 < durka42> does {setca'yga} break up? 00:52 < zipcpi> seTCA'yga 00:52 < durka42> {setca'iga} breaks 00:53 < zipcpi> setca'i'yga 00:53 < durka42> well -y- isn't accepted in zivla right now so we can't check 00:53 < durka42> but as long as {tca'yga} isn't valid... 00:53 < durka42> spagetufa: setca'yga 00:53 < spagetufa> (CU [setcà'y/ga VAU]) 00:53 < durka42> wat 00:53 < zipcpi> I tried coining {zivla} but it clashes with {zivle} 00:54 < durka42> ah 00:54 < zipcpi> So I had to settle for {zevla} :p 00:54 < durka42> vlaste: zivla (janli) 00:54 < vlaste> no results. http://vlasisku.alexburka.com/zivla 00:54 < zipcpi> vlaste: zivla (gimka) 00:54 < vlaste> zo zivla gimkamsmikezypro zo zivle noi catni gismu .i zo zivla gimkamsmikezypro no cipra gismu 00:55 < durka42> vlaste: zivla (janli) 00:55 < vlaste> zo zivla gimkamsmikezypro zo zivle noi catni gismu .i zo zivla gimkamsmikezypro no cipra gismu 00:55 < durka42> vlaste: zevla (janli) 00:55 < vlaste> zo zevla gimkamsmikezypro no catni gismu .i zo zevla gimkamsmikezypro no cipra gismu 00:59 < zipcpi> So yeah now they all have gismu. gismu, cmavo, lujvo, zevla, brivo, cmevo. 01:01 < zipcpi> And apparently {.i} is an experimental cmavo :p 01:01 < zipcpi> http://jbovlaste.lojban.org/dict/.i 01:11 < zipcpi> spagetufa: setca'i'yga 01:11 < spagetufa> (CU [set/ca'ì'y/ga VAU]) 01:12 < zipcpi> spagetufa: spageti'yga 01:12 < spagetufa> (CU [spagetì'y/ga VAU]) 01:12 < zipcpi> spagetufa: sormei'yga 01:12 < spagetufa> (CU [sor/mèi'y/ga VAU]) 01:42 < zipcpi> spagetufa: certyzegasi 01:42 < spagetufa> (CU [certy/ze/gà/si VAU]) 01:46 < zipcpi> http://jbovlaste.lojban.org/comments.html?valsi=29376;commentid=2320;definition=0 <- tu krinu lo nu mi no'e nelci le si'o jvomi'u 01:46 < durka42> spagetufa: spagetufa 01:46 < spagetufa> (CU [spagetùfa VAU]) 01:46 < durka42> is Wuzzy being annoying again 01:47 < zipcpi> ma'i do cau'a go'i 01:49 < durka42> that a search for {pinxydja} should return {pixydja} is a valid point (and it is already working in every search engine besides jbovlaste 01:50 < durka42> but downvoting isn't the solution 01:50 < durka42> ) 01:50 < zipcpi> But... yeah... that's kinda part of the problem. Needing to always use the lowest-scoring jvova'i forms means that e.g. {zi'ervlajvo}, which already has to be mutated to avoid tosmabru, would have to be {zifyvlajvo}. 01:50 < zipcpi> mutce lo ka mabla 01:50 < durka42> jvomi'u doesn't mean you have to use the lowest scoring form 01:50 < durka42> where did you get that idea 01:50 < durka42> you can use the one that sounds best, without fear that it's actually different word! :) 01:51 < zipcpi> Right I mix up jvomi'u vs jvova'i 01:51 < zipcpi> But yeah not being able to search for jvomi'u forms is a problem 01:51 < zipcpi> And jvova'i probably should die except as a kind of suggestion 01:54 < zipcpi> No one can calculate that in their heads anyway 01:54 < durka42> it never was more than a suggestion 01:54 < durka42> that "database sweep" that jvs talks about has never been done and never iwll be 01:54 < zipcpi> lol 01:55 < zipcpi> It will be done some time after CLL v2 and Half Life 3 comes out :p 01:58 < zipcpi> lo zu'o ni'udzau nasai danfu 01:58 < zipcpi> (Does anyone ever use {zu'o}? lol 01:58 < zipcpi> ) 01:59 < zipcpi> I used it a lot when first learning Lojban 01:59 < zipcpi> Mostly in place of {ka} before I understood {ka} 02:01 < gleki> durka42: when i search for {klamygau} in vlasisku it doesnt output klagau 02:02 < durka42> oh was I lying 02:02 < gleki> sutysisku and vlasisku only do the split 02:02 < durka42> vlaste does it 02:02 < vlaste> 113 results: frili, gasnu, mirli, ractu, ceirnalju'o, da'erzu'e, deptsa, flizu'e, jaxpu'a, jibykai… 02:02 < zipcpi> oi ko se tacpe'u lo xagji sofybakni 02:03 < durka42> ue my vlasisku is broken somehow 02:03 < gleki> vlaste is not vlasisku 02:03 < vlaste> 16182 results: notci, bakfu, bakni, bakri, baktu, balji, balni, balre, balvi, bancu… 02:04 < gleki> maybe i should implement such thing for sutysisku 02:04 < gleki> en: klamygau 02:04 < mensi> klagau [< klama gasnu ≈ Come(s) bring about] = x1 brings x2 to x3 02:08 < zipcpi> imo relyle'ujvo is not meant to completely replace -gau and friends. It still does have some use for lujvo like this: http://jbovlaste.lojban.org/dict/rebdesygau 02:08 < zipcpi> And common ones like {muvgau} might still be used 02:08 < zipcpi> But yeah rebdeskyga / rebdeskygasi won't work because then we'd have to define {rebdesku}, which might not be a very useful concept 02:09 < Ilmen> sligau lo rebla? 02:09 < Ilmen> en: tolbirti 02:09 < mensi> [< to'e birti ≈ Polar opposite certain] lo nu mulno sisku zo'u: y no da se tolcri 02:09 < mensi> tolbirti[7937], to'erbirti[9526] 02:11 * nuzba @_Vanessa_sary_: @willingtheweird But a somewhat decent conversation I suppose. I'm not good at Lojban, really. I stuck with French [http://bit.ly/1I88KwY] 02:11 < Ilmen> Why do you think that the polar opposite of "being certain that X is true" is "being certain that X is false"? Why wouldn't it be "not being certain at all"? 02:12 < zipcpi> Hm that's part of the problem with scales isn't it 02:12 < Ilmen> The basis for the proposal of {ju'onai} for incertitude was that to'e birti is incertitude 02:12 < Ilmen> But at any rate to'e birti is undefined 02:12 < zipcpi> Right 02:13 < Ilmen> ta'o coi 02:13 < zipcpi> coi 02:13 < zipcpi> Yeah I have a feeling that scales might be ambiguous in some situations 02:13 < zipcpi> But I'm not very good at actually coming up with a clear example 02:14 < zipcpi> But yeah perhaps it is quite possible for {to'e broda} to be interpreted very differently, depending on what {broda} is 02:15 < Ilmen> I have a feeling that to'e+selbri (polar opposition) is not fully compositional, as if the meaning of to'e+selbri should be part of that of {selbri} or defined separately 02:15 < Ilmen> but maybe I'm wrong 02:16 < zipcpi> In fact that's part of the problem I noted with {ba'a}, because the scale there works very differently from most of the others 02:16 < zipcpi> Well my motivation is comparing it with the other evidentials {UI2} 02:17 < Ilmen> Yes, {ba'a} is an oddball 02:17 * zipcpi nods 02:17 < zipcpi> That's why I made {kai'e}, {lai'i}, {vei'i}, and {moi'i} 02:18 < zipcpi> Though... cmavo proliferation 02:18 < zipcpi> ki'a'au'u'au'inmo fa roda 02:18 < durka42> you will be subject to economic sanctions from the signatories of the Cmavo Non-Proliferation Act of 2010 02:18 < Ilmen> I also wonder whether {cnino} is equal to {to'e slabu} or {to'e fange}. 02:18 < zipcpi> to'e laldo you mean? 02:19 < Ilmen> .u'i 02:19 < Ilmen> to'e laldo is citno. 02:19 < zipcpi> Hm 02:19 < Ilmen> lo ka to'e laldo cu ka citno 02:20 < Ilmen> er, earlier I meant: 02:20 < Ilmen> I also wonder whether {slabu} is equal to {to'e cnino} or {to'e fange}. 02:20 < Ilmen> Silly me. 02:21 < zipcpi> Ah 02:22 < zipcpi> xu do djuno lo du zo .i je'a cipra cmavo // http://jbovlaste.lojban.org/dict/.i 02:22 < Ilmen> I guess that to'e cnino is "to be known since a long time" 02:23 < Ilmen> While "to'e fange" would be just "X is familiar to Y (regardless of how long it has been so)" 02:23 < zipcpi> .i: experimental cmavo 02:23 < zipcpi> Guess before I came along, no one separated their sentences. It all just ran together 02:23 < Ilmen> or something like that 02:24 < zipcpi> Hm 02:24 < Ilmen> jbo: fange 02:24 < mensi> fange = x1 frica lo se tcaci be x2 x3 |>>> frica; tcaci; cizra; rirci; fadni |>>> 02:24 < mensi> xorxes 02:25 < Ilmen> jbo: cnino 02:25 < mensi> cnino = lo nu x1 ckaji x3 cu pu zi co'a se sanji x2 .i x1 to'e slabu x2 x3 |>>> xorxes 02:25 < Ilmen> jbo: slabu 02:25 < mensi> slabu = lo nu x1 ckaji x3 cu pu zu co'a se sanji x2 .i x1 to'e cnino x2 x3 |>>> ckaji; sanji; cnino |>>> 02:25 < mensi> xorxes 02:25 < zipcpi> ca'edai doi xorxes 02:25 < durka42> .i'e 02:25 < Ilmen> en: fange 02:25 < mensi> fange = x1 is foreign/alien/[exotic]/unfamiliar to x2 in property x3 (ka). |>>> For the sense of alien (not from Earth), 02:25 < mensi> see kesfange. See also cizra, jbena, ckini. |>>> officialdata 02:27 * nuzba @willingtheweird: @_Vanessa_sary_ You have yet to comprehend the beauty of lojban. .i do na jimpe le melbi pe la lojban [http://bit.ly/1Jmqkb3] 02:28 < zipcpi> Isn't it {le te melbi be fi la .lojban.} 02:28 < durka42> yes 02:28 < durka42> lo te melbi cu nenri lo kanla be lo melbi 02:29 < zipcpi> :p 02:29 < durka42> no wait 02:29 < durka42> lo te melbi cu nenri lo kanla be lo se melbi 02:29 < durka42> that's how to remember the places, maybe :) 02:29 < Ilmen> .u'i 02:30 < zipcpi> Funnily on the TMC list there is "Red is a pretty color"; I translated it as {lo ka xunre cu melbi}, but I'm not sure if people would be screaming type error lol 02:30 < durka42> how about {lo ka xunre cu te melbi} 02:30 < zipcpi> fi ma 02:31 < durka42> exactly 02:31 < durka42> or just {melbi skari}, I guess that doesn't satisfy the TMC list somehow‽ 02:31 < zipcpi> Dunno lol 02:31 < durka42> lo ka xunre cu te melbi -- redness is a way in which some things are pretty 02:31 < zipcpi> TMC list: https://docs.google.com/spreadsheets/d/1LH_bawafDZBp5h4shGoajnTU7sttOilfpyxqUD4B5BM/edit#gid=0 02:31 < zipcpi> Hm 02:32 < durka42> ooh the first one looks like a good use case for {tankomo} :) 02:32 < zipcpi> lol tankomo 02:32 < zipcpi> ma vlakra 02:33 < zipcpi> lo spanybau 02:33 < durka42> go'i 02:33 < durka42> vlaste: tankomo 02:33 < vlaste> tankomo = x1 (agent) does x2 (property of x1) as much as it/he/she/they x3 (relation between x1 and x2) 02:33 < zipcpi> Wait a minute... 02:34 < durka42> maybe not 02:34 < zipcpi> http://jbovlaste.lojban.org/dict/toltankomo 02:34 < durka42> vlaste: toltankomo 02:34 < vlaste> tankomo = x1 (agent) does x2 (property of x1) as much as it/he/she/they x3 (relation between x1 and x2) 02:34 < zipcpi> ... but the English definition is in "lojban" o.o 02:34 < durka42> ah what 02:34 < durka42> yes 02:34 < zipcpi> oi oi oi oi oi 02:34 < durka42> that should be fixed lol 02:34 < Ilmen> xD 02:35 < durka42> Ilmen want to fix that? 02:35 < Ilmen> vi'o mi'e fuzme 02:35 < durka42> sa'u mi na djica lo nu ki'ugi cikre gi finti simlu 02:35 < durka42> ka* 02:37 < Guest12507> so 02:37 < durka42> coi vitke 02:39 < Ilmen> mulfau 02:39 < durka42> ki'e 02:40 < zipcpi> coi la cmecau vombatu 02:40 < Ilmen> nau zo'u .e'u gau do lo fagri cu carvi lo lojbo glibau velski 02:41 < Ilmen> coi la so rau mei 02:44 < zipcpi> http://jbovlaste.lojban.org/dict/cmecau 02:45 < Ilmen> .i'e terbri krasi notci 02:45 < Ilmen> ja'e lo nu lo velski na se terpa lo cnino vitke 02:46 < zipcpi> Exactly 02:46 < zipcpi> Provides the useful information, without cluttering up the definition 02:46 < Ilmen> zo cmecau pu so'i roi se pilno 02:47 < Ilmen> .i so'i seju so'o 02:48 < zipcpi> ca ciska lo jbobau velcki 02:48 < zipcpi> mo'u 02:48 < zipcpi> x1 claxu lo cmene be x1 bei x2 .i va'i noda cmene x1 x2 02:49 < zipcpi> lol example of xorlo 02:49 < zipcpi> Otherwise that would be a contradiction 02:49 < Amnadya> Attàwîl. 02:51 < Amnadya> x1 claxu lo ka su'o da cmene ce'u x2 02:51 < zipcpi> Yeah I guess that's one way to avoid irrealis-lo 02:54 < zipcpi> Oh you defined claxu-2 to be a ka 02:59 < Amnadya> coi pinxydja zvasku 03:02 < Amnadya> doi la cliva 03:02 < Amnadya> .i ke'u ke'u do mo la durka kunoi zvati lo dotygu'e ca'o 03:09 < cliva> je'e je'e mojgau .i la'a mi lazni ja me'au lo simsa 03:10 < Amnadya> ma smuni .i xu platu su da 03:10 < Amnadya> .i xu tolmo'i 03:10 < Ilmen> la .durkas. nu'o mlauca lodu lifri makau 03:10 < Ilmen> pe lo dotygu'e 03:10 < cliva> .aidji lo ka jungau fe ji'a lo drata ke tsurixi lojbo 03:11 < Amnadya> .i sa'u mi nu'o zu'ejdi lo ka klama .i da'i lo nu do klama cu banzu lo nu lo nu mi klama cu lakne 03:11 < Amnadya> .i lo sorpeka zo'u cacra li bi 03:11 < Ilmen> .i pensi lo nu pensi'u ma 03:11 < Amnadya> la'a la .franfurt. (to pau toi) 03:11 < Ilmen> wa 03:11 < gleki> Ilmen: ma nuzba lo spero te traduki 03:12 < cliva> pau se'i ja'ai 03:12 < Amnadya> ga ja la .franfurt. gi la .tubinen. 03:12 < Ilmen> .i ca pi so'i lo prulamdei mi skami darno .i ja'e bo mi nu'o co'e doi la gleki .i ku'i mi morji ja'a 03:13 < Amnadya> .u'i .oi la .tubinen. cu ba'e zmadu la .franfurt. lo ka zvada'o mi 03:13 < cliva> .ai mi simkarbi lo ni do'e mi darno ja cu kargu 03:14 < cliva> .i ta'o mi pinxe lo pinxyci'a noi jisra vau fa noi pinxyska je cu va'i ciska fau lo nu pinxe ku fi lo pinxyka'i no'oi ka da ce'ai lo ge jibni be lo korbi be da gi nenri be da cu skadu'i soi snuti vau lo bartu be da kei fau noi nu pinxe se pi'o lo pinxykra noi jbini lo moklu jo'u lo kabri no'u lo pinxyra'i 03:14 < Amnadya> su'a lo trene na'o jai kargu .i je lo sorpeka na'o jai to'e kargu 03:15 < Amnadya> ... .u'i 03:15 < Amnadya> ma fasnu 03:15 < gleki> cliva: la .uuZIT. ja'e bazi spoja 03:15 < cliva> mi ji'a ba'o ja ca'o spoja 03:17 < Amnadya> ko mo kau .i ku'i lo zi'ejvo ne mu'a zo pinxydja je zo zvasku cu jai ri'a xruti fai mi lo ka nelci lo jbobau .i ga'a mi ro rinka be ru cu xamgu .i so'o roi pu senpi lo ro da pe lo jbobau je pu zukpei lo ka cliva 03:21 < Amnadya> .i mu'i lo nu xebni ro lujvo pu pilno ji'i no lujvo tu'a lo .itca .i ku'i ca ti mi co'a kufra 03:22 < gleki> uanai 03:23 < Amnadya> .i .ei mi 03:27 < cliva> ro lo re tcadu zo'u mapti mi 03:34 < gleki> fi ma 03:34 < gleki> i la ropnyfesti vau xu 03:34 < cliva> go'i 03:35 < cliva> .i lo ni festi zo'u ba'a ba facki 03:36 < zipcpi> xazdmru = x1 a word in a particular class of zi'evla (x2) that contains syllabic sonorant x3, which when "filled out" by {y}, becomes lujvo x4, which makes it hard to distinguish to x5, which causes Lojbanist x6 (default la zipcpi) to think it should be banned altogether 03:36 < cliva> ku'i zo'oi xazdmru xa'o canci lo te gerna 03:37 < cliva> exp: xazdmru 03:37 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "x" found. 03:37 < zipcpi> i'e 03:38 < gleki> ie lo nalgendra valsi cu banli 03:38 < gleki> i ku'i gendra ma'i lo u zei bangu 03:39 < zipcpi> "Sorry, that isn't actually a word in Lojban." Hmm maybe it should have an appropriate message 03:39 < zipcpi> "This is a xazdmru word and not allowed" 04:01 < cliva> durka42: sei xu xa'o djuno ma noi djedi cu traji co mapti lo ke da'i ronjbofesti 04:01 < durka42> xm xm 04:03 < durka42> gaje lo xavdei ca lo lecydo'i gi lo zeldei cu mapti mi 04:04 < durka42> ji'a lo balvi be lo revomoi co djedi cu mapti 04:04 < zipcpi> http://jbovlaste.lojban.org/dict/kralyma'u 04:04 < durka42> ca ri mi zifre .i za'o na jdice lo ka vitke makau 04:04 < zipcpi> Ah crap 04:05 < cliva> je'e 04:05 < zipcpi> I can't think of a better definition than "x1 is old enough to do x2 according to x3", which kinda loses the clarity of {krali} 04:05 < durka42> po'o mi vofli teka'a la .frankfrt. ca lo resomoi 04:06 < cliva> ti'e la .amnadja cu se cinri lo nu da'i ri kansa 04:06 < zipcpi> Or just $x_1$ is a legal adult (but maybe {flalu} is better than that, because legal-adulthood comes both with rights *and* responsibilities) 04:06 * zipcpi shrugs 04:07 < durka42> la .amnadja cu mo 04:07 < cliva> mo dai no'oi mo 04:08 < zipcpi> kralyma'u might still be useful though... for "old enough to drink alcohol" or something 04:09 < durka42> ma du la .amnadja 04:09 < cliva> la selpa'i 04:14 < zipcpi> http://jbovlaste.lojban.org/dict/fuzma'u // ja'o mi pofygau JVS 04:14 < zipcpi> lo valsi poi se jbovlaste jeku'i cu se ciksi fo no da 04:14 < cliva> pa jinga cu se menre do 04:15 < cliva> (to ge'e selsmudukti ru'e lo se fanva toi) 04:17 < durka42> ua 04:17 < durka42> xunai la selpa'i cu xabju la bergu linji 04:18 < cliva> go'i .i zukpei lo ka (to ju'o cu'i toi) klama la frankoforte 04:18 < durka42> ua 04:19 < zipcpi> By the way has someone come up with a specific brivla for age beyond just {jai nanca}? 04:20 < durka42> ma nilcla F boi B 04:21 < cliva> lo barda .i treki'o li 550 04:21 < durka42> ua lo trene noi jai se ditcu lo cacra be li vo 04:21 < cliva> ku'i za'a li'ai ICE jai kargu 04:23 < zipcpi> What is this Flatland translation on the punctuation page 04:23 < durka42> ua .i mi na se slabu 04:23 < zipcpi> That might be an interesting translation, but come on, we don't even have proper words for dimensionality 04:23 < zipcpi> Which is, kinda central to that story 04:23 < durka42> la rupnrxiiuro be li ji'i panono 04:24 < durka42> judrysni? 04:24 < zipcpi> Which is related to what we're supposed to do with {canlu} 04:24 < durka42> lo lanzu mi .e'ende lo ka vitka la .berlin. 04:25 < zipcpi> I believe that "at least three dimensions" thing should be dropped, and {canlu} just meaning "room / space" in the sense of being able to fit 04:25 < zipcpi> Thus also able to be used for "hard drive space" 04:25 < durka42> ku'i lo cacra be li so cu jdima 04:26 < durka42> zipcpi: ie .i we already use it for "gismu canlu" 04:29 < durka42> cliva: tezu'e lo nu vitke B zo'u da'i .ei mi renli'u ze'a piso'i lo remumoi co djedi jecu lecylerci cliva ca lo resomoi 04:29 < durka42> y 04:30 < durka42> s/lecylerci/lirclira 04:30 < fenki> durka42 meant to say: cliva: tezu'e lo nu vitke B zo'u da'i .ei mi renli'u ze'a piso'i lo remumoi co djedi jecu lirclira cliva ca lo resomoi 04:30 < zipcpi> {mivyselna'a} 04:31 < zipcpi> verba: x1 is a child/kid/juvenile [a young person] of age x2, immature by standard x3. 04:31 < zipcpi> oi 04:31 < zipcpi> "of age x2"? 04:31 < zipcpi> This is like {batke} having a {fi'o kosmu} place without there actually being a word for {kosmu} lol 04:32 < zipcpi> And why must a batke have a kosmu 04:32 < zipcpi> Can't there be buttons that don't do anything 04:33 < cliva> durka42: je'e .i ka'e ka'e (to jo'a toi) pensi'u bu'u lo jbini 04:33 < durka42> zipcpi: that's not the common case 04:33 < gleki> alta: ganai mo gi mo 04:33 < mensi> (I [CU {ga nai} {<FA ZOhE> <CU (¹mo VAU¹)>} gi {<FA ZOhE> <CU (¹mo VAU¹)>} VAU]) 04:33 < gleki> alta: i broda noi mo 04:33 < mensi> (i [FA ZOhE] [CU {FASNU <FA (¹LO [NU {<FA ZOhE> <CU (²broda NF²) VAU>} KEI] KU¹) (¹noi [{FA ZOhE} {CU <mo VAU>}] KUhO¹)> VAU}]) 04:35 < gleki> fixed what i broke earlier 04:35 < gleki> (nobody even noticed) 04:43 < f8l> Was there no test for that? 04:44 < durka42> cliva: ko mi jungau 04:45 < gleki> it's a good question. we need a long text checking for all possible rules 04:46 < Amnadya> ue xu do klama la .berlin. 04:46 < Amnadya> doi la durka42 04:46 < Amnadya> .i xu cumki 04:46 < durka42> mi na djuno 04:46 < durka42> stidi fa lo lanzu 04:46 < durka42> ku'i darno 04:46 < Amnadya> .i la danmo cu cusku fi mi fe lu mi zvati la .berlin. co'a lo re ji'i moi pe la .bivmast. je co'u lo mu ji'i moi li'u 04:46 < Amnadya> .i .ai mi penmi 04:47 < durka42> ua 04:47 < durka42> mi zvati lo merko ca piro lo .bivmast. uinai 04:47 < Amnadya> je'e 04:48 < Amnadya> .i .y ta'o da'i mi sipna zai ma 04:49 < gleki> lo ckana 04:49 < Amnadya> .i tau lo xotli cu kargymau tau lo sorpeka 04:49 < gleki> do ca litru ibo xu 04:50 < Amnadya> ri mo 04:50 < Amnadya> sa sa 04:50 < Amnadya> .i na litru 04:50 < gleki> i ma fasnu 04:50 < durka42> mi na djuno .i ba lo revomoi mi na ternupcpe fi lo xotli 04:50 < Amnadya> la durka cu zvati lo dotygu'e .i je mi je la cirko cu zukpei lo ka penmi 04:50 < gleki> ua ua 04:51 < durka42> xu lo gacri ckana je cersai cu co'e 04:51 < gleki> i je'u logji festi 04:51 < gleki> i e'u do cpedu fi LLG fe tu'a lojdini vau zo'onai 04:52 < durka42> xunai do zgana lo nu mi cikna ca lo ropno co'e iau .u'i 04:52 < Amnadya> .u'e .i mi na pu pensi lo noi nu go'i ku'o cumki 04:52 < durka42> xu jimpe lo tanru .u'i 04:53 < durka42> ua do sitna lo se stidi be la gleki jenai mi 04:53 < Amnadya> go'i 04:55 < gleki> lo nu lo ci do penmi zu'ai cu frica lo LLG nu penmi vau ma poi ralju 04:56 < gleki> i pe'i frica i frica fi lo nu lo do nu penmi cu vajni zmadu u'i 04:56 < Amnadya> ki'u ma do cusku di'u 04:56 < gleki> i ji'a ki'u ma do na invite la nulern tu'a lo dotco gugde 04:57 < gleki> i kargu mleca 05:26 < durka42> iepei 05:27 < mensi> ba'e mi na tugni 05:34 < gleki> la nulern cu jibni zmadu 05:36 < durka42> zvati ma 05:36 < durka42> lo gligu'e xunai 05:42 < gleki> go'i 05:45 * nuzba @bakkasel: Esperanto is now on Duolingo. I'd prefer to learn lojban because it's so different and logical. I'll give it a go though. [http://bit.ly/1J4wmBs] 05:46 < gleki> i lo gligu'e cu jibni zmadu lo merko 06:08 < Amnadya> The recent changes to Ithkuil have made it almost ridiculously easy. 06:11 < Amnadya> (relatively speaking) 06:24 < Amnadya> One thing that apparently only a lojbanist would find (possibly) wrong with Ithkuil is the lack of a question word and the supposed workaround 06:24 < Amnadya> You can't ask "What are you doing?" or "How did you do it?" in Ithkuil 06:25 < Amnadya> you are forced to say "State the thing you are doing" and "State the method by which you did it". 06:26 < Ilmen> je'e 06:26 < Amnadya> In other words, ithkuil conflates "lo broda" and "lo du'u ma kau broda" 06:26 < Amnadya> At least that's what it looks like to me 06:27 < Ilmen> Seems so 06:28 < Ilmen> {.e'o jungau mi lodu'u do makau zukte}, {ca'e mi do cpedu be lo ka jungau mi lodu'u do makau zukte} 06:29 < Amnadya> ie 06:29 < Amnadya> .i ku'i naku drani fa lu ko jungau mi lo zukte li'u 06:29 < zipcpi> Oh... basically questions are restated as commands 06:30 < Ilmen> go'i lu jungau lo se zukte li'u 06:30 < Amnadya> Yes, that part is fine with me, but the raising is ... unsettling 06:30 < Amnadya> "What time is it?" -> "Please indicate the time of day" 06:30 < zipcpi> ko jungau mi ... 06:30 < zipcpi> But then no makau 06:30 < zipcpi> Oh btw just for a lark: http://mw.lojban.org/papri/Eulojban 06:30 < Amnadya> It's logically inaccurate to rephrase it like that 06:31 < Ilmen> How do one say {mi djuno lo du'u ma kau sipna} in Iþkuîl? 06:31 < Ilmen> *does 06:32 < zipcpi> Maybe it's like {mi djuno (tu'a?) lo sipna} 06:32 < Amnadya> I don't know. I think it would become mi djuno lo sipna 06:32 < Ilmen> je'e 06:32 < zipcpi> No tu'a even T.T 06:33 < Amnadya> No tu'a at all. Ithkuil is full of implicit raising. {ri'a do} etc 06:33 < Ilmen> Iþkuîl doesn't like vagueness, so there's probably no co'e/zo'e/do'e/tu'a/... 06:33 < Amnadya> Here is the best example: 06:34 < Amnadya> (example of Qijada trying to be exact but Lojban still being more exact) 06:34 < zipcpi> Though I dunno, I hear you wanted to get rid of {kau} in Toaq Dzu, and I'm not exactly sure how lol 06:34 < zipcpi> Oh hey silly idea 06:34 < zipcpi> Instead of going makau makau makau 06:34 < Amnadya> In Ithkuil, guns, catapults, slings and all other firearms or devices for shooting/hurling projectiles utilize Degree 1 of the UTE suffix ('tool/implement for doing X'), as it is not the firearm/shooting device that kills but the projectile shot/fired from it. Thus a gun is a "bullet-implementer". Care must be taken in assigning appropriate noun C 06:34 < Amnadya> ases or verbal Formats to stems associated with ballistic aerial weapons to distinguish the fact that it is the projectile that is the actual "weapon" and not the object it was fired/projected from. This is the good old "shoot bullets" vs. "shoot guns" vs. "shoot bullets from a gun" vs. "shoot him with a gun" vs. "shoot him with bullets from a gun" 06:34 < Amnadya> problem . This would translate into Ithkuil as "I bulleted him due to/enabled by a gun" or incorparatively as "I gun-bulleted him" using either the RESULTATIVE Format (literally translating as "I bulleted him as a result of concurrently gun-activating") or the INSTRUMENTATIVE Format (literally translating as "I bulleted him by the enabling means o 06:34 < Amnadya> f gun-activating"). 06:34 < zipcpi> Why not just put the {kau} right after {du'u}? 06:35 < Amnadya> So in Ithkuil a bullet kills people, but in Lojban it doesn't. The bullet hitting you kills you 06:35 < zipcpi> Hahahaha 06:36 < Ilmen> Well, I don't think {ri'a do} is nonsensical, but probably most often inappropriate. If events and objects are a single type, then "do" would be like "da doipre" 06:37 < Ilmen> *lonu da doipre 06:37 < Ilmen> ju'o cu'i 06:39 < Amnadya> Yes, sure. 06:40 < Amnadya> zipcpi: Toaq Dzu has no {kau} because it has no {ma}. 06:40 < Amnadya> The word it has is already like {makau} 06:40 < zipcpi> So then you have something to mark the illocutionally "Tell me this fact"? 06:40 < zipcpi> Interesting 06:41 < Amnadya> Yes. "tell me who took my ball" 06:42 < Amnadya> It's still shorter than both English and Lojban zo'o 06:42 < zipcpi> off: lo du'u kau ma lebna lo mi bolci 06:43 < Amnadya> kau is UI so it's fine 06:43 < mensi> ([lo {<du'u kau> <(¹ma CU¹) (¹lebna [lo {mi bolci} KU] VAU¹)> KEI} KU] VAU) 06:43 < zipcpi> Yep 06:43 < zipcpi> makfa 06:44 < zipcpi> I dunno, even though it only saves syllables when there is more than one question word, it's cleaner pe'i 06:44 < Amnadya> Cleaner only if it still turns ma into an indirect question word 06:45 < Amnadya> What if it means {lo du'u xu kau ma lebna} 06:45 < zipcpi> u'i 06:45 < Ilmen> lo zo'u fa gerku fe mlatu cu viska 06:46 < Amnadya> I just opened my Toaq Book for the first time in a year 06:46 < zipcpi> Well, there's also that other use of {kau} though... implicit predication with discursives. {ju'ocu'i makau catra la .tom.} Well, fegli su'o da, but it is used 06:46 < Ilmen> lo purmo co'a vofli 06:47 < Ilmen> co zifre 06:47 < Amnadya> It looks so pretty! 06:47 < Amnadya> zipcpi: Used by you :) 06:47 < zipcpi> u'i 06:48 < Amnadya> So the book shows these two examples: (6.) sún hi dáu moq ? 06:48 < Amnadya> like you which animal INTERROGATIVE 06:48 < Amnadya> „Which animals do you like?" 06:48 < zipcpi> Well the problem is IMO all predicates do have a hidden NU 06:48 < Amnadya> (7.) Sùa sún hi rái moq ? 06:48 < Amnadya> do you which anything INTERROGATIVE 06:48 < Amnadya> „What are you doing?" 06:48 < zipcpi> Usually it's {nu} or {du'u} 06:49 * Amnadya so proud 06:49 < zipcpi> But when talking about stuff like how Lojban words might be defined; well, technically it's {ce'u}, but {ce'u} is inconvenient if you have several variables and want to juggle them like one so often does in definitions 06:50 < Ilmen> ki'u ma no tonga sinxa cu pagbu zo'oi hi .i xu co'e zo mo'oi .enai zo do 06:50 < Amnadya> go'i 06:50 < Ilmen> ja'o zo do .implisite 06:50 < Amnadya> na go'i 06:50 < Amnadya> sún = do 06:51 < zipcpi> So with the old {xyxipa} or {fabu} styles, it is implicit {ka} *and* implicit {ce'ai} 06:51 < Ilmen> ua .i zo nelci ma zvati 06:51 < Amnadya> ua zo nelci na zvati 06:51 < Ilmen> zo nelci ma zvati .i zo nelci na zvati .i zo nelci ba zvati ĭa 06:52 < Amnadya> da'i lu Tcò sún hi dáu moq li'u 06:52 < zipcpi> Sounds like some cross between Cantonese and Vietnamese to me :p 06:52 < Amnadya> .i mi nu'o ro re'u jdice fi lo valsi 06:52 < zipcpi> Though I'm not sure how the tones work yet 06:53 < Ilmen> zipcpi: "q" = [ŋ] 06:53 < zipcpi> I worked with tonal languages before, but not with this many tones 06:53 < zipcpi> I know that 06:53 < Ilmen> jee 06:54 < Amnadya> http://selpahi.de/tonediagram.JPG 06:54 < Amnadya> Actually the light pink tone was removed long ago 06:56 < zipcpi> And so I don't see the problem of using {kau} in the main predicate; sure it's confusing out of context, but the UI2 generally provide sufficient context IMO 06:56 < zipcpi> Of course, I think with Toaq Dzu, UI2 won't be needed 06:56 < Amnadya> Tone# Tone Name Unicode ASCII Function 06:56 < Amnadya> 1 Even ā a- Content clause 06:56 < Amnadya> 2 Rising á a/ Argument phrase 06:56 < Amnadya> 3 Dipping ă a° Sentence relative phrase 06:56 < Amnadya> 4 Falling ả a? Preposition 06:56 < Amnadya> 5 Peaking â a^ Relative clause 06:56 < Amnadya> 6 Low à a\ Predicate 06:56 < Amnadya> 7 Creaky [Dipping] ã a~ Adverbial phrase 06:56 < Amnadya> 8 Neutral a a Function word 06:57 < Ilmen> For writting tonal languages in latin alphabet with diacritics, I'd use dotless i: "ı", so that the tone marks on "i" outstands better 06:57 < zipcpi> Cause UI2 is really just shorthand for {mi se jungau lo du'u...} {mi birti lo du'u...}, etc. 06:57 < zipcpi> Er se jungau fi lo du'u 07:00 < Ilmen> Will Toaq Dzu have something like ri'oi or lo bi'unai? 07:00 < Amnadya> No. 07:00 < Amnadya> sa'e, not as cmavo 07:00 < Ilmen> je'e 07:01 < Amnadya> It has lots of pronouns (which are predicates) 07:01 < Ilmen> Yet, mo'oi is a cmavo, there 07:01 < gleki> yay, ithkuil is now easier than toki pona 07:02 < zipcpi> What's the difference between cmavo and na'e cmavo in Toaq Dzu? 07:02 < Ilmen> maybe mo'oi could be a predicate "x1 is what has property x2" 07:02 < Amnadya> cmavo don't take tones. predicates change with each tone 07:03 < gleki> zipcpi: Ew! Eulojban. 07:03 < zipcpi> lol 07:03 < Ilmen> (x1 mintu mo'oi ckaji be x2) 07:03 < Amnadya> ithkuil: eulojban 07:03 < ithkuil> eulojban: P2S2/DYN-l-ERG-UNI/ASO/U/GRA/NRM-AGC_1/5-EXS-IFL/UNFRAMED 07:04 < zipcpi> Well, erm, yes, I stole selpa'i-s old anaphora system, though I made some changes to it 07:04 < Amnadya> I hope so. 07:04 < Amnadya> E.g. remove all gender stuff 07:05 < zipcpi> Well, it's still has male/female, but {reu} is just "person} 07:05 < gleki> sometimes i can get rid of kau, sometimes i can't. maybe those are different kau 07:05 < Ilmen> There's a fair amount of languages that completely lack gendered pronouns 07:05 < Ilmen> ti'e 07:05 < Amnadya> Turkish 07:05 < gleki> persian 07:05 < Ilmen> Wolof 07:05 < gleki> ~=Mandarin 07:05 < zipcpi> Well, the changes is that deu = danlu, teu = dacti jenai danlu 07:06 < zipcpi> So doesn't distinguish between plants and other inanimate objects 07:06 < Ilmen> animate/inanimate 07:06 < zipcpi> Yep 07:07 < Amnadya> ta'o sai, re: lo no'u ti etc. 07:07 < zipcpi> Also no remna case 07:07 < gleki> there should be an animcay marker denoting lightning. like in Navajo 07:07 < zipcpi> u'i 07:07 < gleki> https://en.wikipedia.org/wiki/Navajo_grammar#yi-.2Fbi-_Alternation_.28Animacy.29 07:07 < zipcpi> No remna pronoun, only prenu 07:08 < Amnadya> (Ithkuil has good pronouns) 07:08 < zipcpi> gleki: Very interesting, but I don't think we want to copy it lol 07:08 < Ilmen> Nice, so it can be as handy for non-humans, like robots etc 07:09 < Ilmen> Now you gotta add some new colors predicate so that it can be spoken by birds, and maybe mantis shrimps 07:09 * Ilmen runs 07:09 < Amnadya> I have found a solution that I will write an article about at some point, which simply involves an unrestrictive {je} (which also solves other problems). 07:09 < gleki> indeed not culturally neutral towards shrimps 07:11 < Ilmen> A solution for "lo ti broda"? 07:11 < zipcpi> Oh Eulojban does have that 07:12 < zipcpi> ti neu broda 07:12 < durka42> zipcpi: how the heck to you pronounce eu 07:12 < Amnadya> [ew] 07:12 < durka42> oi 07:15 < zipcpi> peu = po'u lo, neu = no'u lo. Basically meant to be {poi}/{noi}, but doesn't start a new bridi 07:16 < durka42> Eulojban sounds like a european version, lol 07:16 < durka42> call it Leujban 07:16 < zipcpi> lul 07:17 < zipcpi> The pun is meant to be with the Greek prefix eu-, like eupepsia 07:17 < zipcpi> eukaryote 07:17 < gleki> ew pepsi 07:17 < zipcpi> euphony 07:17 < zipcpi> lol 07:17 < durka42> but it must not be pronounced like those words, since that is {iu} 07:17 < zipcpi> Yeah... lol 07:18 < zipcpi> Well, in Classical Greek, it is [ew] 07:18 < zipcpi> Not sure about modern Greek 07:19 < Amnadya> It turned into ef 07:19 < zipcpi> Right 07:20 < Amnadya> In German it's pronounced {oi} 07:20 < Amnadya> (the diphthong "eu" in general, too) 07:21 < zipcpi> Ah 07:21 * demize is always amused when he hears German people pronounce 'euro'. 07:21 < durka42> oi 07:21 < durka42> .u'i 07:23 < gleki> oi phonia! 07:23 < zipcpi> Hah. What attitudinal should we even assign to {.eu} anyway? :p 07:23 < Amnadya> pain 07:24 < zipcpi> lol 07:24 < gleki> believe me or not but in {ganai mo gi mo} "nai" wasn't specially defined. It was just allowed as any UI in post clause. So the parser thought of {ganai} just as of {ga ui}. 07:25 < gleki> they should've just written "output = function (input){return "{"+input+"}";} 07:26 < gleki> and the parser is done 07:45 < durka42> why is that a problem 07:45 < durka42> seems elegant 07:45 < durka42> who needs another rule 07:45 < durka42> in the original camxes NAI was separate, but then it was unified with UI 07:45 < gleki> it is a problem because in connectives NAI has another meaning 07:46 < gleki> in giheks it was specified differently 07:46 < gleki> which was and is fine 07:46 < gleki> NAI isn't an attitudinal after geks 07:47 < gleki> btw i removed all "free*" from altatufa. the parser is much lighter now 07:47 < gleki> let's hope everything still works 07:47 < gleki> alta: i seimo mi mo 07:47 < mensi> ([i {sei mo SEhU}] [FA mi] [CU {mo VAU}]) 07:47 < gleki> alta: i moseimo mi 07:47 < mensi> (i [FA ZOhE] [CU {mo <sei mo SEhU>} {FE mi} VAU]) 07:48 < gleki> "free" now belong to INTerjections 07:48 < gleki> so no need to specofy them separately 07:48 < gleki> *specify 07:49 < gleki> interjections are specifed in post_clause of almost any construct except UI, NAI, CAI, ROhA and DAI clauses 07:50 < gleki> in the last 5 cases INTerjection rule itself deals with compound interjections 07:55 < durka42> grammars don't capture meaning 07:57 < gleki> in {genai} and {uinai} NAI has different grammar 07:58 < gleki> {uinaisai} is {ui(naisai)} whereas {genaisai} hardly makes sense but if it does then {sai} modifes the clause or {genai} as a whole. 08:00 < gleki> off: genaisai mo gi mo 08:00 < mensi> ([ge {nai sai}] [mo VAU] gi [mo VAU] VAU) 08:00 < gleki> alta: genaisai mo gi mo 08:00 < mensi> SyntaxError: Expected [yY] but "o" found. 08:00 < gleki> and this is intentional 08:00 < gleki> even if i allow {genaisai} it will be implemented differently 08:03 < zipcpi> durka42: Do you still have trouble pronouncing /eu/? lol 08:03 < durka42> zipcpi: in a meeting, haven't listened to the recording yet 08:03 < gleki> in La Alta {eu} means {e'u} 08:05 < zipcpi> Well I would probably call it {la .e'ulojban.} in regular Lojban (as regular as the Lojban that I speak is, anyway :p) 08:05 < gleki> off: mi mo na se gi'e nai mo 08:05 < mensi> ([mi CU] [mo VAU] [na se gi'e nai] [mo VAU] VAU) 08:05 < gleki> off: mi na se e do mo 08:05 < mensi> ([{mi <na se e> do} CU] [mo VAU]) 08:05 < gleki> off: mi se bi'i nai do mo 08:05 < mensi> ([{mi <se bi'i nai> do} CU] [mo VAU]) 08:06 < gleki> off: mi na se bi'i nai do mo 08:06 < mensi> SyntaxError: Expected [yY] but "o" found. 08:06 < gleki> clearly 08:06 < gleki> what could {mi na se bi'i nai do mo} mean? 08:06 < gleki> en: bi'inai 08:06 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 08:06 < gleki> en: bi'i 08:06 < mensi> bi'i = [BIhI] non-logical interval connective: unordered between ... and ... |>>> 08:06 < mensi> officialdata 08:08 < gleki> 16.9) do dicra .e'a mi ca la daucac. bi'onai la gaicac. You disturb (allowed) me at 10 not-from ... to 12 You can contact me except from 10 to 12. 08:08 * gleki notice raising in CLL 08:09 < gleki> so in bi'onai {nai} works not like in {jenai} 08:10 < gleki> because it were like {je} i would understand this phrase as "You disturb me from 10 up to not 12" 08:10 < gleki> i think this behavior of {bi'o} should be changed 08:11 < gleki> maybe just {canai} can do the intended meaning 08:13 < durka42> hmm 08:13 < durka42> yeah 08:13 < durka42> nothing to do with grammar but that is pretty awkward 08:15 < gleki> write to bpfk-list then 08:15 < gleki> but it has to do with the grammar 08:15 < gleki> i noticed that when fixing connectives in altatufa 08:15 < gleki> {nase je nai} allowed 08:16 < gleki> {nase bi'o nai} not allowed 08:16 < gleki> only 08:16 < gleki> {se bi'o nai} allowed 08:16 < durka42> eh? 08:16 < durka42> so bi'o is not in JA? 08:17 < gleki> it was and is in JOI 08:17 < gleki> jek = expr:(NA_clause? SE_clause? JA_clause NAI_clause?) {return _node("jek", expr); } 08:17 < gleki> joik = expr:(SE_clause? JOI_clause NAI_clause? / interval / GAhO_clause interval GAhO_clause) {return _node("joik", expr); } 08:17 < gleki> interval = expr:(SE_clause? BIhI_clause NAI_clause?) {return _node("interval", expr); } 08:17 < durka42> ah 08:17 < durka42> so JA and JOI are not merged 08:19 < gleki> i cant find anything on {JOI nai} in CLL. 08:20 < gleki> indeed i cant 08:21 < gleki> i mean JOI_clause, not intervals 08:21 < gleki> any example of "SE_clause JOI_clause NAI_clause"? 08:25 < gleki> off: se gu'u mo gi mo 08:25 < mensi> ([{se gu'u} mo gi mo] VAU) 08:25 < gleki> off: se gu'u nai mo gi mo 08:25 < mensi> ([{se gu'u nai} mo gi mo] VAU) 08:26 < gleki> off: na se gu'u nai mo gi mo 08:26 < mensi> ([na {se gu'u nai} mo gi mo] VAU) 08:28 < gleki> Currently in {enai} {nai} has its own node called "connective right negator". This node doesnt appear for {nai} in attitudinals or sumtcita. 08:29 < gleki> ^about la altatufa 09:11 * nuzba @_Vanessa_sary_: @MatthewHuang90 @willingtheweird He wrote in Lojban, huh? Lmfaooo [http://bit.ly/1CLOu1I] 09:13 * nuzba @wyoreorm: lojban lo logji bangu 意味はわからんが転写してみた http://t.co/7ZcF8R7g1h [http://bit.ly/1GqBolw] 09:52 * nuzba @soblo_xivan: うぉれおるむさんに続いて、私も「lojban lo logji bangu」を自作文字で転写した。 Arialのようなフォントにするとこんな感じ。 http://t.co/BOfcweWXo0 [http://bit.ly/1Oc7TJQ] 09:58 * nuzba @sashimiwiki: "lojban lo logji bangu"の転写、下は筆記体 http://t.co/6kuzCUPHM7 [http://bit.ly/1MajWJ4] 10:10 * nuzba @palfem_kleet: lojbanを自言語文字で書こうは後でやる気があればやる.今はSDVX芸人タグ見たい. [http://bit.ly/1Mam3MY] 10:12 < bigcentaur> coi ro do 10:17 < zipcpi> Oh I know, I'll make eulojban brivla for what Gleki said. "the sun and the water on a swimmer's chest, the vague tremulous rose color we see with our eyes closed, the sensation of being carried along by a river and also by sleep." 10:20 < zipcpi> k: soljaunma 10:20 < mensi> (CU [Z:soljaunma VAU]) 10:22 < zipcpi> k: xautkangau 10:22 < mensi> (CU [Z:xautkangau VAU]) 10:23 < zipcpi> k: kangaunxe 10:23 < mensi> (CU [Z:kangaunxe VAU]) 10:24 < zipcpi> k: sipnaurxe 10:24 < mensi> (CU [Z:sipnaurxe VAU]) 10:25 < zipcpi> k: rirxaupna 10:25 < mensi> (CU [Z:rirxaupna VAU]) 10:33 < gleki> no different from vanlamigu 10:33 < gleki> or incita 10:34 < zipcpi> I know lol 10:35 < zipcpi> k: limsauljau 10:35 < mensi> (CU [Z:limsauljau VAU]) 10:35 < zipcpi> k: sipnaurxe 10:35 < mensi> (CU [Z:sipnaurxe VAU]) 10:40 < gleki> what could {inai mo} mean? 10:40 < gleki> something similar to {inaku mo} although with the possibility to scale probably 10:40 < Amnadya> "ma na jetnu" 10:41 < zipcpi> k: vanlaumgu 10:41 < mensi> (CU [Z:vanlaumgu VAU]) 10:41 < gleki> if allow it then {na'e i} should also be allowed 10:42 < Amnadya> na nibli .i mu'a lu ui nai li'u je nai lu na'e ui li'u 10:44 < zipcpi> eutca: x1 means whatever the speaker wants it to mean 10:45 < gleki> i'll mark {inai} as wontfix 10:46 < gleki> at least altatufa is the first parser of lojban with all ~200 instances of "free*" removed. 10:48 < zipcpi> eutceu = me'au lo eutca 10:48 < Amnadya> lol 10:59 < gleki> alta: coinai doi nai mi 10:59 < mensi> ([I {<coi nai> <doi nai>} mi DOhU] [FA ZOhE] [CU {COhE VAU}]) 11:01 < gleki> that's interesting. it restored I and then put this vocative into I's post_clause. 11:38 < zipcpi> OK added my own fix to the {boi} problem to Eulojban :p 11:52 * nuzba @volxemo: lojban lo logji bangu の、自作文字転写、やってみるか← [http://bit.ly/1JnvALw] 14:01 < Amnadya> exp: -f mi viska lo cmana 14:01 < mensi> (mi [viska {lo cmana}]) 14:05 < Amnadya> I've been meaning to make a brivla for the smell of rain (petrichor) for a few days now, but nothing sounds pretty enough. 14:06 < Amnadya> Neither cavypanci nor pancyrpetrikora nor petkora satisfy me yet. 14:10 * nuzba @oka_iu_tcan: 創作文字の転写スクリプトがなぜか lojban lo logji bangu なの、めっちゃおもろい [http://bit.ly/1LrKapR] 14:10 * nuzba @aila_sunfa: lojban lo logji bangu リプサム説 [http://bit.ly/1OkQEH5] 14:20 <@xalbo> IIRC, "petrichor" started as a portmanteau from "petra" (rock) and "ichor" (smell). I don't think that really helps, though. 14:20 < Amnadya> Rock and the "blood" of the gods 14:21 <@xalbo> I don't know how I could have forgotten that, you're right. 15:02 < Ilmen> co'o ro do 15:03 < Ilmen> .i .a'o da nu voksnu kei gi'e cabna lo bavlamdei 15:14 < ctefaho> aw crap missed ilmen 16:06 < Amnadya> Case-stacking was added to Ithkui.l "This is necessary to accurately translate sentences like the following: 16:06 < Amnadya> ‘I jog every day except in case of illness.’ or ‘I jog every day except during (an) illness.’ " 16:06 < Amnadya> A Lojbanist should immediately think of another solution 16:06 < durka42> ca ro djedi poi na jai ca bilma fai mi 16:07 < Amnadya> Exactly, or a variation thereof. 16:08 < Amnadya> durka getting all fancy 16:08 < durka42> .u'i 16:09 < durka42> mi milxe co xalbebna ca 16:09 < durka42> zo jai cu zenba lo ka co'e 16:09 < Amnadya> Clearly though you are a Lojbanist 16:09 < Amnadya> je'e .u'i 16:09 < durka42> mi jainmo 16:09 < durka42> jaizmo? 16:09 < keresth> What's the function of the {ca} at the beginning? 16:09 < Amnadya> xu pinxe lo dotco birje 16:09 < Amnadya> {ca} means "when" as in "when it rains" 16:10 < durka42> na go'i .i .ei ich essen keine getreide 16:10 < keresth> I thought it just indicated the present 16:10 < Amnadya> "at the same time as" 16:10 < Amnadya> lol durka 16:10 < Amnadya> je'e je'e 16:10 < durka42> if you don't put in a sumti it defaults to "now" but if you do then it's "at the same time as X" 16:10 < Amnadya> Lojban doesn't have tense. 16:10 < durka42> .u'i 16:11 < Amnadya> (only tense is nau) 16:12 < Amnadya> The rest are just prepositions, "before", "after", "at the same time as" 16:12 < keresth> I see. 16:12 < Amnadya> And they can be followed by "the present" / "now" / "the current moment" 16:12 < Amnadya> as well as anything else. 16:12 < Amnadya> {ca lo cabdei} - "today" 16:13 < Amnadya> {pu lo nu do sipna} - "before you sleep" 16:13 < dutchie> (insert argument about generic you here) 16:13 < dutchie> zo'o 16:27 * nuzba @slaimsan: 何故 lojban lo logji bangu が流行ったし [http://bit.ly/1SsAImw] 18:20 < bigcentaur> coi 18:23 < bigcentaur> .i xu su'o da poi jundi 20:22 < zipcpi> Hey {eutca} might actually be useful. I just added a note observing that {zo prenu cu eutca} :p 20:23 < zipcpi> We might want to give it a standard Lojban variant lol 20:23 < zipcpi> {eutceu}, however... lol 20:28 < zipcpi> {limseuljau} looks like a Korean name lol 20:37 < zipcpi> I mainly invented Eulojban because I kept thinking that "We might have a proper categorical pronoun set if only we had more short cmavo space" 20:37 < zipcpi> And "How I wish /eu/ was a valid diphthong so that can fit"... so, yeah, thought I'd just commit it down on the wiki 20:39 < zipcpi> The {po'u lo} and {no'u lo} thing pushed it over the edge 20:53 < zipcpi> OK I entered {eutca} as {autca} :p 21:42 < zipcpi> mensi: doi xalbo Nah if I added another vowel to the Eulojban class it would probably be /œ/ (assigned to w) instead of /æ/. {la .eulwjbaq.}. mi ja'a kusru lo'e se glibau 21:42 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.xalbo.gy. di'a cusku da 21:46 < zipcpi> mensi: doi xalbo And hey why not make /wu/ a valid diphthong while we're at it (it sounds like an exaggerated British "oh!") 21:46 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.xalbo.gy. di'a cusku da 23:27 < gleki> what was the "manskapi dilemma for NSN"? 23:30 < zipcpi> qeqxreunmrwnmlwu : = x1 is the feeling of bafflement and confusion that English speaker x2 feels when encountering strange new phonemes or usage of phonemes in non-English ways (e.g. syllable-initial q) 23:31 < zipcpi> k: nenxreunmranmlau 23:31 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "n" found. 23:31 < zipcpi> Er 23:31 < zipcpi> Oh left in an /eu/ 23:31 < gleki> itkuiile ga'a mi 23:31 < zipcpi> k: nenxraunmranmlau 23:31 < mensi> (CU [Z:nenxraunmranmlau VAU]) 23:31 < zipcpi> u'i doi gleki 23:58 < gleki> the basic text/text-1/paragraphs/paragraph/statement/sentence structure is completely mindbreaking 23:58 < gleki> i think i will have to reimplement it with NIhO nodes restored --- Day changed Sat Jul 18 2015 00:28 < gleki> off: na'e gu'e mo gi mo 00:28 < mensi> ([na'e gu'e mo gi mo] VAU) 01:02 < cliva> spagetufa: la cmev cu ia ai broda lo uanmo vau ui 01:02 < spagetufa> ([FA {la .cmev. KU}] [{cu <ĭa .ai>} {bròda <(¹FE [lo ŭànmo KU]¹) (¹vau. ŭi¹)>}]) 01:28 < cliva> xu da'i va'o lo nu lo'u .i bai ke le'u co'a gendra kei da pante 01:28 < cliva> fa poi na ta'e pante ro da 01:29 < cliva> side 01:31 < cliva> y ju'i nai .i na sai kompatible 01:31 < cliva> .i da'i lo'u ke .i bai bo le'u 01:33 * nuzba @koridentetsu: ヒッタン語にしても千織語にしても、音節末尾に j が立つことや、jb という子音連続を容認しないので、「ロジバン lojban」という名称を表す方法は、母音を補って lojiban とするのが最適だろうか。 [http://bit.ly/1J6Agd9] 01:43 * nuzba @harmat_noln: lojban lo logji bangu. http://t.co/DwxR2WSUuB [http://bit.ly/1fcRU2x] 01:49 * nuzba @harmat_noln: lojban lo logji bangu. (若葉文字) http://t.co/x7ZiZdNjE4 [http://bit.ly/1J6Bfdh] 01:55 * nuzba @harmat_noln: lojban lo logji bangu. (パラオヌイヤ文字) http://t.co/F4lCfHLDs6 [http://bit.ly/1J6BF3r] 02:45 * nuzba @oka_iu_tcan: 一応みんなやってるから、.lojban. lo logji bangu を凛花で http://t.co/oknjqEjmM4 [http://bit.ly/1StsnPv] 03:02 < zipcpi> Hey I wonder if... 03:02 < zipcpi> vlaste: bemza (gimka) 03:02 < vlaste> zo bemza gimkamsmikezypro no catni gismu .i zo bemza gimkamsmikezypro no cipra gismu 03:03 < zipcpi> vlaste: famza (gimka) 03:03 < vlaste> zo famza gimkamsmikezypro zo fanza noi catni gismu .i zo famza gimkamsmikezypro no cipra gismu 03:03 < zipcpi> vlaste: dumza (gimka) 03:03 < vlaste> zo dumza gimkamsmikezypro no catni gismu .i zo dumza gimkamsmikezypro no cipra gismu 03:03 < zipcpi> vlaste: nemzu (gimka) 03:03 < vlaste> zo nemzu gimkamsmikezypro no catni gismu .i zo nemzu gimkamsmikezypro no cipra gismu 03:03 < zipcpi> vlaste: vumza (gimka) 03:03 < vlaste> zo vumza gimkamsmikezypro no catni gismu .i zo vumza gimkamsmikezypro no cipra gismu 03:04 < zipcpi> Yay 03:04 < zipcpi> lol 03:04 < zipcpi> New gismu for Eulojban 03:10 < gleki> i no longer think that {ni'o TAG bo} makes sense. but {ni'o [I] TAG bo} with I autorestored can make sense 03:21 < gleki> alta: +s mi mo 03:21 < mensi> (NIhO_elided [I:I {FA KOhA:mi} {CU <GOhA:mo VAU>}]) 04:09 * nuzba @nilab: 「ロジバン ( lojban [loʒban] ) は、ログランを元に、さらなる機能性を追求して LLG が開発を引率してきた人工言語である。1987年に公表され、1997年に文法が暫定的に完成」 ロジバン - Wikipedia https://ja.wikipedia.org/wiki/%E3%83%AD%E3%82%B8%E3%83%90%E3%83%B3 [http://bit.ly/1LtnYM2] 05:08 < gleki> alta: ni'o babo mi viska lo cnino gentufa ije ri jai pluka mi i ku'i lo badna cu kukte zmadu 05:08 < mensi> (ni'o [I bridi_elided {I <ba bo> <FA mi> <CU (¹viska NF¹) (¹FE [lo {cnino <gentufa NF>} KU]¹) VAU>} {i je <FA ri> <CU (¹jai [pluka NF]¹) (¹FE mi¹) VAU>}] [i ku'i] [FA {lo <badna NF> KU}] [cu 05:08 < mensi> {kukte <zmadu NF>} VAU]) 05:28 * nuzba @GlekiArxokuna: My @Quora answer to Why doesn't Lojban use the letter H instead of '? http://qr.ae/R8cq6h [http://bit.ly/1Gskymh] 05:30 * mumble jorne co'a fa la'o fa Ilmen fa 05:33 < zipcpi> coi Ilmen 05:33 < zipcpi> mi na kakne lo ka jorne 05:34 < zipcpi> ri'a da poi mi na djuno tau ki 05:36 < cliva> zipcpi: rinka fa sa'e nai lo nu mi lazni .i ko pilno ju'oi zbagazasni.ax.lt 05:36 * mumble jorne co'a fa la'o fa gunro fa 05:37 * mumble jorne co'a fa la'o fa zipcpi fa 05:38 * nuzba @GlekiArxokuna: My @Quora answer to What are the main differences between natural languages and Lojban? http://qr.ae/R8SAfi [http://bit.ly/1LhAOhO] 05:39 < gleki> The most important differences that immediately come to mind are: 05:39 < gleki> 1. In Lojban people start to understand more where metaphors start in their speech. Metaphors are important in our speech but sometimes they can lead us to incorrect reasoning. It's nice that in Lojban you can both use and avoid metaphors when you like. 05:39 < gleki> 2. In Lojban punctuation like e.g. quote marks is pronounceable. Thus, it's easy to know when a quote starts and ends even in colloquial speech. 05:39 < gleki> 3. For me Lojban is like a bridge between programming languages, speakable languages and knowledge representation languages. In no way Lojban can replace all of them just like a bridge between two continents can't replace the continents themselves. 05:40 < zipcpi> Yeah I read it 05:40 < zipcpi> Very interesting answer 05:40 < gleki> one more answer from someone else: For one, many non-native speakers of natural languages can converse easily and spontaneously about everyday matters. From what I've seen and heard of, Lojbanists spend a lot of time hesitating and pondering the exact logical form they want to express. 05:42 * nuzba @oka_iu_tcan: 例としてシダさんの俳句をば https://twitter.com/s_i_d_a/status/621778787722657792 http://t.co/JFULir2orm [http://bit.ly/1LhBkMN] 05:43 < gleki> Ilmen: ma judri lo voksa kumfa 05:44 < zipcpi> zo'oi zbagazasni.ax.lt 05:44 * mumble jorne co'u fa la'o fa Ilmen fa 05:44 * mumble jorne co'a fa la'o fa akti fa 05:45 < Ilmen> 84.73.73.130 05:45 < Ilmen> Port 64738 05:45 * mumble jorne co'u fa la'o fa akti fa 05:46 * mumble jorne co'a fa la'o fa akti fa 05:48 * mumble jorne co'a fa la'o fa Ilmen fa 05:57 * mumble jorne co'a fa la'o fa ctefaho fa 06:23 * mumble jorne co'a fa la'o fa taldar fa 06:26 * mumble jorne co'u fa la'o fa akti fa 06:45 * nuzba @Vanege_EO: Mi ne scias kion #TokiPona faros al mia cerbo, do eble mi devus provi (denove) lerni #Lojban samtempe ^-^" [http://bit.ly/1CNX8Nj] 06:46 < gleki> alta: je mo 06:46 < mensi> (NIhO [I {<FA ZOhE> <CU COhE>} {I je <FA ZOhE> <CU (¹mo VAU¹)>}]) 06:47 < gleki> text/paragraph/statement section almost completely fixed 06:47 < gleki> alta: ni'obabo mo 06:47 < mensi> (ni'o [I {<FA ZOhE> <CU COhE>} {I <ba bo> <FA ZOhE> <CU (¹mo VAU¹)>}]) 06:48 < gleki> not sure if this is what selpa'i meant 06:55 < gleki> alta: bo mo 06:55 < mensi> (NIhO [I {<FA ZOhE> <CU COhE>} {I <DOhE bo> <FA ZOhE> <CU (¹mo VAU¹)>}]) 06:55 < gleki> ^ {ibo} restored into {ido'ebo} 06:58 < cliva> alta: broda ibobrode inajabrodi 06:58 < mensi> (NIhO [I {<FA ZOhE> <CU (¹broda NF¹) VAU>} {i <DOhE bo> <(¹FA ZOhE¹) (¹CU [brode NF] VAU¹)> <i (¹na ja¹) (¹FA ZOhE¹) (¹CU [brodi NF] VAU¹)>}]) 07:00 < gleki> lu inaja li'u gerna dunli lu ije li'u 07:00 < cliva> lo nu jajgau cu simlu lo ka cizra 07:01 < cliva> exp: broda ibobrode inajabrodi 07:01 < mensi> ([{CU <broda VAU>} {i bo <CU (¹brode VAU¹)>}] [i {na ja} {CU <brodi VAU>}]) 07:01 < gleki> xu do te smuni lo nu lo cimoi jufra na lasna lo remoi jufra 07:01 < gleki> exp: broda ido'ebobrode inajabrodi 07:01 < mensi> ([{CU <broda VAU>} {i do'e bo <CU (¹brode VAU¹)>}] [i {na ja} {CU <brodi VAU>}]) 07:02 < gleki> uanai ki'u ma 07:03 < cliva> zo bo ro roi jai rinka lo diklo nu lasna 07:03 < gleki> exp: broda ido'ebobrode ido'ebobrodi 07:03 < mensi> ([CU {broda VAU}] [i do'e bo {CU <brode VAU>} {i do'e bo <CU (¹brodi VAU¹)>}]) 07:04 < gleki> i faulonu mi djica lo ka pilno lu do'e li'u enai lu do'e bo li'u gi'e na jai gau lasna fai lo remoi lo cimoi vau mi zukte ma 07:05 < gleki> [(i mo) (ido'ebo mo) ] (i do'e bo mo) 07:05 < gleki> se djica mi 07:05 < cliva> mi zukpei lo ka jai gau curmi fai tu'a lo'u broda ke .i do'e bo brode [KEhE] le'u 07:05 < cliva> ja lo simsa 07:07 < gleki> ma'i lo catni zo bo jai bapli lo nu lo'u do'e bo le'u co'a diklo lasna i ku'i ma'i lo catni lo ma'oi joi e ma'oi ja terjonma'o cu vlipa zmadu i ja'o na kulnu zo'o nutli co me lo'u do'e bo le'u 07:07 < cliva> ue ru'e xu ja'ai 07:08 < cliva> yacc: broda .i do'e bo brode .i je brodi 07:08 < mensi> ({<broda VAU> <i do'e bo> <brode VAU>} {i je} {brodi VAU}) 07:09 < gleki> yacc: broda .i do'e bo brode .i do'ebo brodi 07:09 < mensi> ({broda VAU} {i do'e bo} {<brode VAU> <i do'e bo> <brodi VAU>}) 07:09 < gleki> ja'a si'au 07:09 < gleki> gerna: mi e tu e do mo 07:09 < mensi> (0[{<mi e tu> e do} CU {mo VAU}])0 07:09 < gleki> gerna: mi e tu ebo do mo 07:09 < mensi> (0[{mi e <tu (1e bo)1 do>} CU {mo VAU}])0 07:09 < cliva> simlu co mleca co vlipa 07:10 < gleki> gerna: mi e tu edo'ebo do mo 07:10 < mensi> (0[{mi e <tu (1e do'e bo)1 do>} CU {mo VAU}])0 07:10 < gleki> mi djica tu'a zoi .gern. (0[{<mi e tu> edo'ebo do} CU {mo VAU}])0 .gern. 07:10 < cliva> .i ku'i mi ro roi simvukna lo simsa[D be zo vlipa 07:11 < cliva> gerna: mi e tu edo'eke do 07:11 < mensi> (0[{<mi e tu> <e do'e> <ke do KE'E>} VAU])0 07:11 < gleki> ue 07:11 < gleki> gerna: mi e tu edo'bo eke do 07:11 < mensi> O_0 07:12 < gleki> gerna: mi e tu edo'ebo ke do 07:12 < mensi> O_0 07:12 < gleki> gerna: mi mo i do'e ke mi mo 07:12 < mensi> O_0 07:12 < gleki> i mabla 07:12 < cliva> je'u 07:12 < gleki> gerna: mi mo i do'e tu'e mi mo 07:12 < mensi> (0[{mi CU <mo VAU>} i {do'e tu'e <mi CU (1mo VAU)1> TU'U}])0 07:13 < gleki> oi 07:13 < cliva> ... ua 07:13 < gleki> i na se djica mi 07:13 < gleki> i nasai lasna 07:13 < gleki> fa lo remoi 07:13 < cliva> .i mi kucli lo du'u ma kau se kosmu ga'a lo banfi'i 07:14 < cliva> .i ka'e ku lu do'e zo'u tu'e li'u ka'e smufrica lu do'e tu'e li'u 07:16 < gleki> gerna: mi mo i do'e zo'u tu'e mi mo 07:16 < mensi> (0[{mi CU <mo VAU>} i {<(1do'e KU)1 zo'u> <tu'e (1mi CU [mo VAU])1 TU'U>}])0 07:16 < gleki> lo'u ije le'u jau gau lasna zu'ai fa lo re jufra 07:16 < gleki> i ku'i lo'u ido'e zo'u le'u na kakne 07:16 < cliva> ie 07:16 < gleki> i ja'o mi ganse no nu slilu fa le nu vlipa doi padavan 07:17 < cliva> .u'i 07:17 < gleki> ikrinu lo nu mi finti lo se finti 07:18 * mumble jorne co'u fa la'o fa zipcpi fa 07:19 * nuzba @qothr: 現行のクレリカ文字で lojban lo logji bangu http://t.co/TXYvaCU1Jm [http://bit.ly/1StI2hV] 07:21 < gleki> la'a ei mi mrilu fi lo BPFK mriste 07:21 < gleki> fe la'e di'u 07:22 < gleki> gerna: mi mo ije do'e zo'u mi mo 07:22 < mensi> O_0 07:22 < gleki> uanai 07:22 < gleki> i ua 07:22 < gleki> gerna: mi mo i do'e zo'u mi mo 07:22 < mensi> (0[{mi CU <mo VAU>} i {<(1do'e KU)1 zo'u> <mi CU (1mo VAU)1>}])0 07:22 < gleki> cizra je ci'izra 07:22 < Ilmen> gerna: mo .i je ti zo'u mo 07:22 < mensi> O_0 07:23 < _mukti_> .i pe'i bu'o lo mriste pe byfy cu binxo lo cnino jboske ki'u lo nu byfy na zukte 07:23 < gleki> off: mo ije ti zo'u mo 07:23 < mensi> SyntaxError: Expected [yY] but "o" found. 07:23 < Ilmen> ĭa na nelci lo nu lo prenexo cu se la'erli'e lu ije li'u 07:23 < mensi> i ji'a mi na nelci mi'e la mensi 07:23 < gleki> ie ie 07:23 < Ilmen> gerna: ge mo gi ti zo'u mo 07:23 < mensi> (0[ge {mo VAU} gi {<ti zo'u> <mo VAU>} VAU])0 07:23 < Ilmen> ŭe ti ku'i gendra 07:23 < Ilmen> .i ku'i .e'i jezdu'i 07:24 < gleki> _mukti_: mi nitcu lo me'oi postgres vreji pe JVS datni sorcu i xu do ka'e benji fi mi 07:24 < _mukti_> gleki: ie .i mi xenru lo nu na spuda 07:24 < Ilmen> sa'e .e'i jetydu'i 07:25 < _mukti_> .i .ai mi fi do benji 07:27 < gleki> mudri: fau lo nu no drata co'a zukte vau mi ba e'idji lo ka si'unai favgau lo cnino jbovlaste i ku'i mi na djica lo ka galfi lo nau datni sorcu formata i su'a mi na djica lo ka junri jbozu'e vau ju'inai ru'e 07:27 < gleki> oi 07:27 < gleki> mudri: u'u ju'inai 07:27 < gleki> _mukti_: fau lo nu no drata co'a zukte vau mi ba e'idji lo ka si'unai favgau lo cnino jbovlaste i ku'i mi na djica lo ka galfi lo nau datni sorcu formata i su'a mi na djica lo ka junri jbozu'e vau ju'inai ru'e 07:28 < mudri> krefu .u'i 07:30 < _mukti_> .i ua mi gleki lo nu do gunka lo cnino jbovlaste .i vajni fa lo nu basti lo tolci'o 07:32 * nuzba @fotono: #lojban zantufa tricu http://guskant.github.io/gerna_cipra/zantufa_farvi.svg i cnino mekso http://mw.lojban.org/papri/zantufa_mekso i ma'oi FA ka'e selbrisle tcita http://mw.lojban.org/papri/zantufa_ze%27oi_FA [http://bit.ly/1StJDUV] 07:32 < gleki> _mukti_: mi nu'o gunka i ku'i fau da'i lo nu no drata cu co'e vau mi co'e 07:33 < gleki> i ei ma'a pilno lo cmavo be zo pu vau u'i 07:34 < gleki> exp: fa joi fe mi'o penmi 07:34 < mensi> ([{fa <joi fe>} mi'o] [CU {penmi VAU}]) 07:35 < _mukti_> ti'e la tsani cu gunka lo cnino jbovlaste .i ru'a na jetnu 07:35 < gleki> i la zantufa cu po'o genturfa'i je cu stidi no smuni 07:35 < gleki> i ku'i ra noi mupli cu ka'e se smuni lo du'u 07:35 < gleki> do joi mi simxu lo ka ce'u ce'u penmi lo barja 07:35 < gleki> i di'u se cusku la guskant 07:36 < gleki> i eipei ma'a tugni 07:38 < _mukti_> .i ue xu lo gunma cu sevzi penmi 07:39 < gleki> la'a zo joi pe la guskant cu du zo ce pe CLL 07:39 < _mukti_> ua 07:39 < gleki> i ku'i mi na tugni 07:39 < gleki> i zo joi pe lo gunma cu plixau 07:40 < _mukti_> .i ku'i xu lo selcmi cu ka'e sevzi penmi 07:40 < gleki> i ku'i mi retsku lo smuni frica 07:40 < gleki> i ie 07:40 < gleki> i la guskant cu xusra lo srana be lo nu simxu 07:40 < cliva> zantufa: by su'i cy mo 07:40 < zantufa> ([{<by BOI> <su'i (¹cy BOI¹)>} mo KU] VAU) 07:40 < cliva> exp: by su'i cy mo 07:40 < mensi> ([{by BOI} {su'i <cy BOI>}] [CU {mo VAU}]) 07:42 < cliva> zantufa: li pa ja re 07:42 < zantufa> ([li {<pa BOI> <ja (¹re BOI¹)>} LOhO] VAU) 07:42 < cliva> zantufa: by ja cy mo 07:42 < zantufa> ([{by BOI} {ja <cy BOI>}] CU [mo VAU]) 07:43 < _mukti_> So this is curious. I see the idea that a ce-set can be decomposed with simxu is in "Lojban for Beginners" 07:44 < _mukti_> I haven't gamed it out, but it seems to me that if a singularizer can be unpacked like that, you have ambiguity if there's more than one singularizer. 07:45 < _mukti_> lo gunma jonai selcmi be lo sonci cu simxu lo ka ce'u damba ce'u 07:45 < _mukti_> lo re gunma jonai selcmi be lo sonci cu simxu lo ka ce'u damba ce'u 07:55 < omni___> c o i 07:56 < omni___> coi 07:59 < omni___> .i do mo 08:00 < _mukti_> coi la omni 08:00 < omni___> coi la mukti 08:00 < _mukti_> za'a do mutce lo ka ce canlu 08:01 < _mukti_> s/ce/se/ 08:01 < fenki> _mukti_ meant to say: za'a do mutse lo ka ce canlu 08:01 < _mukti_> ? 08:01 < _mukti_> za'a do mutce lo ka se canlu 08:02 < omni___> uhi 08:02 < omni___> uhi 08:02 < omni___> juhi nai 08:02 < _mukti_> ue ju'i bu'o 08:04 < omni___> _mukti_: What would you think if someone was writing a novel in Lojban? 08:05 < _mukti_> I'd be glad to hear of it. I'm starting to think that energetic applications of lojban are as important for its future as the more formal projects, like revising CLL. 08:05 < omni___> .i da'i do ma jinvi lo nu su da finti lo cfika be bau lo jbobau 08:06 < omni___> je'e 08:07 < _mukti_> I'm finally making good on the intention that I've had for a long time to read all of «lo se manci te makfa pe la .oz.», and struck by how many lessons are packed into it ... ideas that haven't yet been captured in discussions ABOUT lojban. 08:08 < omni___> That's why I tell everyone to read la .alis. 08:08 < omni___> I doubt many actually follow that advice though 08:08 < _mukti_> I'll have to read that next. 08:09 < _mukti_> I had some misgivings about .alis. since it is, at least by reputation, a community translation. I know xorxes led the way, however, so perhaps he provided a steady hand? 08:10 < omni___> There are two versions. One was a community translations (although xorxes did most of the work) and the other is all xorxes's (although he said that pierre did one or two chapters) 08:10 < _mukti_> Oh! 08:11 < omni___> xorxes' version is gold. 08:12 < _mukti_> ua sai .i .ai mi tcidu 08:12 < omni___> :) 08:13 < _mukti_> .i mi ji'a co'a finti .i ku'i ga'i nai mi finti lo cmalu cukta .i srana lo cinmo cmavo 08:14 < omni___> ua .i'e finti 08:18 < _mukti_> http://vrici.lojban.org/~mukti/ahi.jpg 08:18 < omni___> ue 08:32 < Ilmen> coi 08:32 < _mukti_> coi la .ilmen. 08:33 < Ilmen> coi la mukti .i mukti ma ca lo cabdei 08:33 < Ilmen> .i pu za lo cacra cu voksnu ta'o 08:34 < _mukti_> .i mi pu'o cliva lo zdani te zu'e tu'a lo cersai 08:35 < Ilmen> je'e .i .a'o kukte 08:36 < _mukti_> .i'o 08:44 < gleki> i only translated the preface of Alice putting it into a rhymed poem 08:44 < gleki> zantufa: mi mo ije mo ije mo 08:44 < zantufa> ([mi CU {mo VAU}] [i je {CU <mo VAU>}] [i je {CU <mo VAU>}]) 08:44 < gleki> exp: mi mo ije mo ije mo 08:44 < mensi> ([mi {CU <mo VAU>}] [i je {CU <mo VAU>}] [i je {CU <mo VAU>}]) 08:54 < _mukti_> ue 08:55 < _mukti_> camxes: .i cu broda 08:55 < camxes> SyntaxError: Expected [,] but "b" found. 08:55 < _mukti_> I didn't realized that {cu} was forbidden in that position. Silly. 09:07 < Reid_> coi mi cnino la lojban. 09:07 < Ilmen> coi 09:07 < Ilmen> mi slabu la .lojban. 09:08 < Reid_> au mi ji'a go'i 09:08 < Ilmen> za'a do xa'o certu 09:08 < Ilmen> .i ma bangu do 09:09 < Reid_> la gliban 09:09 < Ilmen> je'e .i mi ji'a se bangu lo glibau ku po'onai 09:10 < Ilmen> .i xu do cilre fo la xunre cukta 09:10 < Ilmen> ku no'u CLL 09:10 < Ilmen> pau 09:11 < Reid_> cilre fo zo'e pe lu la lojban. mo li'u 09:11 < Ilmen> je'e 09:12 < Reid_> sa 09:12 < Reid_> mi cfipu tu'a le cmene 09:12 < Ilmen> .i ca lo nu mi co'a cilre kei mi ji'a pilno la'e lu la .lojban. mo li'u 09:12 < Ilmen> .i do ba'e se cfipu 09:12 < Reid_> la'o gy. Lojban for Beginners gy 09:12 < Ilmen> ŭa 09:12 < Ilmen> je'e 09:12 < Ilmen> .i mi no roi tcidu .i ku'i lo cmene cu slabu 09:13 < Ilmen> (I've never read it; but I know it by its name) 09:13 < Reid_> .ua 09:14 < Ilmen> .i ra'u pu ku mi pilno la boxna ve cilre fi lo ka cilre 09:14 < gleki> do se slabu la lojban i ku'i ji'a je'u do ba'e slabu la lojban to ri se irci toi 09:14 < Ilmen> (I've mainly used the Wave Lessons for learning) 09:14 < Reid_> je'e 09:15 < Ilmen> .i xu la boxna ve cilre do slabu 09:15 < Ilmen> (Do you know of the Wave Lessons?) 09:15 < Reid_> mi pu na snada lonu facki lo se cmene be la boxna ve cilre 09:16 < Ilmen> .i .ai mi judri dunda 09:16 < Reid_> ca mi snada sepi'o le glico cmene 09:16 < Ilmen> http://mw.lojban.org/index.php?title=Lojban_Wave_Lessons 09:16 < Reid_> ca ku 09:16 < Ilmen> .u'a dai 09:16 < Ilmen> pe'i do ca certu tavla 09:17 < Ilmen> .i ca ma do co'a cilre 09:17 < Reid_> lenu zbasu lo jufra kei cu temci barda 09:18 < Ilmen> By the way, "ca mi" is wrong, it means "at the same time as me / my existence" 09:18 < Reid_> .i puza ji'i pa masti 09:18 < Ilmen> You wanted "ca ku mi..." 09:18 < Reid_> .ua 09:18 < Ilmen> pu za lo masti be li ji'i pa 09:18 < Reid_> I realized that after I said it 09:18 < Ilmen> je'e 09:19 < Ilmen> mi zo'u pu za lo nanca be li ji'i ci co'a cilre 09:19 < Ilmen> (As for me, I've begun to learn about 3 years ago) 09:20 < Reid_> xu do ganse lonu le pensi tadji cu cenba 09:20 < gleki> *you began 09:20 < gleki> :P 09:21 < Ilmen> nandu preti 09:21 < Ilmen> ki'e la gleki 09:21 < Ilmen> .i mi pu cuxpei 09:21 < Ilmen> gi'e cuxna lo to'e mapti 09:22 < gleki> y 09:22 < gleki> ije'u frica smuni 09:22 < gleki> i la'cu'i do pu cuxna lo mapti 09:22 < gleki> Scientists have discovered medicines for many diseases this century 09:22 < gleki> i ku'i 09:22 < gleki> Scientists discovered medicines for many diseases in the 19th century 09:22 < Ilmen> lo nu glare cu fanta lo nu mi pensi .u'i 09:22 < Ilmen> .i .ai mi di'a citka 09:23 < Ilmen> (I'll go back eating) 09:23 < Reid_> u'i je'e 09:23 < Ilmen> doi la'oi Reid_ lo nu jbobau cilre cu pluka .a'o do 09:23 < Ilmen> (I hope you'll enjoy learning Lojban!) 09:23 < Ilmen> co'o 09:23 < Reid_> .a'o 09:23 < Reid_> co'o 09:52 < akmnlrse> top 09:52 < akmnlrse> si 09:57 < akmnlrse> .i .e'u lo bavla'i poi nu voksnu zo'u di'a pilno la jinme ke bunre kerfa 10:00 * mumble jorne co'u fa la'o fa Ilmen fa 10:02 < omni___> xagmau fi ma 10:03 < akmnlrse> fi lo ka (to ru'a toi) na xoi suksa co'u .akti 10:03 < omni___> ua 10:04 < omni___> .i mi ja'e lo nu na zgana cu du'e roi jundi lo drata ja cu zvati lo drata 10:04 * mumble jorne co'a fa la'o fa Ilmen fa 10:05 < akmnlrse> ku'i no drata be pa nu lo selfu xoi mentu li 10 cu spofu cu fasnu .i va'i pu lo nu spofu mo'u casnu 10:07 * mumble jorne co'a fa la'o fa gunro fa 10:08 < omni___> pei bau la logbanu ra'oi -ka sinxa lo zilkai .i xendyka .i va'i kamxendo 10:08 < akmnlrse> mi nelci .i ra'u lo jvaprefixi cu milxe lo ka mabla 10:10 < akmnlrse> s/no drata be pa nu lo selfu xoi mentu li 10 cu spofu cu fasnu/pa nu co'e li'o cu ro mei lo fasnu/ 10:10 < fenki> akmnlrse meant to say: ku'i pa nu co'e li'o cu ro mei lo fasnu .i va'i pu lo nu spofu mo'u casnu 10:10 < akmnlrse> .oi 10:11 < Ilmen> ĭe go'oi bunre 10:11 < omni___> ta'o ba'e do ma stidi da'i ra'a lu lo citno ba'o zei jatna li'u 10:12 < omni___> .i mi na se mansa lu lo citno jatna be ba'o li'u .i fegli 10:12 < akmnlrse> mansa mi 10:13 < omni___> .i je li'a lu lo citno ba'o jatna li'u zo'u na tanru (to doi na se slabu be lo se casnu toi) 10:13 < akmnlrse> citno ke be ba'o jatna sei feglymau 10:13 < omni___> .i lu be ba'o li'u zo'u na panra lo stura be lo na se pagbu be zo citno moi 10:14 < omni___> .i su'a ma brivla panra lo tense ja aspekte su'i lo brivla 10:14 < gleki> oi la selpahi cu djica lo nu la itkuiile na ambigu i ku'i na lojbau 10:15 < omni___> s/be lo na/be lo me lo na 10:15 < fenki> omni___ meant to say: .i lu be ba'o li'u zo'u na panra lo stura be lo me lo na se pagbu be zo citno moi 10:16 < omni___> guzmyrmelona 10:16 < akmnlrse> za'a la xorxo cu skicu zo ba'o se pi'o zo mulcabna 10:17 < akmnlrse> .i jai mulcabna jatna sei clani 10:17 < omni___> .i lo ka .ambigu na sarxe la .itku'ile .i JQ pu finti 24 karbi genrkasu ba'e su'i lo so'i srana sufyfikse te zu'e lo nu .ambigu no da (to mupli toi) 10:18 < omni___> .i ku'i do'e su drata pagbu be lo bangu cu sa'u na se vajni xu 10:18 < phma> kurbita 10:19 < gleki> tcilokoiotli 10:19 < phma> ki'a 10:19 < omni___> akmnlrse: ie .i ku'i .i'e pei sarcu da'i fa lo nu gau basti lo sumtcita va'o ro nu tanru djica 10:20 < gleki> phma: lo me'oi pumpkin 10:20 < akmnlrse> li'a na melbi 10:20 < omni___> .i citno ba'e ke mulcabna jatna sei fegli 10:21 < akmnlrse> da'i ka'e pilno zo voi 10:21 < omni___> citno voi ba'o jatna 10:21 < omni___> .i feglyme'a ie 10:21 < omni___> .i za'a zo voi danfu ro da 10:22 < akmnlrse> ta'o zo su'i pe ru noi jufra cu frica fi ma fe zo jo'u 10:22 < omni___> na birti .i jijnu pilno mu'i tu'a lo si'o gunma 10:22 < omni___> .i la'a na vajni 10:22 < akmnlrse> je'e 10:25 < omni___> ui lo cnino me zo zei zo'u mi zifre lo ka cusku lu noi zei je li'u pu lo nu lo panra cu zasti 10:29 < gleki> ga'a mi la lojban co'a itkuiile 10:49 * mumble jorne co'u fa la'o fa ctefaho fa 11:03 * nuzba @oka_iu_tcan: lojbanのwikipediaのlojbanの記事、結構頑張ってるな https://jbo.wikipedia.org/wiki/lojban. [http://bit.ly/1MhBR1r] 12:52 < akmnlrse> omni___: sei kucli zo'u xu zo tu'a ne la zifre ze'e ba dunli lo'u tu'a lo menli si tu'a lo menli le'u lo ka nonseljmi 13:14 * mumble jorne co'u fa la'o fa gunro fa 13:31 * mumble jorne co'u fa la'o fa taldar fa 13:47 * mumble jorne co'u fa la'o fa Ilmen fa 15:18 < omni___> di'a jundi 15:25 < omni___> akmnlrse: lo jai tai retsku zo'u ge'e ni'o no da prane 15:27 < omni___> .i mi lo ka mo kau cu zifre 19:24 * nuzba @slaimsan: mi tavla cei broda la .lojban .i go'i lo banjupunu [http://bit.ly/1KdQKPg] 19:53 < Roflknife34> coi ro 19:54 < phma> coi do 22:18 * nuzba @bobpoekert: @profanegeometry lojban [http://bit.ly/1GtolQ7] --- Day changed Sun Jul 19 2015 01:35 < rlpowell> gleki: Just for the record: < gleki> Why the first line of every paragraph is indented. -- becaus ethat's how books *work* in English? 01:39 < rlpowell> See for example http://practicaltypography.com/first-line-indents.html , and http://english.stackexchange.com/questions/8941/is-an-indentation-needed-for-a-new-paragraph ; the book is already too long to have spaces between paragraphs. 01:41 < rlpowell> Which doesn't mean that you don't have some good points about indentation, I'm just answering that one question. 01:41 < rlpowell> afk. 01:54 < gleki> rlpowell: yes, ofc. it's just that the result is ugly. 02:18 < rlpowell> gleki: OK. It's not your first language, and I've *no idea* what the typographical conventions are in ... Russian? ... so I thought I'd be clear. :) 02:18 < gleki> rlpowell: it's not about the first line itself but in when it's used together with examples. see full issue 02:18 < rlpowell> I stared at it for a while, though, and I think I agree that it's ugly WRT examples. 02:18 < rlpowell> I have some ideas; will try them out maybe tomorrow. 02:19 < gleki> so maybe make the indentation less 02:19 < gleki> and/or increase it for examples 03:43 * nuzba @_Vanessa_sary_: @willingtheweird Huh... Lojban is hard [http://bit.ly/1DomMmz] 03:53 * niek looks att that Twitter conversation. 03:53 < niek> o_O 04:32 < gleki> But for some easy silliness I can accept javascript. Like we do for postprocessing 04:32 < gleki> alta: lo vanci co ca mi citka 04:32 < mensi> (NIhO [I {<ca (¹lo [vanci NF] KU¹)> <FA mi>} {CU <citka NF> VAU}]) 04:33 < gleki> restored back to {ca lo vanci} using javascript 04:33 < gleki> but for e.g. {ni'o i} restorations i would do and actually did in peg 05:09 < zipcpi> "As for {seu} it should be written as {se .u} (two words) although to JVS it will probably get as {se u} without the dot." 05:09 < zipcpi> oi 05:09 < zipcpi> CLL stole an Eulojban word :p 05:10 * zipcpi tries entering {seu} to JVS 05:10 < zipcpi> Nope. 05:14 < zipcpi> vlaste: xismu (gimka) 05:14 < vlaste> zo xismu gimkamsmikezypro zo gismu noi catni gismu .i zo xismu gimkamsmikezypro no cipra gismu 05:14 < zipcpi> x clashes with g? Didn't know that 05:15 < zipcpi> vlaste: brudi (gimka) 05:15 < vlaste> zo brudi gimkamsmikezypro no catni gismu .i zo brudi gimkamsmikezypro no cipra gismu 05:15 < zipcpi> ... This one doesn't clash, but how was it used? 05:15 < zipcpi> vlaste: broda (gimka) 05:15 < vlaste> zo broda gimkamsmikezypro zo brode je zo brodi je zo brodo je zo brodu vu'o noi catni gismu .i zo broda gimkamsmikezypro no cipra gismu 05:15 < zipcpi> lol 05:22 < gleki> those are pseudo-gismu 05:22 < gleki> <cll>as an extreme example, “bradi”, “bredi”, “bridi”, and “brodi” (but fortunately not “brudi”) are all existing gismu. 05:22 < zipcpi> OK 05:24 < gleki> i started incorporating some features from zantufa. 05:24 < gleki> it's hard. 05:25 < gleki> the same language, the same source, but the rules are very different now 05:25 < gleki> and i havent yet looked at spagetufa 05:25 < zipcpi> Yeah... 05:26 < zipcpi> spagetufa has relyle'urafsi; that doesn't change previous grammar though. Not sure what features of it does 05:28 < zipcpi> Ah crap... Curtis broke {i'au} T.T 05:30 < zipcpi> There is sort of a problem with {po'o} and {ji'a} 05:30 < gleki> that they are not UI? 05:31 < zipcpi> In that UI grammar might not be so convenient for it 05:31 < zipcpi> Consider 05:31 < zipcpi> lo broda ku po'o cu brode 05:31 < zipcpi> lo po'o broda cu brode 05:31 < zipcpi> fa po'o lo broda cu brode 05:31 < zipcpi> What do they mean? 05:31 < gleki> 1=3 05:32 < zipcpi> OK, so you say 1=3, which seems to match common usage... 05:32 < zipcpi> But what about 05:32 < zipcpi> ca po'o lo broda cu brode 05:32 < zipcpi> coi la selpa'i 05:32 < zipcpi> You might want to read the upscroll 05:32 < Amnadya> .o'ai 05:33 < zipcpi> Since it deals with your favorite thing... scope 05:33 < Amnadya> In current usage 1=2=3. But 3 only = 1 because everything else is highly inconvenient with those in UI 05:33 < Roflknife34> But does everybody love someone? 05:34 < Amnadya> 3 *should* be distinct 05:34 < Amnadya> 1=2 is not controversial. 05:34 < zipcpi> Amnadya: Exactly, which is clearer when switching {fa} with a sumtcita 05:34 < Amnadya> The other scope is {lo broda po'o} which is != all the others 05:35 < zipcpi> OK that's really interesting 05:35 < zipcpi> Maybe there's actually a use for {sei'o}/NF then :p 05:36 < Amnadya> I've said a million times how inconvenient {ji'a}'s and {po'o}'s grammar is. I proposed BAhE, but maybe NAhE is also possible. Problem is it breaks all past usage. 05:36 < zipcpi> lol 05:36 < gleki> someone could finally add this to le uitki https://gist.github.com/durka/6b01c1e2ff4293457f6b 05:37 < zipcpi> exp: na'e mi cu broda 05:37 < Amnadya> In any case, I agree that {ca po'o ko'a} should not != {ca ko'a po'o} 05:37 < mensi> ([na'e mi LUhU] [cu {broda VAU}]) 05:37 < gleki> alta: na'e mi cu broda 05:37 < mensi> (NIhO [I {FA <na'e mi LUhU>} {cu <broda NF> VAU}]) 05:37 < gleki> :P 05:37 < Amnadya> It's just currently the case because of said annoying grammar of po'o 05:37 < Amnadya> Nobody wants to add in all those {ku} 05:37 < Roflknife34> ma jatna la Lojban I need to complain to them 05:37 < Roflknife34> :P 05:37 < Amnadya> Complain about what? 05:38 < gleki> who is jatna today? 05:38 < zipcpi> po'o'e, ji'a'e (Eulojban peu'o, jeu'a) zo'o 05:38 < Amnadya> {mi dansu ca lo cabdei .i mi go'i pu ji'a lo cabdei} 05:38 < Amnadya> Useful usage where {ji'a} attaches to {pu} 05:39 * zipcpi nods 05:39 < ctefaho> Amnadya: What, in your opinion, would be the ideal grammar for po'o and ji'a? 05:40 < Amnadya> Some prefix selma'o like BAhE or NAhE. 05:40 < ctefaho> je'e 05:41 < menli> camxes: na'e ke mi joi do 05:41 < camxes> SyntaxError: Expected [yY] but "o" found. 05:41 * gleki warning: your free cmavo space is depleted. Please free up more space. Start Cleaner now? 05:41 < zipcpi> lol 05:41 < Amnadya> Another suggestion was to create a cmavo that turns any UI into BAhE 05:41 < gleki> alta: na'e ke mi joi do 05:41 < mensi> (NIhO [I {FA <na'e (¹ke [mi {joi do}] KEhE¹) LUhU>} {CU <COhE VAU>}]) 05:42 < ctefaho> I sort of consider ba'e/za'e to expand to bi'a/zi'a 05:42 < zipcpi> Huh... 05:42 < menli> gerna: na'e bo ke mi joi do 05:42 < Amnadya> lol "expand" 05:42 < gleki> zantufa: na'e ke mi joi do 05:42 < zantufa> ([na'e {ke <mi (¹joi do¹)> KEhE}] VAU) 05:42 < mensi> O_0 05:42 < ctefaho> yes "expand" 05:43 < Amnadya> Btw, as always {voi} is another solution. 05:43 < gleki> now we need another set to go to the left 05:43 < zipcpi> What happend to {LUhU}? 05:43 < ctefaho> but as others seem to abuse primary stress and explicit ja'a as their means of emphasis, not such a big deal 05:43 < zipcpi> How does {voi}/{poi'i} help here? 05:43 < Amnadya> Avoiding the pesky {ku} in {lo broda ku po'o [cu] brode} can be done with {lo broda cu voi ke'a po'o brode}. (I do this now and then) 05:43 < gleki> it's an exchange of captives. {ba'e/za'e} go to the right. {po'o/ji'a} go to the left. 05:44 < gleki> zantufa: na'e ke mi joi do cu mo 05:44 < zantufa> ([na'e {ke <mi (¹joi do¹)> KEhE}] cu [mo VAU]) 05:44 < zipcpi> Amnadya: OK, but what if {lo broda} isn't the first sumti? 05:44 < Amnadya> Then you're fucked. zo'o 05:44 < zipcpi> lol 05:45 < menli> What do you think of using {FA ko'a zo'u...} for shared sumti in a connected bridi? 05:45 < gleki> zipcpi: then as someone said "It surely doesn't. Alas for altást." 05:46 < Amnadya> (*another* but unacceptably verbose option: {fu'e po'o lo broda fu'o}) 05:46 < zipcpi> I don't get it 05:46 < gleki> https://www.reddit.com/r/Ithkuil/comments/3d09vh/is_ca_affix_recursion_possible/ 05:46 < zipcpi> Amnadya: Perhaps if there was a version of {fu'e} that only works for one sumti 05:46 < gleki> gerna: na'e bo mi cu mo 05:46 < mensi> (0[{na'e bo mi LU'U} cu {mo VAU}])0 05:47 < gleki> no idea what guskant did to LUhU 05:47 < Amnadya> zantufa: ke po'o lo broda cu mo 05:47 < zantufa> ([{ke po'o} {lo broda KU} KEhE] cu [mo VAU]) 05:47 < zipcpi> Or like I said, BAhE versions of po'o and ji'a... but what should they be? 05:47 < Amnadya> Well... 05:48 < zipcpi> Huh 05:48 < ctefaho> po'oi'oi si 05:48 < gleki> en: po'a 05:48 < mensi> po'a = [BAI] ponse modal, 1st place: with possessor/owner... |>>> See ponse, po |>>> 05:48 < mensi> spheniscine 05:48 < gleki> someone stole {po'a} 05:49 < zipcpi> ... guskant did it first T.T 05:49 < gleki> en: ja'u 05:49 < mensi> 7 da se tolcri: ja'urjukni, ja'urjuknis, ja'urxe'a, molja'u, dzixe'a, dzixejbo'u, 05:49 < mensi> xantajycalku 05:49 < gleki> luckily not ja'u 05:49 < Amnadya> po'a: fi'o spoja be ri'a lo nu lo jbobau cu plujydu'e fe'u 05:49 < zipcpi> u'i 05:49 < Amnadya> Don't take my ja'u 05:49 < gleki> but then you dot have any more cmavo space on your hard drive for cache file :P 05:49 < gleki> but then you dont have any more cmavo space on your hard drive for cache file :P 05:50 < gleki> zo ja'u mo 05:50 < zipcpi> I like the {ke} thing, but it only works in zantufa? 05:50 < Amnadya> exp: ke mi 05:50 < mensi> SyntaxError: Expected [uU] or [yY] but "i" found. 05:50 < gleki> alta: ke mi mo 05:50 < mensi> (NIhO [I {ke <FA mi> KEhE} {mo VAU}]) 05:51 < zipcpi> alta: ke po'o mi mo 05:51 < mensi> (NIhO [I {<ke po'o> <FA mi> KEhE} {mo VAU}]) 05:51 < gleki> i think it's the issue still lying there in ilmentufa Issues section 05:51 < Amnadya> I already used {ke} when attaching UI to brivla, but it wasn't possible before to use on sumti! 05:51 < zipcpi> Hm yeah it's kinda important for logical connectives too 05:51 < gleki> https://github.com/Ilmen-vodhr/ilmentufa/issues/118 05:51 < gleki> cu'u la xorxes That's inherited from the LALR(1) limitations of YACC. The same applies to 05:51 < gleki> "ke broda gi'e brode ke'e gi'a brodi". It shouldn't be hard to fix with PEG. 05:52 < gleki> so more like a bug 05:52 < zipcpi> je'e 05:52 < gleki> not an official feature 05:53 < gleki> altatufa isn't fixed completely in it in bridi part due to lots of te sumti detector rules 05:53 < zipcpi> I'll probably use {ke po'o} now 05:53 < gleki> alta: ke broda gi'e brode ke'e gi'a brodi 05:53 < mensi> SyntaxError: Expected [,] but "b" found. 05:53 < gleki> clearly 05:53 < zipcpi> T.T 05:53 < gleki> zantufa: ke broda gi'e brode ke'e gi'a brodi 05:53 < zantufa> (CU [ke {<broda VAU> <gi'e CU (¹brode VAU¹) VAU>} ke'e VAU] [gi'a CU {brodi VAU} VAU]) 05:53 < Amnadya> zipcpi: I think I will, too. 05:54 < gleki> Amnadya: xu do jinvi lo du'u zo po'o cu ei cmavo na'e zo ui 05:54 < Amnadya> no'e go'i 05:55 < Amnadya> .i ro lo re cumki cu plixau 05:55 < zipcpi> I think it's kinda like BAhE, both prefix and postfix could be useful 05:55 < Amnadya> fi su'o da 05:55 < gleki> fau lo nu zo po'o cu cupr lo sko'opu cu frica lo drata cnima'o i ja'o na jetnu snima'o i ni'iku cmavo lo drata 05:55 < gleki> i oi mi cikysre 05:55 < gleki> i oi mi ciskysre 05:55 < gleki> fau lo nu zo po'o cu cupra lo sko'opu cu frica lo drata cnima'o i ja'o na jetnu cnima'o i ni'iku cmavo lo drata 05:55 < zipcpi> ue do pilno lo zi'ejvo iau zo'o 05:56 < gleki> ma zi'ejvo 05:56 < gleki> i lo zi'ejvo cu mo 05:56 < zipcpi> zo cikysre 05:56 < gleki> en: zi'ejvo 05:56 < mensi> [< zifre lujvo ≈ Free affix compound] lo nu mulno sisku zo'u: y no da se tolcri 05:56 < mensi> zi'ejvo[6367], zifyjvo[6978], zifryjvo[8008], zifylujvo[9037], zi'erlujvo[9526], ... 05:56 < zipcpi> .y. na se jbovlaste 05:56 < gleki> cikysre = x1 accidentally makes a mistype in the word x2 ({ciskysre} by default) 05:56 < Amnadya> ziflu'o zo'o sai 05:57 < zipcpi> skicu lo lujvo poi selpau lo rafsi poi na se javni 05:57 < zipcpi> la selpa'i cu finti 05:58 < gleki> i havent fixed that {ke} problem in altatufa because im suffering from DCY rule violation that i did to altatufa. 05:58 < Amnadya> zi'ejvo : x1 lujvo x2 gi'e poi'i lo finti be ke'a cu gleua li no lo rafsi javni 05:58 < zipcpi> u'i 05:58 < gleki> tu'a zo gleua vajni sidbo vau zo'onai 05:58 < gleki> pe'i 05:59 < gleki> i ma drata xe fanva fi lo lojbo fo lo lojbo 05:59 < gleki> i lu mi na kurji li'u xu i mabla 05:59 < Amnadya> go'i 05:59 < zipcpi> nasai vajni mi 05:59 < gleki> i lu na vajni mi li'u xu 05:59 < Amnadya> dramau 05:59 < gleki> na mulno dunli 06:00 < zipcpi> no'aiva'e vajni 06:00 < gleki> i da'i lo nu broda na vajni mi ije ku'i mi ca'o zukte lo ka broda 06:00 < Amnadya> x1 dukse lo ka to'e vajni x2 kei lo nu x2 cusku zo .oi ja cu pante vau va'o lo srana be x1 06:06 < zipcpi> me'ei gleua cu ka'ei bo'a mo'ebo'eva'ei se vajni bo'i jecu malsku 06:06 < gleki> la'a la'a 06:07 < zipcpi> .y. bo'a na bi'ai malsku 06:07 < zipcpi> le cusku cu malsku 06:08 < zipcpi> .y. zo malsku je zo malvla na se jbovlaste 06:08 < Amnadya> me'a kulpa 06:08 < zipcpi> u'i 06:09 < Amnadya> koi la su da jmina ro lo re mei 06:09 < Amnadya> .i ba'a nai mi pu troci tu'a lo re moi .i ku'i fliba lo ka finti lo ve skicu poi mansa mi 06:09 < zipcpi> ku'i mi na djuno lo du lo terbri cu ei mokau 06:10 < zipcpi> ie 06:10 < zipcpi> le terbri cu jai nabmi do ji'a la'a 06:11 < Amnadya> .i ji'a xu su da malvla fi lo jbobau 06:11 < Amnadya> ba'e pu'i 06:11 < zipcpi> zo gleua je zo oi'o 06:12 < zipcpi> ju'ocu'i 06:12 < Amnadya> ju'o nai (to na djuno toi) 06:12 < zipcpi> mi na birti tau zo oi'o .i simsa lo nu troci lo ka ca'engau 06:13 < zipcpi> kei ju'ei no'e snada 06:14 < Amnadya> (to ta'o ke'u .i'e sai do jbota'a sei mi je'a .i'onmo toi) 06:15 < gleki> lo'e jbopre na jbota'a 06:15 < zipcpi> u'i 06:15 < Amnadya> (to na ku so'i da snada lo ka galfi lo nu ze'u casnu bau lo glico kei lo nu co'a je'a jbota'a toi) 06:16 * Amnadya ca nitcu je .aidji lo ka jukpa 06:16 < zipcpi> je'e 06:21 < zipcpi> ei mi gunka co'o 06:51 < durka42> coi 07:32 < gleki> coi 07:33 < gleki> dy 07:33 < gleki> to me'o dy sinxa lo dotco toi 07:34 < niftg> .ai mi ji'a rinsa 07:43 < durka42> coi la .niftyg. 07:44 < durka42> to mi ca zvati lo dotygu'e 07:45 < niftg> coi dy ki'a nai bu'o 07:47 < niftg> pu'i xu penmi lo dotygu'e jbopre 07:48 < niftg> .u'a nai jorne spofu za'a 08:01 < gleki> alta: puza lo djedi mi kelci lo gentufa 08:01 < mensi> (NIhO [I {<(¹pu KU¹) (¹za [lo {djedi NF} KU]¹)> <FA mi>} {CU <kelci NF> <FE (¹lo [gentufa NF] KU¹)> VAU}]) 08:02 < gleki> i want to somehow show that {za} is a tense_modal_modifier. but im not sure how 08:07 < gleki> alta: ca vi pu za mi pu za mo 08:07 < mensi> (NIhO [I {<FA ZOhE> <ca KU> <vi KU> <pu za> mi} {CU <(¹pu za¹) mo> VAU}]) 08:07 < gleki> maybe like this 08:19 < gleki> I only see that slotting is needed here but how to present it ... sanji no da 08:19 < gleki> i la'a vajni falonuco'u glita'a 08:23 < durka42> mabla kibterjo'e 08:23 < durka42> xu .akti 08:26 < Amnadya> coi do 08:26 < niftg> coi 08:27 < gleki> i mi co'u djica lo nu zo za cmavo lo na se cmavo be zo pu 08:31 < gleki> i ai mi fatne stika la altatufa 08:33 < durka42> lo sumtcita nunpamebi cu banli 08:36 < gleki> mi pensi lo nu smuni 08:37 < gleki> i zo za cnegau lo ma'oi pu sumtcita 08:37 < gleki> i va'i lu uinai li'u simsa lu puzi li'u 08:38 < Amnadya> mo lu pu ku mi klama zi lo mentu li'u 08:39 < gleki> ji'a simsa lu uinai li'u 08:39 < gleki> i mi na djuno lo du'u ta'i makau gentufa 08:39 < Amnadya> mi za'o jijnu lo du'u lo'u pu za le'u je ro srana cu logji spofu 08:40 < gleki> i ei o gentufa cu cupra lo cnino ke sko'opu linji 08:40 < gleki> sa 08:40 < gleki> i ei lo gentufa cu cupra lo cnino ke sko'opu linji 08:40 < Amnadya> .i ku'i mi lazni ja'e lo nu nu'o troci lo ka je'urja'o 08:40 < gleki> i xu lu puje za li'u logji nunprane zmadu 08:43 < Amnadya> mi jijnu 08:44 < gleki> ma noi jbocei cu djuno lo satci smuni be zo za ibo paunai 08:45 < durka42> u'i 08:47 < durka42> Amnadya: ai pei ba cusku lu pujeza li'u 08:47 < gleki> i tcini falonu ba'e lo nu pilno cu rinka lo nu le bangu cu zenba lo ka jai se skicu fai lo xo kau tcila 08:47 < Amnadya> .ai na'i .i sa'u va'o lo nu da'i lu pu za li'u na'e drani cu ba bilga lo ka cusku lo drata .i ku'i nu'o djuno fa ma'a 08:48 < durka42> je'e 08:48 < durka42> ie 08:48 < gleki> i lo nu ba'e pilno lu puza li'u ba rinka lo nu ma'a zmadu jimpe fi lo smuni be zo za 08:48 < gleki> i cati lu puza li'u pagbu lo rarbau i'a jbobau 08:49 < gleki> i a'o na ambigu 08:49 < gleki> i va'o da'i lo nu na ambigu vau no da nabmi lo pilno i po'o nabmi lo skicu noi va'i jboskepre 08:51 < gleki> i ja'o mi fliba lo ka jimpe tu'a ma'oi zi i ai jai gau spofu sa i ai cikre lo drata pagbu be la altatufa 09:20 * nuzba @Rodericus: @markandeyaka @subetealanutria (como el lojban). Depende… Lo de la estética con el esperanto lo he visto como crítica muchas veces, [http://bit.ly/1HCPELU] 09:44 < Ilmen> I've created this page, feel free to suggest improvements or better wordings: http://mw.lojban.org/papri/BPFK_style_guidelines_for_Lojban_dictionary_definitions_%28proposal_draft%29 09:48 * nuzba @Rodericus: @markandeyaka @subetealanutria Por cierto, la lógica del lojban, por ejemplo, va en otro sentido, en el de la morfosintaxis, con resultados [http://bit.ly/1LkP75s] 09:50 < gleki> i have nothing to add. what i earlier suggested has already been implemented in La Bangu Google Spreadsheet source file: sumti types, splitting of definitions into several cells, glossing each place as nouns, as verbs, semantic families, frames, cross-links, examples, export to different formats ... 09:51 < gleki> take from there. 09:51 < gleki> e.g. attitudinals should clear show what is {ui} and what is {uinai}. not that "attitudinal: happiness - not-happiness" (?!) 09:52 < Ilmen> je'e 09:53 < gleki> it's just that i can repeat every single aspect of it. 09:53 < gleki> e.g. how are you going to generate English-to-Lojban dictionary. 09:53 < gleki> It is also a question what ideal style should be and what style can be supported by current JVS 1.0 09:54 < Ilmen> gleki: As for the jbo→epo translation, I've done more than half of it last night. I will continue later today, and add the result to the Google Doc. 09:54 < gleki> ki'e 09:59 < Ilmen> gleki: Yeah, the best solution for guidelines is to list ideal solutions together with recommendations for dealing with JVS's restrictions 10:01 < Ilmen> for example "Examples translations would ideally have a dedicated field, or for lack of it, be put into the note field." 10:03 < gleki> the problem is even if BPFK adopts some guidelines under JVS 1.0 restrictions who will fix existing defs given that JVS 2.0 will present new possibilities? 10:03 < gleki> who will fix knowing that in future JVS 2.0 will recommend re-fixing again 10:05 < Ilmen> ua nai ru'e 10:06 < gleki> if guidelines are applied then people will probably have to fix existing defs, right? 10:11 < Ilmen> Those would be recommendations, not obligations 10:12 < Ilmen> They would encourage word creators to use them and to try bettering existing definitions 10:14 < gleki> but then when JVS 2.0 is out new recommandations will be out making older one obsolete. 10:14 < Ilmen> I'm not sure what to do for the gismu official definitions, which clearly don't comply to the clarity and readability requirements of the proposed guidelines :p 10:14 < Ilmen> Maybe they'd invite people to propose better competing definitions 10:14 < Ilmen> which may be officially approved later 10:16 < gleki> oh maybe first develop an appropriate ideal realization of the dictionary and then edit them as we wish 10:16 < gleki> or maybe first develop an appropriate ideal realization of the dictionary and then edit them as we wish 10:33 < cliva> spagetufa: +rafcla merdjygau lo zi'evlyvla jo'u zo'e li'o 10:33 < spagetufa> SyntaxError: Expected [.\t\n\r?!\u0020] or end of input but "m" found. 10:33 < spagetufa> Unrecognized option rafcla - use 'spagetufa: +help' for a list of all options 10:33 < cliva> spagetufa: +rafcau merdjygau lo zi'evlyvla jo'u zo'e li'o 10:33 < spagetufa> (CU [mèrdjygau {<FE (¹[lo zi'èvlyvla KU] jo'u [zo'e li'o]¹)> VAU}]) 10:33 < Ilmen> doi la gleki mi toltce se nandu tau lu ko lanli lo se cuxna be fi lo nu sivni do’e lo do skami li'u 10:34 < gleki> y 10:34 < durka42> ma claxu lo rafsi 10:34 < Ilmen> go'i 10:34 < cliva> lo vlaturge'a poi se pilno va'o lo nu zo rafcla cu .optiio se cuxna 10:36 < cliva> y zo rafcau 10:36 < durka42> ua drata vlatuge 10:36 < cliva> milxe co drata 10:36 < durka42> ku'i curmi fa'u tolcurmi lo mintu xu 10:36 < cliva> .i ru'a ro se curmi be lo fadni vlatuge cu se curmi lo rafcau 10:36 < cliva> .i lo fatne zo'u lo rafcau cu curmi zo'oi merdjygau mu'a 10:37 < durka42> ua ue 10:37 < cliva> .i na mulno lo ka jetnu kei fa lo du claxu lo rafsi .i lo valslinku'i javni za'o pagbu 10:37 < gleki> Ilmen: i ko lanli lo nu sivni vau se cuxna pe lo do skami 10:37 < gleki> i pei 10:37 < Ilmen> gleki: "Analizu tion, kio elektitas inter ..." 10:37 < Ilmen> ua 10:37 < Ilmen> je'e 10:38 < gleki> inter privatado de via komputilo 10:39 < Ilmen> .i lanli lo pu se cu'uxna be fo lo ka sepi'o ce'u lo datni pe lo skami cu snura se mipri 10:39 < gleki> clani 10:39 < gleki> u'i 10:40 < Ilmen> lanli lo pu se cu'uxna be fo lo ka nurgau lo sivni datni 10:40 < gleki> sivni datni se cuxna lanli 10:41 < Ilmen> je'e 10:50 < Ilmen> lu ko na pilno pa lerpoijaspu lo vrici jaspu li'u ŭa nai ru'e 10:50 < Ilmen> .i lo jaspu na zilkai 10:51 < cliva> .i za'a ji'i re da pe le vlatuge za'o cfila 10:51 < durka42> ko'oi na ku pa lerpoijaspu fo lo vrici 10:51 < cliva> si zi'e cfila 10:51 < durka42> y 10:51 < cliva> spagetufa: cizdaklybroda 10:51 < spagetufa> (CU [ciz/dakly/bròda VAU]) 10:51 < durka42> ko'oi naku pada lerpoijaspu fo lo vrici 10:51 < cliva> spagetufa: +rafcau cizdaklybroda 10:51 < spagetufa> ([FA {<ci BOI> zdaklybròda KU}] VAU) 10:52 < cliva> spagetufa: +rafcau genrkasu 10:52 < spagetufa> SyntaxError: Expected [oO] but "e" found. 10:52 < durka42> uh oh 10:52 < cliva> .i ku'i ro valsi poi ca da se pilno zo'u cipyzu'e tau ki .i si'au lo re mei ku po'o co'e 10:52 < cliva> .i go'e ba'o 10:53 < durka42> naku pada lerpoijaspu ko fi so'i da 10:55 < Ilmen> mi jmina ca ti doi la gleki 10:55 < Ilmen> fe lo mi xe fanva 10:56 < durka42> xu do fanva la'au slari vanjba 10:56 < Ilmen> .i pei doi la gleki 10:57 < Ilmen> https://docs.google.com/document/d/1iDee0cS55ogA2QSxaNE_nFnFSaYtBMltpK4q6R8n2qU/edit# 10:57 < Ilmen> durka42 ^ 10:57 < durka42> ua 10:57 < durka42> lo mi kibyterjo'e cu mabla .i mi ba viska ca lo za'umoi co nanca 10:57 < Ilmen> .ei pei mi jmina lo mi cmene doi la gleki 10:58 < durka42> ua ba'o mulnybi 10:58 < Ilmen> .i je'e la .durkas. 10:58 < durka42> uabu'o lo spofu fonxa selkei cu co'e 10:58 < Ilmen> Guten Tag :) 10:59 < Ilmen> ki'ai gunta tagji 10:59 < durka42> gu tenta gi co'e 11:00 < Ilmen> ta'o nai xu lo se gasnu be mi cu mapti lo pu se cpedu be do doi la gleki .i va'i ŭo pei 11:01 < Ilmen> .oi se'i lo pa zo lo snada lo ka nerkla lo mi spero vlamei 11:03 < durka42> .u'i 11:03 < durka42> lo spero nu xorlo 11:05 < Amnadya> .u'e Bob participating on lojban-beginners 11:05 < cliva> ... ki'u ma zoi fa stressed_y_less_rafsi? initial_pair y fa pagbu la slinku'i ku noi javni 11:05 < cliva> .i pagbu fe ji'a lo me la catni .camxes. moi 11:07 < durka42> consonant y_less_rafsi* stressed_y_less_rafsi? initial_pair y 11:07 < Amnadya> .itku'ile ga'a mi noi ke po'o lo to'anzu vlaturge'a noi sampu sai cu slabu ke'a 11:08 < durka42> c pat pat pat cly 11:08 < durka42> k: cpatpatpatcly 11:08 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "c" found. 11:08 < durka42> k: cpatpatpatcli 11:08 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "c" found. 11:08 < cliva> spagetufa: +Rslinkuhi_ignore cpatpatpatclybroda 11:08 < spagetufa> (cpatpatpatcly [b r o d a]) 11:08 < cliva> iep 11:08 < durka42> uanai .i no brivla cu se fanmo .ybu 11:09 < Ilmen> ua ISTM = It seems to me 11:09 < cliva> .ai vimcu lo ka'e se pedgycu'a je cu refcfagau lo cipra 11:09 < cliva> .i lo me zo genrkasu moi zo'u pu sampu fa lo nu cikre vau ta'o 11:10 < Amnadya> b r o d a 11:11 < cliva> ciprdebuge po'o .optiio 11:13 < Ilmen> cfikalte te cuxna 11:13 < cliva> ki'e 11:13 < Ilmen> xu naku melbi fa zo cfikalte 11:13 < Ilmen> :) 11:14 < Ilmen> lo cfika kalte cu cfikalte 11:14 < cliva> spagetufa: +Rslinkuhi_ignore cpatpatpatclybroda 11:14 < spagetufa> (cpatpatpatcly [b r o d a]) 11:15 < cliva> spagetufa: +Rslinkuhi_ignore cpatpatpatclybroda 11:15 < spagetufa> SyntaxError: Expected undefined but "c" found. 11:15 < cliva> spagetufa: cizdaklybroda 11:15 < spagetufa> (CU [ciz/dakly/bròda VAU]) 11:15 < cliva> spagetufa: +rafcau cizdaklybroda 11:15 < spagetufa> (CU [cizdaklybròda VAU]) 11:17 < cliva> .i lo se'irci vreji cipra zo'u snada 11:19 < Amnadya> .oi .i ko morjigau .e'o mi lo su'u ma kau tadji lo ka gau ce la korpora-zei-sisku cu cupra lo lujvo liste 11:20 < cliva> mi na pilno KZS te zu'e ku .i gasnu lo nu la .camxes. cu gentufa ro brivla poi pagbu lo vreji 11:21 < cliva> (to ta'o .ei pei lo vlavelcki be zo gentufa cu se stika toi) 11:21 < gleki> Ilmen: ei do jmina lo do cmene 11:21 < Amnadya> je'e .i ku'i mi na tavla fi lo me do moi 11:49 * nuzba @Kohelang: アンケートでlojban書いたのおかゆ? [http://bit.ly/1Icu8iC] 12:10 < clakre> .i mi na jimpe fi lu "me do moi" li'u 12:11 < clakre> because do is not a cardinal number 12:13 < clakre> so can moi be used for things that are not numbers then? 12:13 < Amnadya> {me SUMTI moi} is a special grammar rule 12:14 < clakre> je'e 12:14 < Amnadya> Originally invented so that you can use {moi} with complicated math expressions, .e.g {me li pa su'i pa me'u moi} 12:14 < Amnadya> "The [1+1]th" 12:15 < Amnadya> (where "complicated" is anything that isn't just a string of digits) 12:16 < Amnadya> But {me mi moi} means "mine" 12:16 < Amnadya> {me do moi} "yours" 12:16 < Amnadya> (me-th, you-th) 12:17 < clakre> ki'e! 12:38 < Ilmen> je'e la gleki 12:42 < Ilmen> .i lo pa moi linji zo'u la'a gau mi zo gi'a zo gi'e basti 12:42 < Ilmen> .i linji seju vlali'i 12:43 < Ilmen> .i ta'o ma poi drata cu kansa fi lo linsi nu fanva 12:59 < Amnadya> ma'a co'a jimpe lo du'u lu ma poi li'u drani na ku 12:59 < Amnadya> .i .ei co'a pilno lo drani 13:10 < Ilmen> .a'o naku ba zi facki lo du'u lo'u to'e mu'i nai li'u ja lo'u ki'e sidju le'u ja lo'u ma kau le'u cu mulno lo ka toldra 13:10 < Ilmen> .i sa'e .a'o jitfa .u'i 13:12 < Ilmen> .i ji'a lo du'u ge zo lo drani naku gi zo le drani ro da 13:14 < Ilmen> .i mi di'u xamsku .i .u'i ĭa nai 13:19 < Ilmen> Amnadya: http://mw.lojban.org/papri/BPFK_style_guidelines_for_Lojban_dictionary_definitions_%28proposal_draft%29 --- ŭa pei .i .e'o do'o zifre lo ka stidi 13:20 < Ilmen> .i la'a mi ca lo nu mansa banzuni cu cpedi lo ka cmicu'a ku lo kamni menre 13:20 < Ilmen> se ju cmima 13:20 < Ilmen> .i je'au cpedu 13:30 * nuzba @ishimirai: アポストロフィーはタイピングするうえでデメリットだな。エスペラントはもう古い!プログラミングもできる最も完成された人工言語 - ロジバン - Wikipedia https://ja.wikipedia.org/wiki/%E3%83%AD%E3%82%B8%E3%83%90%E3%83%B3 [http://bit.ly/1Gu529A] 14:38 < durka42> xm 14:38 < durka42> ka'eku mi voikla la .berlin. 14:38 < Amnadya> ue 14:40 < durka42> li ji'i renono cu jdima lo nu vofli litru la .stutygart. ce'o la .berlin. ce'o la .frankfrt. 14:41 < durka42> to li'a lo xotli ji'a cu jai se jdima 14:41 < Amnadya> li'a kargu sai .i pa mei lo krinu be lo nu mi uenmo 14:41 < durka42> ku'i .ei mi catlu la gacri ckana je cersai 14:42 < durka42> do cusku lo sedu'u lo nunvoi ji lo xotli cu kargu sai 14:42 < Amnadya> lo xotli 14:42 < Amnadya> si 14:42 < Amnadya> .u'i 14:42 < Amnadya> .i lo nu voikla 14:45 < durka42> ku'i cumki fa lo nu nalmabla .iki'ubo lo ckule cu pleji fo lo .tubingen. pagbu be lo nunli'u 14:46 < durka42> mi ca'o retsku lo sedu'u .ei mi vitke makau .i noda poi remei cu jinvi lo pamei 14:47 < Amnadya> .y di'u 14:47 < Amnadya> mi co'a snada lo ka smadi lo du'u ma kau smuni se djica 14:48 < durka42> .ei mi klama la .berlin. je la .munytcen. je la xekri cicfoi je la .paris. je je je je 14:48 < Amnadya> .i na djuno .i lo ze'u xabju be la .berlin. co'u manci lo cinri dinju 14:48 < durka42> li'a 14:49 < durka42> .i la cliva cu xabju ma 14:49 < durka42> ku'i li'a tezu'e lo nu litru kei mi pensi lo jboprenunpe'i po'onai 14:50 < Amnadya> xabju lo xelveto ke ralju tcadu 14:50 < Amnadya> .i tsurixi 14:51 < durka42> ie la tsurixi ji'a cu jai se stidi vau .u'i 14:51 < durka42> ku'i la .suuitserland. cu jai kargu ti'e 14:53 < durka42> lo tolkargyrai be lo gairckajvecersai co'e zo'u li rebi cu jdima tu'a lo pa nicte 14:53 < durka42> .i na nalcumki .u'i 14:53 < durka42> .ei mi pensi pensi 14:55 < durka42> to lo smadi jufra zo'u lo ro prenu cu simxu lo ka drata fi lo ka jinvi makau 14:57 < durka42> ua na tolkargyrai .i pada sejdima li repa .iku'i lo dukse zabna simsu ta'e naljetnu ba'a 15:00 < Ilmen> cu dai 15:00 < Ilmen> ni'o di'ai te cuxna lanli 15:02 < durka42> zo ta'e basti zo cu 15:02 < Ilmen> ŭa mi tcidu srera 15:02 < Ilmen> pe tcidu zo na'e fi zo ta'e 15:03 < durka42> je'e 15:05 < durka42> ni'o .ei mi sipna ga'u lo nunjdice 15:06 < Ilmen> Vielleicht bin ich zu sehr müde. 15:07 < durka42> mi na se bangu lo dotco 15:07 < Ilmen> Gute Nacht. 15:07 < Amnadya> lu zu müde li'u drani 15:07 < Ilmen> va'i la'a mi tatpi dukse 15:07 < Ilmen> ŭa ki'e stidi ja dragau 15:08 < Amnadya> .ei ca su da gasnu lo jbonunsla be bu'u lo do cmana 15:08 < Ilmen> .i dukse lo ka simsa lo me lo glibau moi .u'i 15:08 < Amnadya> zo müde ckilu valsi .i se ki'u bo zo sehr na sarcu 15:09 < Amnadya> .i pilno zo sehr va'o lo na'e ckilu valsi no'u mu'a zo zögern 15:09 < Ilmen> je'e .i gau do mi cilre .i'o 15:09 < Amnadya> Du hast zu sehr gezögert 15:10 < Amnadya> You hesitated too much 15:10 < Ilmen> do cuxpei dukse 15:10 < durka42> ba'e mi roroi cuxpei dukse 15:10 < durka42> co'o 15:10 < Amnadya> co'o 15:11 < Ilmen> co'o 15:11 < Ilmen> mi'e si'abzu 15:11 < Amnadya> je'e co'o do ji'a 15:41 < Exodus> Coi! 15:46 < Amnadya> coi 15:46 < raien> coi 15:47 < raien> do mo 15:47 < Amnadya> tatpi ki'u lo nu nicte 15:48 < Amnadya> .i do mo 15:48 < raien> tatpi ji'a 17:01 < noncomcinse> Is Lojbanwiseuse crimethink? 17:02 < noncomcinse> .i mi tcidu la'e la'au pasobivo li 17:02 < noncomcinse> si li'u 17:53 < zipcpi> be'e mensi 18:06 < noncomcinse> mo coi 18:07 < zipcpi> coi 18:08 < zipcpi> xu do pu viska tau la e'ulojban 18:10 < noncomcinse> nago'i 18:11 < zipcpi> http://mw.lojban.org/papri/Eulojban 18:12 < noncomcinse> ua 18:13 < noncomcinse> Speaking of valid cipra dipthongs, is {uai} ([why]) a thing anywhere? 18:13 < zipcpi> It is, but it has a different meaning 18:13 < zipcpi> http://jbovlaste.lojban.org/dict/uai 18:14 < noncomcinse> The meaning isn't as important as the valid dipthong being a thing. 18:14 < zipcpi> Ah, then yes 18:14 < noncomcinse> xu do selsau zo cu 18:14 < noncomcinse> si zo cu'au 18:14 < zipcpi> ja'a 18:14 < noncomcinse> I thought of a better idea than that. 18:14 < zipcpi> Ah... you may also want to look at {i'au} 18:15 < noncomcinse> I've seen it. 18:15 < zipcpi> K 18:15 < zipcpi> So what's the idea? 18:15 < noncomcinse> ca'e: replace current cu'au definition 18:16 < noncomcinse> zo cu'au famyma'o ma'oi cu'ai 18:16 < noncomcinse> And basically, cu'ai lets you set a point in forethought to terminate to. 18:17 < zipcpi> lol cu'ai is taken by one of Curtis' meksyma'o... kinda annoying that... wish we could just get rid of them all and use brimekso instead 18:17 < zipcpi> Hmm... this is different from {fu'e}... {fu'o}? 18:18 < noncomcinse> .i mi pensi lonu mi puza ckaji cu'ai loka cnino loka tadni la .lojban. [ku vaukeiku vaukeiku] cu'au .e loka nitcu lonu lo vlaste cu sidju mi loka tavla fo la .lojban. 18:19 < noncomcinse> fu'e/fu'o sets UI over a space, I thought. 18:19 < zipcpi> Ah 18:19 < zipcpi> So you mean, cu'ai sets a "level" to terminate to 18:19 < noncomcinse> Yeah. 18:19 < noncomcinse> And you can stack them. 18:20 < zipcpi> Hmm... quite similar, but maybe not identical, to my suggestion for tu'e...tu'u 18:20 < noncomcinse> stidi ma 18:20 < zipcpi> https://groups.google.com/forum/#!topic/lojban/NRUIud_OCj8 18:20 < noncomcinse> I thought of tu'e tu'u when I came up with this 18:21 < zipcpi> Longish discussion, but basically I want to be able to use tu'e...tu'u within a NU, so as to include a complex multi-statement passage in e.g. ifle2 or rivbi2 18:21 < noncomcinse> eg, "poi BRIDI soft-{.i} BRIDI ku'o"? 18:22 < zipcpi> Yes 18:22 < zipcpi> Could work in NOI too 18:22 < noncomcinse> I've wanted that, too. 18:22 < noncomcinse> any sub-bridi 18:22 < noncomcinse> xoi 18:22 < zipcpi> There is ju'ei for "soft-i" but not in any parsers yet 18:22 < noncomcinse> xm 18:23 < zipcpi> But tu'e...tu'u still has an advantage over ju'ei, namely, for very complex passages, it's nice to have a "new level 0" for .i to reset to 18:23 < noncomcinse> my idea for cu'ai/cu'au, could be replaced with gu/gi<anything> 18:24 < noncomcinse> Wait, so you want to usw tu'e to make .i not start a new sentence? 18:24 < zipcpi> It starts a new sentence alright, but the entire tu'e...tu'u goes within a sub-bridi 18:25 < zipcpi> Basically, sentences within sentences 18:25 < noncomcinse> mu'a {mi zgana lo nanmu poi tu'e jgari lo badna .i pelxu badna}, zoma famyma'o 18:25 < zipcpi> Not too different from what {lu...li'u} already does 18:25 < noncomcinse> s/zoma/zo'u ma 18:25 < fenki> noncomcinse meant to say: mu'a {mi zgana lo nanmu poi tu'e jgari lo badna .i pelxu badna}, zo'u ma famyma'o 18:25 < zipcpi> tu'u 18:26 < noncomcinse> any vau? 18:26 < noncomcinse> vau tu'u ku'o vau? 18:26 < zipcpi> Seems so 18:26 < zipcpi> zantufa: mi zgana lo nanmu poi tu'e jgari lo badna .i pelxu badna 18:26 < zantufa> (mi CU [zgana {lo <nanmu (¹poi [tu'e {<CU (²jgari [lo badna KU] VAU²)> <i (²CU [pelxu badna] VAU²)>} TUhU] KUhO¹)> KU} VAU]) 18:27 < noncomcinse> so it starts a new sentence, but say I had {.i} at the beginning of that text. The one inside tu'e isn't the same level as the one at the beginning? 18:28 < zipcpi> Yes 18:28 < zipcpi> Like I said, lu...li'u and to...toi already work that way 18:28 < noncomcinse> With my original cu'au as wildcard famyma'o, things like to/lu made it weird if you wanted to use xiro 18:29 < zipcpi> Yeah... 18:29 < noncomcinse> Because xiro would break lu if it didn't stop. 18:29 < zipcpi> i'au does have a similar problem but I just ca'engau that it's not more powerful than .i 18:31 < noncomcinse> if you never opened a subbridi you could just use .i'au every time 18:31 < noncomcinse> not that you'd want to 18:32 < zipcpi> heh 18:32 < noncomcinse> I started reading 1984 and now I 18:33 < noncomcinse> have started talking to people in newspeak 18:33 < zipcpi> lol 18:33 < zipcpi> Not too different from lujvo :p 18:33 < noncomcinse> Honestly, I think that's why it comes so easily. 18:34 < noncomcinse> The only hard part is weird shortenings of words. 18:35 < noncomcinse> Oldspeakful wrds oft bec newspeakful unwrds. 18:36 < zipcpi> lol Eulojban... how I wish there is space for that categorical pronoun set 18:38 < zipcpi> It's exactly why I created Eulojban... it's part "how I wish there is space for this", part "personal madness room where I kill most necessasity for {kei} with NU variants", part satire 18:41 < noncomcinse> Is it the case that /θ/ counts as {'}? 18:41 < zipcpi> Theoretically yes 18:42 < noncomcinse> I had the idea for a dialect that pronounced me'oxy as /θ/ once. 18:42 < zipcpi> Ah. 18:42 < noncomcinse> And then in the last two days I actually properly learned how to do a voiced velar fricatice, as opposed to a uvular. 18:43 < zipcpi> lol 18:59 < zipcpi> It's funny, cause I even see more use for Curtis-zai'e than meksyma'o T.T 19:00 < zipcpi> That I so rudely overriden with my own zai'e :p 19:04 < zipcpi> Though I'd rather just make {lilo'o} or {me'olo'o} grammatical. I have a hard time thinking of when I'd actually want to use it in place of BY or PA 19:04 < zipcpi> And the selma'o XOhEhOhE.... I don't know how it should function 19:04 < zipcpi> Looks like it's meant to be something that can be both BY or PA 19:05 < zipcpi> {bu'o'e} I could see a use for. It's good for spelling out texts where some characters are illegible or unknown 19:06 < zipcpi> alta: lilo'o 19:06 < mensi> SyntaxError: Expected [uU] or [yY] but "i" found. 19:06 < zipcpi> Hm... something that Gleki needs to fix :p 19:07 < noncomcinse> zai'e? 19:08 < zipcpi> My-zai'e is a BAhE, used to mark jargon / "free-space". Even has a rafsi, -zam- 19:08 < zipcpi> Curtis-zai'e is "empty string" 19:09 < zipcpi> http://jbovlaste.lojban.org/dict/zai'e 19:10 < zipcpi> This is a very old idea actually... came up with it during my hiatus... though my first iteration was just assigning a rafsi to {za'e} 19:10 < zipcpi> Then I thought that {za'e} should be reserved for "on-the-spot" terms and slang terms 19:12 < zipcpi> lol {zi'a} is experimental; never noticed that 19:21 < zipcpi> I probably should learn PEG someday and get my own pet parser. If I'm gonna abuse the grammar ten ways to Sunday it's nice to have a proof-of-concept 19:22 < zipcpi> Could probably even train it to recognize Eulojban pronouns lol 19:23 < zipcpi> Can't decide if I should call it maltufa or eultufa :p 19:24 < noncomcinse> en: zi'a 19:24 < mensi> zi'a = [UI3a] nonce-word indicator; indicates previous word is nonce-creation and may be nonstandard |>>> Equivalent to 19:24 < mensi> za'e but in selma'o UI. See also za'e, pe'a, xo'o |>>> gleki 19:24 < zipcpi> ba'e:bi'a za'e:zi'a 19:24 < noncomcinse> when you use tu'a 19:25 < noncomcinse> which has the equivalent jai 19:25 < noncomcinse> which has the FA place fai 19:25 < zipcpi> bi'a is also experimental lol 19:25 < noncomcinse> is there any way to do the fai for tu'a? 19:25 < zipcpi> Oh that one 19:26 < zipcpi> Yeah it's a problem that selpa'i also recognized... he talked about it in an old blog post, using {kai'a} for it. I unwittingly defined {kai'a} differently 19:27 < zipcpi> His suggestion is something like 19:27 < zipcpi> mi djuno tu'a do *kai'a lo ka kanro 19:28 < zipcpi> Something like JOI, though he also noted a use-case where he thought JOI wasn't good enough 19:29 < zipcpi> I'm not sure myself... Maybe {mi djuno tu'a do fe lo ka kanro}? 19:29 < zipcpi> Though it hasn't been quite decided how assigning two sumti to the same terbri place work 19:30 < zipcpi> {kai'i} is open if selpa'i-kai'a should prove useful 19:31 < zipcpi> But I personally think the most exemplary use-case is when you've already gone "tu'a ko'a" but want to go back and clarify it 19:32 < zipcpi> Or deliberately want to front it 19:33 < zipcpi> {kai} is one of the less-useful BAI though 19:35 * zipcpi checks KZS 19:35 < zipcpi> Hm a lot of "sekai" 19:39 < zipcpi> mensi: doi gleki Maybe allow {lilo'o} in altatufa? {me'olo'o} might be useful for "empty string" 19:39 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.gleki.gy. di'a cusku da 19:40 < zipcpi> alta: li ry re dy re 19:40 < mensi> (NIhO [I {FA <li (¹[ry {re dy re}] BOI¹) LOhO>} {CU <COhE VAU>}]) 19:40 < zipcpi> alta: abu za'ure'u cusku 19:40 < mensi> SyntaxError: Expected [,] but "c" found. 19:40 < zipcpi> alta: abu cusku 19:40 < mensi> (NIhO [I {FA <a bu> BOI} {CU <cusku NF> VAU}]) 19:40 < zipcpi> Huh not sure what he did there 19:41 < zipcpi> alta: abu boi za'ure'u cusku 19:41 < mensi> (NIhO [I {FA <a bu> boi} {CU <(¹za'u re'u¹) (¹cusku NF¹)> VAU}]) 19:41 < zipcpi> zantufa: lilo'o 19:41 < zantufa> minajimpe: .i kanpe tu'a [uU] .onai [yY] .i ku'i "i" cu se tolcri 19:41 < zipcpi> spagetufa: lilo'o 19:41 < spagetufa> SyntaxError: Expected [.\t\n\r?!\u0020], [0123456789], [rR] or end of input but "l" found. 19:41 < zipcpi> Hahahaha 19:41 < zipcpi> exp: luli'u 19:41 < mensi> ([lu li'u] VAU) 19:42 < zipcpi> spagetufa: abu za'ure'u cusku 19:42 < spagetufa> (CU [{<(¹[{. <.a bu>} za'u] BOI¹) re'u> cùsku} VAU]) 19:43 < zipcpi> spagetufa: +lerfu abu za'ure'u cusku 19:43 < spagetufa> ([FA {<. (¹.a bu¹)> BOI}] [CU {<(¹[za'u BOI] re'u¹) cùsku> VAU}]) 19:43 < zipcpi> off: abu za'ure'u cusku 19:43 < mensi> SyntaxError: Expected [,] but "c" found. 19:43 < zipcpi> Riiiiight 19:44 < zipcpi> So alta isn't doing anything silly there 19:44 < zipcpi> Alta just never implemented la xornunsei 19:44 < zipcpi> Gleki doesn't like it 19:44 < noncomcinse> whar 19:44 < noncomcinse> s/r/t's that?/ 19:44 < fenki> noncomcinse meant to say: what's that? 19:45 < zipcpi> My own name for the xorxes proposal to prevent letter-strings from containing PA 19:45 < zipcpi> Compare 19:45 < zipcpi> exp: abu za'ure'u cusku 19:45 < mensi> ([{a bu} BOI] [CU {<za'u re'u> cusku} VAU]) 19:45 < noncomcinse> "me'o by cy pa" 19:46 < zipcpi> Oh yes that is affect too, which I'm not too fond of 19:46 < zipcpi> exp: me'o by cy pa 19:46 < mensi> SyntaxError: Expected [yY] but "a" found. 19:46 < zipcpi> exp: me'o by cy pare'u 19:46 < mensi> ([{me'o <(¹by cy¹) BOI> LOhO} {pa re'u} KU] VAU) 19:47 < zipcpi> I really don't like that part. I see the {boi} problem but I use {me'o} (and even {li} for my detri system) for arbitrary character strings 19:47 < zipcpi> I invented {li'ei} purely out of protest. It has 4 votes now though o.o 19:49 < zipcpi> I do have my own suggestion for the {boi} problem in Eulojban... but it's kinda silly 19:49 < zipcpi> Basically even breaks {bypyfyky cusku} in favor of a new cmavo for multi-letteral pronouns 19:50 < zipcpi> Alternatively we could make {tei} auto-close 19:52 < zipcpi> It does have the advantage that letteral pronouns then become basically like KOhA, and won't need {boi} at all 19:52 < zipcpi> Except when you use {leu} 19:52 < zipcpi> To make a multi-letteral pronoun 19:52 < zipcpi> That can't be helped though 19:52 < zipcpi> Cost of wanting multi-letteral pronouns 19:56 < zipcpi> I did enter {leu} as {ko'ai}, but downvoted myself cause it does almost nothing for current grammar 19:58 < noncomcinse> de'a jundi 19:59 < zipcpi> je'e 20:00 < zipcpi> mensi: doi durka42 I don't know why {li'ei} has four votes. It's a protest cmavo o.o 20:00 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.durka42.gy. di'a cusku da 20:04 < zipcpi> mensi: doi durka42 If you are one of the upvoters though I do thank you for joining my protest :p 20:04 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.durka42.gy. di'a cusku da 20:18 < noncomcinse> di'a 20:18 < noncomcinse> en: li'ei 20:18 < mensi> li'ei = [LIhEI] arbitrary character string or irregular number |>>> Accepts any arbitrary string of BY, bu-letterals, PA. 20:18 < mensi> Elidable terminator: lo'o. See li'ai, me'o, mo'e, li. [Tentative proposal: li'ei = li pe'o joi'i / me'o pe'o joi'i] 20:18 < mensi> |>>> spheniscine 20:19 < noncomcinse> co'o mi'e sipna 20:32 < GothAlice> coi 20:32 < mensi> GothAlice: cu'u la'o gy.selpahi.gy.: <GothAlice> Is there a directory or anything of people willing to be paid as 20:32 < mensi> consultants on Lojban? -- No, but put me on that list. I'll do anything Lojban-related for money :) | 20:32 < mensi> 2015-05-03T11:43:51.358Z 20:32 < mensi> GothAlice: cu'u la'o gy.selpahi.gy.: <GothAlice> I want full Lojban translations of the source books we'll be releasing. 20:32 < mensi> -- I just looked at the one you posted. It looks very interesting. How many are there going to be? | 2015-05-03T12:01:31. 20:32 < mensi> 342Z 20:33 < GothAlice> Say what now? 20:33 < GothAlice> A ha. 20:34 < GothAlice> mensi: Interesting auto-responder you have there! There will be, initially, two. The main Apex book, and the spapre one. 20:34 < mensi> xu lo nu casnu la'e di'u cu raktu do 20:36 < zipcpi> Hey crazy idea: What if we had names for all the elements built in a similar style to ISO codes 20:36 < GothAlice> We do. 20:36 < zipcpi> What... {ratnixe} = {cidro}? 20:37 < GothAlice> Ref: Ununseptium as an example of the standard Roman "normal form" of the heavier elements. 20:37 < zipcpi> ratnipapaso 20:37 < GothAlice> I.e. they're the numbers, translated, I would suspect. 20:38 < zipcpi> Well my idea is that it's like the ISO fu'ivla for languages and countries 20:38 < GothAlice> it's unambiguous, since the number of protons is "what matters" in the primary classification. (Secondary classification, "ions", are mathematical variances in the number of electrons, and another secondary classification is in the imbalance of neutrons.) 20:38 < zipcpi> Right 20:39 < GothAlice> :) Math makes things a touch easier, in most cases. 20:39 < zipcpi> So we have two sets, one that uses atomic numbers, one that uses the common abbreviations 20:39 < GothAlice> (So an ISO-ish form would be <protons>-<electron differential>-<neutron differential>) 20:40 < zipcpi> I think electron and neutron differential might be terbri 20:42 < GothAlice> With my understanding of {terbri}, that fits when describing what each aspect of that ISO-ish form itself is, abstractly. I.e. "the definition of this part is defining a predicate relationship with relation {charge} among arguments [protons, electron diff, neutron diff], but I may be wrong. ^_^ 20:43 < zipcpi> E.g. {ratnime'u} (just randomly assigning -me- for empty symbol meant to make it a valid zi'evla) = {ratnisore} = x1 is a quantity of uranium with atomic weight x2 and ionic differential x3 20:43 < zipcpi> lo ratnisore be li recimu 20:44 < zipcpi> Oh wait {ratni} does already have x2 and x3 20:44 < zipcpi> It needs an x4 20:45 < GothAlice> Well, only sorta. 20:45 < GothAlice> It infers the neutron differential from the difference between atomic number and weight. 20:45 < GothAlice> (Which is lame and not explicit.) 20:45 < GothAlice> ;P 20:45 < zipcpi> No, x4 for ionic number 20:45 < zipcpi> Ionic charge 20:46 < GothAlice> http://vlasisku.lojban.org/vlasisku/ratni < has a note about {selci} mass preference 20:46 < zipcpi> Yeah 20:47 < zipcpi> Hmm... ratni is a single atom though 20:47 < GothAlice> Technically… the mass would also incorporate the electron weights. (Which are non-zero, but typically rounded out of most presentations.) You just need very accurate numbers to safely make the ionic inference from the existing arguments. ;) 20:48 < zipcpi> Exactly... well I just think it's just "atomic weight" as defined by convention, rather than literal mass 20:48 < zipcpi> Which treats electrons as negligible 20:49 < zipcpi> While x4 should contain ionic charge 20:50 < zipcpi> Chemical nomenclature is... kinda problematic though 20:50 < zipcpi> We probably can and should make zi'evla for common names like "benzene" 20:51 < zipcpi> While for actually describing chemical composition or structure; well, I have the idea of using li/me'o character strings to basically allow a sort of "free grammar" 20:51 < zipcpi> Then make a special brivla to convert them into the meaning we want 20:52 < zipcpi> For example, SMILES: https://en.wikipedia.org/wiki/Simplified_molecular-input_line-entry_system 20:54 < zipcpi> We could make *'y cmavo all BY (well except for xy'y... I have only myself to blame for that inconsistency, but I can't think of any other suitable cmavo form for that) 20:55 < zipcpi> Or possibly move it to xa'e or something 20:55 < zipcpi> xa~~~'e 21:11 < zipcpi> exp: teibu 21:11 < mensi> ([{tei bu} BOI] VAU) 21:38 < GothAlice> zipcpi: Apologies, my internet decided to get hit by lightning on me. ^_^; Things like the names "benzene" and whatnot are actually a mixture of greek and latin prefixes and suffixes with precise meanings. 21:38 < zipcpi> True 21:39 < zipcpi> But I think fu'ivla is best for them. Pretty much every other language (except maybe Chinese) use fu'ivla for them 21:39 < GothAlice> Worst-case example: Hexanitrohexaazaisowurtzitane 21:40 < zipcpi> Hahaha 21:40 < GothAlice> Reading aid: (these are right-to-left) 21:40 < zipcpi> I've heard of it 21:40 < GothAlice> Looks like a little house. It's explosive, but very stable. 21:40 < GothAlice> Way, way, way, way, way too many nitrogens to make any chemist smile at a correct pronunciation. ;P 21:41 < zipcpi> Um diddle diddle diddle um diddle BOOM! Um diddle diddle diddle um diddle BOOM! 21:41 < zipcpi> Hexanitrohexaazaisowurtzita~~ne 21:41 < GothAlice> Well, no. Very stable. Could hit it with a sledge hammer without causing ill effect. 21:41 < zipcpi> lol 21:41 < zipcpi> Synthethizing it though... 21:42 < zipcpi> Probably creates a lot of nasty by-products 21:42 < GothAlice> wurtz, for example, describes the shape (cage structure) 21:42 < GothAlice> Well, yeah. Getting to the stable point is not something to do unless one owns a good pair of running shoes. 21:42 < zipcpi> Hehe 21:43 < GothAlice> Well, also, specifically, the "stable" form of it is a mixture with TNT. Yes, TNT stabilizes this compound. 21:43 < zipcpi> Chinese name: 六硝基六氮杂异伍兹烷 21:43 < GothAlice> AKA CL-20. 21:44 < zipcpi> -wurtzitane is fu'ivla'ed. The rest is calqued 21:46 < GothAlice> However, before my internet was so curtly interrupted by nature, I had a translation / phrasing question. How would one convey "the avatar / embodyment of the equal balance of life (giving) and death (life-taking)"? 21:47 < zipcpi> Hm only one word for avatar right now, {kibyka'i}, which is probably not what you want 21:47 < zipcpi> We could probably make two brivla for it 21:47 < GothAlice> (Avatar translating pretty literally as {krati}.) 21:48 < zipcpi> Well, mucti + krati maybe 21:48 < zipcpi> x1 is the physical embodiment of abstract concept x2 21:48 < GothAlice> A ha. 21:49 < zipcpi> Then, for the "between life and death", we could have zilkai + jbini 21:49 * GothAlice wishes that first lujvo was incorporated into vlasisku. ;P 21:49 < zipcpi> Hah... I'm making them up right now 21:51 < GothAlice> The sense for "between life and death" I'm going for is difficult to accurately convey. :/ It's a choice between the two, as polar opposites. The avatar of these twin forces can give and take completely, but not much in-between. I.e. it's not "power over life and death", but control over the choice between the two. 21:53 < GothAlice> (If you're alive, they can make you dead in a rather nasty manner. If you're dead, they can return your soul from "literally wherever it is" to a body which will be recreated for you. Mythology-wise, this means they're a cure for vampirism and other soul-related ailments… just in a very unpleasant manner.) 21:53 < GothAlice> This is for the Apex: The Evolved / People of the Thinking Plant thing. ^_^ 21:53 < zipcpi> http://jbovlaste.lojban.org/dict/mutmaika'i 21:54 < GothAlice> Nice. Totally bookmarking that dict. 21:56 < zipcpi> Hmm... may not need that second brivla... it's actually "balance", not "jbini" / no'e 21:59 < zipcpi> le mutmaika'i be lo si'o lanxe fa lo ka jmive kei ce lo ka morsi 21:59 < zipcpi> The embodiment of the concept of balance between living-ness and dead-ness 22:00 * GothAlice parses slowly. ^_^; 22:05 < GothAlice> That's awesome. Now, the follow-up is to create a pseudo-lujvo or pseudo-compound form to use as a noun. The "simple Lojban" version of that, breaking grammar because names can do that, {set of} {livingness} {deadness} {/set} (type-of) {avatar/embodiment} 22:05 * GothAlice ponders. 22:06 < zipcpi> {lo} 22:06 < zipcpi> Turns any "verb" into a noun :p 22:06 < GothAlice> Well, this is an interestingly complex statement, and it's good to have a shorthand for names to improve conversational speed a bit. ;) (Naming full titles in formal style is a first introduction thing, but after that…) 22:07 < zipcpi> Well, does this thing have an actual name? 22:07 < zipcpi> If not you can just {goi} it to something :p 22:08 < GothAlice> Avatars in this universe typically are normal people who then gain possession of an object that "improves" them, so they do have names. The role of being the avatar tends to dominate after picking up the object, though. 22:12 < GothAlice> Lomivmrokumutmaika’is. 22:13 < GothAlice> Loji'emrokumutmaika’is. 22:13 < GothAlice> Shorter wins. ^_^ 22:13 < zipcpi> Haha 22:13 < GothAlice> {.lomivmrokumutmaika’is.} (formatted correctly, and still quite a mouth-ful) 22:14 < zipcpi> Just call it la mivmroka'i 22:14 < GothAlice> Yeeah. 22:15 < GothAlice> Though it was my understanding such mashups insert (kind-of) between the gismu, thus living-kind-of-dead-kind-of-avatar? 22:15 < zipcpi> brivycme ftw 22:15 < zipcpi> http://mw.lojban.org/papri/brivycme 22:15 < zipcpi> Well, lujvo tends not to follow the tanru-grouping doctrine strictly... no one wants to explicity say the kems and keps x.x 22:16 < GothAlice> Aaah, so like ordinary name forms, marked this way the "name" is almost abstract. 22:16 < zipcpi> Yep 22:16 < GothAlice> Nice. 22:16 < GothAlice> That's going to… shorten a fair number of names in the book. XD 22:19 < GothAlice> First, extreme thanks for your assistance with that. I have to run, though, I'm eating up cellular data while I've been waiting for the techs to fix whatever knocked my cable net out, and it's now after 0100. ^_^; 22:19 < zipcpi> co'o 22:23 < rlpowell> If aynone wants to poke at CLL visual stuff: 22:23 < rlpowell> WRT https://github.com/lojban/cll/issues/149 22:24 < rlpowell> My current thoughts are at http://vrici.lojban.org/~rlpowell/media/public/cll_build/cll_prince.pdf 22:25 < rlpowell> As opposed to http://vrici.lojban.org/~rlpowell/media/public/cll_build-2015-06-20/cll_prince.pdf 22:30 < zipcpi> exp: lo broda cu brode jepoi brodi 22:30 < mensi> SyntaxError: Expected [,] but "b" found. 22:30 < zipcpi> exp: lo broda cu brode po'oi brodi 22:30 < mensi> ([lo broda KU] [cu {brode <po'oi (¹CU [brodi VAU]¹) KUhOI>} VAU]) 22:30 < zipcpi> Huh knows NOhOI 22:32 < zipcpi> rlpowell: Interesting; I don't know what to suggest though 22:33 < rlpowell> -_- 22:33 < rlpowell> Well, it's wrong/broken suddenly, so don't suggest anything. :P 22:34 < zipcpi> Oh hey you also wrote that bridi cmaci thing? I have crazy ideas regarding it too 22:35 < zipcpi> Basically, bridi is one excellent replacement for the mekso cmavo space... my idea is to also allow use of character strings or even lo'u...le'u as a sort of "free grammar" defined outside normal Lojban grammar, then having special brivla to convert them 22:35 < zipcpi> Useful for converting such diverse stuff as dates, SMILES notation, Sideswap juggling notation 22:37 < zipcpi> Something like "x1 is the chemical described by jbo-SMILES notation x2" 22:38 < zipcpi> So freeing. You can redefine cmavo and brivla wholesale if necessary 22:39 < zipcpi> Thus no need to dedicate mekso cmavo... they could mean something more suitable for general usage, and also gain a mekso meaning when placed in {lo'u...le'u} and converted with the appropriate brivla 22:51 < rlpowell> There! 22:51 < rlpowell> zipcpi: You can look again if you want. 22:52 < zipcpi> {coi lojban.} should be {coi .lojban.}, even pre-dotside :p 22:52 < rlpowell> Nope, not pre-dotside. 22:53 < rlpowell> Hmm. Maybe? 22:53 < rlpowell> Bah. Don't remember. 22:53 < zipcpi> Just saying that that's a mistake, even by pre-dotside standards 22:53 < zipcpi> pre-dotside only allowed elision of that pause for {la/lai/doi} 22:53 < rlpowell> *nod* 22:53 < rlpowell> http://mw.lojban.org/papri/CLL,_aka_Reference_Grammar,_Errata please 22:56 < zipcpi> Hm one question 22:58 < zipcpi> Even though CLL probably shouldn't encode the "brivla-cmevla merge" proposal yet (if ever), I think it should have {cu} in {la .djan. cu klama}. This is because BCM or not, this elision of {cu} is confusing to beginners 22:58 < zipcpi> When they learn brivla can also be used in names, then they are liable to go {la zipcpi klama} 23:00 < zipcpi> Hm maybe I'll add a section to that page 23:02 < rlpowell> gleki: Please review the spacing in http://vrici.lojban.org/~rlpowell/media/public/cll_build/cll_prince.pdf 23:04 < gleki> any problems with these https related letters? 23:04 < mensi> gleki: cu'u la'o gy.zipcpi.gy.: Maybe allow {lilo'o} in altatufa? {me'olo'o} might be useful for "empty string" | 23:04 < mensi> 2015-07-20T02:39:54.499Z 23:05 < gleki> rlpowell: in some place looks fine but e.g. why some paragraphs dont start with an indented line e.g. on page 8? 23:05 < rlpowell> gleki: Everything after an exmalpe or table or whatever doen't indent. 23:05 < rlpowell> Because it's already clearly seperated frm the surrounding text. 23:05 < rlpowell> Also, may parapgrahps after examples continue the previous sentence. 23:05 < rlpowell> fg 23:06 < rlpowell> Whoops. 23:06 < gleki> hm 23:06 < gleki> looks somewhat strange. 23:06 < zipcpi> Huh does it actually have any examples of la CMEVLA BRIVLA? 23:06 < zipcpi> Can't seem to find any :p 23:06 < zipcpi> Though there are some dot-marking errors 23:06 < rlpowell> gleki: Example: "means that to reach the dog-biting" <-- search on that. 23:07 < rlpowell> If that's indented, it looks weird. 23:07 < rlpowell> gleki: I can put the indents back so you can see; I think it's better without. 23:07 < zipcpi> Ah found one. "la frank. sanli zu'a nu'i la djordj." 23:07 < gleki> rlpowell: maybe ban indenting only for paragraphs that start with a lowercase letter? because there should be some uniformity. some sentences after examples do start a new paragraph but they dont indent. 23:08 < gleki> rlpowell: or maybe as i suggested just indent all first lines of paragraphs but to a small degree. retain indenting of examples as they are now 23:09 < gleki> zipcpi: report all errors and mistypes to https://github.com/lojban/cll/issues please 23:10 < zipcpi> I edited this page: http://mw.lojban.org/papri/CLL,_aka_Reference_Grammar,_Errata#Chapter_1 23:12 < rlpowell> 19-23:07 < gleki> rlpowell: maybe ban indenting only for paragraphs that start with a lowercase letter? -- That's ... probably not impossible, but really really hard. 23:12 < rlpowell> 19-23:09 < gleki> zipcpi: report all errors and mistypes to https://github.com/lojban/cll/issues please -- No, errors of the text go to the errata page, please. 23:13 < rlpowell> gleki: reload please. 23:15 < gleki> rlpowell:now let others share their opinion. i'd rather decreased indenting even more or maybe removed it completely. idk. how it was done originally in paper in the first edition? 23:16 < rlpowell> Way more indented. 23:17 < gleki> huh 23:17 < gleki> GothAlice: please use "mensi: doi selpahi this is a message to you" syntax to send delayed messages 23:23 < rlpowell> gleki: http://snag.gy/7Vpgz.jpg and http://snag.gy/VXLGI.jpg 23:23 < rlpowell> Note that the indent is something like 4 characters wide. 23:32 < rlpowell> gleki: Look at the "means that" para in the new version. 23:32 < rlpowell> I marked it specially so it doesn't indent, which is also the case in the original book (see above pics). 23:36 < zipcpi> mensi: doi GothAlice I also like the semantic-poetic assonance between {mutmaika'i} and {la mivmroka'i}, if all your avatars represent the balance between two opposing concepts. 23:36 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.GothAlice.gy. di'a cusku da 23:39 < gleki> rlpowell: i see. so let's drop this subject until someone else complains. for now let's retain everything as it is now 23:40 < Roflknife34> mi pante . y How do I say next? 23:40 < zipcpi> za'umoi? 23:41 < Roflknife34> ki'e 23:41 < rlpowell> gleki: *nod* 23:41 < rlpowell> I'll post about it. 23:44 < gleki> depends on the meaning of "next" 23:44 < gleki> it can be {za'umai} as well 23:44 < zipcpi> Yeah kinda 23:45 < zipcpi> Or {bavlamji} 23:50 < Roflknife34> I meant it as in. "I want to be the next person to complain (as a joke)" 23:55 < gleki> mi djica lo ka za'umoi lo prenu poi pante 23:57 < gleki> ta'osai English sentence may sound shorter due to a lrage number of consonant clusters 23:57 < gleki> *large 23:58 < gleki> e.g. in last example ican find 7 consonant boundaries in English and only 2 in Lojban translation 23:59 < gleki> if we count each boundary as one syllable then Lojban translation will be one syllable shorter --- Day changed Mon Jul 20 2015 00:03 < zipcpi> Yeah there is basically a trade-off known in the conlang community. 00:03 < zipcpi> Either you include phonemic complexity like tones, consonant clusters, or large phonemic inventory, 00:03 < zipcpi> Or you take more syllables to say something 00:03 < zipcpi> It's like a lump in the carpet :p 00:09 < gleki> I explained this problem http://mw.lojban.org/papri/Are_texts_in_Lojban_lengthy%3F 00:11 * nuzba @uitki: Are texts in Lojban lengthy? http://mw.lojban.org/papri/Are_texts_in_Lojban_lengthy%3F [http://bit.ly/1HKmw3b] 00:11 < zipcpi> coi durka42 00:13 < durka42_> coi coi 00:13 < mensi> durka42_: cu'u la'o gy.zipcpi.gy.: I don't know why {li'ei} has four votes. It's a protest cmavo o.o | 00:13 < mensi> 2015-07-20T03:00:28.373Z 00:13 < mensi> durka42_: cu'u la'o gy.zipcpi.gy.: If you are one of the upvoters though I do thank you for joining my protest :p | 00:13 < mensi> 2015-07-20T03:04:02.264Z 00:13 < durka42_> ue 00:13 < gleki> en: upvote 00:13 < mensi> 5 da se tolcri: ma'udzau, adzau, co'udzau, dzaunlai, nildzau 00:14 < zipcpi> ue zo li'ei cu dzaunlai li vo 00:14 < gleki> where is this -dzau come from? 00:14 < zipcpi> zanru 00:14 < zipcpi> I made a pseudo-rafsi 00:19 < zipcpi> doi durka .i xu do viska la'au loi slari vanjba li'u pe la .e'ulojban. 00:19 < durka42> no I didn't see it 00:20 < durka42> ta'o can you explain this {ke po'o lo broda cu brode} thing? 00:20 < durka42> I think I saw Amnadya/selpa'i use it yesterday and I was quite confused 00:20 < zipcpi> zantufa: ke po'o lo broda cu brode 00:20 < zantufa> ([{ke po'o} {lo broda KU} KEhE] cu [brode VAU]) 00:20 < durka42> ah I see 00:21 < zipcpi> This is an extension of {ke}, first developed to do stuff like this 00:21 < durka42> so how can we fix camxes-exp to parse this :) 00:21 < zipcpi> zantufa ke ko'a ce ko'i ke'e ce ko'o broda 00:21 < durka42> because zantufa has some other problematic features that disqualify it for general use 00:21 < zipcpi> zantufa: ke ko'a ce ko'i ke'e ce ko'o broda 00:21 < zantufa> ([{ke <ko'a (¹ce ko'i¹)> ke'e} {ce ko'o}] CU [broda VAU]) 00:21 < zipcpi> Yeah I know T.T 00:22 < zipcpi> If I ever learn PEG might might adopt a pet parser myself. Can't decide if I should call it {maltufa} or {eultufa} :p (Don't worry, I'll probably keep two versions; one that's suitable for general use, and one that includes all the Eulojban madness) 00:22 < gleki> en: dudvota 00:22 < mensi> dudvota = x1 gives x2 (number) votes (upvotes) to x3 |>>> Positive number in x2 denotes an upvote, negative number a 00:22 < mensi> downvote. Use ca'onai dudvota for cancelling vote. |>>> gleki 00:23 < zipcpi> Did you just make that 00:23 < gleki> yes 00:23 < durka42> mu'i ma 00:23 < gleki> durka42: https://github.com/Ilmen-vodhr/ilmentufa/issues/118 so it's not zantufa's issue. it's our issue 00:24 < zipcpi> Erm, there needs to be a terbri for system too... well there is {ci'e} but I'm not sure how to formalize that 00:24 < gleki> durka42: so that one doesnt have to remember a bunch of brivla. one brivla would be enough. although for "vote score" i need to reread mekso 00:25 < gleki> zipcpi: you didnt provide any examples of such system. i think it's one of things that JVS should force 00:25 < zipcpi> la jbovlaste 00:25 < zipcpi> la flicku 00:25 < zipcpi> Whatever 00:25 < zipcpi> It's kinda an important concept 00:25 < zipcpi> Cause something must be keeping the vote score 00:26 < gleki> it's just a sum of all {lo se dudvota} 00:26 < zipcpi> I know I'm just saying that there is no "upvote/downvote" without a system to do it in 00:27 < gleki> i dont know what is that system unless you provide examples 00:27 < zipcpi> I told you 00:27 < zipcpi> la jbovlaste / la flicku 00:27 < durka42> gleki: it seems like guskant shows how to fix bug 118 :) 00:27 < gleki> alta: ke ui mi e do cu mo 00:27 < mensi> (NIhO [I {<FA ZOhE> <ke ui> <FA (¹mi [e do]¹)> KEhE} {cu <mo VAU>}]) 00:27 < gleki> i think i already stole it into altatufa for sumti 00:28 < durka42> ueu 00:28 < zipcpi> ueu? :p 00:28 < durka42> zantufa: ke ui me e do cu mo 00:28 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 00:28 < durka42> zantufa: ke ui mi e do cu mo 00:28 < zantufa> ([{ke ui} {mi <e do>} KEhE] cu [mo VAU]) 00:28 < zipcpi> ... you use an Eulojban cmavo that I didn't even define :p 00:28 < gleki> and im in the process of rewriting bridi part of that issue 00:33 < gleki> VUhU4 isn't in the list of links to CLL in vlasisku 00:34 < gleki> let's fix it at least for sutysisku 00:34 < zipcpi> http://mw.lojban.org/papri/Sour_Grapes You'll love it durka xo'o 00:34 < durka42> I saw it :) 00:34 < zipcpi> xu zbusufukai 00:35 < gleki> http://mw.lojban.org/papri/BPFK_Section:_MEX_Operators#cmavo:_si.27i_.28VUhU4.29 00:35 < gleki> why {ky ky}? 00:36 < zipcpi> la'a se ci'asre 00:38 < gleki> exp: li si'i ky vei no bi'o ny 00:38 < mensi> SyntaxError: Expected [,] or [rR] but "n" found. 00:38 < gleki> off: li si'i ky vei no bi'o ny 00:38 < mensi> ([li {si'i <(¹ky BOI¹) (¹vei [{no BOI} {bi'o <ny BOI>}] VEhO¹)> KUhE} LOhO] VAU) 00:38 < zipcpi> Eulojban may be mostly satirical, but I am serious about unbanning -mz- :p 00:39 < zipcpi> FREE -MZ- 00:39 < gleki> that will break some things 00:39 < zipcpi> Like what 00:39 < gleki> off: li si'i ky vei no bi'o ro 00:39 < mensi> ([li {si'i <(¹ky BOI¹) (¹vei [{no BOI} {bi'o <ro BOI>}] VEhO¹)> KUhE} LOhO] VAU) 00:39 < gleki> off: li si'i ky du ma 00:39 < mensi> ([{li <si'i (¹ky BOI¹) KUhE> LOhO} CU] [du {ma VAU}]) 00:40 < zipcpi> I mean the "what is not mandatory is forbidden" doctrine for lujvo is dead 00:40 < zipcpi> k: nonpa'a 00:40 < mensi> (CU [L:nonpa'a VAU]) 00:40 < zipcpi> k: nonypa'a 00:40 < mensi> (CU [L:nonypa'a VAU]) 00:40 < gleki> off: li si'i na'u se dudvota 00:40 < mensi> SyntaxError: Expected [,] but "d" found. 00:41 < zipcpi> So {memyze'a} won't break... all I'm suggesting is to accept {memze'a} 00:41 < gleki> off: li si'i mo'e lo se dudvota 00:41 < mensi> ([li {si'i <mo'e (¹lo [se dudvota] KU¹) TEhU> KUhE} LOhO] VAU) 00:41 < gleki> uanai 00:42 < gleki> well, it's {sumji ro se dudvota} of course. as for mekso idk 00:42 < gleki> zantufa: li si'i na'u se mo 00:42 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 00:42 < gleki> zantufa: li si'i mo'e lo se mo 00:42 < zantufa> ([li {si'i <mo'e (¹lo [se mo] KU¹) TEhU> KUhE} LOhO] VAU) 00:43 < zipcpi> Fair warning though, I *will* enter the direction gismu once they are unbanned lol 00:45 < zipcpi> Not sure about rafsi though... 00:47 < rlpowell> gleki: Posted a link to the latest spacing test version to the main list. 00:50 < gleki> let's see what they reply then 00:59 < zipcpi> durka42: Really though, {vanjbacau}? :p 01:00 < durka42> ma nabmi 01:00 < zipcpi> I just find it funny 01:00 < durka42> it's supposed to be sort of a pun 01:00 < zipcpi> Right 01:00 < durka42> vanjbacau ke vanjbacpa nuntoi 01:01 < durka42> using {bo} instead of {ke} would make it more obvious 01:01 * zipcpi shrugs 01:01 < zipcpi> I do tend to avoid nonce-lujvo myself, but, that's the whole point of the exercise 01:02 < durka42> well -cau is an RDA :p 01:02 < durka42> not sure about -cpa sei zbusufukai 01:04 < zipcpi> :p 01:05 < durka42> zipcpi: what's that magic forethought prefix notation mekso cmavo? 01:05 < zipcpi> pe'o 01:05 < durka42> ah yes 01:05 < zipcpi> li'a pendo mi 01:06 < durka42> lo'e mekso cu nonseljmi mi 01:07 < gleki> RDA? 01:07 < zipcpi> "regular derivational affix" 01:07 < durka42> javni smuzba rafsi 01:09 < durka42> vlaste: datru (fapro) 01:09 < vlaste> zo datru gimkamsmikezypro zo tatru noi catni gismu .i zo datru gimkamsmikezypro no cipra gismu 01:09 < durka42> vlaste: datro (fapro) 01:09 < vlaste> zo datro gimkamsmikezypro no catni gismu .i zo datro gimkamsmikezypro zo datru noi cipra gismu 01:09 < durka42> .ei zo datru cu binxo zo datro 01:10 < zipcpi> Meh... {datru} is just {dreika}-lite 01:10 < gleki> e'u genai zo datru ginaizo datro cu se pilno 01:10 < durka42> .u'i 01:10 < durka42> neither of those are relevant 01:10 < durka42> en:dreika 01:10 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 01:11 < durka42> jbo:dreika 01:11 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 01:11 < durka42> lol 01:11 < durka42> so both of you said we should just use some other word that doesn't exist 01:12 < durka42> the problem is {datru} has significant usage 01:12 < zipcpi> {dreika} - x1 is a predicate created to solve problem x2 (default date system) in language x3 (default Lojban) with too many places x4, involving day x5, month x6, year x7, hour x8, minute x9, second x10, of calendar x11, objected by Lojbanist x12, and obsoleted by date system x13 01:12 < gleki> so dont shut it down 01:12 < durka42> but it clashes with an earlier gismu 01:13 < gleki> but it has significant usage :P 01:13 < zipcpi> That completely breaks cmavo set system x14 that is only designed to handle x15 (default 5) places 01:13 < durka42> I think we should just move the definition to {datro} and at {datru} write "dispreferred due to collision -- see {datro}" 01:13 < gleki> similarly you can redirect to {dreiko} or whatever 01:13 < durka42> rules of lojban >> significant usage >> nonexistent words 01:17 < durka42> mi muvgau fi zo datro 01:23 < durka42> JVS is so sloooow 01:24 < zipcpi> Hm {datru} is probably not downvotable though 01:25 < durka42> hum ight we don't want it out of the dictionary 01:25 < durka42> lol 01:25 < zipcpi> vlaste: dadru (fapro) 01:25 < vlaste> zo dadru gimkamsmikezypro no catni gismu .i zo dadru gimkamsmikezypro zo datru noi cipra gismu 01:25 < durka42> just upvote {datro} more 01:25 < durka42> I thought of that but the root of {datru}/{datro} is that it has the same consonants as {detri} 01:25 < durka42> so {datro} is more appropriate 01:25 < zipcpi> No I'm just testing for ka gimpro :p 01:26 < durka42> la vlaste doesn't update that fast :) 01:26 < durka42> none of my recent manual updates have exploded though 01:26 < durka42> perhaps it's time to turn on automatic updates 01:26 < zipcpi> vlaste: famza (fapro) 01:26 < vlaste> zo famza gimkamsmikezypro zo fanza noi catni gismu .i zo famza gimkamsmikezypro no cipra gismu 01:26 < durka42> it just takes 100% CPU for like 5 minutes to rebuild the index :/ 01:26 < zipcpi> vlaste: bemza (fapro) 01:26 < vlaste> zo bemza gimkamsmikezypro no catni gismu .i zo bemza gimkamsmikezypro no cipra gismu 01:27 < durka42> those are morphologically bad, right? 01:27 < durka42> k: famza 01:27 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "f" found. 01:27 < zipcpi> Erm, yeah. 01:27 < zipcpi> -mz- 01:27 < zipcpi> I unbanned it from Eulojban though, so I made new gismu for compass directions 01:27 < gleki> dont use indices zo'o 01:28 < gleki> i da'i zo famza janli zo fanza 01:28 < durka42> better choose a new final consonant 01:29 < durka42> si vowel 01:29 < zipcpi> ii do pilno zo zo va'o tu'a lo nalseljva valsi .i lo xagji sofybakni ba darca 01:29 < durka42> what's the etymology? 01:29 < durka42> they don't seem to correspond to FAhA1 01:29 < zipcpi> FAhA2 01:30 < zipcpi> North, South, East, West 01:30 < durka42> that's FAhA1 01:30 < zipcpi> I intend for them to "replace" the official gismu 01:30 < zipcpi> Oh 01:30 < zipcpi> Really? 01:30 < durka42> vlaste: class:FAhA1 01:30 < vlaste> 4 results: be'a, du'a, ne'u, vu'a 01:30 < zipcpi> bemza, dumza, nemzu, vumza 01:31 < zipcpi> What's the problem, other than -mz- :p 01:31 < durka42> ah ok 01:31 < durka42> what's famza then? 01:31 < durka42> fa'a? 01:31 < zipcpi> Oh just a test 01:31 < zipcpi> Cause I know it "clashes" with fanza 01:31 < durka42> ah ok 01:33 < zipcpi> Well I didn't neglect ortho-Lojban. I also added {be'arna} etc. 01:38 < zipcpi> k: rirxrxrazdani 01:38 < mensi> (CU [Z:rirxrxrazdani VAU]) 01:38 < zipcpi> k: rirxyrxrazdani 01:38 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "r" found. 01:38 < zipcpi> Oh 01:39 < zipcpi> Differs between syllabic /r/ and /y/, not /r/ and /yr/ 01:39 < zipcpi> Thought Pierre might have found a new xazdmru 01:42 < rlpowell> I just made a change that *may* improve overall lojban.org performance. 01:42 < rlpowell> gleki: ^^ 01:42 < rlpowell> durka42: Interested in your thoughts on my spacing test (see main mailing list post). 01:42 < rlpowell> afk chores. 01:55 < gleki> alta: ke broda gi'e brode ke'e gi'a brodi 01:55 < mensi> (NIhO [I {FA ZOhE} {ke <(¹[broda NF] VAU¹) (¹gi'e CU [{brode NF} VAU] VAU¹)> ke'e VAU} {gi'a CU <(¹brodi NF¹) VAU> VAU}]) 01:56 < gleki> looks like i finally incorporated guskant's fixes into alta 01:59 < durka42> rlpowell: je'e 01:59 < gleki> zipcpi: i can implement {li lo'o mo} but not {li mo} 02:01 < zipcpi> Oh right cause of cu li 02:04 < vlaste> sei ca ca'o ningau lo pe mi sorcu 02:12 < durka42> vlaste: datru 02:12 < vlaste> datru = x1 (event) is dated/pertaining to day/occurring on day x2 of month x3 of year x4 in calendar x5 02:12 < durka42> vlaste: datro 02:12 < vlaste> datro = x1 (event) is dated/pertaining to day/occurring on day x2 of month x3 of year x4 in calendar x5 02:13 < durka42> vlaste: datru FULLTEXTSEARCH 02:13 < vlaste> datru = x1 (event) is dated/pertaining to day/occurring on day x2 of month x3 of year x4 in calendar x5 02:13 < durka42> hmm that didn't work 02:13 < durka42> vlaste: DATRU 02:13 < vlaste> datro = x1 (event) is dated/pertaining to day/occurring on day x2 of month x3 of year x4 in calendar x5 02:13 < durka42> heh 02:13 < durka42> vlaste: DATRO 02:13 < vlaste> datru = x1 (event) is dated/pertaining to day/occurring on day x2 of month x3 of year x4 in calendar x5 02:14 < zipcpi> vlaste: *azd* 02:14 < vlaste> xazdo = x1 reflects Asiatic culture/nationality/geography in aspect x2. 02:14 < zipcpi> vlaste: *myz* 02:14 < vlaste> no results. http://vlasisku.alexburka.com/%2Amyz%2A 02:14 < zipcpi> Wait wut 02:14 < durka42> the wildcard search is really finisky 02:14 < durka42> finicky* 02:15 < durka42> ['jemyzda', "kamyzi'e", 'kumyzda', "memyze'a", 'samyzamsmacu', 'samyzetro', 'samyzilkei', 'samyzilkeikei', 'samyzva', 'sumyzmico'] 02:17 < zipcpi> samyzilkeikei lol 02:18 < durka42> vlaste: samyzilkeikei 02:18 < vlaste> samyzilkeikei = x1 plays with video game x2. 02:18 < durka42> .u'i wtf 02:18 < durka42> vlaste: samyzilkeikei (finti) 02:18 < vlaste> samyzilkeikei (finti) = Wuzzy 02:18 < durka42> of course 02:18 < durka42> that was my guess 02:18 < durka42> lo jvapre 02:18 < zipcpi> Bloated lujvo 02:18 < durka42> iesai 02:25 < durka42> py! [entry for entry in database.root.entries.keys() if 'myz' in entry] 02:25 < durka42> vlaste: py! [entry for entry in database.root.entries.keys() if 'myz' in entry] 02:25 < vlaste> No REPL for you! 02:26 < zipcpi> What's that... an injection attempt? 02:26 < durka42> yes 02:26 < durka42> it's supposed to allow me to do it :p 02:27 < durka42> it works in private chat 02:27 < zipcpi> DROP TABLE jbovla 02:28 < zipcpi> vlaste: bu'a'a'); DROP TABLE words;-- 02:28 < zipcpi> lol 02:28 < durka42> phew closed it just in time :p 02:28 < zipcpi> lol 02:28 < durka42> vlaste: coi 02:28 < vlaste> coi = vocative: greetings/hello. 02:29 < durka42> ah 02:29 < durka42> I'm dumb 02:29 < durka42> vlaste: coi 02:29 < vlaste> coi = vocative: greetings/hello. 02:29 < durka42> vlaste: py! [entry for entry in database.root.entries.keys() if 'myz' in entry] 02:29 < vlaste> ['jemyzda', "kamyzi'e", 'kumyzda', "memyze'a", 'samyzamsmacu', 'samyzetro', 'samyzilkei', 'samyzilkeikei', 'samyzva', 'sumyzmico'] 02:30 < durka43> vlaste: py! [entry for entry in database.root.entries.keys() if 'myz' in entry] 02:30 < vlaste> No REPL for you! 02:30 < durka42> ui 02:30 < zipcpi> How does it detect? 02:30 < durka42> username whitelist 02:30 < zipcpi> What if some zbusufukai prenu /nick durka42 02:31 * zipcpi whistles innocently 02:31 < durka42> durka42 is registered 02:31 < zipcpi> lol 02:31 < durka42> also it's a VPN in japan, do whatever you want .u'i 02:31 < durka42> VPS* 02:38 < gleki> zantufa: abuza'ure'u mo 02:38 < zantufa> ([{a bu} BOI] CU [{<za'u BOI> re'u} mo] VAU) 02:38 < gleki> zantufa: li abu mo 02:38 < zantufa> ([li {<a bu> BOI} LOhO] CU [mo VAU]) 02:38 < gleki> zantufa: lipa abu mo 02:38 < zantufa> ([{li <pa BOI> LOhO} {a bu} BOI] CU [mo VAU]) 02:38 < gleki> zantufa: li abu pa mo 02:38 < zantufa> ([{li <(¹a bu¹) BOI> LOhO} {pa BOI} mo KU] VAU) 02:38 < gleki> zantufa: li abu pa mo cu mo 02:38 < zantufa> ([{li <(¹a bu¹) BOI> LOhO} {pa BOI} mo KU] cu [mo VAU]) 02:39 < gleki> alta: li lo'o mo 02:39 < mensi> (NIhO [I {FA <li lo'o>} {mo VAU}]) 02:40 < gleki> jbo: po'e 02:40 < mensi> po'e = [GOI] poi ke'a se jinzi le ka ce'u se steci le ka ce'u srana |>>> poi; jinzi; steci; srana |>>> 02:40 < mensi> xorxes 02:40 < gleki> can't {po'e} be used instead of {po'o}? 02:42 < durka42> in what context? 02:44 < gleki> in any 02:44 < gleki> {po'e} is about steci 02:44 < gleki> {po'o} too 02:52 < durka42> I mean they are different selma'o 02:53 < durka42> mi prami do po'o -- how are you going to rephrase this with {po'e}? 02:53 < gleki> yes, different grammar since some hate UI grammar for po'o 02:54 < durka42> what should it be instead 02:54 < durka42> and you didn't answer my question :) 02:59 < gleki> im just thinking how one is supposed to convert from {do se steci lo kase prami mi} to {mi prami do po'o} in general 02:59 < durka42> mm 03:00 < durka42> mi prami do po'o => broda lo ka mi prami ce'u => do se steci lo ka mi prami ce'u 03:00 < durka42> when it marks non-sumti not sure, need to consider more 03:01 < gleki> i dont understand converting at the first step 03:01 < gleki> you lost {do} 03:01 < durka42> I replaced the marked sumti {do po'o} with {ce'u} 03:02 < durka42> {broda} was {se steci be fa do} 03:02 < gleki> btw robin sent me a copy of jvs.sql 03:03 < durka42> a current database dump? 03:03 < durka42> co'o mi'e ba xrukla 03:03 < Compu-Celebi> jbovlaste uses SQL? 03:06 < gleki> yes, a postgres dump of it with user info deleted 03:06 < Compu-Celebi> Define "postgres." 03:07 < gleki> just sql 03:08 < Compu-Celebi> I do not comprehend your response. 03:11 < xeizlif> Compu-Celebi: postgres is an SQL server; a postgres dump is a dump of its database, presumably either in its own format or as a big list of SQL statements that would create the same databases 03:15 < Compu-Celebi> Why is it called "postgres"? 03:17 < dutchie> "The name refers to the project's origins as a "post-Ingres" database, being a development from University Ingres DBMS (Ingres being an acronym for INteractive Graphics Retrieval System)" 03:17 < dutchie> https://en.wikipedia.org/wiki/PostgreSQL 03:23 < ctefaho> what is the new gismu that means "tolcanci"? 03:23 < Compu-Celebi> It seems that PostgreSQL is a system that one or more servers use, rather than being a server itself. 03:28 < dutchie> "server" is an overloaded term for a piece of software servicing requests and the computer that it runs on 03:48 < durka42> mi nau xrukla 03:49 < Compu-Celebi> go'i ma 03:49 < durka42> ti poi se .irci 03:50 < Compu-Celebi> Oops, I actually needed to use the third place. The second place is for the [i]destination[/i], not the origin. 03:50 < Compu-Celebi> I would have said {go'i fi ma}. 03:50 < durka42> I got the meaning anyway 03:51 < durka42> I mainly announced my return because I thought la gleki might have left me a message, which would only show up if I said something 03:51 < durka42> but he didn't 03:51 < durka42> za'a 03:52 < Compu-Celebi> Yet, your answer was that you returned to here. 03:52 < durka42> oh I see 03:53 < durka42> you meant to ask where I returned from 03:53 < durka42> mi xrukla fi lo nu dormijysai 03:53 < Compu-Celebi> je'a go'i 03:55 < Compu-Celebi> A {nu dormijysai} is an event of being one or more meals, not a meal ingestion event. 03:56 < Compu-Celebi> s/meal/lunch/ 03:56 < fenki> Compu-Celebi meant to say: A {nu dormijysai} is an event of being one or more lunchs, not a meal ingestion event. 03:56 < Compu-Celebi> s/lunchs/lunches/ 03:56 < fenki> Compu-Celebi meant to say: A {nu dormijysai} is an event of being one or more lunches, not a meal ingestion event. 03:57 < durka42> lo puzi nu dormijysai cu nu lo vrici cidja cu dormijysai fi'o citka mi 03:57 < durka42> mi noroi xusra lo sedu'u ba'e mi dormijysai 03:57 < gleki> jbo: tolcanci 03:57 < Compu-Celebi> I intended to say, "A {nu dormijysai} is an event of being one or more lunches, not a lunch ingestion event." 03:57 < gleki> durka42: lo du'u 03:58 < mensi> [< to'e canci ≈ To'e canci*] 8 da se tolcri: solnuntolcanci, berti, cerni, donri, murse, nicte, snanu, 03:58 < mensi> stuna 03:59 < gleki> en:/full canci 03:59 < mensi> 14 da se tolcri: canci, selcanci, snocanci, solnuncanci, solnuntolcanci, tercanci, tolcanci, cliva, ganse, lebna, selkra, 03:59 < mensi> tolyli'a, vimcu, zgana 03:59 < gleki> not invented yet 04:00 < gleki> vlaste: tcusa (gimka) 04:00 < vlaste> zo tcusa gimkamsmikezypro no catni gismu .i zo tcusa gimkamsmikezypro no cipra gismu 04:02 < Compu-Celebi> I did not say "an event of [i]you[/i] being one or more lunches." 04:02 < gleki> vlaste: tcuse (gimka) 04:02 < vlaste> zo tcuse gimkamsmikezypro no catni gismu .i zo tcuse gimkamsmikezypro no cipra gismu 04:03 < gleki> i propose {tcuse} from "chu1xian4" 04:04 < durka42> for what? 04:04 < gleki> for tolcanci 04:04 < Compu-Celebi> Is "gimkamsmikezypro" intended to mean "antonym"? 04:07 < gleki> en: antonym 04:07 < mensi> 11 da se tolcri: selsmudukti, brabra, caibra, capcni, cmacai, cmacma, kancni, nurcni, seicni, smudukti, 04:07 < mensi> tcebra 04:11 < durka42> vlaste: gimkamsmikezypro 04:11 < vlaste> gimkamsmikezypro = x1; (word/quote; probably gismu) conflicts with x2 (word/text; probably gismu) according to rules x3 04:12 < durka42> it refers to the rules in CLL for gismu that would sound too similar 04:13 < gleki> en: gimkamzmikezypro 04:13 < mensi> [< gismu ka zmiku kei fapro ≈ Root word being automatic end abstraction oppose] lo nu mulno sisku zo'u: y no da se 04:13 < mensi> tolcri 04:13 < mensi> gimkamyzmikezypro[16905], gi'urkamyzmikezypro[18494], gimkamyzmikezyfapro[18964], gimkamyzmikykezypro[19035], 04:13 < mensi> gismykamyzmikezypro[19035], ... 04:13 < gleki> en: gimkamsmikezypro 04:13 < mensi> gimkamsmikezypro [< gismu ka simsa kei fapro ≈ Root word being similar end abstraction oppose] = x1 (word/quote; 04:13 < mensi> probably gismu) conflicts with x2 (word/text; probably gismu) according to rules x3 |>>> Neither x1 nor x2 need be " 04:13 < mensi> official and accepted". |>>> krtisfranks 04:15 < durka42> lo se cusku be la mensi ta'e fegli .u'i 04:15 < gleki> en:gimkamsmikezypro 04:15 < mensi> gimkamsmikezypro [< gismu ka simsa kei fapro ≈ Root word being similar end abstraction oppose] = x1 (word/quote; 04:15 < mensi> probably gismu) conflicts with x2 (word/text; probably gismu) according to rules 04:15 < mensi> x3 04:15 < gleki> zmadu xu 04:17 < durka42> ie 04:17 < gleki> Aistlaţervièllîmļ, for example, is the term for “a situation where one lets a normally unavailable opportunity pass by because it is not seen as being the optimal instance or form of that opportunity, despite the likelihood that such an optimal instance/form of the opportunity will likely never come (e.g., letting a bottle of expensive wine go past its prime because one can never decide when would be t 04:17 < gleki> he optimal time to drink it; or letting slip by an opportunity for true love because one hopes someone even ‘better’ may come along.)” 04:17 < gleki> ithkuil: Aistlaţervièllîmļ 04:17 < ithkuil> error 04:18 < gleki> http://www.businessinsider.com.au/the-worlds-smallest-language-has-only-100-words-and-you-can-say-almost-anything-2015-7 04:18 < gleki> probably from ithkuil-2004 04:21 < Compu-Celebi> Ah, it certainly seems to be an Ithkuil word. 05:46 < Amnadya> We still don't have a word for "scene" in the sense of "a part of a play, book, film etc in which events happen in the same place or period of time" 05:47 < Compu-Celebi> Does one exist for "act" in the theatrical sense? 05:47 < gleki> se fanbu 05:48 < Compu-Celebi> en: fanbu 05:48 < Amnadya> "act" might be draci pagbu or draci kantu 05:48 < Compu-Celebi> en:fanbu 05:48 < durka42> Amnadya: seems like we just need a word for "a play/book/film/etc" 05:48 < durka42> then scene can be brodypau 05:48 < Amnadya> I just don't know if pagbu is enough, because some books have parts, chapters, and scenes. 05:48 < gleki> fanbu described here http://mw.lojban.org/papri/ELG._Subjunctives,_imaginary_situations 05:49 < gleki> Amnadya: pagbu lo se cukta enai lo cukta 05:49 < Amnadya> na banzu 05:50 < gleki> pagbu lo tcini poi lo se cukta cu skicu pe'a vau sei d[3~nei zei ge'enmo 05:51 < Amnadya> lo si'o cukta na je lo si'o lisri cu vajni 05:52 < gleki> xu lo se cukta ka'e lisri 05:53 < Amnadya> la'a go'i .i ku'i lo skina ku ji'a co'e .i li'o .i ja'o na sarcu fa lo nu zo cukta pagbu lo se finti valsi 05:53 < gleki> en: /full larcu 05:55 < Compu-Celebi> en:ge'enmo 05:59 < durka42> Amnadya: so like a lispau be bu'u lo pamei 05:59 < Amnadya> mi ba'o jmina fi la nabmyrta'imacusku .i la'a ba finti lo zi'evla 06:06 < Compu-Celebi> Oh, I erroneously believed that {pe'a} preceded, not followed, to what it applied. 06:13 < gleki> la vrici ca masno 06:21 < gleki> i masno i ie masno 06:26 < gleki> en:/full larcu 06:27 < mensi> 15 da se tolcri: larcu, pleplolarcu, artefresko, cnadakfu, finti, lardai, larpra, me'andi, mi'irlarfi'a, pirlarfi'i, 06:27 < mensi> skalpeli, stati, tamtigypre, terxra, zbasu 06:27 < mensi`i> 15 da se tolcri: larcu, pleplolarcu, artefresko, cnadakfu, finti, lardai, larpra, me'andi, mi'irlarfi'a, pirlarfi'i, 06:27 < mensi`i> skalpeli, stati, tamtigypre, terxra, zbasu 06:30 < gleki> ua i JVS cu masno 06:32 < durka42> ma'a'a'a'asno 06:36 < gleki> en: larseku 06:36 < mensi> 2 da se tolcri: ensaio, mlauca 06:36 < gleki> je'e 06:36 < gleki> i pu plixau brivla 06:37 < gleki> i cati JVA cu tcetce zmadu la vrici lo ka masno 06:37 < Compu-Celebi> To what does "JVA" refer? 06:39 < gleki> JVS 06:39 < gleki> i mi pu srera ciska 06:41 < gleki> jbo: larseku 06:41 < Compu-Celebi> .i je'e 06:41 < mensi> larseku = x1 prosa ja pemci ja skina ja draci x2 fi x3 fo x4 .i va'i x1 larselsku x3 noi se pi'o vo'a tavla x4 fi x2 06:41 < mensi> zi'o ku'o x4 |>>> prosa pemci skina draci larselsku |>>> zozeizeizeizeifaho 06:43 < gleki> exp: se fa do cu mo 06:44 < mensi> ([{se fa} do] [cu {mo VAU}]) 06:44 < gleki> je'e 06:44 < gleki> i ma smuni vau u'i 06:54 < durka42> ba'anai la tsani pu finti lo cizra smui 06:54 < durka42> si smuni 07:04 < niftg> la spagetufa cu slilu klama ti ta soi vo'e 07:06 < gleki> coi zo soi 07:06 < cliva> .ei mi co'u favgau lo go'i se pi'o lo ralju fukpi 07:07 < cliva> ie ze'a lo masti be li za'u na se pilno 07:07 < niftg> mi'e dai slabu ja purci me zo soi 07:14 < gleki> alta: na'e mo mo 07:14 < mensi> (NIhO [I {FA ZOhE} {CU <na'e mo> mo} VAU]) 07:14 < gleki> alta: mo nai mo 07:14 < mensi> (NIhO [I {FA ZOhE} {CU <mo nai> mo} VAU]) 07:15 < gleki> alta: na'e mo nai mo 07:15 < mensi> SyntaxError: Expected [yY] but "o" found. 07:15 < gleki> intentional 07:15 < gleki> however 07:15 < gleki> alta: na'e se mo nai mo 07:15 < mensi> (NIhO [I {FA ZOhE} {CU <na'e (¹se [mo nai]¹)> mo} VAU]) 07:15 < gleki> again intentional 07:23 < durka42> whyyyyy 07:23 < durka42> don't we have enough rules... 07:24 < Amnadya> I don't see the problem with {na'e [mo nai]}. 07:24 < Amnadya> NAI should be allowed anywhere. 07:35 < gleki> Amnadya: da'i ma smuni 07:36 < Amnadya> xu do jimpe fi lu mo nai li'u 07:36 < Amnadya> .i broda nai .i na'e broda .i na'e broda nai .i na'e na'e broda 07:37 < Amnadya> .i na nibli lo du ba se pilno 07:37 < Amnadya> .i ku'i .ei cumki fa lo nu pilno .i na sarcu fa lo nu fanta 07:38 < gleki> mu'a lu to'e mo nai li'u i ma smuni i xu lu to'e na'e mo li'u ji lu na'e to'e mo li'u 07:48 < Compu-Celebi> I cannot decipher {lo du ba se pilno}, which seems to be ungrammatical. 07:48 < gleki> yes, some people here use {du} instead of {du'u} and vice versa 07:49 < Compu-Celebi> Why would they do this? 07:50 < gleki> ti'e because du'u is two syllables and that's too many 07:51 < dutchie> zbusufukai zo'o 07:51 < ctefaho> it is part of some cmavo reshuffling, see http://mw.lojban.org/papri/ce_ki_tau_jau 07:53 < ctefaho> me I personally propose ja for ja'a:) 08:02 < dutchie> .u'i mi ganse je nelci lo ka cmene la sutysisku 08:02 < dutchie> (not sure if {cmene} is really the best translation for the html <title>) 08:03 < ctefaho> lo title sisi 08:04 < durka42> gimkamsmikezypro zo titla 08:28 < GothAlice> .ui coi 08:28 < mensi> GothAlice: cu'u la'o gy.zipcpi.gy.: I also like the semantic-poetic assonance between {mutmaika'i} and {la mivmroka'i}, 08:28 < mensi> if all your avatars represent the balance between two opposing concepts. | 2015-07-20T06:36:19.267Z and {la mivmroka'i}, 08:28 < mensi> if all your avatars represent the balance between two opposing concepts. | 2015-07-20T06:36:19. 08:28 < mensi> 267Z 08:44 < durka42> coi la .alis. 08:49 < durka42> jbo:mo'oi 08:49 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 08:49 < ldlework> I really love mo'oi 08:54 < durka42> I'd love to know its expansion 08:54 < durka42> because zipcpi is claiming on the mriste that it fixes a problem and I don't see how it does 08:54 * durka42 looks up selpahi's original article 08:56 < gleki> en: mo'oi 08:56 < mensi> mo'oi = [LE] interrogative gadri: "which" |>>> Roughly equivalent to ma noi, but doesn't start a new bridi. |>>> 08:56 < mensi> spheniscine 08:57 < gleki> wait but {ma noi} does start a new bridi 08:57 < latro`a_> exactly 08:57 < ldlework> That's what he's saying 08:57 < latro`a_> mo'oi broda = ma noi broda ku'o 08:57 < durka42> yeah ok 08:57 < ldlework> It allows you to more easilly relativize the question it self 08:58 < latro`a_> eh, I should put a ku on the left 08:58 < gleki> why is mo'oi needed then 08:58 < gleki> oh 08:58 < gleki> i misread 08:59 < gleki> it's le:mo'oi::zo'e:ma 08:59 < ldlework> do djica mo'oi plise poi zunle 09:00 < latro`a_> that's a little weird 09:00 < latro`a_> how is that different from {do djica mo'oi plise je zunle} 09:00 < durka42> so he definitely is using it wrong on the mriste :p 09:00 < ldlework> latro`a_: I have a bridi and you don't? 09:01 < durka42> {mo'oi plise poi zunle} seems like you're asking me a question and then asserting something about my answer 09:01 < durka42> what if I answer with a plise that doesn't satisfy {zunle} :) 09:01 < ldlework> I don't see why 09:01 < ldlework> durka42: why would that be a problem? 09:01 < latro`a_> it seems like there should be a semantic distinction between {noi ... poi} and {noi ... je ...} 09:01 < durka42> mo'oi plise poi zunle => ma noi plise ku'o zi'e poi zunle ? 09:02 < ldlework> I don't see a problem when a listener answers with a value that the speaker didn't ask for, as a problem with the speaker's speech 09:02 < durka42> ^ honest question, not an attack 09:02 < durka42> je'e 09:04 < gleki> exp: ma noi plise poi zunle 09:04 < mensi> SyntaxError: Expected [,] but "z" found. 09:05 < Amnadya> {plise je zunle} has a non-sensical x2, that's one difference. 09:06 < latro`a_> fair, I guess, but it's the same question as "what's the difference between {ma noi plise gi'e zunle cu ...} and {ma noi plise zi'e poi zunle cu ...} 09:06 < Amnadya> {lo broda poi brode} == {lo broda je poi'i ke'a brode} 09:06 < latro`a_> " 09:06 < ldlework> latro`a_: is it important for there to be a distinction? 09:07 < latro`a_> maybe there isn't one 09:07 < latro`a_> that's part of my question 09:07 < ldlework> latro`a_: what if multiple relativizations are simply conjunctions? 09:07 < ldlework> ah ok 09:07 < latro`a_> also, I disagree with that, amnadya 09:07 < Amnadya> {ma noi plise gi'e zunle} makes more sense than the one with {zi'e poi} 09:07 < latro`a_> the left side is instantiated prior to being relativized 09:07 < latro`a_> the right side is not 09:08 < Amnadya> What do you mean? 09:08 < latro`a_> in other words, there's some brodas we're talking about, I'm restricting to the ones that brode; vs. the right side where there are some things that both broda and brode that we are talking about 09:08 < Amnadya> Ah, no. 09:09 < Amnadya> You are talking about {lo broda *ku* poi brode}. 09:09 < latro`a_> my understanding was that with no inner quantifier there is no difference 09:09 < ldlework> I think that when you go all the way down, its just conjunctions with differences in emphasis. 09:09 < ldlework> IE iterative constraint or realtizivation 09:09 < Amnadya> There is a big difference, though. 09:09 < GothAlice> mensi: doi zipcpi Not all avatars are polar like that; polar avatars are a classification of them. (A "class" a la abominations, aberrations, outsiders, deep ones, the shadow, etc. all having distinct nuance, such as point of origin, power source, and weaknesses, despite all of those being "extradimensional" in general.) 09:09 < mensi> GothAlice: mi ba benji di'u ba lo nu la'o gy.zipcpi.gy. di'a cusku da 09:09 < Amnadya> (with poi anyway) 09:10 < ldlework> But it isn't a logical one 09:10 < durka42> wait I don't know the difference without quantifiers 09:10 < latro`a_> there is actually a logical difference with an inner quantifier 09:10 < durka42> with {po'o} I see the difference 09:10 < Amnadya> You do know! 09:10 < durka42> but with {poi}... not sure 09:10 < ldlework> There is a logical difference between poi and noi? /me raises eyebrow 09:10 < latro`a_> {lo ci broda poi brode} builds a group of three brodas all of which brode 09:10 < latro`a_> {lo ci broda ku poi brode} takes a group of three brodas and takes out some that brode; the result might be fewer than three thingsn 09:11 < durka42> wait let's let Amnadya speak, I want to know what's the difference without quantification :) 09:11 < ldlework> ehh 09:11 < latro`a_> with noi, you have the same effect in emphasis space, but you can't exclude anything in the second case, so the referential aspects are the same 09:11 * durka42 agrees with latro`a_ 09:11 < Ilmen> coi 09:11 < durka42> coi 09:12 < Amnadya> Yes, what latro'a said there is correct. 09:12 < ldlework> Where is that defined? 09:12 < ldlework> lo broda ku poi brode 09:12 < ldlework> The broda that brode's 09:12 < latro`a_> it's really where the first one is defined 09:12 < ldlework> But with quantification, boom it works different 09:12 < latro`a_> the second one works the way you expect 09:12 < ldlework> seems bad 09:12 < Amnadya> That follows from {SUMTI poi broda} == {lo me SUMTI je poi'i ke'a broda} 09:13 < latro`a_> yes, the second one is what you would expect 09:13 < durka42> ldlework: it's the same really 09:13 < latro`a_> the first one is the weird one 09:13 < Amnadya> That's the whole reason why {ma poi} is useless. 09:13 < latro`a_> I'd say "is a hack" rather than "is useless" 09:13 < ldlework> So where is it defined? 09:13 < ldlework> Where can I read about it 09:13 < Amnadya> BPFK. 09:13 < durka42> in the BPFK pages I think 09:13 < ldlework> I'm not too familiar with those, can someone link me? 09:14 < Amnadya> latro`a_: Did you read the blog article about {ma poi} ? 09:14 < durka42> start at mw.lojban.org/BPFK+Sections 09:14 < Amnadya> Because that shows why ma poi is broken. 09:14 < latro`a_> which one is that? 09:14 < Amnadya> http://selpahi.weebly.com/lojban/whichwhat-a-new-gadri 09:14 < latro`a_> again, I agree that it's broken when taken strictly, but I disagree that it doesn't achieve its goals in practice 09:14 < Amnadya> Because we misunderstood it all this time. 09:14 < durka42> ah so {lo broda ku poi brode} => {lo me broda poi brode} ? 09:14 < Amnadya> It's time to move on. 09:15 < durka42> er, => {lo me lo broda ku poi broda} ? 09:15 < zipcpi> AHAHAHAHA Oh man... old {zo'au} is in KOhA 09:15 < mensi> zipcpi: cu'u la'o gy.GothAlice.gy.: Not all avatars are polar like that; polar avatars are a classification of them. (A 09:15 < mensi> "class" a la abominations, aberrations, outsiders, deep ones, the shadow, etc. all having distinct nuance, such as point 09:15 < mensi> of origin, power source, and weaknesses, despite all of those being "extradimensional" in general.) | 09:15 < mensi> 2015-07-20T16:09:48.143Z 09:16 < zipcpi> That's rich... not much better than UI X3 09:16 < Amnadya> .u'i 09:16 < durka42> eh wat? I see a zo'au in LA 09:16 < Amnadya> durka42: Yes 09:16 < durka42> so the difference is that the reference has already happened 09:16 < zipcpi> Erm... yeah someone else stole zo'au 09:17 < zipcpi> ... for a purpose that's only useful pre-BCM T.T 09:17 < latro`a_> I'm not really seeing the difference between selpahi's (4) and (5) 09:17 < zipcpi> And technically one could use lo'u...le'u anyway 09:17 < latro`a_> in practice 09:17 < latro`a_> I see the syntax of course 09:17 < zipcpi> Ah... not sure what we're discussing 09:17 < durka42> don't worry you came in late 09:18 < latro`a_> we were talking about {ma poi} and also about {lo broda poi} (with no {ku}) 09:18 < zipcpi> K 09:18 < durka42> I mentioned {mo'oi} because you tried to use it on the mriste for something that it doesn't actually do, and now somehow we're talking about the difference between {lo broda poi brode} and {lo broda ku poi brode} 09:18 < zipcpi> It doesn't work? o.o 09:18 < Amnadya> Any sumti such that the chosen clothes are among its referents is a truthful answer to (5). In other words, it asks the wrong question. 09:19 < durka42> zipcpi: on the mriste you said {mo'oi pa <broda>} i.e. {mo'oi <sumti>} which... isn't a thing 09:19 < Amnadya> Mo'oi is in LE 09:19 < zipcpi> Grammatically mo'oi is in LE 09:19 < Amnadya> LE pa broda is a thing 09:19 < zipcpi> So it works 09:19 < durka42> ohh 09:19 < durka42> I didn't get it 09:20 < durka42> okay that might actually work 09:20 < durka42> maybe 09:20 < durka42> have to look at the expansion of LE PA :) 09:20 < gleki> Amnadya: fau lo nu na pilno lo cnino cmavo vau ma drata tadji lo nu pilno lo srana be lo ma poi li'u 09:20 < Amnadya> ma poi'i ... 09:21 < durka42> or just manoi? 09:21 < Amnadya> I don't like ma noi, but it's okay with me. 09:22 < Amnadya> It moves the restriction into a presupposition. 09:22 < durka42> yeah 09:22 < zipcpi> ma neu / zo'o 09:22 < gleki> Amnadya: do na'e nelci ki'u lo nu lo famyma'o ibo xu 09:23 < Amnadya> ki'u tu'a zo noi je ki'u lo nu bridi 09:23 * durka42 notes that gleki's new favorite construction "ibo UI" breaks under guskant's mabla change to {.i} 09:24 < gleki> mi gleua li xo'e 09:24 < gleki> Amnadya: xu zo noi na mulno mapti 09:24 < Amnadya> mi pu zi ciksi ru'e 09:24 < Ilmen> pe'i na 09:24 < Amnadya> .i ka'e darlu lo du'u mapti .i ku'i mi zmanei lo drata 09:25 < Amnadya> .i pe'i na dunli lo jai se djica 09:25 < Ilmen> coi 09:25 < Amnadya> coi la menli'i 09:26 < Ilmen> .i xu su da na te tugni 09:26 < niftg> me'enli'i 09:26 < Amnadya> .i xu su da ba'e ja'a te tugni 09:26 < Ilmen> sa'e je cu srana zo mo'oi 09:26 < Amnadya> na djuno 09:27 < gleki> xu ca su'o da lo ro casnu cu simxu lo ka tugni 09:27 < Amnadya> zo'o ru'e nu'o fasnu 09:27 < gleki> i sa'e xu ca su'o da pu'iku lo ro casnu cu simxu lo ka tugni 09:27 < durka42> .u'i ie 09:27 < Ilmen> ja'o casnu te zu'e lo nu djuno lo du xu kau tugni 09:27 < Amnadya> .i ja lo ro casnu cu memkai li pa 09:27 < gleki> i ie ma'oi ca'a vau ei 09:28 < durka42> co'o ei 09:28 < Ilmen> zo'o ru'e 09:28 < Amnadya> co'o 09:28 < Ilmen> co'o do 09:28 < Amnadya> co'o to'e mi 09:28 < Amnadya> zo'o 09:28 < Ilmen> .a'o zandotfri 09:29 < gleki> i ta'o ca lo cabdei je pare'uku mi jungau lo munje fi la altatufa 09:29 < Ilmen> ŭi carvi ca lo cabdei vi 09:29 < gleki> carmi xu carvi 09:30 < Ilmen> carmi pe'a gi'e ku'i ditcma 09:30 < gleki> camcarvi caircarvi 09:30 < gleki> i uidai co'a dunra bu'u do vau zo'o 09:30 < Amnadya> ca lo prulamdei vi carvi .i ku'i ze'i go'i .i to'e ki'u nai pluka sai 09:31 < gleki> alta: pluka sai 09:31 < mensi> (NIhO [I {FA ZOhE} {CU <pluka sai> VAU}]) 09:31 < zipcpi> ... dang it man, Malay has changed... at least colloquial Malay... then again I was never very familiar with colloquial Malay 09:31 < Amnadya> Those damn tinkerers. 09:31 < zipcpi> Exactly 09:31 < gleki> i ta'o ma'oi zo cu ei klesi ma'oi ro'a vau la'a 09:32 < gleki> zipcpi: hire Wuzzy 09:32 < zipcpi> lol 09:32 < Ilmen> Nice to hear. It means Malay is still alive. 09:32 < zipcpi> Malay even has a BPFK :p 09:32 < Ilmen> zo'o ru'e 09:33 < Ilmen> ro rarbau cu voi lo du ki co'u cenba cu xlali be ki nuzba 09:34 < zipcpi> ie 09:34 < gleki> alta: ro'a 09:34 < mensi> SyntaxError: Expected [yY] but "o" found. 09:34 < Amnadya> xu lo runbau cu ckaji lo simsa 09:34 < gleki> no idea how to fix bare {ro'a} 09:35 < gleki> into {ge'ero'a} probably 09:35 < zipcpi> Kamu ini siapa -> Kau ni siapa -> Ko ni sapa 09:35 < zipcpi> Never seen the latter in my entire life 09:35 < zipcpi> But saw it in an advert 09:35 < gleki> but bare {nai} is much more problematic 09:35 < ldlework> coi zipcpi 09:35 < zipcpi> coi 09:36 < zipcpi> Of course being an advert it's probably highly colloquial and not standard spelling 09:36 < gleki> zipcpi: do xabju ma poi si noi gugde 09:36 < zipcpi> lo mejgu'e 09:36 < Amnadya> mejgu'e 09:37 < gleki> zipcpi: xu do se bangu yyy fe la singlic 09:37 < Ilmen> .u'e .i xu ca ti ge xabju lo mejgu'e gi cinmo lo ka na certu tau lo mejbau 09:37 < gleki> i ku'i naku la'a la singlic i lo drata cmene 09:37 < zipcpi> ru'e .i zo'oi Manglish 09:37 < zipcpi> simsa zo malglico u'i 09:38 < gleki> i ie la manglic 09:38 < gleki> i banmu'isu 09:39 < gleki> is there a question word for NAI? 09:39 < Ilmen> xu mixre lo baxso bangu joi lo glibau 09:39 < Ilmen> pei 09:39 < Ilmen> y 09:39 < gleki> selmaho:nai 09:39 < mensi> .i lu nai li'u cmavo zo'oi NAI 09:39 < mensi> http://lojban.github.io/cll/14/3/ 09:39 < mensi> http://lojban.github.io/cll/15/7/ 09:39 < mensi> cmavo: ja'ai, nai 09:39 < gleki> nu'o 09:39 < zipcpi> mixre lo baxsybau joi lo glibau joi lo so'i jugbaukle 09:39 < Ilmen> je'e 09:40 < Ilmen> .i xu tonga bangu 09:40 < gleki> zipcpi: xu la manglic ne ji'a cu sutra farvi gi'e sutra cenba 09:41 < zipcpi> .y. cau'a .i lo'e fu'ivla ta'e tonga co'e 09:42 < Amnadya> jbo:cuxpei 09:42 < mensi> [< cuxna pensi ≈ Cuxna* pensi*] lo nu mulno sisku zo'u: y no da se tolcri 09:42 < mensi> cuxpei[5867], cu'arpei[7456], cuxpensi[7937], cuxnypei[7997], cu'arpensi[9526], ... 09:42 < gleki> jbo: cau'a 09:42 < mensi> gleua = x1 mo'e x2 va'ei se vajni x3 .i ra'a x2 li no so'aroi du li no'ai 09:42 < Amnadya> .. 09:42 < gleki> xu zo cau'a dunli lu ja'a xi xo'e li'u 09:43 < Ilmen> en: cuxtoi 09:43 < mensi> cuxtoi [< cuxna troci ≈ Choose try] = x1 is trying to choose which of x2 (plural) to be in relation x3 with |>>> 09:43 < mensi> Ilmen 09:43 < Ilmen> en: cuxselnandu 09:43 < mensi> cuxselnandu [< cuxna se nandu ≈ Choose 2nd conversion hard] = x1 hesitates on which of x2 (plural) to choose for doing 09:43 < mensi> x3 (relation) with it/them |>>> See also cuxtoi, cuxna. |>>> Ilmen 09:43 < Ilmen> .e'i jmina zo cuxpei 09:43 < zipcpi> mi no'e se slabu le cnino pe la manglic ja lo mejbau .i mi cu ru'esai jikca lo'e se mejbau 09:44 < Ilmen> .i xu naku lo xabju be lo mejgu'e cu se mejbau 09:44 < Ilmen> .i kucli sa'u 09:44 < zipcpi> mi xabju .iku'i mi ralju se bangu lo glibau 09:44 < gleki> velkemjvajvokepcmi-izing of rafsi should start from most stable concepts like e.g. time/space. e.g. we can start from cmalu/barda, gapru/cnita 09:45 < Ilmen> .i xu so'i xabju be lo mejgu'e cu se glibau 09:45 < zipcpi> so'a me lo mi lanzu cu se glibau 09:46 < zipcpi> ja'a 09:47 < Ilmen> cizra tcini .i mi pu zukpei lo ka toltce co cilre fi lo baxsybau .i .a'o na palcau .u'i 09:47 < zipcpi> su'o da se glibau .i su'o de se mejbau .i su'o di se jugbau .i su'o daxivo se banje'u'e 09:48 < zipcpi> .i me'au lo simsa 09:48 < Ilmen> .i'e melbi nu pilno zo me'au 09:50 < zipcpi> o'a 09:50 < zipcpi> .i la cliva cu ctuca 09:50 < Ilmen> .i xu lo mejgu'e vecnu .e lo simsa cu ta'e se glibau si'a 09:50 < Amnadya> .i mi pu zukpei lo ka tadni lo bindo no'u la .ba'asas. 09:51 < Ilmen> ra'o .i ti'e simsa ju simsytce lo mejbau 09:51 < zipcpi> su'o 09:51 < Amnadya> ti'e frili sai 09:51 < Ilmen> je'e la zipcpi 09:52 < gleki> mi pu djica lo ka tadni lo kadno glico i y bazisaibo mi snada u'i 09:52 < gleki> i snada lo ka co'a tadni gi'e mo'u tadni 09:53 < Ilmen> sei dai sadycni 09:53 < gleki> alta: mi pe mo 09:53 < mensi> (NIhO [I {FA <mi (¹pe GEhU¹)>} {CU <mo VAU>}]) 09:53 < zipcpi> lo mejbau zo'u zo'oi lah cnima'o .iku'i jai nandu fai lo ka se fanva 09:53 < gleki> "pe GEhU" because {pe do'e zo'e ge'u} would be really silly 09:53 < zipcpi> xy~'y simsa ru'e zo uai 09:54 < zipcpi> ja zo ko'oi 09:55 < zipcpi> https://en.wikipedia.org/wiki/Manglish#The_.22Lah.22_word 09:55 < zipcpi> u'i 09:56 < gleki> 61 open issues, 64 closed. seems like im starting to win. 09:56 < Ilmen> pe'i mi pu za ju zu tcidu doi la zipcpi 09:56 < gleki> a bit like Chinese "ba" 09:56 < Ilmen> .i'u ru'e 09:57 < zipcpi> "maluation" - embarrassment, from Malay "malu" + English "-ation". 09:57 < zipcpi> Never heard that one 09:57 < gleki> maybe "lah" is like {o'udai} 09:57 < Amnadya> pu za ju pu zu 09:58 < Ilmen> .oi je'e 09:58 < gleki> exp: pu za ju pu zu 09:58 < mensi> ([{<pu za> <ju (¹pu zu¹)>} KU] VAU) 09:58 < zipcpi> "Die/Finish/Gone/Habis/Mampus/Mampui/Sei/See/GG/Pok kai/" - generic exclamations to indicate "trouble", used like the English "dead" or "dead end" - "sei" is usually pronounced as its Cantonese equivalent, "die". (E.g. Today he die because of that loan shark (Today, he is in trouble because of the loan sharks. The word "die" does not mean to die literally)). 09:58 < zipcpi> lol "GG" 09:58 < zipcpi> Probably from "good game" 09:59 < Ilmen> .u'i 09:59 < gleki> it's all needs to be first studied in Chinese dialects 10:00 < Amnadya> mi pu zi jijnu (to fau lo nu va'i na pensi toi) finti zo zongata noi brivla panra zo zo'a (to mi sanji tau zo zonkai toi) 10:00 < gleki> en: zonkai 10:00 < mensi> zonkai [< zo'a ckaji ≈ Tangential to quality] = x1 is alongside of x2 |>>> Predicate form of zo'a; it means whatever 10:00 < mensi> zo'a means. See also zo'a. |>>> Ilmen 10:01 < zipcpi> u'i 10:01 < gleki> mi se mansa zo zonkai 10:01 < zipcpi> la ilmen cu nelci ra'oi -kai- 10:01 < Amnadya> mi djica lo nu lo brivla voksna cu panra lo cmavo voksna 10:02 < Amnadya> zo'a ze'o zo'i 10:02 < Ilmen> mi finti zo zonkai mu'i no drata be ri xi ni'u .e lo nu faumlu fa lo nu no da mapti brivo 10:02 < Amnadya> .i lo rafsi na sidju 10:02 < Ilmen> .i .e'a sai finti lo matmau ja mlemau 10:03 < Ilmen> .i lo velski pe zo zonkai ku ji'a na'e sidju 10:03 < Ilmen> :p 10:03 < zipcpi> zo zomza zo'o 10:03 < Amnadya> lol 10:03 < Amnadya> simsa lo juknyxarju 10:03 < Amnadya> en:juknyxarju 10:03 < mensi> juknyxarju [< jukni xarju ≈ Jukni* pig] = x1 does whatever a spider pig does. 10:03 < dutchie> u'i 10:04 < zipcpi> vlaste: zomzo (fapro) 10:04 < vlaste> zo zomzo gimkamsmikezypro no catni gismu .i zo zomzo gimkamsmikezypro no cipra gismu 10:04 < zipcpi> vlaste: zomza (fapro) 10:04 < vlaste> zo zomza gimkamsmikezypro no catni gismu .i zo zomza gimkamsmikezypro no cipra gismu 10:04 < Ilmen> la'a zo zombi simsyrai 10:04 < zipcpi> melbi 10:04 < zipcpi> ko'oi me'o myzy zifre 10:06 < gleki> somewhere in Amazonia Lojbanists say "uma gata zongata" - a cat walks along. 10:06 < Amnadya> zo'igve ze'ogve zo'agve 10:07 < Amnadya> ( __ + ragve ) 10:08 < Ilmen> .a'u 10:08 < zipcpi> ku'i da'i zo zomzi gimpro zo zomza 10:08 < Amnadya> xu smisimdu'e 10:09 < zipcpi> zomzyzomza 10:09 < Amnadya> .i li'a na smisimzma lo cmavo panra 10:09 < zipcpi> amzamzamzamzamzamza 10:09 < Ilmen> en: zomzi 10:09 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 10:09 < Ilmen> .i'u nai cai 10:09 < zipcpi> .y. selpau me'o myzy 10:10 < zipcpi> .i ja'ebo na se javdi 10:10 < zipcpi> si javni 10:10 < Amnadya> .e'e pei mi jmina lo mi zi'evla 10:10 < Amnadya> .i da'i su melmau cu zasti 10:10 < zipcpi> mi na pante 10:11 < zipcpi> da'i zo zomzi cu dunli lu xo'i zo'i li'u 10:12 < zipcpi> doi ilmen 10:12 < Amnadya> zo li'armi ne do cu plixau .i la selckiku pu finti zo va'arga lo simsa 10:12 < Ilmen> je'e zy 10:13 < zipcpi> k: bemza 10:13 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "b" found. 10:13 < Amnadya> spagetufa: amza 10:13 < spagetufa> (CU [.àmza VAU]) 10:13 < Amnadya> xe'e'e 10:13 < zipcpi> u'i 10:13 < gleki> ue 10:14 < gleki> spagetufa: bemza 10:14 < spagetufa> (CU [bèmza VAU]) 10:14 < Amnadya> la spagetufa cu jinga 10:14 < gleki> i ei la altatufa cu zukte da 10:14 < zipcpi> spagetufa: memze'a 10:14 < spagetufa> (CU [mem/zè'a VAU]) 10:15 < zipcpi> spagetufa: samzamsmacu 10:15 < spagetufa> (CU [sam/zam/smàcu VAU]) 10:15 < zipcpi> ui~~~~~ 10:15 < zipcpi> ba'adai'i la uuzit cu pante 10:16 < zipcpi> spagetufa: pacxlu 10:16 < spagetufa> (CU [pàc/xlu VAU]) 10:16 < zipcpi> .y. 10:16 < gleki> alta: xu na bamzu 10:16 < mensi> (NIhO [I xu] [FA ZOhE] [CU {na banzu} VAU]) 10:16 < gleki> :P 10:16 < gleki> alta:kibcu'a 10:16 < mensi> (NIhO [I {FA ZOhE} {CU <kibycu'a VAU>}]) 10:17 < zipcpi> spagetufa: kibcu'a 10:17 < spagetufa> (CU [kiby/cù'a VAU]) 10:17 < Amnadya> lo gentufa cu jamna se'e 10:17 < zipcpi> xy'y 10:17 < gleki> zu'ainmo 10:18 < gleki> alta: 6٦۶६૬੬৬୬౬೬൬௬༦၆๖៦໖ 10:18 < mensi> SyntaxError: Expected [yY] but "a" found. 10:18 < zipcpi> alta: samzamsmacu 10:18 < mensi> (NIhO [I {FA ZOhE} {CU <sanzamsmacu VAU>}]) 10:18 < zipcpi> .y. 10:18 < gleki> alta: 6٦۶६૬੬৬୬౬೬൬௬༦၆๖៦໖ da 10:18 < mensi> (NIhO [I {FA <(¹xa [xa xa xa xa xa xa xa xa xa xa xa xa xa xa xa xa]¹) BOI> da} {CU <COhE VAU>}]) 10:18 < gleki> spagetufa: 6٦۶६૬੬৬୬౬೬൬௬༦၆๖៦໖ da 10:18 < spagetufa> SyntaxError: Expected [.\t\n\r?!\u0020] or end of input but "x" found. 10:19 < zipcpi> alta: memze'a 10:19 < mensi> (NIhO [I {FA ZOhE} {CU <menze'a VAU>}]) 10:19 < zipcpi> pe'i zo memyze'a cu mapti semau 10:21 < zipcpi> le spagetufa cu to'e za'e banmrulygau keji'a me'o cy xy 10:22 < Amnadya> un-banhammers je'e je'e 10:22 < zipcpi> :p 10:23 < zipcpi> spagetufa: axca 10:23 < spagetufa> (CU [.àxca VAU]) 10:23 < zipcpi> spagetufa: axka 10:23 < spagetufa> SyntaxError: Expected [.\t\n\r?!\u0020] or end of input but "a" found. 10:24 < zipcpi> xa'a'a'a'a suda dukse lo ka zbusufukai kei sy 10:25 < gleki> https://github.com/holtzermann17/toki-pona-emoji/blob/master/word_list.org 10:28 < gleki> https://github.com/holtzermann17/toki-pona-emoji/blob/master/sample_phrases.org 10:29 < gleki> Thank you :smile: 10:29 < gleki> (Response) :smile: 10:29 < zipcpi> u'i 10:32 < gleki> alta: lo lojbo ne po'o mo 10:32 < mensi> (NIhO [I {FA <lo (¹lojbo [ne po'o] GEhU¹) KU>} {CU <mo VAU>}]) 10:33 < gleki> still doesnt work for names 10:33 < zipcpi> alta: la djan klama 10:33 < mensi> (NIhO [I {FA <la (¹djan klama¹) KU>} {CU <COhE VAU>}]) 10:33 < zipcpi> Oh cause it'd be part of the LA clause 10:34 < Amnadya> Isn't {la} already irregular with regards to relative phrases 10:34 < zipcpi> Yeah that's why I made {voi'i} 10:35 < zipcpi> Which... sorta clashes with new-voi lol 10:35 < Amnadya> Afterall, people seem to say that {la poi brode ku'o broda} has the relative clause as part of the name, even though CLL says that {lo poi brode ku'o broda} is the same as {lo broda ku poi brode} (I'm not taking sides with this statement) 10:36 < zipcpi> Hm 10:36 < zipcpi> Regardless it's an ugly solution for relative-clause titles 10:37 < zipcpi> That's why I wanted to wring one last kosmu out of voi for that 10:37 < zipcpi> But nope, new-voi, so I had to make {voi'i} 10:38 < Amnadya> (BPFK repeats CLL's view with regards to relative clause placements) 10:38 < Amnadya> (But BPFK is free to change their view in face of strong arguments!) 10:38 < Amnadya> tl;dr {la} is an exception. 10:39 < Amnadya> However, I'm not sure everyone would agree that {lo mi gerku} is {lo gerku ku pe mi} 10:39 < Amnadya> But that's CLL and BPFK. 10:40 < zipcpi> Hm yes, seems problematic if one wanted to be pedantic about {pe} 10:40 < gleki> ei la noi mu'a me BPFK jatna su'o da cu stidi 10:40 < gleki> sa 10:40 < gleki> ei la noi mu'a me BPFK jatna vau su'o da cu stidi 10:40 < zipcpi> Not to mention it breaks old po and po'e 10:40 < Amnadya> Never heard of those. :P 10:41 < durka42> lojbanists never want to be pedantic, don't worry about it 10:41 < zipcpi> I did actually sometimes use {po} to mean {be fi'o ponse}, but I don't really ever use {po'e} 10:41 < Amnadya> Another fegli option: {lo mi gerku} != {lo pe mi gerku} 10:42 < zipcpi> Regardless I added {pesai} and {pecai} to JVS just as a "drop-in" replacement with a cleaned-up meaning. Not sure if they'd be used often 10:42 < Amnadya> {po} and {po'e} are so rare that nobody should complain about having to use brivla for them instead 10:43 * zipcpi shrugs 10:43 < Amnadya> Not to mention {po'e} is {be}. 10:43 < zipcpi> Most of the time 10:43 < durka42> {be do'e} 10:43 < zipcpi> pe = bedo'e 10:43 < gleki> y mi pu pilno zo po ca lo ka traduki zo'ei lo linji pe la mupli 10:44 < zipcpi> http://jbovlaste.lojban.org/dict/pesai 10:44 < zipcpi> http://jbovlaste.lojban.org/dict/pecai?bg=1;langidarg=2 10:44 < Amnadya> pesai = poi ke'a co'e sai 10:44 < zipcpi> Heh 10:46 < zipcpi> I just added those on a lark, based on what people tell me {po} and {po'e} might mean 10:46 < zipcpi> May have some poetic use 10:46 * Amnadya de'a jundi 10:46 < zipcpi> And doesn't take up any cmavo space, so it's zabna that way 10:47 < zipcpi> So is like, use it, or don't; doesn't really matter 10:49 < gleki> a new function in mensi 10:49 < gleki> mensi: aigne 10:49 < mensi> CommonSenseError: Expected normal word but Curtis found. 10:49 < zipcpi> lol 10:49 < zipcpi> mensi: toi'e 10:49 < mensi> xu lo nu casnu la'e di'u cu raktu do 10:49 < gleki> works only for aigne 10:49 < zipcpi> OK 10:50 < gleki> finti: gleki 10:50 < mensi> 8738 da se tolcri: admine, advarka, adverbu, aftobuso, aftostopi, agmasto, aigne, aitco, aitxero, akrobat, akti, albatro, 10:50 < mensi> albuma, aldeia, alga, algoritma, alkanse, alna, altare, altrici, ambigu'o, amble, ambuci, ambulance, amfite'atro, ampute, 10:50 < mensi> anci, ankabuta, ankla, anli, ... 10:51 < gleki> due to some error im now the author of it 10:51 < zipcpi> ca'edai 10:59 < zipcpi> But yes there is a potential problem with the {lo broda pe mi} vs the {lo broda ku pe mi} distinction; intentionality isn't obvious, and closing a sumti can be difficult if NU is involved 11:04 < durka42> zantufa: ke lo broda ke'e pe mi 11:04 < zantufa> minajimpe: .i kanpe tu'a [uU] .onai [yY] .i ku'i "i" cu se tolcri 11:04 < zipcpi> spagetufa: ke lo broda ke'e pe mi 11:04 < spagetufa> SyntaxError: Expected [uU] but "i" found. 11:04 < zipcpi> spagetufa: ke po'o lo broda cu brode 11:04 < spagetufa> SyntaxError: Expected [,] but "b" found. 11:05 < zipcpi> ... something that Spagetufa didn't fix yet :p 11:07 < zipcpi> Well there's always i'au. But Curtis kinda broke i'au for UI lol 11:13 < zipcpi> I mean it could still work via deep magic. But it means that it won't be useful for attaching UI to sumti 11:21 < phma> If I want to say "you can also get there by Amtrak" with "klama", does Amtrak go in the route place or the vehicle place? 11:21 < zipcpi> Ugh I can never remember the five-place brivla T.T 11:22 < Amnadya> Is this a joke question? 11:22 < phma> no, this is serious. I'm translating the jbonunsla announcement. 11:23 < Amnadya> Then I'd tend to go with the vehicle place. 11:23 < Amnadya> Or are the rails themselves called Amtrak as well? 11:24 < phma> Sounds good. Amtrak doesn't have exclusive use of the rails. I was once on an Amtrak train that had to wait for the Tropicana train. 11:24 < Amnadya> Or reword with pilno la .amtrak. 11:28 * nuzba @garydale: Keyboard Shortcuts for Novelists:   Or you could just mash the keys together in a fit of anxious rage ... http://bit.ly/1IfnCtx #lojban [http://bit.ly/1fhJHKL] 11:29 < Amnadya> Suspicious lojban hashtag 11:29 < phma> mi platu le nu zvati lo berpau be la .delyuer. ze'a lo djedi be li so'u pu'o lo jbonunsla 11:29 < phma> is that correct use of "pu'o"? 11:29 < zipcpi> exp: le pa lu'i lo tcati ce lo ckafi 11:29 < mensi> ([le {<pa BOI> <lu'i (¹[lo tcati KU] [ce {lo ckafi KU}]¹) LUhU>} KU] VAU) 11:30 < Amnadya> I don't think so. I think {pu} is better there 11:30 < Amnadya> Or say {ca lo nu lo jbonunsla pu'o fasnu} 11:31 < phma> If I put "pu" there I could be in Delaware any time before jbonunsla. I'm leaving my sister's on Friday and will go straight to jbonunsla, unless there's someone to pick up. 11:31 < Amnadya> Experimental {se pu'o} could work, but nobody ever says that 11:32 < phma> {pu'o} is a tense marker, not a proper preposition. You can't se-convert tense markers. 11:32 < Amnadya> Not under CLL, no. But you can if tenses are merged with BAI. 11:33 < Amnadya> The main point however is that {pu'o} there doesn't mean what you want 11:33 < phma> What does {pu'o lo jbonunsla} mean? 11:33 < Amnadya> It means that at the time of jbonunsla, the main bridi is about to happen. 11:34 < phma> Then do I want {ba'o}? 11:34 < Amnadya> {ba'o} works, then it'd mean that at the time of jbonunsla you will have stayed at your sister's for a while. 11:34 < Amnadya> People usually avoid {ba'o} and {pu'o} as sumtcita because they are so confusing. 11:36 < Amnadya> What about {pu lo jbonunsla zi}? 11:36 < Amnadya> It's a short time before jbonunsla 11:36 < phma> How about " ba'o lo co'a jbonunsla "? 11:36 < zipcpi> Aaaaaaaaa 11:37 < zipcpi> ii~~~~~~~ 11:37 < zipcpi> sko'opu nu kalsa 11:37 < Amnadya> I think it would be much clearer not to use {ba'o} or {pu'o}. {ca lo nu pu'o} or {ca lo cfapru be lo jbonunsla} work 11:38 < phma> sko'opu ki'a 11:38 < zipcpi> "scope" 11:38 < Amnadya> There seems to be disagreement about {pu'o}'s meaning though. You are using it as "going to happen soon", while some use it as "it's going to happen, but not necessarily soon" 11:38 < phma> a silly scope? 11:38 < zipcpi> It's not in JVS because we don't know what the place structure should be T.T 11:39 < zipcpi> Nah, scope as in logical scope 11:41 < Amnadya> You could also use {prula'i}. 11:42 < Amnadya> Your stay precedes your going to jbonunsla 11:42 < zipcpi> Actually it's probably not the fault of scope, but of {ba'o} and {pu'o} lol 11:43 < Amnadya> They are the most mindbreaking tags. 11:43 < zipcpi> broda ba'o ko'a = ba'o broda ca ko'a ? 11:43 < Amnadya> I have yet to find a common use case 11:44 < Amnadya> ca ko'a ba'o broda 11:44 < zipcpi> Mm 11:44 < gleki> alta: ke lo broda ke'e pe mi 11:44 < mensi> (NIhO [I {<FA ZOhE> <ke (¹FA [lo broda KU]¹) ke'e>} {CU <FASNU (¹FA [LO {NU <(²FA ZOhE²) (²CU [COhE VAU]²)> KEI} KU] [pe {DOhE mi} GEhU]¹) VAU>}]) 11:44 < gleki> unexpected parse, isnt it? 11:44 < Amnadya> ke fa oO 11:45 < gleki> alta: ke gau lo broda ke'e pe mi 11:45 < mensi> (NIhO [I {<FA ZOhE> <ke (¹gau [lo broda KU]¹) ke'e>} {CU <FASNU (¹FA [LO {NU <(²FA ZOhE²) (²CU [COhE VAU]²)> KEI} KU] [pe {DOhE mi} GEhU]¹) VAU>}]) 11:45 < Amnadya> zantufa: fa ke lo broda 11:45 < zantufa> ([fa {ke <lo broda KU> KEhE}] VAU) 11:45 < Amnadya> zantufa: ke fa lo broda 11:45 < zantufa> ([ke {fa <lo broda KU>} KEhE] VAU) 11:45 < Amnadya> zantufa: fa ke fa lo broda 11:45 < zantufa> ([{fa KU} {ke <fa (¹lo broda KU¹)> KEhE}] VAU) 11:45 < Amnadya> je'e 11:46 < Amnadya> .i la'a plixau 11:46 < gleki> zantufa: ke lo broda cu mo 11:46 < zantufa> ([ke {lo broda KU} KEhE] cu [mo VAU]) 11:46 < gleki> y 11:47 < Amnadya> Anyone confused about ZAhO might want to read this: selpahi.de/ZAHO_Explanation.txt 11:47 < gleki> ei ma'a jijnu ma pe lu ke lo broda cu mo li'u 11:47 < gleki> i xu lu fa ke lo broda li'u 11:47 < durka42> ZAhO or #ZAhO? 11:47 < durka42> :p 11:47 < Amnadya> ZAhO vs ZA'O 11:48 < ecraven> coi ro do :) 11:48 < Amnadya> coi 11:49 < Amnadya> .i doi la gleki mi jijnu tu'a lu fa ke lo broda li'u 11:49 < Amnadya> .i do ma jijnu doi drata 11:51 < zipcpi> ie lu fa ke lo broda li'u 11:51 < gleki> Amnadya: mi na jijnu i mi kucli mi kakne lo ka stika la altatufa 11:52 < gleki> xorxe's formula for TAG ko'a is easy to understand although i prefer not {ca} but {fau=fi'o fanbu} there. 11:54 < gleki> oh no 68 issues opened, 63 closed. im losing this battle... 11:55 < zipcpi> Issues where 11:55 < gleki> https://github.com/lagleki/glekitufa/issues?q=is%3Aopen+is%3Aissue 12:05 < zipcpi> exp lo broda cu brode zo'u ko'a 12:05 < zipcpi> exp: lo broda cu brode zo'u ko'a 12:05 < mensi> SyntaxError: Expected [yY] but "o" found. 12:05 < zipcpi> I wonder if we can overload zo'u.... then again that doesnt' mean we *should* 12:06 < zipcpi> Might be semantically confusing 12:08 < zipcpi> Ugh my little thread is out of control... how many topics are there? 12:08 < zipcpi> NU tu'e...tu'u 12:08 < zipcpi> tu'e...tu'u + TAG 12:09 < zipcpi> Logical connectives... 12:09 < zipcpi> mo'oi 12:09 < zipcpi> Gah 12:10 < zipcpi> ja'a mi mutmaika'i lo ka kalsa kei va'o lo jbobau 12:16 < durka42> you started it :p 12:16 < durka42> "by the way" questions are always a bad idea 12:16 < durka42> they end up becoming the main topic 12:16 < zipcpi> :p 12:25 < zipcpi> spagetufa: lo nu tu'e broda .i brode tu'u 12:25 < spagetufa> ([FA {lo <nu (¹tu'e [{CU <bròda VAU>} {.i <CU (²bròde VAU²)>}] tu'u¹) KEI> KU}] VAU) 12:25 < zipcpi> spagetufa: tu'e broda .i brode tu'u ca ko'a 12:25 < spagetufa> SyntaxError: Expected [,], [lL], [mM], [nN], [rR] or [xX] but "k" found. 12:25 < zipcpi> spagetufa: ca ko'a zo'u tu'e broda .i brode 12:25 < spagetufa> ([{ca ko'a} zo'u] [tu'e {<CU (¹bròda VAU¹)> <.i (¹CU [bròde VAU]¹)>} TU'U]) 12:25 < zipcpi> So it's just the post-next thing that has a problem 12:26 < zipcpi> spagetufa: tu'e broda .i brode tu'u sei ca ko'a 12:26 < spagetufa> SyntaxError: Expected [,], [lL], [mM], [nN], [rR] or [xX] but "k" found. 12:26 < zipcpi> spagetufa: tu'e broda .i brode tu'u sei ca ko'a cu no'a 12:26 < spagetufa> (tu'e [{CU <bròda VAU>} {.i <CU (¹bròde VAU¹)>}] tu'u [sei {<ca ko'a> cu} no'a SE'U]) 12:27 < zipcpi> alta: tu'e broda .i brode tu'u sei ca ko'a cu no'a 12:27 < mensi> (NIhO [I {tu'e <NIhO (¹I [FA ZOhE] [CU {broda VAU}]¹) (¹i [FA ZOhE] [CU {brode VAU}]¹)> <tu'u (¹sei [ca ko'a] cu¹) no'a SEhU>}]) 12:27 < zipcpi> alta: tu'e broda .i brode tu'u sei ca ko'a 12:27 < mensi> SyntaxError: Expected [oO] but "a" found. 12:27 * zipcpi presses the buzzer 12:29 < zipcpi> Hm we *could* just allow {sei} to work like {zo'au}? 12:32 < Amnadya> Don't think so. 12:33 < zipcpi> Well, elidable selbri, altatufa staile :p 12:33 < zipcpi> But maybe that's mabla 12:33 < durka42> that way lies dragons si ambiguity 12:33 < zipcpi> lol 12:34 < Amnadya> I'm copying my handwritten plot notes to my computer right now, and there is a word I don't remember what I wanted it to mean 12:34 < zipcpi> ambigudrakono 12:34 < Amnadya> {carmirai} 12:34 < Amnadya> I'm pretty sure it has something to do with carmi and traji 12:34 < Amnadya> .o'i drakono 12:36 < zipcpi> Someone stole zo'au though, though I'm not sure JVS-zo'au is useful 12:37 < zipcpi> Well, has no use with BCM. Pre-BCM there might be an argument, but there's always lo'u...le'u 12:42 < zipcpi> ei mi cliva co'o 12:48 * nuzba @dogu_beer: しぇるにしょい。 #ohayo #Lojban [http://bit.ly/1TMIIBs] 12:54 < niftg> za'a da cernicoitu'itsku 13:27 < ecraven> is there a collection of lojban resources in .epub format anywhere? 13:28 < Amnadya> I don't know of any. 13:29 < Amnadya> PDFs are easier to come by 13:30 < ecraven> Amnadya: indeed, but less handy on ebooks, they'd need to be layouted for a5, not a4 :( 13:30 < Amnadya> Yeah 13:31 < durka42> I think the CLL can be built as an epub but I dunno if that code has bitrotted 13:31 < durka42> it's in the list of build targets 13:31 < ecraven> I just found cll 1.0 as an epub, I'll start with that :) 13:31 < ecraven> thank you ! 13:33 < durka42> bleh 13:33 < durka42> we need to get some more modern resources available as epubs then :) 13:36 < durka42> are there any good ways to convert wiki pages to epubs 13:36 < durka42> the wave lessons, say... 13:37 < durka42> mensi: doi gleki what do you think about instaling this? https://www.mediawiki.org/wiki/Extension:EPubExport 13:37 < mensi> durka42: mi ba benji di'u ba lo nu la'o gy.gleki.gy. di'a cusku da 13:38 < ldlework> Write everything in ReST 13:38 < durka42> or docbook! 13:38 < ldlework> Import into export formats such as mw and pdf, html, etc 13:39 < durka42> markdown is where it's at 13:39 < ldlework> lol markdown is terrible 13:39 < ldlework> docbook: way too verbose 13:39 < ldlework> markdown: not even a standard 13:39 < ldlework> rest: has a standard and exports to everything 13:40 < ldlework> human readable, and typeable 13:40 * Amnadya uses paint for everything. 13:40 < ldlework> The problem is there is no rst->docbook 13:40 < ldlework> and docbook is required for publishing 13:40 < ldlework> But only the CLL requires that 13:40 < durka42> if you can import from mw to rst and then export to epub... then we got a solution :p 13:40 < ldlework> if publishing can be forgiven, nothing beats rst, imo 13:41 < ldlework> durka42: there isn't 13:41 < ldlework> That's the problem with using app-specific markdowns as first source 13:42 < durka42> agreed 13:46 * nuzba @464161niftg: lojban › baby words, but general relevance: dai-like cmavo https://groups.google.com/d/topic/lojban/-6IIW2fn0mo/discussion / 自分は未だ読んでゐないものの [http://bit.ly/1HMa3fk] 17:18 * nuzba @usagipi314: lojban lo logji banguをなんとなく転写してみた なんだろう...めっちゃ創作文字っぽい [http://bit.ly/1JtmP2x] 17:20 * nuzba @usagipi314: @usagipi314: lojban lo logji banguをなんとなく転写してみた なんだろう...ちょっと無理があるけどめっちゃ創作文字っぽい http://t.co/JRIfvqu4ff [http://bit.ly/1JtmZHk] 19:13 < noncomcinse> camxes: ia nai cai 19:13 < camxes> (ia [nai cai]) 19:13 < noncomcinse> camxes: ia ru'e 19:13 < camxes> (ia ru'e) 19:14 < noncomcinse> camxes: ia ru'e nai 19:14 < camxes> (ia [ru'e nai]) 19:14 < noncomcinse> camxes: ia ru'e cai 19:14 < camxes> (ia [ru'e cai]) 19:14 < noncomcinse> camxes: ia nai ru'e cai 19:14 < camxes> (ia [nai {ru'e cai}]) 19:19 < noncomcinse> ki'e .camxes. 20:05 < _mukti_> ui mi mo'u jorne lo tcana poi se cmene zo lojban 21:40 < zipcpi> Crap who voted up cmene-quote-zo'au while I was away? I just added postnex-zo'au 21:46 < zipcpi> lol now I have two votes to their two. The fight is on 21:55 < zipcpi> Challenge: Translate Fermat's Last Theorem (not the proof, of course, only the statement) 21:56 < zipcpi> Without VUhU D: 22:01 < zipcpi> exp: a'y jo'u by jo'u cy vu'o poi cmima lu'i ro kacna'u 22:01 < mensi> SyntaxError: Expected [,] but "k" found. 22:01 < zipcpi> ... 22:02 < zipcpi> exp: a'y 22:02 < mensi> SyntaxError: Expected [.\t\n\r?! ], [0123456789], [eE] or [oO] but "a" found. 22:02 < zipcpi> Ah 22:02 < zipcpi> exp: abu jo'u by jo'u cy vu'o poi cmima lu'i ro kacna'u 22:02 < mensi> ([{<(¹a bu¹) BOI> <jo'u (¹by BOI¹)> <jo'u (¹cy BOI¹)>} {vu'o <poi (¹CU [cmima {lu'i <ro BOI> kacna'u KU} LUhU] VAU¹)> KUhO}] VAU) 22:03 < zipcpi> exp: ro abu jo'u ro by jo'u ro cy vu'o poi cmima lu'i ro kacna'u ku'o 22:03 < mensi> ([{<(¹ro BOI¹) (¹a bu¹) BOI> <jo'u (¹ro BOI¹) (¹by BOI¹)> <jo'u (¹ro BOI¹) (¹cy BOI¹)>} {vu'o <poi (¹CU [cmima {lu'i <ro BOI> kacna'u KU} LUhU] VAU¹)> ku'o}] VAU) 22:04 < zipcpi> exp: ro abu jo'u ro by jo'u ro cy vu'o poi cmima lu'i ro kacna'u ku'o ro ny poi mulna'u zi'epoi dubmau li re zo'u 22:04 < mensi> ([{<(¹[ro BOI] [a bu] BOI¹) (¹jo'u [ro BOI] [by BOI]¹) (¹jo'u [ro BOI] [cy BOI]¹)> <vu'o (¹poi [CU {cmima <lu'i (²ro BOI²) kacna'u KU> LUhU} VAU]¹) ku'o>} {ro BOI} {ny BOI} {poi <CU (¹mulna' 22:04 < mensi> u VAU¹)> KUhO} {zi'e <poi (¹CU [dubmau {li <re BOI> LOhO} VAU]¹) KUhO>}] zo'u) 22:06 < zipcpi> exp: no lu'i abu ce by ce cy vu'o poi cmima lu'i ro kacna'u ku'o ro ny poi mulna'u zi'epoi dubmau li re zo'u 22:06 < mensi> ([{<no BOI> <lu'i (¹[{a bu} BOI] [ce {by BOI}] [ce {cy BOI}]¹) (¹vu'o [poi {CU <cmima (²lu'i [ro BOI] kacna'u KU²) LUhU> VAU}] ku'o¹)> LUhU} {ro BOI} {ny BOI} {poi <CU (¹mulna'u VAU¹)> KUhO} 22:06 < mensi> {zi'e <poi (¹CU [dubmau {li <re BOI> LOhO} VAU]¹) KUhO>}] zo'u) 22:07 < zipcpi> Hm no the terms should be reversed 22:22 < zipcpi> exp: ro ny poi mulna'u gi'e dubmau 22:22 < mensi> ([{ro BOI} {ny BOI} {poi <(¹CU [mulna'u VAU]¹) (¹gi'e [CU {dubmau VAU}] VAU¹)> KUhO}] VAU) 23:10 < gleki> mensi: do durka i think that no new extensions should be installed 23:10 < mensi> gleki: cu'u la'o gy.durka42.gy.: what do you think about instaling this? https://www.mediawiki.org/wiki/ 23:10 < mensi> Extension:EPubExport | 2015-07-20T20:37:42.098Z 23:10 < mensi> You are being a bit negative. 23:10 < gleki> mensi: doi durka i think that no new extensions should be installed. 23:10 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.durka.gy. di'a cusku da 23:13 < gleki> mensi: doi durka as for exporting wiki pages to other formats i suggest that all of them are formatted under strictly formalized guidelines, then you could understand how those MediaWiki templates (that YOU know how they work), syntax is related to the language you need. 23:13 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.durka.gy. di'a cusku da 23:14 < gleki> mensi: doi durka then via MediaWiki API your scripts will download the source and convert it. seriously. mediawiki formatting is too rich. no such extensions work well with mediawiki templating 23:14 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.durka.gy. di'a cusku da 23:31 < gleki> mensi: doi zipcpi {sei ca ko'a} doesnt work alone since it must have a selbri 23:31 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.zipcpi.gy. di'a cusku da --- Day changed Tue Jul 21 2015 02:32 < gleki> en:some 02:32 < mensi> 270 da se tolcri: su'o, bu'a, bu'e, bu'i, co'e, da, da'au, datpre, de, di, falgau, farja'o, gincilta, gledji, jifnarju'o, 02:32 < mensi> karpo'i, kukselzba, kukselzbasu, la su'oda, mlibra, na'ebo, nerbi'o, pipcarna, piso'o, pisu'o, 02:32 < mensi> samjavyfonxypliduskemjuglerci'analka'ebi'ojaxyja'e, samyuan, seljunri, selskutidmulbi'o, selteptce, suzroi, 02:32 < mensi> vibnrbarpinji, xrucpe, zo'ei, adzau, aigne, anste, banli, barkla, befcaucreka, benre, besto, blurdrakono, bongu, 02:32 < mensi> bruneleskia, bu'ai, bu'oi, campu, cau'i, cau'o'e, ci'ai, ci'oi, cisma, cistrxrudu, ckajrdinti, cmeta, cnemu, co'udzau, 02:32 < mensi> crarai, da'ei, daigno, da'oi, dau'a, de'au, debrfaba, dekpu, di'au, di'ei, digno, di'oi, dizlo, djalazani, drata, dricai, 02:32 < mensi> drudi, dzaunlai, e'elxam, etxani, etxanoli, e'urfausiace'a, ... 02:36 < durka42> hmm, different LE have different expansions for LE PA 02:36 < mensi> durka42: cu'u la'o gy.gleki.gy.: i think that no new extensions should be installed. | 2015-07-21T06:10:49. 02:36 < mensi> 529Z 02:36 < mensi> durka42: cu'u la'o gy.gleki.gy.: as for exporting wiki pages to other formats i suggest that all of them are formatted 02:36 < mensi> under strictly formalized guidelines, then you could understand how those MediaWiki templates (that YOU know how they 02:36 < mensi> work), syntax is related to the language you need. | 2015-07-21T06:13:50.040Z 02:36 < mensi> durka42: cu'u la'o gy.gleki.gy.: then via MediaWiki API your scripts will download the source and convert it. seriously. 02:36 < mensi> mediawiki formatting is too rich. no such extensions work well with mediawiki templating | 2015-07-21T06:14:32. 02:36 < mensi> 670Z 02:37 < gleki> strange. now "en: some" now works correctly 02:37 < gleki> mensi: gadri 02:37 < mensi> lo [PA] broda = zo'e noi ke'a broda [gi'e zilkancu li PA lo broda] 02:37 < mensi> la [PA] broda = zo'e noi lu [PA] broda li'u cmene ke'a mi 02:37 < mensi> lo PA sumti = lo PA me sumti 02:37 < mensi> la PA sumti = zo'e noi lu PA sumti li'u cmene ke'a mi 02:37 < mensi> loi [PA] broda = lo gunma be lo [PA] broda 02:37 < mensi> lai [PA] broda = lo gunma be la [PA] broda 02:37 < mensi> loi PA sumti = lo gunma be lo PA sumti 02:37 < mensi> lai PA sumti = lo gunma be la PA sumti 02:37 < mensi> lo'i [PA] broda = lo selcmi be lo [PA] broda 02:37 < mensi> la'i [PA] broda = lo selcmi be la [PA] broda 02:37 < mensi> lo'i PA sumti = lo selcmi be lo PA sumti 02:37 < mensi> la'i PA sumti = lo selcmi be la PA sumti 02:37 < mensi> PA sumti = PA da poi ke'a me sumti 02:37 < mensi> PA broda = PA da poi broda 02:37 < mensi> piPA sumti = lo piPA si'e be pa me sumti 02:37 < durka42> yeah... I'm not going to do that and maintain such scripts 02:37 < durka42> I just thought we could try it to see if epub export works 02:37 < gleki> and im not going to increase memory load for lmw 02:37 < durka42> ku'i a'icu'i 02:37 < gleki> even tweetanew is disabled 02:38 < durka42> and nothing of value was lost zo'o 02:38 < gleki> something was lost since people retweeted what uitki posted 02:39 < durka42> ua 02:39 < gleki> i need to write a script to read rss of le uitki and post news to twitter 02:40 < durka42> but then where do you run it .u'i 02:42 < gleki> https://github.com/pxg/Node-Rss-to-Twitter 02:42 < gleki> the script is so small that i dont even want to use it 02:47 < gleki> zantufa: na'e bo mi mo 02:47 < zantufa> ([{na'e bo} mi LUhU] CU [mo VAU]) 02:47 < gleki> zantufa: na'e mi mo 02:47 < zantufa> ([na'e mi] CU [mo VAU]) 02:49 < gleki> zantufa: na'e bo noi mo vau mi mo 02:49 < zantufa> ([{na'e bo} {noi <CU (¹mo vau¹)> KUhO} mi LUhU] CU [mo VAU]) 02:49 < gleki> zantufa: na'e noi mo vau mi mo 02:49 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 02:49 < gleki> je'e 02:50 < gleki> no, it makes no sense 02:52 < gleki> how should it even be understood 02:52 < gleki> na'e modifies only NOI ? 02:54 < Amnadya> coi stati remna 02:54 < Amnadya> Could a smart Lojbanist tell me what they think about https://www.reddit.com/r/Ithkuil/comments/3d07k4/how_to_specify_destination_another_missing_case/ 02:55 < durka42> na'i 02:55 < durka42> would you accept a dumb lojbanist's opinion? 02:55 < Amnadya> I don't see any, but yes. 02:55 < durka42> :) 02:56 < Amnadya> Sometimes Ithkuil has very straightforward methods for saying stuff, but sometimes JQ comes up with rather contrived-seeming solutions. 02:56 < gleki> yy mi jinvi da 02:56 < gleki> i mi tugni la selpa'i 02:57 < Amnadya> ua 02:57 < gleki> i la versiio ka'e vlipa 02:57 < Amnadya> However, maybe I just don't understand it well enough 02:58 < gleki> ka'eku lo do menli co'a lojbo i ku'i ibu na jbobau gi'e na lobykai gi'e la'acu'i na bridi bangu 02:58 < Amnadya> lakne 02:59 < durka42> are there short definitions of CPT, PRG and ICP somewhere? 03:00 < Amnadya> CPT = completive version, PRG progressive aspect, ICP inceptive extension 03:00 < gleki> http://ithkuil.net/abbreviations.html 03:00 < Amnadya> PRG = ca'O 03:00 < Amnadya> CPT is related to mo'u 03:01 < Amnadya> ICP is lo tolfa'o be 03:01 < gleki> there is also a never complete notes for lojbanists https://docs.google.com/document/d/1SI7G0C-MoJIZ51N8jhL_7rLd6mIHA4Waq3F7w159RRs/edit 03:02 < gleki> no'i do ka'e xusra lo du'u ibu spofu i ku'i mu'a la nutka i na bridi bangu i ku'i ca'a se pilno 03:03 < durka42> vlaste: nutka 03:03 < vlaste> nutka = x1 talks in a non-sumti-or-predicate manner x2 (property of x1) 03:03 < gleki> number = expr:(PA_clause (PA_clause / lerfu_word)*) {return _node("number", expr); } 03:03 < gleki> lerfu_string = expr:(lerfu_word (PA_clause / lerfu_word)*) {return _node("lerfu_string", expr); } 03:03 < durka42> wat 03:03 < Amnadya> .i ku'i va'o ku la .itku'ile cu nandu ki'u nai lo nu lo gerna cu pluja je ki'u ja'ai lo nu lo smuni ciste na .y jai li'armi sa'e nai ru'e 03:04 < gleki> en: baurnu'uku 03:04 < mensi> baurnu'uku = x1 is the language with ISO 639-3 code ''nuk'' (Nuu-chah-nulth, Nootka) |>>> See also bangu |>>> 03:04 < mensi> gleki 03:04 < gleki> Amnadya: #ermenefti xu 03:04 < durka42> now I'm thinking about how to translate those examples to lojban 03:05 < gleki> this number/lerfu_string distinction is just lol 03:05 < durka42> seka'a and fa'a, I guess 03:05 < Amnadya> do'e lo jbobau lo nu cilre fi zo nu je zo lo cu se jalge lo nu jimpe fi lu lo nu li'u .i jicmu lo jbobau .i frili .i la .itku'ile zo'u mo .i na va'arga 03:05 < Amnadya> .i lo gerna na nabmi 03:05 < gleki> off: abu za'ure'u mo 03:05 < mensi> SyntaxError: Expected [yY] but "o" found. 03:05 < gleki> because number must start with a PA 03:06 < gleki> off: za'uabu re'u mo 03:06 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "z" found. 03:06 < gleki> off: za'u abu re'u mo 03:06 < mensi> ([{<za'u (¹a bu¹)> re'u} mo] VAU) 03:06 < gleki> so we can say ">a" times but not "a times" 03:07 < Amnadya> ie .i mi na nelci 03:07 < mensi> mi xebni 03:07 < gleki> re do xebni 03:07 < durka42> camxes: +exp vei abu re'u mo 03:07 < camxes> (CU [{<vei (¹[a bu] BOI¹) VEhO> re'u} mo] VAU) 03:08 < gleki> "Alas for altást." (C) 03:08 < gleki> altasta = x1 (text) is a word or construct having grammatical limitations x2 (property of x1) that make no sense 03:11 < durka42> according to x3 by logic system x4 03:11 < Amnadya> xu na betri fa lo nu facki lo du'u lo pluja gerna ba'e na je lo smuni nu kalsa cu ralju krinu lo nu la .itku'ile cu nandu je ba jai se fliba fai lo ka co'a certu 03:13 < durka42> .u'i 03:13 < gleki> y i ku'i mi nelci je se pluka lo ka tadni lo smuni i na vjni mi falonu na ka'e akti 03:14 < gleki> i mu'a xo da se bangu la lojban i so'u da i simsa ibu 03:14 < gleki> i ko pilno lo glibau e lo bancumunu e PHP i ki'u bo ralju bangu 03:15 < durka42> ni'o mi puzi co'a tcidu la'o sy la luna es una cruel amante sy .i lo skami cu se bangu la loglan 03:16 < gleki> durka42: ( VEI_clause mex VEhO_elidible / number ) ROI_clause i fegli 03:17 < gleki> durka42: http://www.hollywoodreporter.com/heat-vision/bryan-singer-tackling-sci-fi-778949 03:17 < Amnadya> so'o remna ca'o troci lo ka certu tau la .itku'ile kei je ja'a djica lo nu se bangu .i lo nu da'i spofu cu se badri ca lo nu da'i facki 03:18 < gleki> ke'u gau ko e'u ibu gletu la to'anzu 03:18 < gleki> i za'adai JW na permite fi lo ka cnegau 03:18 < gleki> sa'ai JQ le'ai 03:19 < Amnadya> la zgana be la .ie'ovas. 03:20 < durka42> xu zo be ji zo pe 03:20 < durka42> .wik Jehovah's Witness 03:21 < fenki> "Jehovah's Witnesses is a millenarian restorationist Christian denomination with nontrinitarian beliefs distinct from mainstream Christianity." - https://en.wikipedia.org/wiki/Jehovah's_Witness 03:21 < Amnadya> be ja pe oO 03:23 * nuzba @oka_iu_tcan: http://jbovlaste.lojban.org/dict/ifle if x1 (proposition) is true then x2 (event) happens else x3 (event) happens [http://bit.ly/1IhhP6u] 03:23 < durka42> aizeias bu voci pi'e pano mai lu do'o me lo mi zgana sei la .ie'ovas. cu cusku je lo mi selfu noi mi ba'o cuxna tu'a ke'a 03:24 < gleki> lo vedli be la ie'ovas 03:25 < durka42> sumlafi 03:25 < Amnadya> .oi .u'i xu lo cevni na ka'e fasnu ja bridi 03:25 < Amnadya> .i kakne ro da 03:25 < durka42> ie 03:25 < durka42> xu la ie'ovas cu cevni 03:26 < Amnadya> li'a 03:26 < Amnadya> .i datcme 03:26 < Amnadya> .i cevni lo se cevni 03:26 < gleki> o'i brod lo se broda 03:27 < Amnadya> cevni ma kau 03:27 < gleki> i la'oi DCY vajni javni 03:28 < durka42> xu zoi gy I am that I am gy cu xe fanva lu mi me ma kau li'u 03:28 < Amnadya> .u'i 03:29 < durka42> lo cevni ku po'o cu jai se curmi fai lo ka pilno lu ma kau li'u poi na nenri lo mutybri 03:29 < Amnadya> .i lo krici be lo xriso cu ze'u sai dambysi'u tu'a lo smuni 03:29 < gleki> je'usai ermenefti 03:29 < durka42> lojbo panra 03:29 < gleki> alta: abu re'u mo 03:29 < mensi> (NIhO [I {FA ZOhE} {CU <(¹a bu¹) re'u> mo} VAU]) 03:30 < durka42> alta: abu za'u re'u mo 03:30 < mensi> (NIhO [I {FA ZOhE} {CU <(¹[a bu] za'u¹) re'u> mo} VAU]) 03:30 < gleki> :P 03:30 < durka42> .u'i 03:30 < durka42> alta: abu boi za'u re'u mo 03:30 < mensi> (NIhO [I {FA <a bu> boi} {CU <za'u re'u> mo} VAU]) 03:30 < durka42> alta: abu boinai za'u re'u mo 03:30 < mensi> (NIhO [I {FA <(¹a bu¹) boi> nai} {CU <za'u re'u> mo} VAU]) 03:30 < Amnadya> la .ba'amut. cu zanru 03:30 < gleki> oi mi co'u morji lo du'u la ba'a mut cu mokau 03:30 < durka42> LA BAHAMUT CU CUSKU LO SE CUSKU BE LA BAHAMUT 03:30 < gleki> i la'a tolxorxe 03:31 < gleki> abuboi nai ~=- na'e abu boi 03:32 < gleki> i va'oku zo nai pritu je ckilu cmavo 03:32 < gleki> gi'enai cnima'o stika 03:32 < gleki> ibo ma'i la altatufa 03:34 < gleki> i cinri falo nu lo mekso pe la zantufa pu jai se facki mi fai lo ka mulno mapti ro drata pagbu be lo gerna pe la altatufa 03:36 * nuzba @oka_iu_tcan: そういえば、昨日nifさんが dancari って言ってたので、思わずそれを文字った zancari を定義してしまた http://jbovlaste.lojban.org/dict/zancari [http://bit.ly/1IhjARn] 04:15 < gleki> either 1. {vei ny} and {pa} are numbers, {ny} and {pabu} are pronouns, or 2. unification, {pa} works as a pronoun except for {PA SELBRI/SUMTI/ROI/MAI/MOI} 04:40 < zipcpi> I prefer 1; PA as pronouns give me a headache 04:40 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: {sei ca ko'a} doesnt work alone since it must have a selbri | 2015-07-21T06:31:12.607Z 04:40 < mensi> mediawiki formatting is too rich. no such extensions work well with mediawiki templating | 2015-07-21T06:14:32. 04:40 < mensi> 670Z 04:40 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: {sei ca ko'a} doesnt work alone since it must have a selbri | 2015-07-21T06:31:12. 04:40 < mensi> 607ZMediaWiki API your scripts will download the source and convert it. seriously. mediawiki formatting is too rich. no 04:40 < mensi> such extensions work well with mediawiki templating | 2015-07-21T06:14:32.670Z 04:40 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: {sei ca ko'a} doesnt work alone since it must have a selbri | 2015-07-21T06:31:12. 04:40 < mensi> 607Z 04:40 < zipcpi> Er, yes, I know... but alta is alta :p 04:41 < zipcpi> Anyway I added {zo'au} to JVS, though it has a competing def 04:42 < zipcpi> And lol Zantufa... she repurposed FA for to'ai/vo'ai/xo'ai 04:42 < durka42> wat 04:42 < zipcpi> Yes 04:42 < zipcpi> http://mw.lojban.org/papri/zantufa_ze%27oi_FA 04:43 < zipcpi> What would the rafsi be even :p 04:43 < durka42> ua 04:43 < durka42> .u'i 04:43 < durka42> rafsi: faq feq fiq foq fuq 04:43 < zipcpi> u'i 04:44 < durka42> nah I don't like that 04:44 < durka42> it seems useful at first but it's way inconsistent 04:44 < durka42> conflicts with tag unification 04:44 < durka42> maybe I'll come around later though 04:44 < zipcpi> Problem with Eulojban-q is that CVq rafsi has this little problem of being ambiguous before k/g/x. Unless we force it to always have a -y- 04:45 < zipcpi> Or have a -y- before k/g/x, but that's a little more complicated 04:45 < mensi> y no da se tolcri 04:46 < zipcpi> .y. ri'a ma la mensi cu spuda 04:46 < gleki> do pu cusku zo'o raf: 04:46 < gleki> sa 04:46 < gleki> i do pu cusku zo'oi rafsi: 04:47 < gleki> any other opinions on lerfu vs. numbers? 04:47 < zipcpi> I have my own Eulojban suggestion but it's a little crazy 04:48 < zipcpi> Basically ca'e all lerfu pronouns can only have one letter, unless you use {leu} (selma'o LI, suggested ortho-Lojban variant {ko'ai}) 04:49 < durka42> zipcpi: I thought you said qk/qg/qx would be a disallowed pair, then there is no problem 04:49 < zipcpi> Hm... 04:49 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: {sei ca ko'a} doesnt work alone since it must have a selbri | 2015-07-21T06:31:12. 04:49 < mensi> 607ZMediaWiki API your scripts will download the source and convert it. seriously. mediawiki formatting is too rich. no 04:49 < mensi> such extensions work well with mediawiki templating | 2015-07-21T06:14:32.670Z 04:49 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: {sei ca ko'a} doesnt work alone since it must have a selbri | 2015-07-21T06:31:12. 04:49 < mensi> 607Z 04:49 < zipcpi> ... 04:49 < gleki> sorry 04:49 < gleki> it's because i reloaded mensi 04:49 < gleki> alta: pa cu mo 04:49 < mensi> (NIhO [I {FA <pa BOI>} {cu <mo VAU>}]) 04:49 < gleki> alta: pa mo 04:49 < mensi> (NIhO [I {FA <pa BOI> mo KU} {CU <COhE VAU>}]) 04:49 < durka42> ambigu 04:49 < durka42> alta: pa da mo 04:49 < mensi> (NIhO [I {FA <pa BOI> da} {CU <mo VAU>}]) 04:50 < gleki> na ambigu 04:50 < durka42> alta: pa lo broda cu mo 04:50 < mensi> (NIhO [I {FA <pa BOI> <lo broda KU>} {cu <mo VAU>}]) 04:50 < durka42> alta: lo broda pa mo 04:50 < mensi> (NIhO [I {<FA (¹lo broda KU¹)> <FE (¹pa BOI¹) mo KU>} {CU <COhE VAU>}]) 04:50 < durka42> alta: lo pa broda be re mo 04:50 < mensi> (NIhO [I {FA <lo (¹[pa BOI] [broda {be <FE (²re BOI²) mo KU>} BEhO]¹) KU>} {CU <COhE VAU>}]) 04:50 < gleki> i ku'i simlu lo ka ambigu i lo nu go'i ka'e nabmi lo remna 04:50 < durka42> alta: lo pa broda be re cu mo 04:50 < mensi> (NIhO [I {FA <lo (¹[pa BOI] [broda {be <FE (²re BOI²)>} BEhO]¹) KU>} {cu <mo VAU>}]) 04:50 < durka42> maybe 04:51 < durka42> I mean "this PEG is not ambiguous" is a trivial statement :p 04:51 < gleki> u'i ie 04:51 < Amnadya> I think I'd rather see {lo brode be re cu mo} as {lo brode be re lo cohe cu mo} 04:52 < durka42> obviously it's {lo broda be re me zo cu be'o mo} 04:52 < Amnadya> Option 3: BY becomes part of PA. Neither are pro-sumti. (Just listing it for completeness) 04:53 < zipcpi> Another thing about Eulojban is that it'd more accurately transcribe Euclid's name. Though you can probably say that about any phonemic expansion, then we end up with Ithkuil :p 04:55 < zipcpi> http://mw.lojban.org/papri/The_North_Wind_and_the_Sun#la_zipcpi.27s_version.2C_2015 04:55 < zipcpi> Gonna have to define {selkufygla} 04:55 < Amnadya> by LLG ? 04:55 < durka42> comfortably warm? 04:56 < zipcpi> Yes 04:56 < Amnadya> When has LLG ever translated something 04:56 < zipcpi> I'm not sure o.o 04:56 < durka42> pusaiku 04:56 < gleki> ba lo nu la gleki co'a cmima 04:56 < durka42> ue 2015 xu 04:56 < Amnadya> do du la gleki je nai LLG .i la gleki pu fanva 04:56 < durka42> ah it's an "everybody has a share" kind of thing then 04:56 < durka42> xu doi la .mailos. 04:56 < gleki> i ku' ma'a nitcu lo catni xe fanva 04:57 < durka42> shouldn't you ask others before calling yours catni, .u'i 04:57 < Amnadya> ie sai 04:57 < gleki> mi pu cpedu 04:57 < Amnadya> What is pre-Lojban? 04:57 < durka42> tolclite 04:57 < gleki> i ku'i no drata be la lojbab pu spusku 04:57 < durka42> gleki: .i ma pu danfu 04:58 < gleki> i ko tcidu lo LLG mriste 04:58 < durka42> lo nu na spusku na sai nu zanru 04:58 < durka42> li'a 04:58 < zipcpi> I dunno... but the style is that really old style I remember, like the preponderance of rafsi cmene 04:58 < zipcpi> An actual use of jei lol 04:58 < durka42> since it's the LLG's translation and I'm a member I can edit it to add pauses, right? 04:59 < gleki> li'a 04:59 < zipcpi> I think I can drop the sel- 05:00 < gleki> i mi pu nitcu lo catni xe fanva xau la uikipediias 05:00 < durka42> na'i .i mi glare je kufra 05:00 < durka42> zo'o 05:00 < zipcpi> u'i 05:00 < gleki> i ku'i no pilno be la uikipediias pu pante 05:00 < gleki> i ja'o na vajni lo nu xukau lo catni cu zasti 05:00 < gleki> *fa 05:01 < Amnadya> There needs to be a mechanism by which a text becomes officially LLG approved. But until LLG actually votes or whatnot it's misleading to call anything an LLG translation. 05:01 < durka42> mi morji lo mristeka'u .i ku'i .ei gaje bitygau lo drata lo du'u zabna gi ba'e nai sevzi catni xusra 05:01 < zipcpi> ganaida'i do glare je kufra gi li'a lo nu do glare cu se kufra do 05:02 < durka42> mi glare prenu iau ro'u 05:02 < gleki> Amnadya: e'u do fanva gi'e mrilu fi lo LLG mriste 05:02 < Amnadya> no'e jetnu .i ka'e glare je kufra be lo ka zvati ma kau .i na nibli lo du'u se pluka lo ka glare 05:03 < Amnadya> .i na cinri mi doi la gleki .i lo lisri cu to'e cinri .i je mi senpi lo du'u vajni fa lo nu LLG cu zukte lo srana 05:04 < zipcpi> http://mw.lojban.org/papri/fancylojban_challenges xu lemedeimoi cinri 05:04 < Amnadya> cu * 05:04 < zipcpi> si cu cinri 05:04 < Amnadya> ka'e cinri lo skami se cinri 05:07 < gleki> do pante i ku'i na dudvota 05:08 < zipcpi> adzau ma 05:08 < gleki> lo xe fanva 05:41 < zipcpi> exp: .i ku'i .ei gaje bitygau lo drata lo du'u zabna gi ba'e nai sevzi catni xusra 05:41 < zipcpi> alta: .i ku'i .ei gaje bitygau lo drata lo du'u zabna gi ba'e nai sevzi catni xusra 05:41 < mensi> ([i {ku'i ei}] [CU {ga je} {CU <bitygau (¹lo drata KU¹) (¹lo [du'u {CU <zabna VAU>} KEI] KU¹)> VAU} {gi <ba'e nai>} {CU <sevzi catni xusra> VAU} VAU]) 05:41 < mensi> (NIhO [i {ku'i ei}] [CU {ga je} {CU <bitygau (¹[FA {lo drata KU}] [FA {lo <du'u (²[FA ZOhE] [CU {zabna VAU}]²) KEI> KU}]¹) VAU>} {gi <ba'e nai>} {CU <sevzi catni> xusra} VAU]) 05:42 < durka42> seems the same 05:42 < zipcpi> I have no idea what cogas means by {zancari} 05:44 < durka42> the lojban def seems straightforward 05:44 < gleki> https://en.wiktionary.org/wiki/%E6%96%AD%E6%8D%A8%E9%9B%A2 05:45 < gleki> xm 05:45 < zipcpi> durka42: Oh yeah didn't think to look there 05:46 < gleki> it's like making something without unnecessary decorations 05:47 < gleki> sampu isnt enogh imo 05:47 < zipcpi> Probably 05:47 < gleki> wow, so slow 05:47 < mensi> gleki: cu'u la'o gy.zipcpi.gy.: Oh yeah... forgot about that. Must've been a brain-fart.... though the web version does 05:47 < mensi> lack the "alta" version though that marks the place structure | 2015-05-13T19:48:23. 05:47 < mensi> 719Z 05:48 < zipcpi> ... laldo laldo 05:48 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: try describing the priority of nai/toi'e/cai etc. in an article like it's done for ke/bo 05:48 < mensi> etc. in Waves. | 2015-05-22T06:08:35.669Z 05:48 < zipcpi> oi oi oi oi oi 05:48 < gleki> mensi should trim dots automatically 05:52 < gleki> mensi: doi zipcpi.... coi 05:52 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.zipcpi.....gy. di'a cusku da 05:53 < zipcpi> coi 05:54 < gleki> oi mi bebna i na cikre 05:54 < gleki> i ku'i mi cikre lo drata 05:55 < gleki> alta: pa mo 05:55 < niftg> denpAbu so'i mei 05:55 < mensi> (NIhO [I {FA <pa BOI> mo KU} {CU <COhE VAU>}]) 05:55 < gleki> alta: ny mo 05:55 < mensi> (NIhO [I {FA <ny BOI>} {CU <mo VAU>}]) 05:55 < gleki> alta: pabu mo 05:55 < mensi> (NIhO [I {FA <pa bu> BOI} {CU <mo VAU>}]) 05:55 < gleki> alta: vei ny mo 05:56 < gleki> zo'o nandu preti 05:56 < mensi> (NIhO [I {FA <(¹vei ny VEhO¹) BOI> mo KU} {CU <COhE VAU>}]) 05:57 < gleki> ie 05:57 < gleki> alta: vei ny pa mo 05:57 < mensi> (NIhO [I {FA <(¹vei ny pa VEhO¹) BOI> mo KU} {CU <COhE VAU>}]) 05:57 < gleki> li'o 06:04 < zipcpi> mi fanva fi lo tcekitaujei .i xu srera 06:04 < zipcpi> .y. la jbouitki cu spofu ga'ami 06:05 < zipcpi> go'ijamasno 06:10 < zipcpi> zantufa so breaks my use of {ibabo} there 06:10 < zipcpi> I'd be needing {iau} everywhere 06:11 < zipcpi> .i le remei cu tugni zu'ai lo du'u, lo traji be lo ka clira fa lo nu ce'u snada lo ka gau ce'u le litru co'u dasni le kosta, cu ei jai se pajni fai lo ka traji lo ka vlipa <-- so much nesting 06:13 < Amnadya> traji lo ka broda could be condensed 06:14 < zipcpi> How? tanru? relyle'ujvo? 06:14 < Amnadya> brodyrai 06:14 < zipcpi> lol 06:14 < Amnadya> What? :O 06:15 < zipcpi> :p 06:15 < ctefaho> ӏИ шоы ро до ӏИ до мо 06:15 < niftg> шоы 06:15 < gleki> ми лифри ло ка банзу забна. и ма нузба дой ла штефа'о 06:15 < zipcpi> I guess {vlipyrai} isn't too bad though 06:15 < gleki> и зо шой драни 06:15 < zipcpi> Not as if that lujvo could mean anything else 06:16 < gleki> и зо'ой шои жи'а драни 06:16 < Amnadya> It's a completely normal CLL lujvo. 06:16 < gleki> interesting that under this lerfu/namcu split {li abu mo} becomes ungrammatical 06:17 < gleki> it could theoretically be parsed as {li lo'o abu mo} 06:17 < ctefaho> (in other news I now hate Microsoft's keyboard layouts too) 06:18 < gleki> this of course makes sense since {li} accepts only mex 06:18 < zipcpi> exp: li abu mo 06:18 < mensi> ([li {<a bu> BOI} LOhO] [CU {mo VAU}]) 06:18 < niftg> mi darsi punji po'o lo fukpi be lo lerpoi poi cy ke'a ciska 06:18 * ctefaho brb 06:18 < zipcpi> No, letters can stand in for numbers. All the split does, is that "letter strings" can no longer contain PA 06:19 < zipcpi> off: li pa ry 06:19 < mensi> ([li {<pa ry> BOI} LOhO] VAU) 06:19 < gleki> the problem is that complicates other parts of the grammar 06:19 < zipcpi> exp: li pa ry 06:19 < mensi> ([{li <pa BOI> LOhO} {ry BOI}] VAU) 06:19 < gleki> i now implemented {vei ny} and {pa} as numbers 06:20 < zipcpi> exp: pa ry 06:20 < zipcpi> off: pa ry 06:20 < mensi> ([{pa BOI} {ry BOI}] VAU) 06:20 < mensi> ([pa ry] BOI) 06:20 < zipcpi> Oh crap 06:20 < gleki> and {ny} and {pabu } as pronouns 06:20 < zipcpi> That means my fancylojban thing relies on xornunsei 06:21 < gleki> what is gendra now in alta is {li vei abu} 06:21 < zipcpi> I don't see why {li lo'o abu} should be a preferred interpretation over {li abu lo'o} 06:22 < gleki> we can assume that {vei} can be elided after {li} but for that we need just another mex that allows first {vei} be elided. it means creating another mex grammar for {li} only 06:22 < zipcpi> Famyma'o should be delayed as long as possible 06:22 < ctefaho> gleki: idea is ĭ <-> й, ŭ <-> ў, ị<->ы ụ<->ю 06:22 * ctefaho brb2 06:22 < gleki> ctefaho: ы is {y} 06:22 < gleki> ю is {iu} 06:23 < gleki> lerfu fi lo nu prami 06:23 < ctefaho> I know I have overridden them 06:24 < ctefaho> Unless you think some other letters are less bad for ị and ụ ? 06:24 < gleki> the first is definitely й 06:24 < gleki> the second doesnt exist in Russian 06:24 < ctefaho> what is ĭ then? 06:24 < gleki> й 06:24 < ctefaho> I don't use Russian cyrillic I use unicode cyrillic 06:25 < gleki> [w] does exist in Belarussian 06:25 < gleki> ў yes 06:25 < gleki> what is ụ for?[D 06:26 < gleki> i think i dont understand your symbols 06:26 < ctefaho> just [ʊ], ideally 06:26 < ctefaho> or [w] too 06:26 < ctefaho> ị = [ɪ]/[ʏ] ([j]) 06:26 < gleki> and ŭ ? 06:26 < ctefaho> [w] 06:26 < ctefaho> ị as [j] 06:26 < ctefaho> .. 06:27 < ctefaho> ĭ as [j]* 06:27 < gleki> there is only one letter for [w] in Cyrillic scripts 06:27 < gleki> and one for [j] 06:28 < ctefaho> I just didn't see any that would fit ị and ụ well so I just took ы and ю 06:28 < ctefaho> (а́ е́ и́ о́ у́ ә́ for stress marking) 06:29 < gleki> i think not stressed ы is lojban {y} 06:29 < ctefaho> I did cyrillic schwa for "y" 06:29 * ctefaho hides 06:30 * ctefaho also brbx3 06:31 * nuzba @464161niftg: .u'i zo'oi minimalist noi la'a ca bu'u se smuni lo cizra co'a https://twitter.com/oka_iu_tcan/status/623439221840179200 [http://bit.ly/1Lq7ypd] 06:33 < gleki> at least now {li vei} is a workaround, sounds like {li'ei} O_0 06:34 < zipcpi> li vei ry re dy re ? 06:34 < gleki> alta: li vei ry re dy re 06:34 < mensi> (NIhO [I {FA <li (¹[vei {ry <re dy>} re VEhO] BOI¹) LOhO>} {CU <COhE VAU>}]) 06:35 < gleki> otherwise how would you say "Empty string is loved by Bergu"? 06:36 < gleki> alta: li by se nelci 06:36 < mensi> SyntaxError: Expected [,] but "b" found. 06:36 < gleki> i can fix that :) 06:36 < gleki> alta: li lo'o by se nelci 06:36 < mensi> (NIhO [I {<FA (¹li lo'o¹)> <FE (¹by BOI¹)>} {CU <se nelci> VAU}]) 06:36 < zipcpi> I dunno, explicitly say lilo'o? Or use Curtis-zai'e (move it to zai'u or something) :p 06:36 < gleki> en:zai'e 06:36 < mensi> zai'e = [BAhE] jargon word indicator; indicates next word is a jargon word 06:36 < zipcpi> No, that's my zai'e 06:36 < gleki> mensi: aigne 06:36 < mensi> CommonSenseError: Expected normal word but Curtis found. 06:36 < zipcpi> My zai'e won out over his zai'e 06:37 < zipcpi> http://jbovlaste.lojban.org/dict/zai'e 06:37 < zipcpi> But XOhEhOhE is a lol-worthy selma'o 06:38 < zipcpi> And I have no idea how it would work 06:38 < zipcpi> From what I can tell it's supposed to be able to replace both PA and BY 06:45 < gleki> okay what if {li} accepts a lerfu string but lerfu are either PAbu or {ny}? 06:46 < Amnadya> mi na nelci 06:46 < mensi> i mi i mi i mi mo i mi na nelci 06:46 < gleki> alta: li abu repapapabu mo 06:47 < mensi> (NIhO [I {FA <li (¹[a bu] [re pa pa] [pa bu]¹) LOhO>} {CU <mo VAU>}]) 06:47 < gleki> alta: li abu re mo 06:47 < mensi> (NIhO [I {<FA (¹li [a bu] LOhO¹)> <FE (¹re BOI¹) mo KU>} {CU <COhE VAU>}]) 06:47 < zipcpi> My own Eulojban suggestion, *and* also my jo'aucu'i xornunsei suggestion, does not touch LI...LOhO 06:47 < gleki> so basically {bu} is a terminator here 06:47 < gleki> Amnadya: do nelci ma 06:48 < gleki> i xu do djica tu'a lo valsi poi se gerna tu'a zo li'ei 06:48 < gleki> *te gerna 06:48 < gleki> *te gerna fi tu'a zo li'ei 06:48 < zipcpi> But my Eulojban suggestion breaks {bypyfyky cusku} in favor of {leu bypyfyky cusku} / {ko'ai bypyfyky cusku} 06:49 < Amnadya> mi so'o roi ciksi lo se nelci be mi 06:49 < zipcpi> It will interpret bypyfyky cusku as faby fepy fify foky cusku 06:49 < zipcpi> :p 06:49 < zipcpi> pai'e zbusufukai 06:50 < Amnadya> Then there are some 5 proposal so far 06:50 < zipcpi> My way though, it'll make single letterals a completely serviceable stand-in for ko'a-set without needing any {boi}s 06:51 < zipcpi> Except I did not spread my gift to mekso grammar... I don't care about mekso grammar anymore anyway T.T 06:51 < zipcpi> All I want is arbitrary charstrings 06:52 < zipcpi> Heck it's just a matter of time before I delete all VUhU from Eulojban :p 06:54 < ctefaho> gleki: this is what I use http://pastebin.com/RFz7jNw2 06:55 < ctefaho> ...change that Notepad++->Pastebin 06:56 < ctefaho> If you can think of better characters for ị and ụ do tell 06:58 < gleki> Ы is never pronounced that way 06:58 < gleki> it's never a semi-vowel 06:59 < gleki> Amnadya: mi na mulno jimpe lo se nelci be do i do pu xusra lo du'u do djica lo nu cikre lo cumka'i ciste i ku'i aupei lu abu mo li'u co'a se smuni lu vei abu mo li'u 07:00 < ctefaho> gleki: ị and ụ are not semivowels in my speech 07:00 < ctefaho> or shouldn't be 07:00 < gleki> i fau lo nu mi'o na sanji ro lo se stidi vau mi stidi lo nu zbasu lo cnino ke uitki papri 07:01 < Amnadya> Here's a summary poll: http://strawpoll.me/4985847 07:03 < zipcpi> lol no option for xo'eucu'i? 07:03 < Amnadya> What's that? 07:03 < zipcpi> xorxes suggestion, except not toching LI...LOhO 07:04 < Amnadya> That's option 2 07:04 < zipcpi> Option 2 is full xorxes suggestion... 07:04 < Amnadya> Ah 07:05 < Amnadya> Well then there needs to be 2a and 2b 07:05 < zipcpi> Mhm 07:06 < zipcpi> And maybe there should be a "nothing" option as well 07:06 < durka42> Amnadya: okay, I voted :p 07:06 < durka42> isn't 1 the "nothing" option 07:06 < Amnadya> The only options I would currently give my consent to are 1 and 4. 07:06 < Amnadya> Yes 07:07 < zipcpi> lol 5 is too crazy T.T 07:08 < durka42> Amnadya: xu do na cusku fo lo BPFK co'e 07:08 < Amnadya> .i'a 07:08 < zipcpi> 3 is even worse though; I'm not even sure how that would work 07:09 < zipcpi> And I don't know how 4 would work unless we bring in my Eulojban categorical pronoun set, and even then I'd still like something to {goi} to every now and then 07:11 < zipcpi> And the categorical pronoun set is *the* #1 reason I invented Eulojban. Because it takes up so much cmavo space, and they all want to be monosyllabic T.T 07:12 < zipcpi> Can theoretically save two by dropping {feu} and {meu}, but still 07:13 < zipcpi> Hm without 2b I can only vote for 5. Or reluctantly 1, but only to keep LI...LOhO from breaking T.T 07:13 < durka42> xu ca su'o da do na titinmo 07:14 < Amnadya> This poll isn't anything official, just a list of stuff. I'll add 2b on the mriste 07:14 < durka42> oi 07:14 < durka42> xu ca su'o da do na tirtinmo 07:14 < zipcpi> tirtinmo ki'a 07:14 < Amnadya> Are there any other options I forgot? 07:15 < Amnadya> Or maybe you have some last-minute proposals? 07:15 < zipcpi> No... my Eulojban proposal is quite enough lol 07:16 < zipcpi> Wait wut 07:16 < zipcpi> Oh 07:16 < zipcpi> It reordered them 07:16 < zipcpi> Thought it somehow registed my vote for 5 as a vote for 2 07:18 < gleki> i cant vote 07:18 < gleki> my brain already exploded 07:18 < zipcpi> #1 is "nothing" 07:18 < zipcpi> If you'd prefer that 07:19 < gleki> i want {abu mo} retain its current meaning,as for lerfu detri i'd like {li jydy re} parse but if it's so neede then add {bu} to the end 07:19 < gleki> which option is that? 07:19 < gleki> or options 07:19 < zipcpi> #1 and #5 from what I can tell 07:19 < zipcpi> But #5 will break bypyfyky cusku 07:20 < Amnadya> #2 07:20 < zipcpi> No, #2 will break li jydy re 07:20 < zipcpi> Unless 2b 07:20 < Amnadya> He said he'd add bu 07:20 < gleki> no, option 1 is not for me 07:20 < zipcpi> Xorxes proposal will add {bu} to every single nacle'u 07:21 < zipcpi> Gleki: Then I think none of the options are suitable for you, unless maybe 2b is added 07:21 < gleki> when mensi arrives ... 07:21 < Amnadya> li jdydy rebu = #2 07:21 < Amnadya> jydyrebu 07:21 < gleki> alta: li jydy re mo 07:21 < mensi> (NIhO [I {<FA (¹li [jy dy] LOhO¹)> <FE (¹re BOI¹) mo KU>} {CU <COhE VAU>}]) 07:21 < gleki> se? those are two sumti now 07:21 < gleki> *see 07:21 < zipcpi> Amnadya: But he wants {li ny renopamubu} 07:21 < gleki> alta: li jydy rererere bu 07:21 < mensi> (NIhO [I {FA <li (¹jy dy [re re re] [re bu]¹) LOhO>} {CU <COhE VAU>}]) 07:21 < zipcpi> Not {li ny rebunobupabumubu 07:22 < zipcpi> Like what xorxes wants <.< 07:22 < Amnadya> This is starting to get annoying .u'i 07:22 < gleki> i'i 07:22 < gleki> i mi tatpi 07:22 < gleki> alta: li jy re bu no bu 07:22 < mensi> (NIhO [I {FA <li (¹jy [re bu] [no bu]¹) LOhO>} {CU <COhE VAU>}]) 07:22 < gleki> O_0 07:22 < gleki> who wrote altatufa ? 07:23 < zipcpi> Erm, it *will* work 07:23 < zipcpi> It always worked 07:23 < Amnadya> That's CLL 07:23 < gleki> my goal was to make as few rules as possible while retaining most current usage 07:23 < gleki> off: li jy re mo 07:23 < mensi> ([{li <(¹jy re¹) BOI> LOhO} CU] [mo VAU]) 07:23 < gleki> aha. here is the difference 07:23 < gleki> alta: li jy re mo 07:23 < mensi> (NIhO [I {<FA (¹li jy LOhO¹)> <FE (¹re BOI¹) mo KU>} {CU <COhE VAU>}]) 07:24 < gleki> two sumti 07:24 < zipcpi> I don't really like it. It does the least damage to my detri system, but it would make reading arbitrary charstrings difficult 07:24 < gleki> but i can of course make {li} more powerful as in CLL 07:25 < gleki> alta: li ry re ry re bu 07:25 < mensi> (NIhO [I {FA <li (¹ry [re ry] [re bu]¹) LOhO>} {CU <COhE VAU>}]) 07:25 < gleki> you just terminate using {bu} in the very end 07:25 < zipcpi> Right 07:26 < gleki> i dont propose anything. im just trying to shrink PEG 07:26 < Amnadya> alta: li ry aigne bu 07:26 < mensi> (NIhO [I {FA <li (¹ry [aigne bu]¹) LOhO>} {CU <COhE VAU>}]) 07:26 < zipcpi> u'i 07:26 < zipcpi> alta: li ry re dy re aignebu 07:26 < mensi> (NIhO [I {<FA (¹li [ry {re dy}] LOhO¹)> <FE (¹re BOI¹) aignebu KU>} {CU <COhE VAU>}]) 07:27 < zipcpi> Oops 07:27 < zipcpi> alta: li ry re dy re aigne bu 07:27 < mensi> (NIhO [I {FA <li (¹ry [re dy] [re {aigne bu}]¹) LOhO>} {CU <COhE VAU>}]) 07:27 < Amnadya> This is confusing. 07:27 < zipcpi> ie 07:28 < gleki> it only has the problem of gardenpathing 07:29 < gleki> someone starts dictating a password and everybody thinks "Will he put {bu} in the end?" 07:29 < zipcpi> The other to implement 5 is to rework {tei} so that it autocloses. Except I suggested {tei}<>{ne'e} in CKTJ T.T 07:29 < zipcpi> s/other/other method 07:29 < fenki> zipcpi meant to say: The other method to implement 5 is to rework {tei} so that it autocloses. Except I suggested {tei}<>{ne'e} in CKTJ T.T 07:30 < gleki> en: tei 07:30 < mensi> tei = [TEI] composite letteral follows; used for multi-character letterals. |>>> 07:30 < mensi> officialdata 07:30 < gleki> did xorxes want to put {bu} after each digits or only as {parecivomu bu}? 07:30 < zipcpi> After each digit 07:31 < gleki> *digit 07:31 < gleki> oh 07:31 < zipcpi> mabla mi ji'a 07:31 < gleki> with {li} in front? 07:31 < Amnadya> fi do 07:31 < gleki> or only with {me'o}? 07:31 < Amnadya> li = me'o 07:31 < zipcpi> li and me'o are both in LI 07:33 < durka42> zipcpi: lo ka ti'inmo cu ka cinmo la'e zoi zoi T.T zoi .i za'a do ze'e ti'inmo 07:33 < zipcpi> u'i 07:34 < durka42> ji'a mi na djuno lo du'u lo ka ti'inmo cu mo 07:34 < zipcpi> ca'e eutca 07:34 < gleki> 4 total votes 07:35 < Amnadya> 4 votes by tinkerers and the conservative option is winning 07:35 < zipcpi> Selpa'i voted for two I think 07:35 < zipcpi> lol 07:35 < gleki> oh my someone voted for 3 07:36 < gleki> i didnt vote 07:36 < Amnadya> Did I say something wrong? guskant seems upset :( 07:36 < zipcpi> o.o 07:37 < zipcpi> I dunno, I only ever talked to her through the mriste 07:37 < zipcpi> Maybe her zantufa is getting too much hate? *shrugs* 07:38 < gleki> what i agree with is that sumti tail and {sei} clause are described in PEG superugly 07:39 < durka42> where did you post? uanai 07:39 < Amnadya> main list 07:40 < durka42> oh about xoi/soi? 07:40 < durka42> I thought you were posting this straw poll thing 07:40 < Amnadya> yes 07:40 < Amnadya> Not yet. 07:40 < durka42> je'eju'inai 07:40 < zipcpi> durk42: lo ka ti'inmo cu ka ce'u cinmo lo mukti be lo nu lo'e dracyrxanime prenu cu klaku .iku'i na sarcu fa lo nu fatci fa lo nu klaku 07:41 < durka42> je'e 07:44 < gleki> "lerfu = lerfu ( lerfu+ / digit+ )*" wat. okay but what about {pabu}? are they lerfu? 07:44 < zipcpi> Yes 07:44 < Amnadya> go'i 07:44 < gleki> can this resulting lerfu string be used in LI? 07:44 < gleki> a question not answered in the poll 07:44 < zipcpi> ca'eda'oi leu cylyly 07:45 < gleki> zipcpi: your eulojban looks like loglan. beware of copyright infringements! 07:45 < zipcpi> What, loglan has /eu/? 07:46 < gleki> loglan: feu 07:46 < mensi> feu = in fact/actually, a discourse operator (UI). 07:46 < gleki> not necessarily UI, there were rants 07:46 < gleki> ... 07:46 < gleki> :) 07:47 < gleki> unlike in this channel 07:47 < zipcpi> How is it pronounced? [ew]? 07:47 < gleki> mi na cmima la loglo kamni 07:47 < zipcpi> Ah 07:47 < zipcpi> Eulojban {feu} = "she" 07:47 < gleki> to zo cmima malgli toi 07:48 < zipcpi> Meh, it's kinda why I made {setmima} 07:48 < zipcpi> So that there's a word specifically to talk about sets 07:48 < gleki> cmima lo te kamni 07:49 < zipcpi> That won't depend on how {cmima} is used laically 07:49 < gleki> what was that mukti wanted? beheaded bridi? let's see if i can make them 07:49 < durka42> AFAIK <feu> was ambiguous between fe,u or fe'u or fe.u and that's why we have ' now 07:49 < zipcpi> Hm 07:50 < gleki> ' didnt exist in loglan 07:50 < gleki> loglan: hu 07:50 < mensi> hu = who/whom/what, the argument interrogative. E.g., Hu fa godzi? Who will go?; Da merji hu? X is married to whom?; Hu 07:50 < mensi> fa durzo hu? Who will do what? 07:50 < gleki> fehu then? 07:50 < gleki> coi: hu fa godzi i da merji hu i hu fa durzo hu 07:50 < mensi> ma ba klama i ko'a speni ma i ma ba gafygau ma 07:51 < durka42> man, those loglygismu are painfully english-based 07:51 < gleki> iu doi la mensi 07:51 < durka42> loglan: durzo 07:51 < mensi> durzo = K do V to D with B. 07:51 < durka42> za'a na mapti zo gafygau 07:51 < zipcpi> dumza 07:51 < durka42> coi zbusufukai 07:51 < gleki> mapti no lojbo vau la'a 07:52 < durka42> ie 07:53 < gleki> e'u yyyy zo durzo 07:53 < gleki> sei pavmeigau co bangu 07:54 < gleki> i lo gerna ca durzo lo nu lo stedu cu cortu vau mi lo lerfu javni 08:13 < Amnadya> Okay, seems it wasn't my fault. 08:15 < Amnadya> Btw, the xoi distinction I brought up should be of interest to everyone 08:15 < Amnadya> (everyone who uses xoi) 08:16 < durka42> oh wow throwing shade on camxes there 08:16 < Amnadya> (more scope nuances, yay!) 08:16 < durka42> she really is going rogue 08:16 < durka42> ei mi tcidu piro lo mristeka'u 08:16 < Amnadya> She might be having a bad day 08:16 < durka42> happens to the best of us 08:17 < Amnadya> We're in the same boat though, many of us are trying to fix the illogical parts of Lojban 08:18 < Amnadya> .u'i se'i 08:19 < zipcpi> I saw your xoi thing but I can't remember where 08:20 < Amnadya> I just wrote it in the latest mriste post 08:20 < zipcpi> Oh yeah. It was interesting. I've never used xoi that way or seen it used that way but it seems useful 08:21 < durka42> I don't see the problem with guskant's option 1 08:21 < durka42> after reading just the first post 08:22 < zipcpi> Similarly, you can eliminate LE and move it to LOhOI saving even more rules. OI 08:22 < zipcpi> NUNUNUNUNUNU 08:23 < zipcpi> Oh the Pxoi and Nxoi? 08:23 < zipcpi> Sounds a lot like po'oi and no'oi 08:23 < durka42> huh? 08:24 < zipcpi> Except it works somewhat differently 08:24 < zipcpi> Newest selpa'i post 08:24 < zipcpi> NOhOI is even implemented into camxes 08:24 < zipcpi> exp: so'i verba cu krixa po'oi fanza 08:24 < mensi> ([{so'i BOI} verba KU] [cu {krixa <po'oi (¹CU [fanza VAU]¹) KUhOI>} VAU]) 08:26 < durka42> ah I think xoi is Nxoi 08:27 < durka42> otherwise it seems to be scope-escaping 08:27 < durka42> this is a tangent though, heh 08:27 < durka42> after decrying tangents above 08:27 < zipcpi> ... I've been using it like Pxoi more often 08:27 < durka42> really? 08:28 < durka42> maybe I'm not understanding the difference 08:28 < durka42> can we give some more examples maybe 08:28 < zipcpi> {le trene cu sutra xoi pi'ai minli fei'u cacra li munono} 08:28 < durka42> well for that it doesn't matter which it is 08:28 < zipcpi> ... I keep going back to that example 08:28 < durka42> can you translate or expand it 08:28 < Amnadya> They haven othing to do with no'oi and po'oi 08:28 < zipcpi> Even though {pi'ai} is a berserk button for some 08:28 < durka42> yeah I don't know what that was about :) 08:28 < durka42> I'm fine with pi'ai 08:29 < durka42> I want to know what the example means to you 08:29 * durka42 rarely gets a response when asking this kind of question 08:29 < zipcpi> The train is quickly going at 500 miles an hour 08:29 < durka42> ui ie 08:29 < Amnadya> Pxoi is not scope escaping. 08:29 < zipcpi> While NOhOI does 08:29 < zipcpi> Yeah I get it 08:29 < Amnadya> This was @durka 08:30 < Amnadya> It's not a tangent. This thread is about xoi. 08:30 < durka42> so 08:30 < durka42> dude, zipcpi, come on 08:30 < zipcpi> So you perhaps want to change xoi vs soi so that one is Pxoi and one is Nxoi? 08:31 < zipcpi> durka42: What? 08:31 < ldlework> blah blah we need more rap 08:31 < durka42> you started a tangent! 08:31 < Amnadya> {soi} cannot be either. 08:31 < durka42> because "Pxoi" kinda looks like "po'oi" if you squint! 08:31 < durka42> what's the point 08:31 < durka42> anyway 08:31 < zipcpi> ... crap 08:31 < durka42> I was going to try to expand {so'i verba cu krixa xoi fanza} 08:31 < Amnadya> Expanding them reveals their difference yes 08:32 < zipcpi> I told you, I'm the Avatar of Chaos 08:32 < durka42> so'i da poi verba zo'u da krixa xoi fanza 08:32 < zipcpi> mi mutmaika'i lo si'o kalsa 08:32 < durka42> actually that's the wrong direction to expand 08:33 < durka42> lo su'u so'i verba cu fanza cu krixa 08:33 < durka42> which xoi am I using there? 08:33 < durka42> whichever one it is, that's my interpretation of xoi :) 08:33 < durka42> oops, I also reversed the selbri 08:33 < durka42> lo su'u so'i verba cu krixa cu fanza 08:33 < Amnadya> Hold your horses. 08:33 < durka42> je'e 08:34 < zipcpi> Sounds like Pxoi 08:34 < Amnadya> First of all, {xoi} is a normal bridi operator (Pxoi or Nxoi doesn't matter here), so {so'i} being to the left has scope over {xoi}. 08:35 < durka42> hmm true 08:35 < Amnadya> Your first one with {zo'u} was correct 08:35 < Amnadya> Now you just have to remove xoi 08:35 < durka42> so'i da poi verba zo'u da krixa xoi fanza 08:35 < Amnadya> ie 08:35 < durka42> so'i da poi verba zo'u lo nu da krixa cu fanza 08:36 < Amnadya> so'i da poi verba zo'u lo nu da krixa cu fasnu gi'e fanza (fasnu to be safe) 08:36 < durka42> heh ok 08:36 < durka42> or stick a da'inai somewhere 08:37 < Amnadya> It's important to retain a {je} (or gi'e) somewhere 08:37 < Amnadya> in the expansion for Pxoi 08:38 < Amnadya> Because that's the difference to Nxoi, which has no {je} 08:38 < Amnadya> (It's exactly as with poi vs noi) 08:39 < durka42> well yeah 08:39 < durka42> so I get that difference 08:39 < durka42> which is about selecting vs annotating reference 08:39 < durka42> but with {xoi} the referents are... instances of the bridi? 08:39 < durka42> s/reference/referts 08:39 < fenki> durka42 meant to say: which is about selecting vs annotating referts 08:39 < durka42> s/referts/referents 08:39 < fenki> durka42 meant to say: which is about selecting vs annotating referents 08:40 < durka42> malylercu'aca'a 08:40 < Amnadya> With {je}, {poi} and {Pxoi} you make a claim about the conjunction only 08:40 < zipcpi> cu'aca'acu'aca'acu'aca'a 08:40 < durka42> so the expansion for Nxoi would be the same except gi'e => gi'u? 08:41 < Amnadya> I would use {.i} 08:41 < zipcpi> .i lau'u fanza 08:41 < durka42> so'i da poi verba zo'u lo nu da krixa cu fasnu .ibo xo'e lo nu go'i cu fanza 08:42 < durka42> er 08:42 < Amnadya> zo xo'e ua nai 08:42 < durka42> xo'e lo go'i * 08:42 < durka42> well I'm not sure how you are going to un-restrict it 08:42 < durka42> so that it doesn't refer to the entire bridi 08:42 < durka42> also, I have an inkling that the parallel is backwards and Pxoi corresponds to noi 08:43 < durka42> but I could be wrong 08:43 < Amnadya> It doesn't. 08:43 < zipcpi> zo lau'u cu sance simsa lo jugbau valsi be lo si'o tigra 08:43 < durka42> the way I understood the original example, {xoi} does not affect the set of children we are talking about 08:43 < durka42> it just notes that the situation is annoying 08:43 < zipcpi> lau'u ba citka do 08:43 < durka42> but I'm beginning to think that I was thinking about {soi} 08:44 < durka42> yeah I was definitely thinking about {soi} 08:44 * durka42 re-evaluates the example 08:44 < Amnadya> It doesn't attach to the children. {xoi} is a *bridi* relative clause. It restricts the "predicate" that the children have to satisfy 08:44 < Amnadya> They satisfy the *conjunction* 08:44 < durka42> right 08:44 < zipcpi> So {soi} is {sei}+ke'a+bridi tail? 08:44 < Amnadya> (Pxoi I mean) 08:45 < durka42> zipcpi: in semantics, yes 08:45 < durka42> Amnadya: aha you betrayed your vote :p 08:45 < zipcpi> Huh? how? :p 08:45 < durka42> Amnadya: okay I have vacillated over to liking (E)/Pxoi better 08:45 < Amnadya> But both are needed anyway. 08:46 < durka42> Nxoi seems more like {soi} to me 08:46 < durka42> unfortunately there are no other Coi left, are there 08:46 < durka42> except qoi 08:46 < Amnadya> {Nxoi} differs from {soi} in that {Nxoi} is a bridi operator. The {so'i} has scope over it, but wouldn't have scope over a {soi} 08:46 < Amnadya> Big big difference 08:47 < durka42> I didn't say it was the same 08:47 < durka42> I said it seems more like {soi} 08:47 < durka42> and for maximum utility they should be maximally different :) 08:47 < zipcpi> simlu .i kinda sorta 08:48 < zipcpi> me'ei kinda cu me'ei simlu ja milxe 08:48 < zipcpi> me'ei sorta ku ji'a go'i 08:49 * Amnadya afk 08:49 < ldlework> Has anyone here read GEB? 08:51 < durka42> mi nai 08:52 < durka42> jbo: xoi 08:52 < mensi> 9 da se tolcri: ca'emsku, cutci, fuzma'u, guglypevgri, guglyzamgri, indika, kralyma'u, xluzerma'esazri, 08:52 < mensi> ze'irsaivi 08:52 < zipcpi> lol dang 08:53 < zipcpi> {ca'emsku}? 08:53 < zipcpi> The zi'evla hyphens are so random 08:53 < durka42> it's a hack to avoid making a lujvo since ra'oi ca'e rafsi lo na se djuno be mi 08:53 < zipcpi> catke 08:53 < durka42> I mildly disapprove of those because they violate the principle of least surprise 08:53 < zipcpi> li'a mi rolrafcre 08:54 < zipcpi> m can never be a lujvo hyphen 08:54 < zipcpi> It's n that might be confusing 08:55 < durka42> m is slightly less confusing than n 08:55 < durka42> very very slightly 08:55 < durka42> consider noisy channels 08:57 < zipcpi> Apparently I use {xoi} a lot in definitions 08:57 < zipcpi> jbo: ni'ai 08:57 < mensi> cu'alvu'u = x1 cu'alni le ni'ai ke'a vujnu x2 boi x3 08:58 < zipcpi> I changed it to {le vujnu be x2 bei x3} 09:00 < zipcpi> http://jbovlaste.lojban.org/dict/rolrafcre 09:01 < durka42> en:rolrafcre 09:01 < mensi> [< ro rafsi certu ≈ Each of affix expert] lo nu mulno sisku zo'u: y no da se tolcri 09:01 < mensi> rolrafcre[8827], rolrafcertu[10886], rolrafsycre[10957], rolrafsycertu[13016] 09:01 < zipcpi> mi puru'e jmina 09:02 < durka42> je'e za'a 09:03 < durka42> vlaste: XOI 09:03 < vlaste> 4 results: xoi, gi'ei, i'au, adverbi 09:03 < durka42> en: /full xoi 09:03 < mensi> 7 da se tolcri: xoi, olgoixorxoi, xoi'u, adverbi, banxu'o'i, gi'ei, i'au 09:03 < durka42> lol wat 09:04 < durka42> en: olgoixorxoi 09:04 < mensi> olgoixorxoi = x1 is an olgoi-khorkhoi/Mongolian death worm of type/variety x2 |>>> 09:04 < mensi> krtisfranks 09:04 < durka42> new favorite word 09:04 < zipcpi> lol 09:07 < zipcpi> Don't like zi'evla hyphens? {camze} :p 09:07 < zipcpi> Though we can't do that to everything 09:07 < zipcpi> It will clash 09:08 < zipcpi> Like zomza clashed with zomzi 09:08 < zipcpi> caqne 09:08 < zipcpi> X3 09:09 < zipcpi> Got a problem? Use more phonemes. Still got a problem? MORE PHONEMES 09:09 < zipcpi> And soon you'll get Ithkuil 09:10 < durka42> ithkuil: caqne 09:10 < ithkuil> caqne: P1S1/STA-c-OBL-AGG/COA/A/GRA/RPV-RPS-IFL/UNFRAMED 09:12 < zipcpi> ithkuil: eulojban 09:12 < ithkuil> eulojban: P2S2/DYN-l-ERG-UNI/ASO/U/GRA/NRM-AGC_1/5-EXS-IFL/UNFRAMED 09:12 < durka42> ithkuil: eulojbaq 09:12 < ithkuil> eulojbaq: P2S2/DYN-l-ERG-UNI/ASO/U/GRA/NRM-NTR_1/5-EXS-IFL/UNFRAMED 09:12 < durka42> ah changed from AGC to NTR naturally 09:12 < zipcpi> ithkuil: eulwujbaq 09:12 < ithkuil> eulwujbaq: PPT2-P1S3/STA-CNG-jb-OBL-UNI/COA/U/DEL/NRM-EXS-IFL/UNFRAMED 09:13 < zipcpi> Now the root changed 09:13 < zipcpi> ithkuil: qeunxrwnmrwinmlwu 09:13 < ithkuil> error 09:14 < zipcpi> uesai da na'e valsi fo la itku'ile 09:14 < zipcpi> .y. fi 09:15 < zipcpi> ithkuil: qeunxrönmröinmlöu 09:15 < zipcpi> ... 09:16 < zipcpi> ithkuil: eulojbaň 09:16 < ithkuil> eulojbaň: P2S2/DYN-l-ERG-UNI/ASO/U/GRA/NRM-COO_1/5-EXS-IFL/UNFRAMED 09:17 < zipcpi> ithkuil: ňeunxrönmröinmlöu 09:17 < ithkuil> error 09:17 < gleki> alta: vei pa mo 09:17 < mensi> (NIhO [I {FA <(¹vei pa¹) BOI> mo KU} {CU <COhE VAU>}]) 09:18 < gleki> olgoixorxoi! 09:18 < gleki> ldlework: is there anything lojbanic there? 09:19 < gleki> alta: li ny pa mo 09:19 < mensi> (NIhO [I {FA <li (¹[{ny VEhO} pa] BOI¹) LOhO>} {CU <mo VAU>}]) 09:19 < gleki> alta: li pa ny mo 09:19 < mensi> (NIhO [I {FA <li (¹[pa {ny VEhO}] BOI¹) LOhO>} {CU <mo VAU>}]) 09:23 < zipcpi> Durka42 have you seen this? http://mw.lojban.org/papri/fancylojban_challenges 09:24 < zipcpi> Oops I replaced the wrong link 09:25 < Ilmen> "… but we'd probably need a new brivla for while-loop" 09:25 < zipcpi> Hi Ilmen 09:25 < Ilmen> zvailu 09:25 < zipcpi> lol 09:26 < zipcpi> But: Question: Can Lojban, as designed, actually handle this? Can KOhA and letteral variables be reinterpreted in an existing du'u-proposition, or does the proposition remain the same even if the variables get goi-ed to something else? 09:26 < zipcpi> en: ra'o 09:26 < mensi> ra'o = [RAhO] flag GOhA to indicate pro-assignment context updating for all pro-assigns in referenced bridi. |>>> 09:26 < mensi> officialdata 09:26 < Ilmen> That's a question I've wonder myself. 09:26 < Ilmen> I'm not sure. 09:27 < Ilmen> In the very worst case, it could be given lu-quotes, which would be interpreted at each iteration. 09:27 < zipcpi> lol 09:28 < Ilmen> ta'o coi 09:28 < Ilmen> jbo: zvailu 09:28 < mensi> zvailu = go x1 jetnu gi jetnu fa x2 gi'e ba bo zvailu fa x1 fe x2 |>>> mukti fa lo nu djica lo nu da jbobau brivo la'o 09:28 < mensi> gy. while loop .gy .i srana fa zo .ifle .e zo rapli .e zo krefu |>>> Ilmen 09:29 < Ilmen> za'a xrekursiva velski 09:29 < Ilmen> dai 09:29 < Ilmen> s/za'a/za'adai 09:29 < fenki> Ilmen meant to say: za'adai xrekursiva velski 09:31 < Ilmen> Also, if you haven't read it already, this page may interest you: http://mw.lojban.org/papri/Writing_algorithms_in_Lojban 09:31 < Ilmen> You may want having a look at my second translation attempt, shown on this page. 09:32 < Ilmen> .e'i mi de'a jundi 09:32 < Ilmen> co'o 09:32 < zipcpi> Hm {gi} 09:34 <@xalbo> Does Lojban have a good way of returning tuples? 09:34 < mensi> xalbo: cu'u la'o gy.zipcpi.gy.: Nah if I added another vowel to the Eulojban class it would probably be /�/ (assigned to 09:34 < mensi> w) instead of /�/. {la .eulwjbaq.}. mi ja'a kusru lo'e se glibau | 2015-07-18T04:42:54. 09:34 < mensi> 616Z 09:34 < mensi> xalbo: cu'u la'o gy.zipcpi.gy.: And hey why not make /wu/ a valid diphthong while we're at it (it sounds like an 09:34 < mensi> exaggerated British "oh!") | 2015-07-18T04:46:19.853Z 09:35 < zipcpi> off: li pa ce'o re ce'o ci 09:35 < mensi> ([li {<pa BOI> <ce'o (¹re BOI¹)> <ce'o (¹ci BOI¹)>} LOhO] VAU) 09:35 < zipcpi> ? dunno 09:36 < zipcpi> And yeah I went ahead and added {w} and a bunch of cmavo of... somewhat limited utiliy; mostly just reassignments and bridi-tail hate :p 09:36 <@xalbo> I'm thinking that the proposed form for "while" there doesn't work, for the reasons you stated. But I feel like there's a functional version that would work. 09:38 <@xalbo> Take x3 and feed it into x2. Feed the result into x1. if the result of that is true, then you're done. Else do x1(x2(x2(x3)), etc. 09:38 <@xalbo> I don't know. When I think in Lojban I tend to think in functional/recursive paradigms anyway. 09:38 < zipcpi> Is there a way to define a while loop recursively 09:39 <@xalbo> I'm not sure there's a reason to, really. 09:39 < zipcpi> Well, one good reason might be what I said about "informal algorithms" 09:39 < zipcpi> I believe while-loops are easier on human parsing for informal algorithms than recursive functions 09:41 <@xalbo> That makes some sense. 09:41 < zipcpi> Of course, for the ultimate in recursion... there is the Towers of Hanoi solution algorithm 09:41 < zipcpi> Well not really ultimate 09:41 < zipcpi> But each one branches twice 09:44 <@xalbo> x1 is the sequence of moves that move the right discs from tower x2 (sequence of ring sizes) to tower x3 (sequence of ring sizes) using temporary storage tower x3 (sequence of ring sizes). 09:44 < zipcpi> lol 09:47 <@xalbo> Great, I don't even know how to get "the empty sequence" as a sumti to check the base case. 09:47 < zipcpi> Hahaha 09:48 < latro`a_> can you get the empty set? 09:48 < zipcpi> Curtis has {zai'o}, though I'm not sure how you disambiguate it 09:48 < latro`a_> .i ca'e no da cmima ko'a .i lo porsi be fi ko'a cu ... 09:48 < latro`a_> {no da cmima ko'a} defines {ko'a} (provided we've agreed on extensionality of sets) 09:49 < latro`a_> (and provided da can't get screwed over by universe of discourse adjustments) 09:49 < zipcpi> lo setmima be noda 09:49 < zipcpi> Right 09:49 <@xalbo> fo'a ko'a ko'e ko'i ca'ei ga fo'a ko'a vu'i zai'o du gi ... 09:51 <@xalbo> I need to define more primitives first. 09:51 < zipcpi> Mm... 09:52 <@xalbo> Or not... 10:05 <@xalbo> http://pastebin.com/mR37Pup8 10:05 <@xalbo> The actual move is a little bit of a cheat, and I'm not sure of the types, or that I didn't get some of the arguments backwards. 10:05 < zipcpi> ce'ai, not ca'ei 10:07 < zipcpi> Once you get that fixed feel free to add to the fancylojban challenges page 10:11 <@xalbo> Thanks! 10:12 < durka42> hmm I should add towers of hanoi as an example in my lisp dialect 10:12 < durka42> currently I only have factorial and fibonacci examples 10:12 < durka42> like 6 variants each 10:15 < durka42> la .camxes. ta'e co'u kibjorne ca lo cabdei Nxoi fanza mi 10:15 < durka42> crap, they are both necessary aren't they 10:15 < durka42> maybe we need a UI 10:15 < durka42> then we can nix noi and ne and no'oi as well! 10:15 < durka42> compositionality FTW 10:16 < zipcpi> lol 10:16 < durka42> no seriously 10:16 < zipcpi> And {neu} 10:16 < durka42> well not serious about removing {noi} 10:16 < zipcpi> Maybe it could be {neu} :p 10:16 < durka42> but a UI for restrictiveness would solve the Pxoi/Nxoi problem 10:17 < durka42> and then noi gets defined in terms of {poi ui'i'i} 10:17 < zipcpi> Or {ueu} :p 10:17 < durka42> or the other way around 10:17 < durka42> I'm talking about a real lojban cmavo! 10:17 < durka42> :p 10:17 < zipcpi> lol 10:17 <@xalbo> I should write the problem more clearly, but I'm lazy. 10:17 < durka42> Eulojbaq is for unrealistic if-we-didn't-have-usage-history proposals :) 10:17 <@xalbo> And that's not the interesting part :) 10:20 < zipcpi> {ta'o}? 10:20 < zipcpi> {poi ta'o} = {noi}? 10:21 < zipcpi> I dunno. {noi} is a bit like parenthetical statements 10:21 < zipcpi> But might not be exactly the same 10:23 < ldlework> Its a penultimate assertion 10:24 < ldlework> err that's the wrong word 10:24 < ldlework> Its a 'secondary' assertion. 10:25 < zipcpi> Sounds a bit like {ta'o} but I might be wrong 10:26 < ldlework> ta'o is a metalinguistic force indicator 10:26 < ldlework> noi results in a conjunction 10:27 < ldlework> mi prami lo bruna noi gleki => da zo'u da bruna je se prami be mi .i je da gleki 10:29 < gleki> So for mekso I reimplemented powerful {li} 10:29 < gleki> now im reimplementing sumti_tail, sei-clause as modifications of bridi 10:30 < gleki> alta: lo ge camo gi mo cu mo 10:30 < mensi> SyntaxError: Expected [yY] but "o" found. 10:30 < gleki> O_0 10:31 < durka42> yeah noi is basically another sentence 10:31 < durka42> so maybe UI isn't powerful enough 10:32 < durka42> on the other hand po'o is in UI so 10:32 < zipcpi> I'm still not exactly sure how {ta'o} would be wrong, or what else {ta'o} might mean in that context 10:33 < ldlework> durka42: hehe 10:33 < ldlework> zipcpi: its an operator on the discourse 10:33 < durka42> zipcpi: yeah you might be right 10:33 < durka42> we would just want to specify it 10:33 < durka42> well 10:33 < durka42> I dunno 10:34 < ldlework> Its appearance is of no logical consequence 10:34 < durka42> would it make sense to be able to restrict the domain of referents while simultaneous claiming that the restriction is unimportant 10:34 < ldlework> Its like introducing a "xu" into a sentence 10:34 < ldlework> What's the important of a restriction if not to restrict the domain of referents? 10:34 < durka42> so'i verba poi krixa cu fanza 10:34 < durka42> so'i verba poi ta'o krixa cu fanza 10:34 < durka42> ldlework: right that's my thinking 10:35 < durka42> is there any? 10:35 < zipcpi> THe thing is {poi} is defined as "restrictive", vs {noi} 10:35 < durka42> right 10:35 < ldlework> Well it depends on whether or not you're trying to prove the truth-value of descriptions along with the surface proposition 10:36 < ldlework> Whether or not "the restriction is important" is the difference of whether or not the description appears in a logical expansion 10:36 < durka42> actually I do think {ta'o} has a slightly different meaning 10:36 < durka42> because {ta'o nai} is "returning to main point" 10:36 < durka42> so I think this UI would be different 10:36 < zipcpi> Hm 10:36 < gleki> alta: lo ge camo gi mo cu mo 10:36 < mensi> (NIhO [I {FA <lo (¹ge [ca mo] gi mo¹) KU>} {cu <mo VAU>}]) 10:36 < gleki> alta: sei ge mo gi mo 10:36 < mensi> (NIhO [I {sei <CU (¹ge mo gi mo¹)> SEhU}] [CU {COhE VAU}]) 10:36 < ldlework> broda ko'a noi brode => da zo'u da se broda 10:37 < zipcpi> en: noi 10:37 < mensi> noi = [NOI] non-restrictive relative clause; attaches subordinate bridi with incidental information. |>>> 10:37 < mensi> officialdata 10:37 < ldlework> broda ko'a poi brode => da zo'u da se broda je brode 10:37 < zipcpi> "incidental" / "incidentally" 10:37 < durka42> so'i verba poi ta'o krixa cu fanza .i ta'onai mi xagji 10:37 < durka42> it's more like {fu'e sei na'e te tavla} :p 10:37 < durka42> vlaste: incidentally 10:37 < vlaste> 10 results: ra'unai, ne, noi, no'u, ra'u, ra'ucu'i, noi'a, no'oi, jibri, liste 10:37 < zipcpi> ma'oi UI kalsa 10:38 < durka42> ie 10:38 < durka42> ra'u? 10:38 < zipcpi> ra'unai? o.o 10:38 < gleki> oi'oi = oh-my-god relative clause ; attaches subordinate bridi with ACCIDENTAL information 10:38 < durka42> lol 10:38 < zipcpi> ui'au'ei'oi'oi = Complaint about the disordered state of attitudinals 10:39 < gleki> hey, havent i fixed them 10:39 < zipcpi> lol 10:39 < zipcpi> Well, "we", but not completely, there's still evidentials 10:39 < zipcpi> And {toi'e} 10:39 < gleki> although ithink that for new UI (interjection roots) we need a new name not to confuse with the previous UI. I think OI is a fine name for the new selmaho. 10:40 < zipcpi> lol 10:40 < zipcpi> u'isai 10:41 < durka42> UI42ab✗* 10:41 < gleki> how do you want to organize UI2 then if those are evidentials you mentioned? 10:41 < zipcpi> Hm I'm not sure. The main problem with UI2 is whether they should affect UI1... probably not unless modified with {toi'e} 10:42 < durka42> ie 10:43 < gleki> they shouldnt imo 10:45 < ldlework> What do you mean affect? 10:45 < gleki> how can one feel {ui toi'e ba'a}? 10:45 < durka42> whether {ie ca'e} implies {ie toi'e ca'e} or not 10:46 < durka42> ui toi'e ba'a = mi kanpe lo nu mi gleki ? 10:46 < zipcpi> By the way I use {uo} for "Finally" / "At last" 10:47 < zipcpi> zabna ji mabla 10:47 < gleki> {mi kanpe lo nu mi pu gleki} or {mi kanpe lo nu mi ba gleki} ? 10:47 < gleki> jb: uo 10:47 < mensi> uo = uo [interjection] — Voila! (completion), uo nai — incompleteness 10:47 < mensi> :uo mi facki lo danfu — Voila, I found a solution! 10:47 < mensi> :uo nai do tu na ru'e darxi — Oh, you almost hit it. 10:47 < mensi> :Related words: mulno, mansa, fanmo, snada 10:47 < durka42> ya that's what uo means 10:47 < durka42> gleki: or ca. not specified 10:47 < gleki> "finally", is more like {za'oku broda uo} 10:48 < gleki> durka42: it's a drawback. fix that by adding tenses 10:48 < durka42> mi kanpe lo nu mi gleki .i ku'i mi cinmo lo ka badri .i ma nabmi mi 10:48 < durka42> no just use {sei} and cinmo selbri in that case 10:49 < ldlework> (zipcpi any progress on Speech Acts? (I finished it again)) 10:50 < zipcpi> Sigh... I really need to catch up on some things... 10:51 < ldlework> In the edition of GEB I have, there is a long preface by the author 10:51 < ldlework> In it, he goes on for quite a while regarding his fascination with translation 10:51 < ldlework> Such as the difficulty in translating GEB to french and the many plays on English words and the difficulty there 10:52 < ldlework> He mentions that he wrote a book on the facinating realm of translation and I had the thought that it would be a very nice gift for selpa'i 10:52 < gleki> Actually previous generation Lojbanists read it 10:52 < ldlework> ua 10:53 < ldlework> gleki: do you recall any impressions? 10:53 < gleki> http://mw.lojban.org/papri/G%C3%B6del,_Escher,_Bach 10:53 < gleki> http://mw.lojban.org/papri/G%C3%B6del_Numbers_and_Lojban 10:54 < ldlework> Oh I thought you meant the book he wrote about translation 11:00 < zipcpi> Ugh I think I'll define "illocutionary", "perlocutionary", and "locutionary" all as frame LOJBO 11:00 < zipcpi> x1 pertains to "concept" in aspect x2 11:00 < zipcpi> It's the vaguest way to define a word, but sometimes that's the only awy 11:05 < durka42> ldlework: so who's going to translate GEB to lojban :) 11:07 < gleki> copyright 11:08 < durka42> ight 11:19 < ldlework> zipcpi: what 11:21 < zipcpi> Well, that's kinda how most adjectives with a complex meaning are defined 11:21 < ldlework> But none of those words have anything to do with defining concepts 11:21 < zipcpi> Like cultural "adjectives" 11:21 < zipcpi> They are "adjectives" that "pertain" to a certain concept in some way 11:22 < zipcpi> I just want to make tanru out of them 11:22 < ldlework> zipcpi: I thought you were suggesting a definition for one of those words or something 11:22 < zipcpi> Oh 11:22 < zipcpi> I meant the definition of the Lojban equivalents 11:23 < ldlework> Would never be something like "x1 pertains to "conept" 11:23 < zipcpi> It's the simplest way to convert a complex adjective into a Lojban word. We can think of what lujvo to make with it later 11:24 < ldlework> zipcpi: I mean, "x1 pertains to "concept" in aspect x2" is not a good definition for "illocutionary" "perlocutionary" or "locutionary" 11:24 < ldlework> If you're saying something else, sorry for misunderstanding. 11:24 < zipcpi> And what else could it be? If we have to describe every single actor within those concepts, we end up with a kurtynomvla 11:25 < ldlework> zipcpi: Those words mean specific discrete things. They are not so general as to require such a floaty unsubstantial definition. 11:25 < ldlework> They are technical terms. 11:26 < zipcpi> Yes, but they relate so many concepts that it's best to explain them in terms of a bridi with {ka} 11:26 < zipcpi> Not having terbri for every single actor 11:26 < durka42> xu do ralte lo mupli 11:26 < ldlework> I don't understand why you think lojban definition for "illocutionary" is vastly complex 11:26 < ldlework> or has many "actors" 11:26 < durka42> za'a da'asnu lo mucti .u'i 11:26 < ldlework> What do you think that illocutionary means? 11:27 < zipcpi> I'm just going by the dict definition 11:27 < zipcpi> ""pertaining to a linguistic act performed by a speaker in producing an utterance, as suggesting, warning, promising, or requesting." 11:27 < ldlework> And so why do you think that definition implies many argument places? 11:27 < ldlework> x1 is the illocutionary force intended by x2 in communicating utterance x3 11:27 < zipcpi> Because we already have a four-place brivla, {skuselzu'e} 11:28 < durka42> vlaste: skuselzu'e 11:28 < vlaste> skuselzu'e = x1 is a speech act, performed by speaker x2 to listener/audience x3, for purpose x4 11:28 < ldlework> That is a bad definition 11:29 < ldlework> The x4 is the speech act 11:29 < zipcpi> <.< 11:29 < ldlework> ca lo nu pilno zo xu ku ma do illocutionary 11:30 < ldlework> And the listener shouldn't be a part of it 11:30 < ldlework> Superflous information 11:30 < rlpowell> lojban.org is going up ad down because testing a backup fix. 11:30 < zipcpi> It's informed by the root words 11:30 < ldlework> And doesn't even include the actual utterance, wtf 11:30 < ldlework> zipcpi: makes sense 11:30 < zipcpi> jvojva and all... I don't know how to define else to define it 11:30 < ldlework> indeed 11:30 < ldlework> what? 11:31 < zipcpi> Nevermind 11:31 < durka42> je'e camgusmis 11:31 < zipcpi> It doesn't need to, because the utterence can be {lo nu cusku X} 11:32 < zipcpi> It's within the x1 11:34 <@xalbo> If -sku- is already in there, then making someone get to {lo nu cusku} to fill the cusku2 seems pointlessly complicated. 11:34 < zipcpi> Where else to put the terbri though? 11:35 < zipcpi> That's the problem with lujvo... sometimes we can't include all the terbri we want, or we end up with extremely complicated place structures T.T 11:36 < rlpowell> gleki: You awake? 11:36 < rlpowell> Can people see if the site is super slow right now? 11:37 < durka42> I mean, it's working... 11:37 < zipcpi> Maybe between x2 and x3 11:37 <@xalbo> {lo nu ce'u ce'u ce'u ce'u cusku cu se zukte ce'u ce'u} seems like the right basis, once you align the right {ce'u}s with each other and eliminate the ones that don't matter. 11:37 < zipcpi> So that x2 ~ x4 are all cusku1 cusku2 cusku3 11:37 < ldlework> There doesn't have to be a problem with lujvo 11:38 < zipcpi> $x_1$ is a speech act, characterized by speaker $x_2$ uttering text $x_3$ to listener/audience $x_4$, for purpose $x_5$ 11:38 < ldlework> I just don't see how x1 and x4 are different 11:38 < zipcpi> x1 is the act itself, the x5 is the end goal 11:39 < ldlework> What's the difference? 11:39 < zipcpi> Same difference as the difference between zukte2 and zukte3 11:39 < gleki> rlpowell: here 11:39 < ldlework> zipcpi: zukte2 here is the utterting of text 11:40 < zipcpi> Yes, the "act" itself 11:40 < ldlework> That means that all speech acts are defined by the same act 11:40 < ldlework> the utterance of speech 11:40 < ldlework> which is wrong 11:40 < zipcpi> Then I don't know what is right 11:40 < durka42> oi, those definitions for lotkuti and perlokuti... 11:40 <@xalbo> I probably came in too late; what is the real goal we're attempting to achieve? What's the finti3 for this whole word? 11:41 < ldlework> zukte2 should not manifest in the definition, because zukte2 is implictly filled by cusku'ing text. 11:41 < durka42> you gotta provide examples with definitions that vague 11:41 < zipcpi> I just want to make tanru with them 11:41 < zipcpi> I don't know exactly what they mean T.T 11:41 < ldlework> They are very simple concepts 11:42 < ldlework> What the speaker intends 11:42 < zipcpi> Simple to you... 11:42 <@xalbo> ldlework: I think what zipcpi wants is x1=lo nu x2 cusku x3 x4 cu se zukte x2 x5. 11:42 < ldlework> And what the listener concludes 11:42 < ldlework> That's all it is. 11:42 < ldlework> then you need nuncusku in there somewhere 11:43 < zipcpi> Yes, I *could* make it nunskukezyselzu'e, but I hate that sort of lujvo 11:43 < durka42> zipcpi: then maybe crowdsource the definitions somewhere instead? 11:43 < ldlework> x1 is the illocutionary force (intended cognitive result) of x2 as expressed in text x3 11:43 < durka42> instead of coming here all defeated "gosh I guess I just have to do define them this way" :/ 11:44 < durka42> JVS is a terrible discussion platform if you haven't noticed 11:44 < zipcpi> That one is a separate brivla, ldlework. There's also "illocutionary act" 11:44 <@xalbo> ldlework: Could you provide a concrete example with all three of those places filled? 11:45 < ldlework> ma illocutionery do lo se du'u do prami mi 11:45 < Ilmen> .e'u zo .itlokuti 11:45 < Ilmen> noi ca su'o da mi pilno 11:45 < ldlework> ma .itlokuti fi zo xu 11:45 <@xalbo> {ma} feels like cheating; what would a possible answer be? I'm trying to understand the type. 11:46 < ldlework> You can provide the name of a force like "Question" 11:46 < ldlework> Or you can describe the constitutive rules that allow for question asking 11:46 <@xalbo> No, I don't think I can. 11:46 < durka42> .u'i 11:47 < ldlework> lo nu do spuda cu .itlokuti mi zo xu 11:47 < gleki> i think i will redefine {incita} based on {bridi} definition 11:47 < zipcpi> Well, that's the thing. It's like how {lojbo} magically attaches to {bangu} to describe {jbobau} while having the same structure as {bangu} 11:48 < Ilmen> How about "x1 (relation between x2 and x4) is the illocution (intended result relation) of x2 expressing x3 (text) to addressee x4"? 11:48 < zipcpi> Similarly these words simply import the concept, then we qualify it with another word. If you don't like tanru, we can make specific lujvo later 11:48 < ldlework> Ilmen: I really don't think that the listener is involved. 11:48 < gleki> for "illocution" I suggest (strangely to here that from la gleki) something purely Lojbanic like {mu'umgu},{kreivi}. since you are creating something that isnt specified nice in natlangs. 11:48 < gleki> jbofuvi i mean 11:48 < ldlework> I don't see how illocutionary isn't nicely specified in say, English 11:49 < Ilmen> Do you think the addressee is irrelevant to the utterance purpose? 11:49 < gleki> no, not jbofuvi 11:49 < ldlework> Ilmen: yes 11:49 < ldlework> Its superflous 11:49 < ldlework> And doesn't change our understanding of the intention 11:49 < gleki> ldlework: i mean the terms themselves in their place structure 11:49 < ldlework> Wellllllll 11:49 < Ilmen> Is the speaker relevant to the illocution? 11:49 < ldlework> I guess you can say something, say during a speech, and care to illicit different perlocutionary effect in different subsets of your listeners 11:50 < ldlework> Ilmen: I'm convinced. 11:51 < ldlework> It terms of an adjective I guess you just change it to something like 11:51 < ldlework> x1 is related to the desires of speaker x2 expressing x3 to x4 11:52 < ldlework> without saying what those desires are 11:52 < ldlework> But it seems like, that is kind of the transformation tanru do implicitly 11:53 < ldlework> "to-do-with" or "kind-of" 11:53 < Ilmen> x1 srana me'ei .itloku'i 11:53 < ldlework> So you'd just use the formal definition in a tanru and get a srana sense by default 11:53 < Ilmen> ~ x1 srana lo si'o .itlokuti 11:53 < ldlework> It seems strange to need literal adjectives in lojban 11:53 < ldlework> and adverbs 11:53 < ldlework> since we have a grammar for it 11:54 < ldlework> just verbs and grammar 11:54 < zipcpi> If you really want to front "purpose", we could have {skuterzu'e} as well. But I won't know how to define it. 11:55 < Ilmen> nunsku terzu'e 11:55 < ldlework> zipcpi: its just that other than a name, there really isn't anything that *is* a speech act 11:55 < ldlework> you have questions, assetions, promises and so on, which a words that point to protocols 11:56 < ldlework> s/which a/which are 11:56 < fenki> ldlework meant to say: you have questions, assetions, promises and so on, which are words that point to protocols 11:56 < ldlework> Like, if you want to define the "kind of thing" that is a Touchdown, how do you do that? 11:56 < ldlework> How do you define the thing that is both a Touchdown and an Offsides Violation? 11:56 < zipcpi> That sounds like a job for zai'e 11:57 < ldlework> They are both sets of constitutive rules. 11:57 < ldlework> Beyond designating a name to those sets of rules, they are not substantiated. 11:57 < durka42> you could write the rules down though 11:57 < durka42> in any language 11:57 < ldlework> durka42: right, but what is the verb 11:57 < durka42> if you have enough patience 11:57 < zipcpi> I dunno, one of the things that I kinda sorta intend to do, is to translate the rules of a game 11:57 < ldlework> that denotes the kind of thing 11:57 < ldlework> that is both a Touchdown and an Offsides Violation 11:57 < ldlework> other than x1 is the set of rules that allow for... 11:57 < durka42> er 11:58 < durka42> are there such things? 11:58 < durka42> seems like if you are offsides you can't make a touchdown 11:58 < durka42> but I dunno anything about football 11:58 < ldlework> What... we're talking about classifying those things 11:58 < ldlework> Not how they relate 11:58 < ldlework> Or how they actually word 11:58 < ldlework> work 11:58 < durka42> anyway, I would say "by describing what it is" but I don't get the question 11:58 < ldlework> well that's the answer I was looking for 11:58 < durka42> .u'i 11:58 < ldlework> The only property both of them share, is that they are allowed for by constitutive rules. 11:59 < durka42> yes 11:59 < durka42> a ciste 11:59 < ldlework> The only way to say what a constitutive rule is, is to state the rules that it consists of 11:59 < durka42> {ciste} 11:59 < ldlework> Or provide the name for it 12:00 < ldlework> Right, so now, how would you define the verb "to touchdown" 12:00 <@xalbo> A touchdown is a kind of play (which is itself a kind of event). 12:00 < zipcpi> That one would be defined by the rules of the game itself 12:00 < ldlework> xalbo: the point is, a touchdown is not defined by restrictive rules 12:00 < ldlework> IE, you cannot perform a touchdown outside the context of a football game 12:00 <@xalbo> x1 (play) results in a touchdown for team x2 (with some yadda about the match etc) 12:01 < durka42> "to catch the ball in the end zone during a football game"? (+ whatever other rules of football I'm forgetting) 12:01 < ldlework> This seems to presuppose knowledge of touchdowns 12:01 < zipcpi> Like I said, to define it, one would have to define the rules of the game 12:02 <@xalbo> Or, less of a name, "x1 is an offensive play that ends with team x2 possessing the ball and entering their opponent's end zone, (and isn't a foul, and all the rest of the rules)" 12:02 < ldlework> zipcpi: that's a tautology though 12:02 < zipcpi> It isn't 12:02 < ldlework> since touchdowns are a rule of the game 12:02 < ldlework> if we replace the word touchdown with "rule of the game" 12:02 < ldlework> you have produced a tautology 12:03 < zipcpi> The rules of the game itself define what a touchdown is. If we translated all the rules into Lojban, it would say exactly what the touchdown-brivla means 12:03 < ldlework> the verb "to touchdown" either denotes events which qualify as touchdowns 12:03 < ldlework> or it denotes the rules that constitute a touchdown 12:03 < ldlework> That's the problem 12:03 < Ilmen> X touchdown vs E nu X touchdown 12:04 < durka42> you want x1 is a touchdown under rules x2? 12:04 < ldlework> So either your illocutionary force verb, will either simply denote "things" which qualify as an illocutionary force or it will denote thos eforces 12:04 <@xalbo> "the verb "to touchdown" either denotes events which qualify as touchdowns" -- Yes, I don't see why that's a problem. 12:04 <@xalbo> The verb {gerku} denotes animals that qualify as gerku. 12:04 < ldlework> Right but what constitutes things such as constitutive rules 12:04 < ldlework> IE, things that qualitfy and can be denotes as examples of such 12:05 < ldlework> Touchdown is like Question 12:06 < durka42> so... "x1 is illocutionarily a question"? 12:06 < ldlework> Well we have a word for Question 12:06 < durka42> right 12:06 < ldlework> It denotes something that qualifies under the rules of asking a question, but it does not tell us what a question is 12:06 < durka42> so zipcpi just wants to say {lotkuti preti} 12:06 < durka42> what a Question is... lo ka lotkuti preti cu ka ... 12:07 < ldlework> Right what what is the verb, that denotes the thing that is a question. 12:07 < ldlework> Other than a retrograde denotion of it being a question 12:07 < durka42> you lost me 12:07 < durka42> too many big words 12:07 < ldlework> IE, if, as a young person in lojbanistan, if you want to know what a question is, what it means, how do you perform a question, how do you construct something for which preti is true 12:07 < durka42> pray take pity 12:08 < zipcpi> Like I said, it's like {lojbo}, it is meant to bring in the semantic so that we can define tanru or lujvo based on them, nothing more 12:08 < ldlework> Which verb's definition do you look at? 12:08 < Ilmen> .itlokuti mi lu do ze'u na jai se ganse li'u fa lo ka brireisku lo ka se zvati 12:08 < durka42> preti? 12:08 < ldlework> preti just tells you what things *are* questions 12:08 < zipcpi> But I downvoted myself in case you want to crowdsource some better definitions 12:08 < ldlework> it doesn't teach you how to figure out which things the predicate is true of 12:08 < ldlework> or how to create preti1's yourself 12:08 < durka42> jbo:preti 12:08 < mensi> preti = x1 vlapoi gi'e se cusku x3 fi x4 tezu'e lo nu ri x1 da jungau ku gi'e selsmu lo srana be x2 je nibli be lo du'u 12:08 < mensi> x3 brireisku da x4 12:09 < durka42> actually it does 12:09 < durka42> if you know lojban 12:09 < durka42> whichi s the problem with all dictionaries :p 12:09 < durka42> they are all sepulka 12:09 < ldlework> Okay I didn't realize that the lojban definition actually provided some semblance of how questions are defined in LP 12:09 < zipcpi> ? 12:09 < ldlework> So 12:09 < durka42> we could back translate that to english but it'd be more confusing probably 12:10 < Ilmen> LP ki'a 12:10 < durka42> for us mortals who use the dictionary as a translator 12:10 < ldlework> language-philosophy 12:10 < Ilmen> je'e 12:10 < Ilmen> banpijyske 12:11 < ldlework> is-illocutionary1 can only provide the actual intended force desired 12:11 < niftg> xu nu banlanli 12:11 < durka42> {preti} is recursively defined... 12:11 < ldlework> imo 12:11 < durka42> also that "ri x1 da jungau" seems wrong 12:11 < Ilmen> Where is there recursion? 12:11 < zipcpi> {brireisku}? 12:11 < durka42> Ilmen: breireisku 12:11 < Ilmen> brireisku is not based on preti. 12:12 < durka42> jungau2 is a person but x1 of preti is a question... 12:12 < durka42> oh okay 12:12 < durka42> jbo:brireisku 12:12 < mensi> brireisku [< bridi preti cusku ≈ Bridi* preti* cusku*] = x1 cpedu fi x3 fe lo ka jungau x1 lo du'u x2 brije'u ma 12:12 < mensi> kau 12:12 < durka42> there's the speech act 12:12 < zipcpi> Ah... one of Ilmen's magic untranslatable predicates, naturally :p 12:12 < Ilmen> brireisku = to ask (on of the meanings) 12:13 < ldlework> We should go through and create LP based definitions of the speech-acts 12:13 < ldlework> And encodify all the rules 12:13 < Ilmen> It is for indirect quote asking. 12:13 < ldlework> sounds boring as hell, nevermind 12:14 < Ilmen> He asked "Do you see me?" = reisku lu xu do mi viska li'u 12:14 < zipcpi> The more pertinent question is how to translate the word "force" in "illocutionary force" 12:14 < Ilmen> He asked whether I see him = brireisku lo ka jei mi vo'a viska 12:14 < Ilmen> .i .e'e karbi 12:15 < zipcpi> And whether any tanru based on {itlokuti broda} is enough, or whether we want {itlokuti} to *be* illocutionary force, and build the rest of the concepts from there 12:15 < durka42> zo bapli li'a 12:15 < Ilmen> ask + quote marks ---> reisku; ask + question word ---> brireisku 12:16 < ldlework> What's interesting to me is that preti denotes text uttered with specific illocutionary force 12:16 < ldlework> it denotes an instance 12:16 < ldlework> But if you wanted to answer a question such as "ma itlokuti fi go'e" 12:17 < ldlework> You could never answer with a description denoting _a question_ 12:17 < durka42> lo si'o preti 12:17 < Ilmen> A preti (question sentence) may have another illocution that requesting information, for example if it is a rethorical question. 12:17 < ldlework> That's what I was about to get at is that 12:17 < ldlework> we almost need a way to refer to predicates themselves 12:17 < durka42> the idea of a question/Question ... lo si'o preti 12:17 < ldlework> And this reminds me very deeply of modern discussions of indefinite descriptions 12:17 < zipcpi> Crap... bapli... now that's a problem-word. The problem is that it's about "forcing" someone to do something, which is a semantic that might not be desired when we say "force" 12:17 < latro`a_> what do you mean a way to refer to predicates? we have that already 12:18 < ldlework> If you say so 12:18 < latro`a_> ka is the longhand, si'o is the shorthand 12:18 < latro`a_> unless I'm misunderstanding what you mean 12:18 < ldlework> Really, does that refer to the predicate itself? 12:18 < latro`a_> yes 12:18 < ldlework> In what sense do you mean? 12:18 < Ilmen> You have {me'ei}, too 12:18 < latro`a_> me'ei is another shorthand, yes 12:18 < ldlework> Because "referring to predicates" is a expression that means many things 12:18 < latro`a_> lo ka ce'u nelci ce'u is "the predicate "x1 likes x2"" 12:19 < Ilmen> me'ei citka = the concept/predicate of "X eats Y" 12:19 < durka42> zipcpi: I was joking about zo bapli .u'i 12:19 < durka42> zipcpi: it's much closer to terzu'e, isn't it? 12:19 < ldlework> What verb describes a "lo me'ei broda" 12:19 < latro`a_> me'ei broda cu ka broda 12:19 < Ilmen> me'ei is a LE 12:19 < ldlework> So we have no verb 12:19 < Ilmen> (predicate article) 12:19 < ldlework> Ilmen: ua 12:19 < Ilmen> si'o citka kei 12:19 < ldlework> je'e 12:19 < latro`a_> sure we do; it's an abstractor 12:19 < durka42> si'o 12:20 < zipcpi> durka42: The problem is that I want some predicate that can describe the thing that goes into bapli1 without the "force someone to do something" semantic 12:20 < ldlework> si'o isn't a predicate 12:20 < zipcpi> Like "force of nature" 12:20 < ldlework> NU are not predicates 12:20 < latro`a_> si'o itself isn't a predicate, but it's a predicate builder 12:20 < ldlework> So what is the verb that describes all predicates constructed from abstractions? 12:20 < zipcpi> http://dictionary.reference.com/browse/force?s=t 12:20 < latro`a_> what do you mean exactly? 12:20 < zipcpi> Seems to be #4 12:21 < latro`a_> what is the property that all abstractors have in common? 12:21 < latro`a_> or all predicate abstractors? 12:21 < zipcpi> su'umdu? 12:21 < ldlework> latro`a_: what predicate is true and only true of all predicates that can be constructed from abstractions. 12:21 < zipcpi> I made that word quite recently lol 12:21 < latro`a_> any abstractions or just function abstractions? 12:21 < durka42> any predicate can be constructed from abstractions 12:22 < ldlework> latro`a_: any abstraction 12:22 < durka42> zipcpi: jitro then? 12:22 < Ilmen> kai'u is an abstractor. 12:22 < latro`a_> so {lo nu broda} does the thing you have in mind? 12:22 < ldlework> right 12:22 < zipcpi> xlura ja jitro ja... "affect" 12:22 < latro`a_> that's a word we don't have, as far as I know 12:22 < ldlework> latro`a_: fair enough 12:22 < zipcpi> Problem is that if we put enough jas together we'd want a new word 12:23 < latro`a_> I've never needed it, but I would probably need it to teach a small child the language 12:23 < latro`a_> well, not to get them to communicate 12:23 < ldlework> en: astrato 12:23 < mensi> astrato = x1 is a grammatical abstraction of type x2 |>>> All Lojban constructs can be divided either to astrato/ 12:23 < mensi> naly'astrato or to bridi, fasnu, selsku, namcu, dacti, gunma |>>> gleki 12:23 < Ilmen> {bu'a} is the word you need. 12:23 < zipcpi> Ilmen: lol 12:23 < zipcpi> See jonse 12:23 < latro`a_> but to *teach* it in the sense of what you study in the early school years 12:23 < ldlework> latro`a_: I think I understand you 12:23 < zipcpi> My Lojban definition is the best definition (xo'o) 12:24 < latro`a_> but apparently gleki made the thing 12:24 < zipcpi> http://jbovlaste.lojban.org/dict/jonse 12:24 < ldlework> Okay so do answer the question "ma itloki fi zo xu" you might say "lo si'o preti" ? It doesn't seem to fit. 12:24 < durka42> that is exactly what I would say 12:24 < Ilmen> lo ka brireisku 12:24 < ldlework> Is the concept of questions, the actual intended force? 12:24 < durka42> or lo si'o jetrei 12:25 < zipcpi> Look if you hate my words I'll just add a -kai to them. Ilmen's favorite suffix 12:25 < latro`a_> zo itloki mo 12:25 < durka42> {itloki} sounds like {itku'ile} lol 12:25 < ldlework> But is your intended perlocutionary force in your listener The Concept of Questions? 12:25 < zipcpi> Then {itlokuti} can be defined in a more specific manner 12:25 < Ilmen> "illocution" @ latro`a 12:25 < latro`a_> it's not really "concept" 12:25 < latro`a_> that's a bad english gloss 12:25 < ldlework> Still 12:25 < latro`a_> (si'o doesn't match the ma'oste anymore anyway) 12:25 < ldlework> Even if you say something more advanced 12:25 < ldlework> like 12:25 < durka42> I don't understand the objection sorry 12:25 < latro`a_> (ka does, sorta) 12:25 < durka42> I haven't read Speech Acts 12:26 < ldlework> You defer this way a lot but I don't think you need specialized knowledge to answer the question 12:26 < ldlework> Its a typing question 12:26 < latro`a_> I'd need to know what sort of thing itloki1 wants 12:26 < durka42> ldlework: I didn't understand the words 12:26 < durka42> "perlocutioary force" 12:26 < ldlework> latro`a_: that's what we've been trying to find out 12:26 < durka42> what's wrong with the answer {lo si'o jetrei} 12:26 < Ilmen> I've only read that one -- http://online.sfsu.edu/kbach/spchacts.html Indeed it certainly has nothing to do in length in comparison with the book of the same name. 12:27 < ldlework> durka42: the actual result in your listener, intended or not 12:27 < durka42> if you say {xu} then you'd better intend the listener to hear a question 12:27 < latro`a_> durka42, sure, that's what it *means*, but what is one of those 12:27 < ldlework> durka42: What is the intended cognitive effect you wish to instil in your listener by asking them a question. 12:27 < durka42> I would like them to tell me whether what I asked is true or not 12:27 < ldlework> I don't want an event that describes the effect 12:27 < ldlework> I want a thing that denotes the effect 12:28 < latro`a_> I don't think you two are using the word the same way, then 12:28 < durka42> which word, heh 12:28 < rlpowell> gleki / others: So it turns out that something like 3/4 of our database backup was giant cache tables in mediawiki. 12:28 < rlpowell> That we don't need to backup. 12:28 < rlpowell> So. 12:28 < ldlework> latro`a_: more like, I think it is easy to provide a description of what is meant rather than to produce words that denote them 12:28 < Ilmen> {lo ka ce'u co'a sanji lo nu ce'u brireisku...} 12:28 < durka42> rlpowell: lol, and the backup was running in the background and slowing stuff down? 12:28 < rlpowell> THe nightly backups should have much less impact now; it's tstill a total outage, but it takes about 2 minutes. 12:28 < durka42> ah 12:28 < rlpowell> durka42: Nope; it requires a table lock. So. 12:28 < Ilmen> "the relation of [listener] starting to be aware that [speaker] wants them to inform whether... 12:28 < Ilmen> " 12:28 < Ilmen> pei 12:29 < rlpowell> 2 minutes instead of "WTF is going on I give up", whichi s the previous timeframe. 12:29 < durka42> lo du'u lo se cusku cu bilga lo ka jungau lo cusku ... 12:29 < durka42> rlpowell: I am familiar with that unit of time :D 12:29 < rlpowell> :D 12:29 < durka42> it is the basic unit of time used by the Rust compiler 12:29 < rlpowell> LOL. 12:31 < Ilmen> .itlokuti =da'i kai'u ko'a ko'o ko'u ce'ai da zo'u ko'o cusku ko'u da te zu'e lo nu ko'o da ckini ko'a 12:31 < Ilmen> (simple suggestion) 12:31 < durka42> no place for the da? 12:32 < Ilmen> Cadgu'a don't want an explicit addressee slot 12:32 < Ilmen> so I use da instead 12:32 < durka42> je'e 12:32 < ldlework> Ilmen: I changed my mind on that point 12:32 < ldlework> I think for the same statement different illocutionary forces can be directed at different listeners 12:32 < ldlework> For example, if you state that you will do something in the future, during a speech 12:32 < ldlework> Those who want you to do that thing will hear a promise 12:33 < Ilmen> je'e 12:33 < ldlework> Those who don't want you to do that thing will hear a threat 12:33 < durka42> unless you are a politician 12:33 < durka42> in which case the listeners will hear nothing 12:33 < ldlework> :) 12:33 < zipcpi> {kai'u}? 12:33 < ldlework> is ko'a the x1? 12:33 < Ilmen> At least take note of my lojban definition suggestion, it might be a good starting point 12:34 < durka42> x1=ko'a, x2=ko'o, x3=ko'u 12:34 < ldlework> is the kai'u stuff superflous? 12:34 < durka42> I dunno why the nonstandard choice :) 12:34 < ldlework> in the assumption that stuff is true out of context? 12:34 < zipcpi> Oh it's {me'au lo ka} 12:34 < durka42> kai'u is one way to write definitions 12:34 < durka42> another way is {lo ka ko'a ko'e ko'i itlokuti cu ka ...} 12:34 < durka42> or you can use fabu 12:34 < durka42> or whatever, seems everyone has their own style 12:35 < zipcpi> I prefer du'u if we're gonna fill them in with ko'a-series 12:35 < ldlework> Seems like you could just use the da series direct 12:35 < ldlework> without any abstraction 12:35 < ldlework> and just give the definition 12:35 < Ilmen> I'm the one who first suggested fabu, so ja'odai I have several definition styles :P 12:36 < zipcpi> Do we also need {ka'ei'u}? lol 12:36 < ldlework> oh there are only three da 12:36 < Ilmen> Technically "ce'u" is the way to go for writting x1...xN 12:36 < ldlework> seems like ko'a is a fine choice 12:36 < ldlework> like 12:36 < Ilmen> x1 citka x2 ----> ce'u citka ce'u 12:36 < ldlework> yeah but that's silly 12:36 < ldlework> oh 12:36 < ldlework> but for definitions you need something that distinguishes 12:37 < Ilmen> But ce'u has disadvantage 12:37 < Ilmen> s 12:37 < latro`a_> so in official jbo you bind ce'u to bindable variables 12:37 < ldlework> like seems like you could just state your definition as the second part 12:37 < durka42> ce'ai is shorthand for zo'u with sequential ce'u bindings 12:37 < zipcpi> ldlework: Erm, *someone* took bo'V exactly for that purpose 12:37 < Ilmen> (the reader has to count the slots etc) 12:37 < latro`a_> and in experimentals you can avoid the explicit bindings with ce'ai 12:37 < durka42> fabu is clearest for me 12:37 < ldlework> zipcpi: nice, seems the cleanest 12:37 < durka42> but apparently a true lojbanist can here "zo'u" and immediately memorize an unbounded list of bindings :p 12:37 < durka42> hear* 12:37 < ldlework> What does the da standfor in ilmen's definition? 12:38 < durka42> ldlework: see the internal {zo'u} 12:38 < durka42> it stands for "someone" 12:38 < Ilmen> ldlework: the addressee 12:38 < ldlework> so x5? 12:38 < ldlework> I realize da was used to side-step positionality 12:38 < Ilmen> well with the aforementioned def there is only three slots 12:38 < ldlework> but would it be x5? 12:38 < durka42> Ilmen wrote a definition that involves an addressee without exposing it in the place structure 12:38 < ldlework> oh right so x4 12:39 < durka42> function itlokuti(x1, x2, x3) { let da = new Addressee(); let bridi = ...; return bridi; } 12:39 < ldlework> Ilmen: can you state your definition plainly with just koha4? 12:39 < ldlework> including the addressee 12:39 < Amnadya> I think {.itlokuti} could have a minimalistic (#zancari) 2-place definition, and a 3-place definition if the speaker matters. 12:40 < zipcpi> Maybe {ka'ei} should be just {ka} + some UI 12:40 < durka42> which 2 places, Amnadya? 12:40 < Amnadya> The utterance, and its purpose. 12:40 < ldlework> Both the speaker and the listener matter though. 12:40 < Amnadya> (in reverse order) 12:40 < ldlework> No utterance has an implicit purpose 12:40 < durka42> hmm 12:40 < Amnadya> You mean "inherent" 12:40 < ldlework> And given a speaker, no utterance has an implicit purpose 12:40 < durka42> the speaker can be an attribute of an utterance though 12:40 < ldlework> sure 12:40 < zipcpi> Then that's just {itlokuti skuselzu'e} 12:41 < durka42> ma itlokuti lo se cusku be do... 12:41 < ldlework> The bigger question is wtf an itlokuti1 is 12:41 < Ilmen> .itlokuti =da'i kai'u bo'e cusku bo'i bo'o te zu'e lo nu bo'e bo'o ckini bo'a 12:41 < durka42> {le'o ma itlokuti lo se cusku be do} would be the nerdy translation of "what did you mean by that, *******?!" 12:41 < Ilmen> @ ldlework 12:41 < durka42> oi 12:41 < ldlework> Ilmen: ki'e 12:41 < Ilmen> with bo'V = xN 12:41 < durka42> ko na pilno lo mabla cmavo pe la zipcpi 12:41 < durka42> .u'i 12:42 < zipcpi> u'i 12:42 < Amnadya> loi mabla cnino ! 12:42 < Amnadya> pe le gerna 12:42 < Amnadya> those bloody new inventions 12:42 < durka42> _re_inventions 12:42 < zipcpi> mi remoi kurti 12:42 < Amnadya> (well nick was referring to {bu'u}...) 12:43 < durka42> haha 12:43 < Ilmen> ldlework: In my proposed def, an itlokuti-1 is a relation (binary-ka) that the speaker wants to have with the addressee as a result of uttering the specified utterance 12:44 < ldlework> Ilmen: that seems strange since not all illocutionary forces have to do with binding the speaker and listener in anyway 12:44 < zipcpi> "binding" is very loose here 12:44 < ldlework> Even so 12:44 < ldlework> If those variables are bound, they appear in the abstraction somewhere 12:44 < Ilmen> this could be for example the relation "[addressee] begins to be aware than [speaker] wants them to tell them whether [proposition] is true", etc 12:44 < ldlework> Which might not always make sense 12:44 < zipcpi> All it means is "speaker predicates the listener" in some way 12:44 < ldlework> If anything 12:44 < ldlework> I would say 12:44 < ldlework> It is a set of properties bound to the listener 12:45 < ldlework> IE, the intended perlocutionary effects 12:45 < ldlework> The speaker is irrelevant, unless specifically addressed by those effects. 12:45 < Ilmen> je'e 12:45 < latro`a_> it doesn't have to be plural, you can roll "properties" in the english sense into one block in lojban 12:45 < Ilmen> Then it may be an unary/monadic property of the addressee 12:45 < latro`a_> interesting perspective, though 12:45 < ldlework> We have other words that denote sets though 12:46 < latro`a_> sure, but it's not necessary 12:46 < latro`a_> it can just as well be {lo ka ... gi'e ... gi'e ...} 12:46 < ldlework> Sure but the sets definition allows for actual sets 12:46 < latro`a_> unless there is something like "each one of these is a *possible* effect" or something 12:46 < latro`a_> why do you want a set? 12:46 < ldlework> Including the set of a singular conjunctive property 12:46 < ldlework> latro`a_: its more general? 12:46 < ldlework> more accomodating 12:46 < latro`a_> in what respect is it more general? 12:47 < ldlework> I don't have to tie up all my properties if they exist separately with a conjunction just to fit it into the x1 12:47 < latro`a_> unless you're actually doing something to the elements separately in the predicate, it's just syntactic noise 12:47 < ldlework> If I happen to have a single property, I can just submit it 12:47 < ldlework> What 12:47 < latro`a_> the predicate to which you pass the set, that is 12:47 < ldlework> There's no syntax effects based on the definition of the predicate 12:47 < ldlework> it simplly *allows* for non-singular references 12:47 < ldlework> It still allows for the singular reference 12:47 < Ilmen> if .itlokuti-1 is passed to ckini-2, then giving a property plural as itlokuti-1 probably doesn't pose a problem 12:47 < ldlework> IE the set with one item 12:47 < ldlework> I can't understand how you can't cede that being more general 12:48 < latro`a_> I'm saying, when would the plural form be different from the singular conjunctive form 12:48 < latro`a_> semantically 12:48 < Ilmen> if ckini allows for plural of properties as its x3 12:48 < latro`a_> because the singular conjunctive form involves fewer words 12:48 < latro`a_> in jbo 12:48 < ldlework> latro`a_: never semantically, but in terms of what types we pass to our predicates it matters 12:48 < ldlework> If I *already have* seprarte properties, as the language allows, I don't have to provide a singular conjunction 12:48 < ldlework> As your definition would require 12:48 < latro`a_> ah, so you mean if you have separate properties bound 12:48 < ldlework> The alternative perfectly supports all cases 12:49 < latro`a_> though in practice the situation is still the same, it's just that we already have the set-building combinator but not the property-conjuncting combinator 12:49 < ldlework> I may have a thing, a discrete set, of properties that I want to pass to the predicate 12:49 < ldlework> But with your definition, I have to reach inside, and join them, and pass that 12:49 < latro`a_> unless you have a combinator to conjunct properties 12:49 < ldlework> Right which is extra work 12:49 < latro`a_> which we don't now, but could easily define 12:50 < ldlework> For literally no other difference 12:50 < Ilmen> What tell you my definition prevent plurals of properties from being acceptable as itlokuti-x1? 12:50 < ldlework> Other than requiring one, so that you can pass a singular property 12:50 < ldlework> When you can do that anyway 12:50 < ldlework> when a set is required 12:50 < ldlework> Ilmen: type fetishism 12:50 < Amnadya> What are you even talking about... 12:50 < latro`a_> it adds the syntactic noise of putting it in a singleton set, if you actually want a set rather than the usual "default plural reference" that we use 12:50 < ldlework> latro`a_: true 12:51 < ldlework> damn that's even more extremist than I was imagining 12:51 < ldlework> okay lets go back to ignoring the problem 12:51 < Ilmen> I don't even see what's the problem 12:51 < ldlework> Ilmen: nevermind its not important 12:52 < ldlework> I fell into an argument I tend to think is stupid anyway :3 12:52 < ldlework> or rather started it 12:52 < Ilmen> I don't think {mi ckaji lo ka sipna jo'u lo ka senva} is bad 12:52 < ldlework> that itlokuti-1 is properties of the listener is pretty nice lojbanic way of describing illocutionary force 12:53 < Ilmen> (giving a plural of ka's to ckaji2) 12:53 < durka42> and here I thought illocution was about the speaker 12:53 < ldlework> durka42: its the effects in the listener, that the speaker intends to bring about 12:53 < ldlework> communication is all about getting the listener to know what we what them to know 12:54 < durka42> ah I see 12:54 < ldlework> the definition of peritlokuti should be just a small mutation to the definition of itlokuti 12:54 < ldlework> Here's a question 12:54 < Ilmen> x2 cusku x3 fo x4 te zu'e lo nu x4 ckaji x1 12:55 < Amnadya> tosmabru alert 12:55 < ldlework> Is a perlocutionary force dependent on the speaker who uttered it? 12:55 < ldlework> I think so. 12:56 < Ilmen> {perlokuti} @Amnadya 12:56 < Amnadya> ie 12:56 < Ilmen> naljudyselsre 12:56 < ldlework> Sorry, I don't have the morphology within me :( 12:57 < Amnadya> lo do ve skicu cu simlu lo ka na mabla doi la .ilmen. 12:57 < Ilmen> The word shapes aren't very relevant of the discussion anyway 12:57 < Ilmen> so ju'inai jbopli 12:57 < Amnadya> mi'e jboplise 12:59 < Ilmen> Whoops, meant to say {cusku x3 fi x4...} 12:59 < Amnadya> .i ka'e plixau fa lo nu da'i x4 ba'e ckini x2 fi x1 12:59 < Ilmen> mabla se ju jai se bebysre 13:00 < Amnadya> .i ku'i lo te ckini sumti ka'e jai fanza 13:00 < ldlework> en: preti 13:00 < mensi> preti = x1 (quoted text) is a question/query about subject x2 by questioner x3 to audience x4. |>>> See also nabmi, 13:00 < mensi> danfu, ciksi, frati, spuda, cpedu. |>>> officialdata 13:00 < Ilmen> .i mi pu stidi tau me'ei ckini .i ku'i la cadgu'a cu jinvi lodu ka'e xlali fe mu'a tau zo ca'e 13:01 < Amnadya> mi ji'a pu pensi zo ca'e 13:01 < ldlework> Thanks for the discussion 13:01 < Amnadya> .i ku'i ie pei zo ca'e zo'u lo nu cusku zo ca'e fau lo nu no da tirna ja sanji cu se jalge na ku lo te zukte 13:01 < mensi> mi tugni i ie mi tugni 13:01 < Amnadya> .i sarcu fa lo nu su da tirna je cu .y tugni 13:01 < Amnadya> ja cu se mansa 13:12 < Ilmen> Are you satisfied with the definition of {.itlokuti} we've come up with so far? If so, we probably should add it to JVS, even if it'll be adjusted/corrected later, so that the fruits of this discussions don't get lost. 13:12 < durka42> ko na cirko lo grute be lo dei nuncasnu 13:15 < zipcpi> u'i 13:17 < Ilmen> Well then I'll add it; feel free to later edit it, mangle its argument structure or propose competing definitions if a vote is desirable. 13:17 < zipcpi> Ilmen: I'm not sure he's here anymore... just add it 13:19 < Ilmen> mi vi'onde 13:19 < durka42> what time is it in california anyway 13:19 < durka42> la'a lunchtime 13:19 < Ilmen> .y za'a do xa'o finti lo velski prosa pe zo .itloku'i 13:19 < Ilmen> .i ku'i smuni frica 13:20 < Ilmen> .i .au pei dunda lo dei smuni lo drata brivo 13:20 < durka42> .ai mi tavla fo lo .itlo ku'i 13:21 < Ilmen> lo .itlokuti cu .itlo kunti 13:21 < Ilmen> li'a dai 13:21 < durka42> oi'o 13:23 < Amnadya> .itlo kurti 13:23 < Ilmen> ju'i la zipcpi 13:23 < Ilmen> .u'i 13:23 < zipcpi> coi 13:24 < zipcpi> Ilmen: No, go ahead and overwrite my definition 13:24 < Ilmen> .i ke'u .i zo .itlokuti zo'u do ba'o smudu'a .i .au pei lo cnino smuni zo'u smudu'a lo drata poi brivo 13:24 < zipcpi> Or just add yours 13:24 < Ilmen> .i je'e 13:24 < zipcpi> I downvoted myself 13:25 < zipcpi> Cause I defined them with the LOJBO frame just so I can use them in tanru 13:25 < Ilmen> .i ku'i lo nu dizlo cmicu'a cu banzu lo nu na ka'e galtu cmicu'a lo cnino zo'o 13:25 < Ilmen> sei zangli si malgli 13:25 < zipcpi> lol 13:26 < zipcpi> ma'udzau, ni'udzau. Or use gleki's word {dudvota} but I don't like having to specify dudvota2, or use a sumtcita for the system 13:27 < durka42> I like ma'udzau and ni'udzau, they're more lojbanic 13:27 < durka42> being jbofuvi 13:27 < zipcpi> Heh 13:28 < Ilmen> $x_1$ (property of $x_4) is the illocution of $x_2$ uttering/expressing $x_3$ (quote) to audience $x_4$; $x_2$ utters/expresses $x_3$ to $x_4$ in order that $x_4$ has property $x_1$ 13:28 < Ilmen> .i'e pe 13:28 < Ilmen> i 13:28 < zipcpi> i'e 13:30 < Amnadya> Speaking of illocution, in Ithkuil illocution can be shown in an affix together with mood. There are 8 moods and 6 illocutions, so there are 48 affix forms for this slot. Some of the combinations are rather odd, so I'm making a table 13:31 < Amnadya> (well actually 41) 13:32 < Ilmen> mo'u smudu'a 13:32 < Ilmen> .i jmina lo lojbo ku ji'a 13:32 < Ilmen> http://jbovlaste.lojban.org/dict/itlokuti 13:34 < Ilmen> Maybe an illocution isn't specific only to linguistic utterances. How about handwaves and other syntaxless gestures? 13:34 < zipcpi> Hm that's related to the problem of what cusku2 should be o.o 13:36 < Ilmen> jbo: vi'o 13:36 < mensi> vi'o = [COI] tcita lo cmene ja ve skicu le du'u sinxa le du'u makau te cusku .e le du'u le cusku ba cpezu'e fi le te 13:36 < mensi> cusku |>>> doi; cpezu'e |>>> xorxes 13:36 < Amnadya> (off-topic/ithkuil: https://docs.google.com/spreadsheets/d/1NmTjSi51M8LCvjEJUWro03tS5AdPJNDLCeyWDOmTWMo/edit#gid=0 ) 13:37 < Ilmen> mi benji lo curmi cpedu notci 13:37 < Ilmen> do doi la .amdadiĭas. 13:37 < zipcpi> Problem is then defining {lotkuti} and {perlokuti} :p 13:38 < Ilmen> pe'i pei lotkuti ≠ cusku 13:38 < Amnadya> ua nai doi la .ilmen. 13:38 < zipcpi> "pertaining to the act of conveying semantic content in an utterance, considered as independent of the interaction between the speaker and the listener." 13:38 < Ilmen> "You need permission" 13:38 < zipcpi> Hmm... 13:39 < Ilmen> lo sfaile cu se stela doi la .amnadıǐas. 13:39 < Amnadya> .u'u https://docs.google.com/spreadsheets/d/1NmTjSi51M8LCvjEJUWro03tS5AdPJNDLCeyWDOmTWMo/edit?usp=sharing 13:40 < zipcpi> Amnadya: Hm related is the problem that Gleki and I are currently in a "fight" over what {ju'anai} should mean 13:40 < zipcpi> Problem is I don't understand what he wants it for 13:41 < zipcpi> He mentioned something related to Ithkuil 13:41 < Amnadya> I think he uses it for representative essence 13:41 < Amnadya> Instead of da'i 13:42 < zipcpi> The problem is that {ju'a} is defined as related to {xusra}. .iku'i gy xusra lo drata :p 13:42 < Amnadya> I need a unique example in each cell, because I want to translate them into Ithkuil and memorize the 41 combinations that way. 13:42 < zipcpi> I can't help you there T.T 13:44 < zipcpi> Admonitive/subjunctive: "This may hurt"? 13:45 < Amnadya> The admonitives are easy to combine for me, because it's just an added warning. 13:45 < Amnadya> But combining ko with ju'ocu'i is weird 13:46 < zipcpi> Heh 13:46 < Amnadya> "Make it the case that you might go there" 13:46 < zipcpi> Declarative = {ca'e}? 13:46 < Amnadya> Yes 13:46 < Ilmen> What is no'e birti, again? Mild certitude? 13:47 < Ilmen> Is Declarative synonymous to Performative in Iþkuîl? 13:47 < Amnadya> Yes 13:47 < Ilmen> je'e 13:48 < Ilmen> (To me, þ is easier to type than t-cedilla) 13:48 < zipcpi> ca'e lo redo co'a speni zu'ai 13:48 < Ilmen> spesi'u 13:48 < Ilmen> .e'u 13:48 < Ilmen> je'au zo 13:48 < zipcpi> I tend to avoid most RDA 13:49 < Ilmen> ki'a 13:49 < zipcpi> Regular derivational affixes 13:49 < zipcpi> I'm saving them for when relyle'ujvo is finished 13:49 < zipcpi> -gau is really hard to avoid though I'd admit 13:49 < Ilmen> gau X broda 13:50 < Ilmen> simxu could be remade as a SE 13:50 < zipcpi> Yes, I use that. But sometimes that is inconvenient 13:50 < zipcpi> Oh, gleki did make it a SE, but he abandoned it in favor of zu'ai 13:50 < Ilmen> selmaho: zu'ai 13:50 < mensi> .i lu zu'ai li'u cmavo zo'oi KOhA7 13:50 < mensi> http://lojban.github.io/cll/7/7/ 13:50 < Ilmen> wanaɪ 13:51 < Ilmen> ithkuil: iþkuîl 13:51 < ithkuil> error 13:52 < zipcpi> Mostly I avoid RDA because they have potentially mabla interactions, and also making sure all lujvo we want to make follow that rule is hard 13:52 < zipcpi> I mean 13:52 < zipcpi> Look at {rebdesygau} 13:52 < Ilmen> Some allographes or alternative letters should be added :3 13:53 < zipcpi> Is it rebdesku + gasnu or rebla + deskygau? 13:53 < zipcpi> And there's also an elided sefsi 13:54 < zipcpi> That's why I like the relyle'ujvo idea. We bring in affixes that are guaranteed to only expand one way 13:54 < zipcpi> Thus allowing regular lujvo to be defined however is convenient 13:56 < Ilmen> en: tolkufygla 13:56 < mensi> [< to'e kufra glare ≈ Polar opposite comfort hot] lo nu mulno sisku zo'u: y no da se 13:56 < mensi> tolcri 13:56 < mensi> tolkufygla[9927], tolkufrygla[10957], to'erkufygla[11516], tolkufyglare[11966], to'erkufrygla[12546], ... 13:57 < Ilmen> .i'e finti .i benri si plixau valsi 13:58 < zipcpi> Hmm might want {lenku} variants too 14:00 < Ilmen> en: olgoixorxoi 14:00 < mensi> olgoixorxoi = x1 is an olgoi-khorkhoi/Mongolian death worm of type/variety x2 |>>> 14:00 < mensi> krtisfranks 14:00 < Ilmen> xD 14:00 < zipcpi> lol 14:00 < sorpaas> lujvo: pensi gunma 14:00 < mensi> peirgunma[9006], pesygunma[9037], pensygunma[10067] 14:03 < sorpaas> xu pamai lo lujvo cu drani? 14:03 < zipcpi> What do you want it for? 14:03 < sorpaas> A collection of thoughts? 14:04 < zipcpi> I'd tend to just have {loi selpei} for that... 14:04 < sorpaas> What if I want a selbri... 14:05 < Ilmen> {selpei gunma}, {gunma (be) lo selpei} 14:06 < sorpaas> lujvo: selpei gunma 14:06 < mensi> gunma = x1 is a mass/team/aggregate/whole, together composed of components x2, considered jointly. 14:06 < sorpaas> lujvo: se pensi gunma 14:06 < mensi> selpeigunma[10855], selpesygunma[11986], selpensygunma[13016] 14:07 < zipcpi> Why do you need a lujvo for that? {gunma} isn't an RDA 14:07 < sorpaas> What is an RDA? 14:07 < zipcpi> Regular derivational affix 14:07 < Ilmen> .e'e na pilno lo ki'a zei valsi 14:07 < zipcpi> u'i 14:07 < Ilmen> First say "Regular derivational affix goi RDA" :3 14:08 < zipcpi> :p 14:08 < sorpaas> Can you illustrate that or point me to some links? 14:08 < zipcpi> Erm, we mean things like -gau 14:08 < zipcpi> Which *tend to* (but not always) expand one way when used in lujvo 14:09 < sorpaas> But I found {gungunma} exists in dict. 14:09 < sorpaas> Because I want {la selpeigunma} 14:09 < zipcpi> Hey go ahead and use anything with {la} without defining the brivla 14:09 < Ilmen> He means that you'd need to define {selpeigunma} in the dictionary, because it's not made of semantically compositional elements 14:10 < Ilmen> (rafsi are only there as mnemonic hints, and any lujvo should ideally be defined in the dictionary) 14:10 < sorpaas> Sorry to ask again... what's semantically compositional elements? 14:11 < Ilmen> The meaning of "barda zdani" can be infered from that of "barda" and that of "zdani", so "barda zdani" is compositional 14:12 < zipcpi> I use nonce-lujvo with {la} all the time: http://mw.lojban.org/papri/brivycme but that's different from using it as a real word with an implied place-structure 14:12 < Ilmen> However, the meaning of ".i bu" cannot be inferred from that of ".i" and that of "bu", so it's not compositional and must be learnt in the dictionary or while learning the language 14:13 < sorpaas> @Ilmen but "selpei gunma" makes perfect sense for me: a collection of thoughts? 14:13 < Ilmen> "selpei gunma" as a tanru is fine 14:13 < sorpaas> So basically I just want a name without any whitespace so that I can put it in an url and use it as a Lojbanic name... 14:14 < Ilmen> If you go for a lujvo, however, it would be preferable to define the lujvo in the dictionary 14:14 < Ilmen> Ah 14:14 < zipcpi> Names are fine 14:14 < zipcpi> http://mw.lojban.org/papri/brivycme 14:14 < Ilmen> As a name "selpeigunma" is fine 14:14 < Ilmen> la selpeigunma 14:15 < zipcpi> Many of the words on that page don't have a defined meaning otherwise 14:15 < zipcpi> Like {la xorlo} even 14:16 < zipcpi> The main problem even with "obvious" lujvo as a brivla, is the place structure 14:17 < zipcpi> Like what should {lo se selpeigunma} be. If you just use it as a name though, it isn't a problem 14:18 < sorpaas> So the reason why we can't (yet) use selpeigunma as a selbri is that this lujvo is not defined, and If (in imaginary case) I do want that, I need to claim to have defined it first? 14:20 < Ilmen> Well people will not be able to find {selpeigunma} in the dictionary and will have to guess its meaning 14:20 < Ilmen> (a lujvo isn't equal in meaning to a tanru, indeed) 14:21 < sorpaas> Understood. Thanks for all the explanations! 14:21 < Ilmen> je'e ŭi 14:27 < Ilmen> mi de'a jundi co'o 14:27 < zipcpi> co'o 14:34 < sorpaas> lujvo: se gunka 14:34 < mensi> selgu'a[6387], selgunka[7937] 14:34 < mensi> selgu'a [< se gunka ≈ 2nd conversion work] = x1 is a task/work done by x2 with goal/objective x3. 14:36 < zipcpi> {se gunka} tends to be preferred these days.... less rafsi to memorize or look up :p 14:37 < zipcpi> I gotta go though. Night 14:38 < sorpaas> @zipcpi Thanks! I'm just creating another name :) 14:38 < zipcpi> ua je'e 14:38 < sorpaas> co'o 14:39 < zipcpi> Amnadya: Subjunctive/directive might be "You may want to do this"? 14:39 < zipcpi> Something like e'uru'e ko... 14:40 < zipcpi> Or gajanai da'i do ... gi ko ... 14:40 < Amnadya> I believe illocution has scope over of mood (although the grammar doesn't say .oi sai), and the subjunctive expresses that the factuality of the statement is uncertain. 14:41 < zipcpi> What would be really funny is if you use {cau'a} 14:41 < zipcpi> ko cau'a plipe 14:41 < Amnadya> It's either "Make it the case that X may be the case" or "It may be the case that I direct you to do X". I believe it's the former. 14:42 < Amnadya> The second one doesn't really make sense 14:42 < zipcpi> {ko ka'e plipe}? 14:43 < Amnadya> Soooort of 14:43 < zipcpi> lol 14:43 < zipcpi> OK gotta go 14:43 < zipcpi> co'o 14:43 < Amnadya> co'o :) 15:02 < Sonjaaa> coi! 15:02 < Amnadya> coi 15:04 * nuzba @oka_iu_tcan: i carmi sicpi i cerni ruble gusni i a'o viska #xaiku #lojban [http://bit.ly/1TOINEO] 15:12 < ctefaho> лах лах и скән жаг сер 15:12 < ctefaho> упп ошһ нер ошһ нер ошһ упп 16:56 * nuzba @slaimsan: How do I say "お大事に" in Lojban? [http://bit.ly/1Idr6GH] 17:26 * nuzba @oka_iu_tcan: i cerni gusni i aunaisai balcu'e i sipna pacna #xaiku #lojban [http://bit.ly/1COIxRW] 17:44 < sorpaas> lujvo: logji gunka 17:44 < mensi> lojgu'a[6387], lojgunka[7937], logjygu'a[8517], logjygunka[10067] 20:08 < zipcpi> mensi: doi durka42 Don't like bo'V space being taken? Well, if we implemented the {leu} suggestion then FAbu would be just as good as KOhA, then just ca'e ka'ei / ka'ei'u uses that set. But... 20:08 < mensi> zipcpi: mi ba benji di'u ba lo nu la'o gy.durka42.gy. di'a cusku da 23:19 < gleki> in la bangu interjections are now split into interjections, suffixes, ending and right scalers 23:19 < gleki> *endings 23:20 < zipcpi> Cool 23:20 < zipcpi> Gotta go. See ya 23:26 < gleki> If the meaning is preserved, then {na} cannot be a variation of {fi'o se natfe} or {xoi se natfe}, as that would make {ro da} an argument of {na} and remove it from the x1 place of {zvati}, which is not what's happening here.</quote> 23:26 < gleki> wat 23:26 < gleki> this means NA is still a new selmaho 23:27 < gleki> obviously im not going to add one more member into FAM,BAM, ZAM set --- Day changed Wed Jul 22 2015 00:35 < isd> coi rodo .i do mo 00:38 < gleki> irci 00:40 < durka42> sampla 00:40 < mensi> durka42: cu'u la'o gy.zipcpi.gy.: Don't like bo'V space being taken? Well, if we implemented the {leu} suggestion then 00:40 < mensi> FAbu would be just as good as KOhA, then just ca'e ka'ei / ka'ei'u uses that set. But... | 2015-07-22T03:08:29. 00:40 < mensi> 325Z 00:40 < isd> .u'i 00:41 < durka42> jeez 00:41 < durka42> {boi} is just one syllable 00:41 < gleki> alta: li ny pa ny mo 00:41 < mensi> (NIhO [I {FA <li (¹[VEI {ny <pa ny>} VEhO] BOI¹) LOhO>} {CU <mo VAU>}]) 00:42 < gleki> alta: li ny pa ny re ny re ny mo 00:42 < mensi> (NIhO [I {FA <li (¹[VEI {ny <pa ny> <re ny> <re ny>} VEhO] BOI¹) LOhO>} {CU <mo VAU>}]) 00:42 < gleki> no idea how it should look like 00:43 < gleki> restore VEI in front of each letteral? 00:43 < durka42> do troci ma 00:44 < gleki> lo nu lo prane meksyci'e co'a zasti 00:44 < gleki> btw implement this for Lojban http://www.cs.ru.nl/~freek/pubs/holcas.pdf 00:46 < durka42> .u'i 00:58 < gleki> exp: lo broda noi mo 00:58 < mensi> ([lo {broda <noi (¹CU [mo VAU]¹) KUhO>} KU] VAU) 00:58 < gleki> the same as {lo broda ku noi mo}, right? 00:58 < gleki> semantically i mean 01:02 < durka42> camxes: lo broda noi mo 01:02 < camxes> ([lo {broda <noi (¹mo VAU¹) KUhO>} KU] VAU) 01:02 < durka42> camxes: lo broda ku noi mo 01:02 < camxes> ([{lo broda ku} {noi <mo VAU> KUhO}] VAU) 01:03 < durka42> yeah probably the same since {noi} can't affect the set of referents 01:05 < gleki> im completely lost. then why is it parsed so? 01:05 < gleki> why not put {ku} before {noi}? 01:08 < gleki> PEG says literally "selbri relative_clauses?" wth. are we having a bridi relative clause here? ;D 01:10 < durka42> quantifiers 01:13 < gleki> http://lojban.github.io/cll/8/6/ 01:13 < gleki> 6.5 = 6.7 01:17 < durka42> iep 01:17 < gleki> but that's when quantifiers exist 01:18 < gleki> because the next two rules make sense 01:18 < gleki> quantifier selbri relative_clauses? / quantifier sumti 01:19 < gleki> if one considers "quantifier selbri" as one construct 01:19 < durka42> without quantifiers it still matters with {poi} 01:19 < durka42> as selpa'i was talking about yesterday 01:19 < durka42> restrictive/unrestrictive 01:20 < durka42> so you can't just put the {ku} inside every time there is no quantifier 01:21 < gleki> what is the difference and where? 01:22 < gleki> 6.1=6.2=6.3 01:22 < durka42> that's where there is a difference 01:23 < gleki> difference from CLL? 01:23 < durka42> a xorlo difference I guess 01:23 < durka42> it's still fuzzy for me 01:23 < durka42> though selpa'i claims I understand it 01:23 < durka42> you'll have to scroll up 01:25 < durka42> actually selpa'i would claim that {lo broda ku poi brode} is useless I think 01:25 < durka42> because {lo broda ku} identifies the referents already 01:28 < gleki> i just dont like that relative clause attaches to selbri directly 01:28 < gleki> sumti_5, sumti_tail and sumti_tail_1 is a complete mess 02:07 < gleki> what is useless, what is not useless 02:07 < gleki> ... 02:12 < gleki> how are those examples expanded using {me} or whatever ... 02:59 < phma> lo ckunu na ckini lo ckana 03:38 < Amnadya> coi 03:38 < gleki> coi prije 03:39 < Amnadya> .i mi co'a lo prula'i vanci cu bilma fi lo crisa .influ'enza 03:39 < gleki> oi dai 03:39 < gleki> i e'uru'e do vreta gi'enai pilno lo skami 03:40 < Amnadya> mi ba'o ze'u vreta .i nitcu lo ka na vreta 03:42 < Amnadya> .i ui nai la durka ba'o cliva 03:42 < gleki> lo nu pilno lo skami a lo jbobau ka'e milxe ckape lo bilma 03:43 < Amnadya> xu go'i 03:43 < gleki> lo do stedu ka'e spoja vau zo'o 03:43 < Amnadya> pe'i spoja ri'a lo nu lo nazbi cu culno 03:44 < gleki> i e'u do na pilno lo skami ije fau lo nu do djica co jbozu'e vau finti lo lojbo sepi'o ke po'o lo stedu 03:45 < Amnadya> .u'i zo'o mi ji do prije .i do stidi lo jai se prije 03:46 < Amnadya> .i je'e je'e .i ku'i .ai cusku pa da 03:46 < Amnadya> .i va'o lo nu da'i sai do na se vajni lo nu spofygau lo purci jboselsku zo'u zo poi poi jbini lo selbri joi zo ku ka'e se vimcu lo bangu 03:49 < Amnadya> (to cumki ju xamgu toi) 03:50 < gleki> sa'u mi nitcu lo ve ciksi be ci jufra 03:51 < gleki> i re lo poi ninmu vau mu prenu cu broda 03:51 < gleki> i re lo mu prenu poi ninmu cu broda 03:51 < gleki> i re lo mu prenu ku poi ninmu cu broda 03:51 * Amnadya cu bilma je na se .e'ande fi lo ka pensi lo pluja 03:51 < gleki> i oi za'u ci jufra 03:51 < gleki> Amnadya: ie zo'onai e'u do zei'a kanro i ta'i bo na pensi 03:52 < gleki> i ai mi co'u retsku 03:52 < Amnadya> je'e .i .ai mi ba zi cliva .i ku'i ro mai do ma djica .i xu djica tau lo jufra poi na se pagbu zo poi 03:53 < Amnadya> .i va'i ma tarmi lo zo poi ve ciksi je cu jai se djica do 03:53 < gleki> mi djica tu'a lo xe fanva be le ci jufra 03:54 < gleki> to ta'o caku mi spofygau lo PEG gerna toi 03:54 < Amnadya> xu mi zifre lo ka fanva .i glico 03:55 < gleki> zifre 03:55 < gleki> i mi pu'i tcidu CLL 03:56 < Amnadya> je'e .i mi ba sidju do za ku .ai 03:56 < Amnadya> mensi: doi durka42 {lo broda ku poi brode} is not useless, but simply distinct from {lo broda poi brode}. The former has the same form as e.g. {do poi brode}, which isn't useless at all (it implies a non-singular {do}, and the referent(s) of {do poi brode} is/are merely among the referents of {do}, but usually not the same.) 03:56 < mensi> Amnadya: mi ba benji di'u ba lo nu la'o gy.durka42.gy. di'a cusku da 03:56 < Amnadya> co'o ru'e 03:57 < gleki> co'o 05:01 < gleki> this non ASCII symbols JVS bug is terrible 05:37 < gleki> wow, Curtis wrote a new page to the wiki! 06:54 < gleki> aha now i see why guskant wanted {nu tu'u mi mo ije do mo tu'u kei} 07:46 * nuzba @uitki: Proposal: distinguishing digits from numbers #lojban http://mw.lojban.org/papri/Proposal:_distinguishing_digits_from_numbers [http://bit.ly/1HRC3OH] 07:46 * nuzba @uitki: fancy #lojban challenges http://mw.lojban.org/papri/fancylojban_challenges [http://bit.ly/1MokJr6] 07:46 * nuzba @uitki: The North Wind and the Sun in #lojban The North Wind and the Sun [http://bit.ly/1MokJY8] 07:46 * nuzba @lojban_org: The North Wind and the Sun in #lojban The North Wind and the Sun [http://bit.ly/1HRCecP] 07:46 * nuzba @uitki: Eulojban #lojban http://mw.lojban.org/papri/Eulojban [http://bit.ly/1MokKuZ] 07:50 * nuzba @uitki: Sour Grapes story in #lojban http://mw.lojban.org/papri/Sour_Grapes [http://bit.ly/1MokNa5] 08:14 * la_kristan may be suddenly disconnected at any moment. 08:14 < la_kristan> .i coi 08:15 < la_kristan> .i xu lo prenu cu zvati 08:15 <@xalbo> mi zvati 08:15 <@xalbo> .i ca ku mi jundi 08:15 < clakre> coi la kristan! 08:16 < la_kristan> :-) 08:17 < la_kristan> the end bracket on my name makes the smiley face look like a mean face at first glance... 08:17 < sorpaas> Is whitespace necessary in Lojban? 08:17 < la_kristan> sometimes 08:18 < la_kristan> it helps one see where the stresses syllables are. 08:18 < la_kristan> *stressed 08:19 < la_kristan> .i zoi gy. futile .gy. xe fanva ma la .lojban. 08:23 < la_kristan> .i xu do djuno 08:25 < la_kristan> .i do djuno .i ku'i do na djica lonu do cusku fi mi 08:26 < la_kristan> .i uinai 08:27 < la_kristan> .i doi @xalbo do na jundi .i la'a la clakre na djuno 08:28 <@xalbo> .u'u mi pu de'a jundi 08:28 < clakre> mi na djuno ma? 08:28 < clakre> (answer: almost everything) 08:28 < la_kristan> .u'i 08:29 <@xalbo> zoi gy. futile .gy srana zo fliba .e zo troci .e zo plixau to dukti toi .i ku'i mi sanji lo prane panra 08:29 < la_kristan> .i doi clakre do na djuno ledu'u zoi gy. futile .gy. xe fanva ma kau la .lojban. 08:30 < Thorne> Is it possible to make a Double entendre in Lojban? 08:30 < clakre> ua 08:31 < ldlework> A double-entendre is helped by polysemy but doesn't require it 08:31 < zipcpi> be'e mensi 08:31 < zipcpi> What have I done... I just invented pointers lol 08:32 * la_kristan looks up each word of @xalbo 's response... 08:32 <@xalbo> Now I need to figure out how to use mekso together with {la'e} and {lu'e} to do pointer arithmetic in Lojban. 08:33 < zipcpi> Heh... I don't use mekso 08:33 <@xalbo> .i'e 08:33 < Thorne> Would you say Lojban would be an ideal first language in a future society? 08:34 <@xalbo> I think it would definitely be an interesting one, in the scifi sense. In the world we live in, I don't see it ever actually happening. 08:34 < zipcpi> I use {le sumji be li pa bei li pa cu du li re} instead 08:34 <@xalbo> I use {li re sumji li pa li pa}. 08:34 < Thorne> But ambiguity is not always disadvantageous, I suppose. 08:35 < zipcpi> Or that could work too 08:35 < Thorne> Especially to literature. 08:35 < zipcpi> I was thinking too hard 08:35 < ldlework> Thorne: Lojban is only grammatically unambiguous 08:35 < ldlework> Thorne: it has plenty of the semantic kind 08:35 < zipcpi> {sumji li pa li pa fa li re} works too 08:35 < sorpaas> How to say "at a time" in the sense of "step by step" or "not worrying about the future" like in "finish a bit work a day at a time"? 08:35 <@xalbo> jbopre tend to use the words "ambiguous" and "vague" in particular ways, so that we can claim that Lojban is never ambiguous, but it is possibly vague, sometimes far more so than natlangs. 08:35 < Thorne> Yeah, I speak of grammatically ambiguity.. 08:36 < Thorne> I think it would make an ideal second language for certain needs. 08:37 < Thorne> Hehe, My scrollback says [20:54] * Loaded log from Fri May 17 23:58:19 2013. 08:37 < Thorne> Last time I came here to discuss my presentation. 08:38 < la_kristan> futility of futility; all is futile! 08:38 < zipcpi> Anyway with a few magic brivla for variables and while-loops I came up with this: http://mw.lojban.org/papri/zipcpi:_solution-gcd-2 08:39 < la_kristan> leave it to zipcpi to have magic brivla... 08:39 < Thorne> I should google for some puns in Lojban. 08:40 < zipcpi> la_kristan: Psht. I don't hold a candle to Ilmen for inventing magic brivla :p 08:40 <@xalbo> sorpaas: I'm not thinking of a good direct equivalent. I can think of ways to paraphrase it, like {ko ca gunka lo cabna .e nai lo balvi}. 08:41 < Thorne> BBS. 08:41 <@xalbo> And krtisfranks for magic cmavo. 08:41 < zipcpi> Haha 08:41 < la_kristan> @xalbo so, like "vain"; "all is vanity" 08:42 < ldlework> xalbo: I was thinking more of a preposition of some kind 08:42 < la_kristan> My translating of Ecclesiastes can't progress very well without it. 08:42 < zipcpi> I dunno, they have pretty good lojban defs imo. Except maybe for goilka'i... not sure what {virtu'ale} means formally :p 08:42 < zipcpi> Oh... you should talk to phma 08:42 < zipcpi> He's working on translating the Bible too 08:43 < zipcpi> la_kristan: The word {kosmu} might be helpful there 08:43 < phma> You called? 08:43 <@xalbo> Seems like {narplixau} or something. 08:44 < la_kristan> I got pretty far in translating John 1 into toki pona... 08:44 < sorpaas> @xalbo thanks! I found the definition of "at a time" in dictionary is "during any one period". Is {ca lo paroi} a valid translation for that? 08:44 < sorpaas> Or {ca'o lo paroi}? 08:44 < la_kristan> phma : they say you're translating the Bible too. 08:45 < la_kristan> I was trying to say "vanity of vanities; all is vanity." 08:45 < la_kristan> but not succeeding. 08:46 < phma> Someone's done Ecclesiastes 1. Let me see if I can figure out who it was in the git log. 08:46 < sorpaas> Oh, should it be {ca'o pa loi paroi}? 08:47 < la_kristan> zipcpi : "no such valsi in the dictionary " 08:47 <@xalbo> sorpaas: Those don't parse. {paroi} is a tense. 08:47 <@xalbo> {ca [ro] lo [pa] krefu} might be about right. 08:47 < ldlework> ko jundi pa da ba tu'a le bavla'i 08:47 < phma> It's la mukti, aka Shunpiker. 08:48 <@xalbo> sorpaas: ldlework's suggestion is nice. 08:48 < la_kristan> phma : link? 08:48 < ldlework> ki'e 08:49 < phma> https://github.com/phma/lojban-bible/ 08:50 < ldlework> hmm, I wish I could suggest to them to use diacritics(??) instead of captilization 08:50 < zipcpi> danmo... hm that's pretty metaphoric 08:50 < phma> Are you going to jbonunsla? 08:50 < ldlework> doi ma 08:51 < phma> la kristan 08:51 < ldlework> ba'a la kristan 08:51 < ldlework> jinga 08:51 < zipcpi> Reminds me of a Chinese idiom. Literally meaning something like {fomsle draci} 08:51 < ldlework> heh 08:52 < la_kristan> phma : no, I won't make it. 08:53 < zipcpi> ldlework: Yeah a lot of what is there is somewhat dated in style. Stuff like .xaviLOT. was obviously made pre-dotside too 08:53 < zipcpi> (Havilah) 08:54 < sorpaas> Thanks! I spent sometime understanding that and it sounds pretty! 08:55 < la_kristan> .i doi la'o gy. Idlework .gy. xu do djica lonu do cusku zo be'e .enai zo ba'a 08:57 < la_kristan> "everything is smoke/smog/air pollution"... 08:57 < la_kristan> fascinating 08:59 < zipcpi> {jajgau} for teacher though? Not sure about that? 09:00 <@xalbo> zipcpi: IIRC, there was a user who went by Imami, who said that the literal translation was something based jmaji. 09:00 * la_kristan could still be suddenly disconnected at any time 09:00 < zipcpi> Oh... the source text 09:00 < zipcpi> I don't know Hebrew or Classical Greek 09:01 < zipcpi> Makes me wonder about the {danmo} metaphor 09:01 < la_kristan> me neither. 09:02 < la_kristan> I usually rely on the Esperanto translation. 09:02 < zipcpi> I tend to use the Amplified Bible when trying to translate to Lojban 09:03 < zipcpi> And hm, it uses "vapor". Interesting 09:03 < la_kristan> http://www.steloj.de/esperanto/biblio/libro_pre.html 09:04 < zipcpi> I don't know Esperanto though 09:04 < la_kristan> vantaĵo 09:04 < la_kristan> not any more helpful than the kjv 09:05 < zipcpi> Amplified Bible: https://www.biblegateway.com/passage/?search=Ecclesiastes+1&version=AMP 09:05 < ldlework> la_kristan: no I was saying "I expect you, doi la kristan" 09:06 < sorpaas> And how to say "I'm only one [person]"? Can "one" itself be a selbri? 09:06 < ldlework> sa'u mi du pa prenu 09:06 < zipcpi> mi pa po'o mei 09:06 < la_kristan> afk 09:06 < sorpaas> Yeah, thanks! 09:07 < zipcpi> {mei} turns a number into a selbri, meaning "numbering this many" 09:07 < ldlework> It seems wierd to use po'o with a number, but sure 09:07 < zipcpi> {po'o} means "only", attaches to last lexical unit, in this case {pa} 09:08 < gleki> i'd rely on glossed version in original languages than on any translations 09:08 < ldlework> sorpaas: do gunka ma 09:10 < sorpaas> mi troci co finti lo zgike 09:11 < sorpaas> se se se fanva lo zgike 09:11 < ldlework> ua je'e 09:11 < ldlework> a'o snada 09:11 < la_kristan> se se se? 09:11 < sorpaas> What's the word for earsing the last word? 09:11 < ldlework> si 09:11 < la_kristan> si 09:11 < la_kristan> all together now... 09:12 < la_kristan> Idlework : no point in expecting me... 09:13 < la_kristan> I have prior reservations 09:13 < la_kristan> in the opposite direction. 09:14 < la_kristan> I'm going to Oregon; wish I could stay there forever. 09:15 < la_kristan> except to go to jbonunsla 09:15 < gleki> if {po'o} is like DA not only in scope but in existentiality then who knows 09:16 < la_kristan> gleki : to what are you referring? 09:16 < sorpaas> lujvo: galtu traji 09:16 < gleki> 09:06 < zipcpi> mi pa po'o mei 09:18 < mensi> galrai[5867], ga'unrai[7456], galtraji[7917], galtyrai[7997], ga'urtraji[9506], ... 09:18 < gleki> {serai galtu} might also work 09:19 < cliva> exp: mo gi'e ba ke mo 09:19 < mensi> ([CU {mo VAU}] [gi'e ba] ke [CU {mo VAU}] KEhE VAU) 09:19 < cliva> exp: mo gi'e cu ba ke mo 09:19 < mensi> ([CU {mo VAU}] [gi'e {cu <ba (¹ke mo KEhE¹)> VAU} VAU]) 09:19 < cliva> gerna: mo gi'e ba ke mo 09:19 < mensi> (0[{mo VAU} {gi'e ba} ke {mo VAU} VAU])0 09:20 < cliva> gerna: mo gi'e cu ba ke mo 09:20 < mensi> O_0 09:20 < gleki> zo cu sepli cmavo 09:20 < zipcpi> Another possibility might be {mi za'unai pa mei}, if you rely on my somewhat unorthodox use of NAI/CAI :p 09:21 < gleki> i zo cu te sepli lo stedu lo rebla 09:21 < cliva> gerna: mo gi'e ja'a ba ke mo 09:21 < mensi> (0[{mo VAU} gi'e {<ja'a (1ba [ke mo KE'E])1> VAU} VAU])0 09:21 < gleki> zipcpi: that sounds more like {su'e} 09:21 < cliva> ua 09:21 < zipcpi> Oh 09:21 < zipcpi> Er, no, not exactly 09:22 < cliva> ja'o zo ja'a zo'u ba'e re da fegli nu zviki 09:22 < zipcpi> Maybe? But what I mean is "one, and not more than one" 09:22 < zipcpi> Though poetically {mi su'epa mei} works too 09:22 < gleki> cliva: lo zviki cu mo 09:23 < cliva> me'oi hack ba'a nai ru'e 09:23 < cliva> en: zviki 09:23 < zipcpi> noroi se jmina fi la jbovlaste 09:23 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 09:24 < zipcpi> so'e selckikyvla cu go'i 09:24 < zipcpi> .y. selckikynomvla 09:24 < noncomcinse> coi lo se .irci 09:25 < cliva> lo selfu cu ma'asno 09:25 < zipcpi> coi 09:25 < noncomcinse> .i .u'i zo'o pilno fi dei fe lo selckikynomrafsi 09:25 < noncomcinse> coi zy coi cy 09:26 < cliva> coi noncomnompre 09:26 < noncomcinse> ca lo cabdei mi reze'u cilre tau zo zu'i 09:28 < noncomcinse> .ije co'a sanji lodu zasti fa lo so'i se cinri cmavo poi na fadni 09:29 < noncomcinse> .i se mu'i bo mi se cinri lonu da'i ca lo ro djedi zo'u lo pa cmavo cu jai se djuno lo se .irci 09:30 < cliva> co'a ku xu 09:31 < zipcpi> u'i mi ji'a jmina da loi selckikynomrafsi 09:31 < noncomcinse> .i milxe .i co'a zo'e 09:31 < zipcpi> ra'oi -zam- noi rafsi zo zai'e 09:32 < cliva> da'i va'o lo nu pilno lo mekso cmavo cu cnino ba'e ro lo .irci 09:33 < noncomcinse> .i .ei ba cliva ca lonu co'u klama la .seint.lu,is. 09:33 < noncomcinse> ja'a go'oi mekso 09:34 < noncomcinse> co'o uinai .i'a 09:34 < cliva> co'o 09:40 < la_kristan> su'e and su'o always confused me 09:41 < zipcpi> su'o alone means "some", but can modify numbers to mean "at least" 09:41 < zipcpi> su'e means "at most" 09:41 < la_kristan> because "at l/e/ast" and "at m/o/st", I always got them switched. 09:41 < zipcpi> Oh 09:43 < la_kristan> but I think I've got it through my thick skull now... 09:49 < gleki> some SW'Ords in my pocket, but not more than 10 SW'Eaters on me 10:04 < zipcpi> {lo faubganke be lo jetnu} would be a really nerdy way to say "something that happens repeatedly and unceasingly" :p 10:04 < durka42> en:faubganke 10:04 < mensi> durka42: cu'u la'o gy.Amnadya.gy.: {lo broda ku poi brode} is not useless, but simply distinct from {lo broda poi brode}. 10:04 < mensi> The former has the same form as e.g. {do poi brode}, which isn't useless at all (it implies a non-singular {do}, and 10:04 < mensi> the referent(s) of {do poi brode} is/are merely among the referents of {do}, but usually not the same.) | 10:04 < mensi> 2015-07-22T10:56:48.408Z 10:04 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 10:05 < durka42> ua 10:05 < durka42> jbo:faubganke 10:05 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 10:05 < durka42> uanai 10:05 < zipcpi> Er, sorry 10:05 < zipcpi> I just added it and it hasn't updated li'a 10:05 < zipcpi> Do-while loop 10:05 < durka42> ah 10:05 < durka42> mensi: ko ningau 10:05 < ldlework> Do we have a list of Lojban blogs? 10:05 < durka42> vlaste: update! 10:05 < vlaste> auto-update temporarily disabled due to memory bugs 10:05 < mensi> sei ca ca'o jai gau cnino be fai lo pe mi sorcu 10:06 < zipcpi> Yeah I made several magic zi'evla for algorithms 10:06 < durka42> do they have jbojbovelcki? :D 10:06 < zipcpi> Yes 10:07 < ldlework> I'm thinking of creating a shared lojban blog. What do you guys think about that? 10:07 < zipcpi> faubganke is defined in terms of ganfauke. ganfauke is defined recursively 10:07 < zipcpi> Sounds interesting 10:07 < zipcpi> We'd each talk about the crazy stuff we do :p 10:08 < ldlework> Basically, it would be based on git. So you'd author a post, and submit a PR and hopefully someone else participating gives you a LGTM and it gets merged, which causes a docker image to be built and automatically deployed to refresh the website. 10:08 < ldlework> So basically all you'd need to know how to use is git. 10:09 < ldlework> And how to write RestructuredText 10:09 < ldlework> Basically I want to write about Lojban, but not often enough I want to maintain a blog 10:09 < ldlework> And not often enough anyone would pay attention to it 10:10 < mensi> i ba'o jai gau cnino 10:12 < ldlework> I kind of want to make it about conlangs in general though 10:12 < durka42> en:faubganke 10:12 < mensi> faubganke = x1 happens, then if x2 is true, x1 repeatedly happens as long as x2 is 10:12 < mensi> true 10:12 < durka42> jbo:faubganke 10:12 < mensi> faubganke = x1 fasnu .ibazibo x2 x1 ganfauke 10:12 < ldlework> x1 is sustained by x2 10:12 < durka42> .u'i za'a linsi 10:12 < durka42> en:ganfauke 10:12 < mensi> ganfauke = if x1 is true, x2 happens, then repeats as long as x1 is true 10:12 < durka42> jbo:ganfauke 10:12 < mensi> ganfauke = ganai x1 jetnu gi ge x2 fasnu gi baziku x1 x2 ganfauke 10:13 < durka42> ua 10:13 < durka42> iteration => recursion transformation 10:13 < durka42> nice 10:13 * zipcpi nods 10:14 < gleki> ldlework: share that blog on github 10:14 < ldlework> gleki: I have to make it first 10:14 < gleki> ldlework: will there be some kind of converter of it into readonly static html pages? 10:15 < durka42> something like jekyll I'd imagine 10:16 < gleki> not necessarily to gh-pages, just that it needs to be preserved 10:17 < la_kristan> I don't know how to use git, but I could probably learn. 10:17 < la_kristan> is this supposed to be a blog in Lojban, or about it? 10:18 < gleki> la_kristan: just register on github and try to create your own repository. just make a single file and write something into it. there is a text editor there. 10:18 < gleki> no knowledge of programming is required 10:21 < zipcpi> durka42: They're jbofuvi too. ganfauke = ganai + fasnu + krefu. faubganke = fasnu + ba + ganai + krefu 10:22 < durka42> ua 10:22 < durka42> i'e na kognato 10:22 < durka42> zo druuaile zo'o 10:22 < zipcpi> u'i 10:24 < gleki> ma ge'esai ka'e kognato 10:24 * la_kristan can find a link to sign in, but not to register. 10:24 < dutchie> https://github.com/join 10:25 < ldlework> I will probably use Pelican 10:26 < gleki> when i open https://github.com/ incognito it just shows me the form to register 10:26 < dutchie> mi ji'a 10:30 < la_kristan> hmmm... should I be "Christa627" as on other websites, or "kristan", since that's my Lojban name? 10:30 < gleki> ldlework: no commenting, not logging in via facebook or something? 10:30 < zipcpi> durka42: Using ganfauke, and two more magic brivla, {goilka'i} and {goilbi'o}, I translated the iterative version of Euclid's algorithm: http://mw.lojban.org/papri/zipcpi:_solution-gcd-2 10:31 < ldlework> gleki: There's no need to 'log in' since the site is static html 10:31 < zipcpi> Quite a bit longer than the recursive translation though 10:31 < ldlework> gleki: you edit the source on your own website, then submit a Pull-Request to the shared git repo 10:31 < zipcpi> Oh yeah and also uses tu'e...tu'u in NU 10:31 < ldlework> When your changes are merged, the website will automatically be rebuilt and deployed 10:31 < durka42> comments can be done using disqus if that's desired 10:31 < gleki> or some facebook plugins 10:32 < durka42> facebook is gross :) 10:32 < zipcpi> Now {goilka'i} is really a magical brivla... it's meant to work like variables or pointers 10:32 < zipcpi> (Yes I accidentally invented pointers) 10:32 < ldlework> durka42: yeah disqus for sure 10:33 < durka42> en:goilka'i 10:33 < mensi> goilka'i = x1 (text / virtual-text-object) is a variable representing x2 10:33 < durka42> en:goilbi'o 10:33 < mensi> goilbi'o = x1 (goilka'i variable) changes value to now represent x2 10:33 < durka42> eww imperative programming :) 10:33 < durka42> jbo:goilka'i 10:33 < mensi> goilka'i = x1 sumka'i x2 gi'a virtu'ale lo sumka'i be x2 .i va'o lo nu da goi ko'a goilka'i kei lu «la'e ko'a» li'u cu 10:33 < zipcpi> lol 10:33 < mensi> sinxa le se goilka'i be ko'a 10:33 < durka42> jbo:goilbi'o 10:33 < mensi> goilbi'o = x1 noi goilka'i cu binxo lo goilka'i be x2 10:33 < durka42> pe'i lojban is much more suited to functional programming with immutable variables 10:33 < durka42> but I'm biased of course 10:34 < zipcpi> Yeah and I'm biased the other way cause I never worked with functional programming 10:34 < durka42> you should 10:34 < durka42> it will warp your mind 10:34 < zipcpi> :p 10:34 < ldlework> Except all those variable/mode binding and reset cmavo... 10:34 < ldlework> :) 10:35 < durka42> {goi} is just a label 10:35 < durka42> you mean curtis' mabla slate of zmico? 10:35 < ldlework> That's like saying, all variables are just labels. 10:36 < ldlework> In Python that's true, but we don't consider it a functional language (directly anyway) 10:36 < ldlework> Also beyond the variables, there are mode cmavo too 10:37 < ldlework> Like cmavo that change the grammar going forward, until they are reset and so on 10:37 < durka42> true 10:37 < durka42> but at the core it's a predicate language 10:37 < ldlework> Sure, as most languages are 10:38 < la_kristan> durka42 : Lojban warps my mind enough as it is... 10:38 < ldlework> Even object oriented languages are built ontop of predicates 10:38 < gleki> Are predicate languages predicatable 10:38 < ldlework> What does that mean? 10:38 < gleki> pun 10:38 < durka42> object oriented languages are often built on top of objects 10:38 < ldlework> durka42: objects are just a way of organizing state though 10:39 < durka42> that's true 10:39 < ldlework> the action of object oriented languages is still predicates 10:39 < ldlework> I've been reading GEB and its fascinating his insight that predicates don't truly have definitions 10:39 < la_kristan> "kristan" is already taken: -( 10:39 < durka42> have you read "Execution in the Kingdom of Nouns"? 10:39 < ldlework> no 10:40 < durka42> it's about java and objects 10:40 < ldlework> So the idea is that predicates really don't have definitions 10:40 < ldlework> And that their definition really arises out of the set of things that are true of that predicate 10:41 < ldlework> And all the predicates, in turn, true of those things 10:41 < la_kristan> but "la-kristan" isn't... 10:41 < gleki> i never understood and never understand how lojban is a predicate language. 10:41 < ldlework> IE, if the only Dad I've encountered is my dad, I might think that being a cool guy who loves his family is what makes a dad 10:42 < la_kristan> but I don't really like that... 10:42 < ldlework> But as I encounter other Dad's that might not be as cool, I come to a richer understanding of dad-ness 10:42 < ldlework> By learning what predicates are also true of other things known to be true of the predicate dad 10:42 < gleki> in fact im lagleki on github because someone took {gleki}. maybe it wwas even me :/ 10:43 < la_kristan> maybe... 10:43 < ldlework> I thought it was interesting anyway 10:43 < gleki> la_kristan: ko pilno zo krista vau zo'oru'e 10:43 < durka42> I read an interesting post/study about that recently 10:43 < durka42> they claimed to show that actually language is truly more abstract than that 10:44 < durka42> by measuring reaction times with priming by words vs pictures 10:44 < la_kristan> I have two usernames on Wikipedia, because I forgot the password to the first one. 10:44 < ldlework> durka42: I'm not sure what you're saying 10:44 < la_kristan> and had failed to register an email address. 10:45 < durka42> ldlework: well they started with the hypothesis that the meaning of "dad-ness" is just a collection of examples, and found evidence against that hypothesis 10:45 < durka42> was just a poster though 10:45 < durka42> hardly definitive 10:46 < la_kristan> gleki : nah, maybe I'll just go by "janKulisa". 10:46 < ldlework> durka42: the more fundamental lesson in what I had read is that language, and really knowledge itself is the act of identifying isomorphisms. 10:46 < la_kristan> then I can blog in toki pona. 10:46 < ldlework> s/knowledge/meaning 10:46 < fenki> ldlework meant to say: durka42: the more fundamental lesson in what I had read is that language, and really meaning itself is the act of identifying isomorphisms. 10:47 < gleki> yesterday i started pontufa project, a parser for "tokipona enhanced" language 10:47 < la_kristan> oh no... 10:47 < durka42> Alta Pona 10:47 < ldlework> durka42: If I know that that guy over there is my dad, and you tell me that guy over there is your dad, by understanding the isomorphism between those two guys, I further my understanding of the predicate that unites them 10:47 < durka42> yeah 10:47 < la_kristan> enhanced language is always trouble... 10:47 < ldlework> Rather than understanding those guys by way of the predicate... 10:47 < durka42> you somehow get meaning out of a finite set of examples 10:48 < durka42> so about language learning 10:48 < durka42> language learning is deep magic 10:48 < gleki> la_kristan: non enhaced too. jan Mato had problems with parsing prepositions 10:48 < ldlework> This is the kind of stuff I want to write about 10:48 < durka42> ja'o lo cifnu cu makfa 10:48 < gleki> makfa ji temakfa[D[D[D[D 10:48 < durka42> toki pona is parseable? 10:49 < la_kristan> durka42, it's even deeper magic, with the reformemuloj around... 10:49 < durka42> what? 10:49 < gleki> oh, toki pona parser is no longer online 10:50 < gleki> oh no it's in C# 10:51 < la_kristan> jen estis reformemulo 10:51 < la_kristan> kiu diris, "mi estus stultulo 10:51 < la_kristan> se mi taksus la lingvon 10:51 < la_kristan> jam granda atingo, 10:51 < la_kristan> sen ŝanĝi al ĝi la regulon!" 10:52 < gleki> if smeone can read an instruction on how to run this ... https://github.com/matthewdeanmartin/tokipona.parser 10:52 < la_kristan> something I wrote awhile back. 10:53 < gleki> kiu estis la nomo de la reformemulo 10:53 < durka42> it runs online 10:53 < gleki> durka42: online, yes. how to run it 10:53 < durka42> I can't bring myself to clone a project with such terrible commit messages :) 10:54 < la_kristan> gleki: "amigueo", mi pensas, tiufoje... 10:54 < gleki> maybe i should ask him on facebook 10:54 < la_kristan> sed estas multaj. 10:55 < la_kristan> sed feliĉe en Esperantujo ili ne efikas. 10:56 < la_kristan> almenau, malmulte efikas... 10:58 < jsgrant> Has anybody tried training a dog, with Lojban-based commands yet? 10:58 < zipcpi> I have a dog, but I never taught her Lojban 10:58 < ldlework> I speak lojban to my cat 10:58 < ldlework> well 10:58 < ldlework> more like I speak in lojban about my cat, to my girlfriend :/ 10:58 < ldlework> ju'inai 10:59 < la_kristan> "There were problems creating your account " 10:59 < la_kristan> but they won't say what problems... 10:59 * jsgrant assumes it would work fairly well if teaching because they don't really need to understand structure, grammar, etc, just x means do y. Like how we US police sometimes uses German commands to train Dogs. 10:59 < la_kristan> :-( 11:01 < zipcpi> jsgrant: Interesting. Is it because then it won't be confused with normal speech? 11:01 < jsgrant> Also, I haven't been on the Lojban Site/Wiki for a bit -- is it actually on Mediawiki now? It vaguely looks like it. 11:01 < gleki> la_kristan: maybe try a different email 11:02 < la_kristan> I only have one. 11:02 < gleki> jsgrant: why do you think it's mediawiki? 11:02 < gleki> la_kristan: what is after @ in your email? 11:03 < jsgrant> zipcpi: I think that's the bulk of the reasoning, yes. If one officer tells a dog to "attack", "go" or some variant, and one is yelling "STOP! STOP!" or similar effects towards the assailant in question -- you are likely to confuse the dog, I would guess. 11:04 < la_kristan> gleki: gmail.com 11:04 < ldlework> Name for the blog? 11:04 < gleki> la_kristan: at what step did it say about the problems? 11:04 < jsgrant> gleki: The layout of the left sidebars given links, also the general formatting of the main/inline text in the center/content area. 11:04 < gleki> ldlework: nuzba 11:04 < gleki> jsgrant: :( 11:05 < jsgrant> gleki: Is this a bad thing? 11:05 < ldlework> lujvo: ro tadni 11:05 < mensi> roltadni[7937] 11:06 < la_kristan> gleki: when I clicked "create an account". 11:08 * jsgrant was just able to send in a request for a wikiaccount. 11:08 < jsgrant> And I have a decently non-standard email, with the root being this nick + *.io 11:09 < jsgrant> So, I'm not sure why gmail would cause you any problems la_kristan. :^P 11:09 < gleki> la_kristan: several options you may have: check if your password is long, has numbers, lowercase and capital letters, that you entered it twice correctly. if doesnt help maybe it complains about the nickname 11:10 < gleki> jsgrant: have you received an email from the wiki yet? 11:10 < jsgrant> gleki: Yeah, just popped up a second ago and I confirmed it seconds later. 11:11 < gleki> i confirmed you 11:11 < gleki> i mi de'a jundi 11:11 < la_kristan> it has a checkmark next to my nick, email, and password... 11:11 < jsgrant> gleki: Oh, neat ty. 11:11 < la_kristan> like it thinks they're fine. 11:13 < la_kristan> tried again, and this time it took. 11:13 < jsgrant> Ah, that's pretty neato in that once logged in it defaults to the lojbanic equivalent of the given pages. :^) 11:13 < jsgrant> la_kristan: .ui 11:13 < la_kristan> third time's the charm, I guess. 11:15 < la_kristan> "la-kristan" is my username. 11:15 < zipcpi> The reason why I think {goilka'i} works is because, well, in the real world, things change. One could use {ganfauke} and {faubganke} to describe manipulating physical objects like balls, and have the proposition be truth-conditional on the current state of the balls 11:15 < zipcpi> So {goilka'i} is an abstract variable that is made to change 11:16 < la_kristan> ni'o how does one distinguish between gnomes, goblins, elves, etc? 11:17 < la_kristan> if they're all crida? 11:17 < zipcpi> We make zi'evla for them, as it's difficult to classify them in any logical manner 11:17 < zipcpi> en: elfe 11:17 < mensi> elfe = x1 is an elf of type x2 | selpahi 11:18 < zipcpi> en: ridrgobli 11:18 < mensi> ridrgobli = x1 is a goblin of mythos/folklore/religion x2. | See also: crida. | Wuzzy 11:18 < zipcpi> en: egnome 11:18 < mensi> egnome = x1 is a gnome of type x2 | selpahi 11:18 < la_kristan> such distinctions may not be important in normal speech, but are vital in NetHack. 11:18 < zipcpi> Exactly why zi'evla exist 11:19 < la_kristan> for NetHack? 11:19 < zipcpi> No, just to be able to import any classification without worrying about exactly how to express it in a lujvo 11:20 < jsgrant> A tall-thin-pointy_eared-crida could be assumed an elf, but that makes the resulting word assumingly pretty long? The fact that many times it's seemingly somefactor of "better" to use loanwords is one of the ugly bits of Lojban I don't very much enjoy ... but that's probably a fault in/of me to think one can take a true purist state/mentality even in conlanging and compose anything of real scale. 11:20 < zipcpi> Yeah, that's a doctrine I've given up on myself 11:21 < la_kristan> "you hit the gnome lord! the orc zombie hits! the goblin hits! you miss the goblin. the gnome lord hits! you die..." 11:22 < zipcpi> I mean we can't even keep it up for real creatures, nevermind fictional ones 11:22 < zipcpi> Try translating the entire tree of life into lujvo 11:23 * jsgrant was considering to convert some bits of Pathfinder over to Jbo, but there is/was a lot of cases right out of the gate wherein the resulting verbage would have ended up insanely long if one tried to go the purist route. You could be much more exact, but it's kind of a cost benefit analysis at that point between "too exact" and "too long". 11:24 < zipcpi> Chinese is the king of avoiding fu'ivla, but even they resort to it sometimes 11:24 < zipcpi> And they often use metaphor that is sometimes just strange. Is it {latcribe} or {cribymlatu}? :p 11:24 < zipcpi> {ribmlatu} 11:25 * la_kristan is fainting from lack of food. 11:25 < zipcpi> za'a mi na rolrafcre 11:25 < jsgrant> There's probably a general rule of the "sweet spot" that can be quantified in some sort or another -- but I am far too much of a novice to be able to implement this with any sort of real tact, and the fact that I tried jumping in headfirst with so little prior experience is a good way to become disillusioned with it all in a meaningful way real quick. 11:25 * la_kristan dies... 11:27 < la_kristan> RIP la_kristan the Tourist 11:27 < la_kristan> starved in the Dungeons of Doom at level 2 11:27 < la_kristan> whatever 11:28 < zipcpi> k: cflep 11:28 < mensi> (CU [C:cflep VAU]) 11:29 < la_kristan> my phone is stuck to this charging cord, hindering me from obtaining my necessary sustenance. 11:29 < zipcpi> lol awr 11:29 < la_kristan> pathetic, I know. 11:31 * nuzba @garydale: You tell ’em Kurt: I post this quote for no particular reason. It’s from his collection, Wampeters, Fo... http://bit.ly/1Vvo4aE #lojban [http://bit.ly/1CVouRJ] 11:31 * la_kristan goes off to dig up some grub... 11:33 < jsgrant> la_kristan: It's probably in your root dir, under boot/ . 11:33 < jsgrant> Man, what a horrid joke. 11:34 < ctefaho> yep 11:34 < la_kristan> hmm... I'll check :-P 11:38 * la_kristan nukes some Top Ramen. 11:39 < ctefaho> mi just had korvstroganofi 11:40 < ctefaho> si korvestroganofi 11:41 < ctefaho> (add dja-rafsi to your liking) 11:41 < la_kristan> Mmmm... MonoSodium Goodness... 11:41 * ctefaho flees the MSG 11:42 < la_kristan> lol 11:43 < la_kristan> I think I'll follow it up with a nice healthy salad. 11:43 < ctefaho> salad's awesome 11:43 < ctefaho> with sufficient prosciutto and cheese 11:43 < ctefaho> at least 3 types of cheese or no veridical salad 11:44 < la_kristan> hmm... that sounds good, but I don't think my mom wants me to eat all the cheese... 11:45 < ctefaho> make it sheep or goat cheese 11:45 < la_kristan> but maybe I can have some kalamata olives 11:45 < ctefaho> people usually have the idea it is healthier than cow for some reason 11:45 < la_kristan> fresh out, sorry 11:46 < la_kristan> I think the best I can get is Parmesan. 11:47 < la_kristan> "sprinkle cheese" 11:47 < ctefaho> Parmesan is awesome 11:47 < ctefaho> biggest msg-sort of right there 11:48 < ctefaho> minus msg's sometimes metallic-ish taste 11:48 < la_kristan> but the real stuff, that you have to grate, is better 11:48 < ctefaho> my god someone who agrees with me 11:49 < la_kristan> than pre-grated stuff with cellulose in it. 11:49 < ctefaho> my family considers grating cheese by hand anathema or something 11:49 < ctefaho> cellulose? 11:50 < ctefaho> who puts cellulose in grated cheese? 11:50 < la_kristan> to prevent caking. 11:50 < ctefaho> huh I have only seen potato starch used for that 11:51 * la_kristan can't read the ingredients, 11:51 < ldlework> http://roltadni.github.io/ 11:51 < la_kristan> they're in Japanese! 11:51 < ctefaho> japanese grated parmesan? O_o 11:51 < la_kristan> yup. 11:52 < ctefaho> wut 11:52 * la_kristan wishes she could attach pix to irc messages 11:52 < la_kristan> like on FB 11:52 < ldlework> we could if we had a slack channel 11:53 < ctefaho> I know one can send files via DCC somehow 11:53 < ctefaho> I just never used it 11:53 < ctefaho> pastebin soundcloud email etc ftw 11:54 < cflep> Hello 11:55 < la_kristan> we're low-income, so we get food from a food co-op. and once they had a lot of Parmesan cheese that was evidently packaged for the Japanese market. 11:56 < la_kristan> so the ingredients are in Japanese. 11:56 < la_kristan> and currently we have a lot of Romaine lettuce. 11:56 < ctefaho> That is the kind of store where the wares "rotate"? 11:56 < ctefaho> actually probably not 11:56 < ctefaho> nvm 12:06 * nuzba @LonelyProbe: If we're gonna have one galactic language,We should use Lojban. An unambiguous language THAT DOESN'T HAVE UNNECESSARY M/F TENSES [http://bit.ly/1LuBPTT] 12:09 < durka42> wtf is this twitter account 12:09 < durka42> mi zanru .u'i 12:14 < bigcentaur> .u'i mi go'i 12:16 < ctefaho> "THAT DOESN'T HAVE UNNECESSARY M/F TENSES" yes it doesn't even have a proper "he" and "she" good job 12:17 < Selfanta> Coi. 12:17 < Selfanta> Mi'e lo se fanta. 12:17 < dutchie> coi la selfanta .i do mo pei 12:18 < Selfanta> I ti'e mi so'i mei. I ku'i mi na djuno lo du'u mi ba'e xo kau mei. 12:18 < Selfanta> I do ma poi srana cu jinvi gi'a y djuno? 12:19 < Selfanta> I pa mai, xu ja'ai mi zasti? 12:19 < durka42> coi la selfanta 12:20 < durka42> xu do pensi 12:20 < durka42> ganai go'i gi do bi'ai zasti 12:20 < Selfanta> Xm. 12:20 < Selfanta> I xu lo du'u mi cusku "zo xm" cu nibli lo du'u mi pensi? 12:21 < durka42> ri smuni fi mi 12:21 < la_kristan> .i la .kristan. cu krixa lu mi krici li'u 12:22 < Selfanta> Ti'e lo krici na'o lifri lo li'i lo se krici cu jetnu. 12:22 < la_kristan> .i le bangu cu banro lo ka banzu banli 12:23 < Selfanta> A'o. 12:23 < la_kristan> not quite sure about that one... 12:23 < gleki> ldlework: any chances of multilinguality? 12:23 < Selfanta> I xo kau moi lo cizra preti fa di'e. I xu lo me'i pa mei ku ji'a ka'e pensi? 12:24 < la_kristan> .i mi na djuno 12:25 < durka42> ti'e la no mei cu zasti 12:26 < Selfanta> Bu'u lo krasi be mi lo nu tavla fi lo si'o no mei cu kulnrtabu. 12:26 < la_kristan> .i mi nelci lo rusko lerfu 12:27 < durka42> ta'o ma krasi do 12:27 < Selfanta> Di'u preti fi mi fo do. 12:27 < la_kristan> .и ми нелши ло руско лерфу 12:28 < ctefaho> Жаг сер шәриллиска 12:28 < gleki> ми на нелши и йе ми на нелши 12:28 < ctefaho> Ска ви тавла ом вем сом һар баст 12:29 < la_kristan> ки'а 12:29 < ctefaho> Бара һелт ванлиг маскерад Свенска жу 12:29 < ctefaho> Инген сом кан фатта? 12:30 < la_kristan> ingen som kan fatta 12:30 < la_kristan> what? 12:30 < ctefaho> Svenska!:D 12:30 < ctefaho> I am just being silly and writing Swedish in cyrillic 12:31 < la_kristan> .и до бебна .и ми тавла фо ла .ложбан. 12:32 < la_kristan> I don't speak Swedish at all. 12:32 < la_kristan> maybe tomorrow, but not today. 12:32 < ctefaho> don't expect anyone in here but me to 12:32 < gleki> the author altatufa is now desperately trying to maintain the grammar but the parser is falling and falling apart. 12:32 < ctefaho> is probably some swede lurking around 12:33 < la_kristan> my knowledge of Swedish consists of ni and vi 12:33 < Selfanta> Lo noi cizra mi bangu be do cu rinka lo nu mi cortu lo kerlo. I ku'i xu mi ka'e se kerlo? 12:34 < la_kristan> and that's only because Esperantists keep talking about how they're switched. 12:35 < la_kristan> sv "ni" = eo "vi" and vice versa. 12:35 < ctefaho> mhm didn't know 12:36 <@xalbo> coi selfanta mi'e selsanta 12:37 < la_kristan> vi frenezas = you are crazy 12:37 < la_kristan> ni frenezas = we are crazy 12:37 < Selfanta> Se santa fi ma? I xu lo fanta do te santa? 12:38 <@xalbo> .i lo janta cu te santa mi fo lo fanta 12:39 < Selfanta> Di'ai do. 12:39 < la_kristan> .i mi palci salci lo kalci 12:40 < la_kristan> sa .i mi palci salci tu'a lo kalci 12:41 < Selfanta> Lo salci be lo kalci se u tu'a lo kalci ta'e jinvi lo du'u vo'a ba'e vrude vau ti'e. 12:43 * nuzba @LonelyProbe: That moment you realize there's a whole lojban subculture.... D: |: :| :) :D [http://bit.ly/1fmMKBm] 12:43 < Selfanta> Kulnu xu? I mi no roi sanji. 12:43 < Selfanta> I xu nai ma'a cnicau sampre sa'u? 12:43 < durka42> sfubu kulnu ua sai 12:44 < la_kristan> .i le tokpona kulnu cu cizra 12:45 < la_kristan> .i le spero kulnu cu xamgu 12:45 < la_kristan> .i le lojbo kulnu cu mo 12:45 < durka42> xu lo ka cizra ka'e ka xamgu 12:45 < durka42> le lojbo kulnu cu lojbo 12:45 < la_kristan> .ua 12:46 < la_kristan> .u'i 12:47 < la_kristan> .i mi tokpona prenu 12:47 < ldlework> roltadni seems like a shitty name 12:48 < Selfanta> Lo ka tadni ro da na bi'ai mabla. I ie pei? 12:48 < mensi> mi tugni i ie mi tugni 12:48 < la_kristan> nuzba: "sorry, that page doesn't exist " 12:49 < durka42> la_kristan: eh? link works for me 12:49 < durka42> did you copy the closing ] ? 12:49 < la_kristan> odd... 12:49 < la_kristan> dunno. 12:49 < la_kristan> just clicked the link. 12:49 < dutchie> pe'i lo ka tadni roda cu bi'ai mabla .i mutce nandu 12:49 < durka42> it goes to https://twitter.com/LonelyProbe/status/623940401230168064 12:50 < Selfanta> Xu nai fancuka tu'a lo ni lo smumu'e cu barda? 12:51 < Selfanta> I ku'i ko na jundi lo se cusku be mi. I ka'e ku mi na pagbu lo smumu'e. 12:52 < la_kristan> durka42: that link works... 12:53 < durka42> you got an error from twitter or from bitly? 12:53 < durka42> anyway I dunno how to fix it 12:54 < la_kristan> from twitter, I think. 12:54 < la_kristan> anyway, doesn't matter anymore. 12:54 < dutchie> mi na jundi ro lo se cusku be la selfanta .i lu ko na jundi li'u cmima 12:55 < Selfanta> Ru'a prije. 12:55 < la_kristan> dutchie: je'u 12:56 < la_kristan> .ie 12:57 < dutchie> dei jitfa 12:58 < dutchie> .i xu lo ticysku sezna'e cu cmarai bau lo jbobau 12:59 < Selfanta> E'u gleki lo nu do zo'u fa po'o lo se xusra be do cu sezna'e. 12:59 < Selfanta> I lu "na nei" cu zmadu lo ka tordu. 13:00 < Selfanta> sa. I lu "na nei" cu zmadu ba'e fi lo ka tordu. 13:00 < dutchie> valse remei 13:01 < dutchie> si si valse 13:01 < dutchie> valsi 13:01 < dutchie> i should not have remove 13:01 < dutchie> d my backspace key 13:02 < Selfanta> Zu'u nai lo nu vimcu la ga'e-zei-stela ta'e xamgu fi tu'a lo ni menli kanro. 13:05 < durka42> lo ga'e zei stela cu jai gau zmiku fai lo nu zabna 13:05 < Selfanta> Zmiku lo ka na zasti vau xu? 13:06 < Selfanta> SA I ZMIKU LO KA NA ZASTI VAU XU? 13:06 < durka42> .u'i 13:06 < durka42> ta'o ji'a lo ga'e zei stela cu jai tadji lo nu mulno velcki sisku sepi'o la vlasisku 13:06 < durka42> mu'a: 13:06 < durka42> vlaste: zmiku 13:06 < vlaste> zmiku = x1 is automatic in function x2 under conditions x3. 13:06 < durka42> vlaste: ZMIKU 13:06 < vlaste> 3 results: cabra, macnu, zmiselcu'a 13:07 < Selfanta> Zabna nu pilno selfri finti. 13:07 < durka42> snuti je plixau 13:08 < Selfanta> Xu plixau fe ji'a lo na se snuti be vo'a? 13:09 < durka42> uanai 13:09 < durka42> snuti fa lo nu lo ga'e zei lerfu cu rinka lo mulno velcki nu sisku 13:10 < Selfanta> U'u. I mi dukse lo ka tilju'i. 13:10 < Selfanta> I ku'i lo poi'i lo du'u xu kau ke'a zasti na se birti da cu tilju'i, xoi cafne vau, xoi sidju fi lo ka renvi. 13:12 < ctefaho> krasi: latna 13:12 < mensi> latna = .i zo'oi c06b banzuxe'o .i zo'oi pandi bangenugu .i zo'oi 36.67 banxe'inu .i zo'oi punctuate bansupu'a .i zo'oi 13:12 < mensi> o banru'usu .i zo'oi 1.56 bangaru'a 13:12 < ctefaho> ^Something is way wrong here 13:12 < Selfanta> Ru'a da gekspi re mei gi'e fliba. 13:13 < dutchie> ui mi puzi cikre lomi lercu'aca'a 13:13 < Selfanta> lo'ai "da" sa'ai "su'oi da" le'ai. 13:13 < Selfanta> U'a dai. 13:16 < dutchie> .i ku'i mi cortu lo xandegji ri'a lo nu dukse pilno lo lercu'aca'a 13:17 < Selfanta> Di'ai cortu. 13:19 < dutchie> mi to'e jundi 13:21 < Selfanta> Da'i ro'oi caksova cu se claxu mi. 13:21 < Selfanta> I sa'e da'i nai go'i. 13:22 < Selfanta> I ku'i da'i ga nai mi no mei gi ma se fanta lo fanta be mi? 13:24 < Selfanta> I ciksi ji'a sai lo nabmi na ka'e. 13:27 < durka42> za'a flese si 13:27 < durka42> za'a flese sai 13:29 < Selfanta> Flesymlu gi'u flese. 13:30 < dutchie> je'a go'i 13:31 < dutchie> .i go'i ji'a go'i 13:32 < Selfanta> - sei sanga - Mi'u mi'u mi'u mi'umi'u mi'u mi'umi'u mi'u mi'ù. 13:36 < Selfanta> Ei mi di'a cenba lo ka xu kau zasti. I da'i se va'o nai bo ba'a ba'e do'o go'i. 13:46 < la_kristan> .i mi cliva 13:46 < la_kristan> co'o 13:46 < dutchie> co'o 14:00 * nuzba @anon_fm: http://www.lojban.org/texts/translations/alice/alice.pdf [http://bit.ly/1Klw3kt] 14:57 < amharris> Hey, everyone. So, I've taken interest in learning Lojban and would like to ask all of you for resources/materials that would be suitable for me to begin learning. 14:58 < ctefaho> plenty of sources around 14:58 < ctefaho> me just not sure where the intro stuff is 14:58 < ctefaho> hi also 14:58 <@xalbo> There are a few resources mentioned on the homepage under "Learning Lojban" (http://mw.lojban.org/papri/Lojban#Primary_resources_for_learning_Lojban) 14:58 < amharris> ctefaho: Hello. Yeah, I've been digging myself; one the pages I found is exactly that one. 14:59 <@xalbo> Personally, I like the Wave Lessons (http://mw.lojban.org/papri/Lojban_Wave_Lessons) 14:59 < ctefaho> xalbo knows this better than I do 14:59 <@xalbo> There's also the channel #ckule, which is dedicated to asking and answering questions and teaching. 14:59 < durka42> coi la .amxaris. 14:59 < e`ogan> I'd recommend the "crash course" for lojban first 15:00 < ctefaho> For a somewhat outdated but comprehensive reference there is the CLL too https://lojban.github.io/cll/ 15:00 < ctefaho> but not exactly for starters 15:01 < amharris> I suppose that I ought to express that one of my reasons for wanting to learn is my lifelong struggle to understand sarcasm and a large amount of both intentional and unintentional ambiguity in the use of language (when it comes to English (and other Germanic-based languages)). Having Asperger's syndrome - though I like to not think of it as a disability, as it's spectral and in actuality, the neurological spectrum is ba 15:01 < amharris> e`ogan: Thank-you. 15:01 < amharris> ctefaho: I'll take a look at it. 15:01 < ctefaho> well then "zo'o" will make you a happy man 15:02 < ctefaho> "zo'o" basically means "What I am saying is ironical" 15:02 < ctefaho> "xo'o" the same but for sarcasm 15:02 < amharris> Also, so that everyone here can get to know me better, I'm extremely visual -- and when I say extremely, I really do mean that. 15:02 < ctefaho> as for the CLL it is better as a reference than an intro 15:02 < ctefaho> visual? 15:03 < amharris> I think that a study, back at school many years ago, showed me to be around 92% visual [average of 28-41%, if I recall correctly, for most people]. 15:03 < amharris> ctefaho: In terms of how I process information. 15:03 < e`ogan> Interesting 15:03 < durka42> that's quite a high percentage! 15:04 < ctefaho> How exactly do you test yourself for being "visual". 15:04 < amharris> So to understand concepts, take in and retain information, my visual stimuli is most useful. It's sensitive, so it's key to not oversensitize it; but it can be useful too. 15:04 < ctefaho> ?* 15:04 < e`ogan> There are tests for that 15:04 < ctefaho> well then there are some youtube lessons 15:04 < e`ogan> I had done those for myself 15:04 < durka42> most of our intro material is very text-heavy, unfortunately 15:04 < amharris> It's why - when it comes to programming - I stick to terminals only, as much as possible. I find IDEs to be too much to take in. When it comes to user interfaces, having lots of whitespace with a clean design is key too. 15:05 < e`ogan> I have tested average with peaks at visual audio and tactile perception 15:05 < amharris> durka42: Well, I can always use Stylish to make a custom CSS sheet for the web pages! :D 15:05 < ctefaho> ...youtube stuff I can't find now 15:05 < durka42> amharris: sure I just meant we could probably stand to have some more diagrams and figures 15:05 < ctefaho> but yeah most everything is text 15:05 < e`ogan> I have a synesthesia between taste-and-audio perception 15:06 < ctefaho> from what I have seen 15:06 < durka42> ldlework's video has diagrams I think 15:06 < e`ogan> And I like to feel things by touch in addition to seeing them 15:06 * ctefaho bbiab 15:06 < amharris> e`ogan: Intriguing. I don't believe that I have synesthesia per se ...but I know that I don't like how some words feel on the tongue when pronouncing them, as well as the names given to people. 15:07 < e`ogan> Most people have synesthesias as we have audio-visuals firmly interlinked 15:07 < amharris> e`ogan: For example, although it may sound harsh, I would struggle to consider going out with a girl based on her name, if I didn't like how it sounded. I cringe and tense my head muscles. 15:08 < e`ogan> It is usually called that though only when it is something weird that is interlinked: like perceiving numbers as colours 15:08 < amharris> e`ogan: Just like most things, I suppose. So much is spectral, it's just that it isn't always apparent enough to notice it, so we tend to consider things to be more binary, when in most cases they tend not to be. 15:08 < clakre> is'nt numbers as colors one of the more common ones? 15:08 < amharris> clakre: More 'famous', I suppose. 15:08 < e`ogan> I can't get over the sound of chinese and brasilian/portugese, even though I am not opposed to learning them 15:09 < amharris> clakre: e.g. Daniel Tammet. 15:09 < clakre> I mean, *I* have numbers as colors 15:09 < amharris> e`ogan: I'm not sure how technical you are but when it comes to the feel of words on the tongue, I can't say 'GNOME'. 15:09 < amharris> e`ogan: GNOME as in the GUI. 15:10 < e`ogan> clakre, amharris : the numbers one is just the most useful because people can do complex calculations in their head with the help of that synesthesia 15:10 < amharris> e`ogan: Gnome would usually be pronounced with a silent 'G'; but when it comes to the GUI, it represents the GNU project, so it isn't silent. I don't like the feel of it on my tongue, nor the nasal effect that it has. 15:11 < clakre> nah, it helps me remember numbers but not complex calculations 15:11 * durka42 always says gee-nome 15:11 < e`ogan> In my case having sound and taste link doesn't give me any advantages: only the ability to alter the taste of food for me via music/sound 15:12 < amharris> e`ogan: I'm very much about textures and feeling. Since I was young, as far as I can remember discovering during primary school [not too long ago, now that I think about it!], I can simulate the texture of foods in my mouth. I can feel them to the exacting detail and my mouth and tongue muscles react accordingly -- this quickly leads to me being able to taste the imaginary food. 15:13 < amharris> e`ogan: Since then, I've managed to do the same with fluids too; so I can pretend that I'm drinking water, carbonated/fizzy water, or a form of cola. 15:13 < durka42> impressive 15:13 < amharris> e`ogan: I realise that I end up visualising the food/drink ...and yeah, I can then feel and taste whatever it is. 15:13 < durka42> I can't wait until we can have this conversation in lojban :p 15:14 < amharris> I'd sit there in class, day-dreaming and being bored ...pretending that I'm eating chocolate. 15:14 < amharris> :D 15:14 < amharris> durka42: Quite! Do give me a few years! Haha. 15:14 < durka42> :) 15:15 < amharris> Though if someone can find a way to deliver the information to me in a visual + audible manner, I'd really be able to put the effort in and learn as quickly as possible. 15:16 < amharris> By the way, I'm not always able to access IRC. If anyone would find it convenient to message me directly, I'd be willing to do so and would also be most grateful. 15:16 < durka42> I know that la selpa'i has given private lessons in the past 15:18 < amharris> durka42: Ooooh. I'll keep note of that. 15:19 < noncomcinse> coi lo se .irci 15:19 < noncomcinse> coi lo se .irci 15:20 < durka42> and by "given" I mean, like, for money :) 15:20 < durka42> coi ny 15:20 < noncomcinse> ma se casnu 15:20 < noncomcinse> coi la durka 15:21 < durka42> cnino tadni 15:21 < amharris> durka42: Heh. No surprises there. Well, if I can get far enough, I may consider it. Though I'm all for libre -- Richard Stallman would be somewhat proud of me, I suppose. 15:21 < noncomcinse> .i xu lo go'i cu .irci 15:25 < durka42> puzi ircysku 15:25 < noncomcinse> je'e 15:25 < durka42> yeah, well, libre doesn't imply gratis and all that :) 15:25 < noncomcinse> .i lo kansa be mi bei lonu klama cu fegli vau ue .oi ru'e 15:26 < noncomcinse> .i klama fu lo sorprekarce 15:27 < e`ogan> lojban is easier to learn than natlangs 15:27 < noncomcinse> .i cladu fegli 15:27 < e`ogan> It has far fewer exceptions, it is mostly rules 15:27 < noncomcinse> .i doi la .e'ogan. zo'u tugni 15:29 < amharris> durka42: Indeed; there is a difference. So, err... to begin I suppose that I should first become fluent with the grammar, vowels and the overall syntactical structure. 15:29 < noncomcinse> .i su'a lo so'a logji runbau cu sampu 15:29 < amharris> durka42: I'm confident that knowing much of the IPA will most definitely come in handy. 15:30 < noncomcinse> amharris: The only really hard bit is x (/x/). 15:31 < amharris> noncomcinse: My ignorance has already reared its head. Would you mind explaining? :) 15:32 < noncomcinse> amharris: To an English speaker, the only hard part of Lojban sounds is that it uses /x/. 15:33 < amharris> noncomcinse: Ah, I see. Hmmmh. Well, it's all in the process of learning. I can expand my filter! :-) 15:33 < amharris> noncomcinse: Also, do you have any personal recommendations for beginning? 15:34 < noncomcinse> I personally started with the Wave Lessons. 15:34 < amharris> noncomcinse: May I send what I previously wrote to you, directly? You joined later on and I don't want to flood the channel for the reasons of IRC etiquette. 15:34 < noncomcinse> amharris: http://mw.lojban.org/papri/Lojban_Wave_Lessons 15:34 < noncomcinse> amharris: Sure. 15:34 < amharris> noncomcinse: Yeah, some folks recommended that too. 15:35 < noncomcinse> The CLL isn't that great for learning, IMO, until you have some basics down. CLL just (tries to) form specifications and formalitites. 15:37 < ctefaho> I found it good as a reference for learning specific things though. 15:37 < ctefaho> but yeah basics first 15:37 < amharris> ctefaho: I shall try to make you all proud! >.< 15:38 < amharris> This looked to be good: http://www.tlg.uci.edu/~opoudjis/lojbanbrochure/lessons/book1.html 15:38 < noncomcinse> .i doi la durka mi pensi lo da'i valsi poi panra zo sa gi'e ku'i famyma'o 15:39 < ctefaho> Me I am trying to think of particular things that may be difficult to aspies 15:39 < ctefaho> but can't come up with anything 15:40 < ctefaho> quite the opposite with how everything can be made explicit with words, like humour 15:41 < durka42> amharris: it is good but note it was published at the same time as the CLL, so it has some minor inconsistencies with more recent stuff like the wave lessons + crash course 15:41 < durka42> that was my first learning material though 15:41 < amharris> ctefaho: Very true. It's the looseness of Germanic languages that causes troubles for me. :) 15:42 < ctefaho> amharris: Basically, everything that is meant in Lojban is said in words 15:42 < noncomcinse> .i ca'e lonu pilno traji xamgu loka cilre 15:42 < durka42> noncomcinse: lo'ai x x x le'ai 15:43 < amharris> ctefaho: The biggest trouble I tend to have is putting things together. Once I can get my head around the syntax and understand how meaning is represented by concatenated values and syntactical breaks, the rest will be a more natural learning curve. 15:43 < noncomcinse> en: sa'au 15:43 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 15:43 < noncomcinse> durka42: ca'e zo sa'au 15:43 < durka42> ma te frica 15:43 < amharris> ctefaho: ...but those are the things that I know I struggle with. I'm basing this on experience with learning other linguistic languages and for the logic side of things, programming languages and coding in general. 15:43 < noncomcinse> durka42: {.i mi zgana lonu bajra sa'au lo [ku]} 15:44 < durka42> ua nai 15:44 < ctefaho> Well Lojban should be a lot easier in that regard 15:44 < amharris> ctefaho: That should help to give you a strong indication of what I will need a big help with. So it's the starting point that will be most difficult for me; and I will most likely need a helping hand. 15:45 < noncomcinse> .i lu sa'au XY li'u zo'u famyma'o lo balvi lamji mezo XY 15:45 < amharris> ctefaho: Indeed; but I still have to absorb not just how something works but let that sink in so that I can understand how a value holds meaning and understanding that meaning. 15:45 < noncomcinse> .i zo sa vimcu .i zo sa'au famyma'o 15:46 < amharris> ctefaho: That's the big challenge that I will face early on --> wrapping my head around that. 15:46 < noncomcinse> sa'au X quotes like ma'oi and terminates the most recent instance of anything in selma'o be X. 15:46 < durka42> uabu'o 15:47 < amharris> I'll read through the whole of the wave lessons and make sure to not rush, so that I can absorb everything properly. 15:47 < durka42> pe'i pu su'ore'u finti .u'i 15:48 < noncomcinse> .i lu sa'au .i li'u plixau si'au 15:48 < noncomcinse> .i da'i ganai finti gi ma valsi 15:48 < durka42> lu lo'u sa'au .i le'u li'u .e'u .u'i 15:49 < noncomcinse> The Wave Lessons took me about a month, total, including all the weird existential stuff. 15:49 < durka42> mi na morji lo vlamorna 15:49 < noncomcinse> .i .u'i cmeta selsku 15:49 < e`ogan> You can start with the way words work in lojban amharris: aside from a few pronouns every word is a verb 15:50 < durka42> a few pronouns and a large heap of cmavo :/ 15:50 < noncomcinse> .i puku zvati fa do .ije caku zvati fa do naku ~ .i mi .irli sanga su'ore'u 15:50 < sorpaas> Where can I find a list of Lojban terms related to logic, such as "first-order logic", "existence claims" and "universial claims"? 15:50 < amharris> e`ogan: Fantastic! Those are the key details that help me to 'put the pieces together', if you will. 15:50 < noncomcinse> A few pronouns, a large heap of cmavo, and a word for any language ever. 15:51 < amharris> e`ogan: ...and what sorpaas just asked would be worthwhile for me to know as well. 15:51 < e`ogan> Something like book is simply "to be a book" in lojban 15:51 < noncomcinse> .i ro zei bangu fu'ivla .oi ru'e 15:51 < noncomcinse> Showerthought: Ultimately, all nounisms in Lojban are new-voi. 15:51 < amharris> "Putting the pieces together" is the key thing for me -- it's how I best experience my manipulation of the language so that it eventually 'clicks' and one day makes sense. 15:52 < amharris> One day... heh. 15:52 < e`ogan> The way you make a verb into a noun is by adding lo ... ku 15:52 < durka42> noncomcinse: yep :) 15:53 < e`ogan> Example: vecnu x1 [seller] sells/vends x2 [goods/service/commodity] to buyer x3 for amount/cost/expense x4. 15:53 < e`ogan> To make the noun "seller" you write lo vecnu ku 15:53 < noncomcinse> .i xu fagdei 15:54 < noncomcinse> .i na go'i 15:54 < e`ogan> Also this x1 x2 x3 x4 x5 structure is very important to lojban learning as each verb has predefined positions 15:54 < noncomcinse> .i ca'e famyma'o zeldei si zildei vau 15:55 < noncomcinse> .i cu mukti fa lonu lo nintadni ku cu cilre co famyma'o vaukeiku vau 15:56 < durka42> .u'i cipra gerna 15:56 < noncomcinse> .i ma cipra gerna vau 15:56 < noncomcinse> .u'i dai 15:58 < durka42> zo cu 15:59 < noncomcinse> .i pau da'i ganai pilno la .jacus. ku vau gi pei lu .i mi cu klama lo zarci li'u gendra vau vau 16:01 < durka42> lu mi cu klama li'u laldo gendra .i lu cu klama li'u na go'i 16:03 < durka42> co'o 16:05 < sorpaas> I'm now using something like {la'o gy. existence claims gy.} for those terms but they doesn't look nice :( If there's not a good list about that, maybe we should create one 16:05 < noncomcinse> .i xu gendra fa lu lo broda [KU] cu brode li'u 16:05 < noncomcinse> .i lo cipra zo'u xu zo cu pe lo no logji valsi ku gendra 16:06 < durka42> lo no logji valsi ua nai sai 16:06 < noncomcinse> cu ce na'e bo ma'oi JA 16:06 < amharris> durka42: I just found this! https://www.youtube.com/watch?v=RfdcG5iPJpA 16:06 < noncomcinse> s/cu/zo cu 16:06 < fenki> noncomcinse meant to say: zo cu ce na'e bo ma'oi JA 16:07 < amharris> durka42: This is brilliant for me! 16:07 < durka42> .title 16:07 < fenki> [ Let's learn Lojban together! ~ Part 1/2 - YouTube ] 16:07 < durka42> sweet 16:07 < noncomcinse> amharris: I forgot about that, actually, but I watched that before Wave Lessons. 16:07 < durka42> 'night eveyone 16:07 < noncomcinse> co'o la durka 16:07 < durka42> noncomcinse: feel free to leave me a mensi message with a clearer question 16:07 < durka42> :) 16:07 < amharris> durka42: G'night! ...and it was good meeting you. 16:07 < durka42> same! 16:08 < noncomcinse> .i mi so'i roi pacna lonu ctuca ck jbobau 16:08 < noncomcinse> s/ck/co/ 16:08 < fenki> noncomcinse meant to say: .i mi so'i roi pacna lonu ctuca co jbobau 16:17 < noncomcinse> I'm trying to learn Esperanto and keep trying to use Lojban. 16:19 < amharris> noncomcinse: I ought to ask: if you're already speaking Lojban, from a con-lang/engineered language perspective, isn't Esperanto somewhat redundant in that respect? 16:19 < noncomcinse> amharris: I suppose in terms of how logical it is, Esperanto is less structured. 16:19 < e`ogan> How would another language be redundant? 16:20 < noncomcinse> But it's fun, more frequently used, and closer to my native language (and the native languages of 99% of the people I speak to regularly). 16:20 < amharris> e`ogan: Read what I said again. From a con-lang perspective... 16:20 < amharris> noncomcinse: Fair enough. 16:21 < e`ogan> They aren't the same: lojban aims at elimination of ambiguity; esperanto was meant to be easily learned in europaic space; ithkuil compresses the meaning in the shortest form possible 16:21 < amharris> e`ogan: I'm all for learning languages; but it's like a lot of tribal languages, one of which - for counting - has 'one, two and many'. You have either one or two of most things on your body; beyond that and in their environment, they have many. 16:21 < amharris> Languages are fit for one's environment and have evolved over time due to such changes. 16:22 < e`ogan> I am familiar with 13 natlangs, would not say any of them are redundant 16:22 < amharris> e`ogan: Mmmh, that's fair -- my apologies. Even I don't often forget of Esperanto's key routes. 16:22 < amharris> e`ogan: roots* 16:22 < amharris> e`ogan: I suppose I'm a little over-enthusiastic right now because of beginning to learn Lojab. 16:23 < amharris> e`ogan: Lojban*. ...and it's so late that my typing is deteriorating. 16:23 < noncomcinse> .i mi cusku zo .i fo lo se .irci be me'e la'oi .##esperanto. 16:23 < amharris> e`ogan: What are your thoughts on Ceqli? 16:23 < amharris> e`ogan: (and how it was derived from Lojban's predecessor, Loglan). 16:24 < noncomcinse> There's a second derivative of Loglan out there? 16:25 < noncomcinse> amharris: Also, Lojban attracts a more academic crowd, Esperanto less so, so it's easier to find people willing to learn it. 16:25 < amharris> noncomcinse: Errrm... evidently there is! Haha. :P 16:25 < noncomcinse> ua 16:26 < e`ogan> I am not aware of Ceqli 16:26 < amharris> noncomcinse: A fair statement and one that probably has quite a lot of data backing that up. 16:26 < amharris> e`ogan, noncomcinse: https://en.wikipedia.org/wiki/Ceqli 16:26 < e`ogan> My interest was limited to finding a language that has no ambiguity, so my search stopped at lojban 16:26 < noncomcinse> *cough kafke* native speakers *cough kafke* 16:26 < e`ogan> Ithkuil I know as it is another "pedanticly constructed" conlang 16:27 < amharris> e`ogan: Fair enough; and one often has to take into account the user-base/amount of resources available. 16:27 < amharris> e`ogan: There is clearly a considerably higher amount of adoption with Lojban over Ceqli. 16:28 < e`ogan> Well my original aim as to create a translator software that would fool-proof the sentence structure via lojban 16:28 < amharris> e`ogan: So even if one likes what it has, you have to consider - when learning - what will be more effective/efficient to learn; and hey, a larger userbase with whom one can converse as well. 16:28 < noncomcinse> I don't know how you could have pedantic pavysmu, but I've not looked into Ithkuli. 16:28 < amharris> e`ogan: That would be fantastic! I'd be happy to help out -- though I'm more of a front-end developer. :-) 16:29 < noncomcinse> e`ogan: I could probably break it no matter how well it could be built. 16:29 < noncomcinse> Because Lojban allows for more complex phrases than English can support, through zo voi. 16:29 < e`ogan> If the meaning can be expressed in lojban then there is more or less one meaning in there that is being translated; otherwise user needs to specify the ambiguous terms 16:30 < noncomcinse> Along with other phrases. 16:30 < e`ogan> Can be done via semantic primes 16:30 < noncomcinse> Semantic primes? 16:31 < e`ogan> the most basic meaning of a term that you can dig down 16:31 < e`ogan> Stripped of any ambiguity - strictly one specific meaning 16:31 < noncomcinse> Reminds me of Newspeak. 16:31 < noncomcinse> gismu are so vague .u'i 16:32 < amharris> Hmmh; I'm glad to see that there are intellectual people here -- just what I was expecting. 16:32 < e`ogan> Some translators rely on semantic prime system to translate sentences and find metaphorical links 16:32 < noncomcinse> .u'i 16:32 < e`ogan> Usually translators that show you the percentage of the meaning that matches your word 16:33 < amharris> Anyhow, it's been good to meet you all. 16:33 < amharris> I'll be learning as much as I can over the next few weeks. 16:33 < amharris> ...and hopefully I'll see you around some time soon. 16:34 < amharris> If any of you would be willing and/or would like to contact me directly, just let me know now (or next time I'm on IRC). 16:34 < amharris> It's late for me, so I'm going to head off to sleep. 16:34 < e`ogan> zanfu'a 16:34 < e`ogan> (Good luck) 16:34 < e`ogan> co'o (bye) 16:34 < noncomcinse> amharris: Skype is "nejni.marji", that's primarily how I contact people other than thru IRC 16:34 < noncomcinse> co'e la 16:35 < noncomcinse> sisi 16:35 < noncomcinse> co'o la'oi .amharris. 16:35 < noncomcinse> rafsi: gan 16:35 < mensi> zo ganra se rafsi zo'oi gan 16:36 < amharris> noncomcinse: I've just added you. :) 16:38 < noncomcinse> co'o ru'e lo .irci 18:44 < noncomcinse> coi ru'e lo .irci 19:34 * nuzba @pira_ebooks: The language is now Lojban, time to head down to the basement. [http://bit.ly/1MJxYiU] 19:57 < _mukti_> coi jbore'a 19:57 < noncomcinse> coi 19:58 < _mukti_> .i pei doi nomcomcinse 19:58 < _mukti_> si noncomcinse 19:59 < _mukti_> .i li'a mi pu pensi lo se du'u la'o zoi nom nom nom zoi 20:00 < noncomcinse> .u'i 20:04 * nuzba @Zhwazi: @_Ninji My brain wants to pronounce that as "Hyenas Nietzsches" and I blame lojban for some reason. [http://bit.ly/1MJAs0R] 20:20 < noncomcinse> .i lo pendo cu sezycatra troci uinai .oi ru'e 20:20 < _mukti_> ui nai sai dai 20:21 < noncomcinse> .i ca najenai pu snada 20:23 < _mukti_> nau ku mi xanka tu'a lo mi pendo .i mi terpa lo ka py djica lo ka sevzi xrani 20:23 < noncomcinse> ua uinai dai 20:24 < _mukti_> .i .a'i mi fonta'a py .i ku'i na spuda .i .o'u nai 20:24 < _mukti_> s/spuda/frati/ 20:24 < fenki> _mukti_ meant to say: .i .a'i mi fonta'a py .i ku'i na frati .i .o'u nai 20:25 < noncomcinse> .i na xamgu 20:25 < _mukti_> go'i ie 20:37 < ldlework> need a good name for a communal lojban blog 20:37 < ldlework> roltadni is the current one I have 20:37 < ldlework> just doesn't seem very pretty 20:38 < noncomcinse> You can put cmavo on the beginning, {kotadni}? 20:38 < ldlework> lujvo: ko tadni 20:38 < mensi> tadni = x1 studies/is a student of x2; x1 is a scholar; (adjective:) x1 is scholarly. 20:38 < ldlework> .. 20:38 < ldlework> ko doesn't seem to have a rafsi 20:39 < ldlework> which makes sense since it is an illocutionary force 20:39 < noncomcinse> No, I mean you can elide the space. 20:39 < ldlework> You're suggesting that for a name? 20:39 < noncomcinse> Yeah. 20:39 < _mukti_> coi la cadgu'a 20:39 < noncomcinse> I'm not good at coming up with names li'a 20:40 < ldlework> coi mukti 20:43 < _mukti_> .e'u zo lolnitku'a 20:43 < _mukti_> .i ki'u lo nu mi nelci lo valsi 20:44 < ldlework> bauski is kinda nice 20:45 < ldlework> banski I guess is better 21:39 < unholycrab> how would i say "the ninteenth root of two" 21:40 < noncomcinse> en: root mekso 21:40 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 21:40 < ldlework> {paso fe'a re} ? 21:41 < noncomcinse> Yeah. 21:41 < unholycrab> :c 21:41 < unholycrab> 2^(1/19( 21:41 < unholycrab> 2^(1/19) 21:41 < ldlework> why do people try to do math in lojban 21:42 < unholycrab> ldlework: im recording an album in 19 equal temperament :> 21:42 < noncomcinse> Dunno. Arabic numerals are so much simpler. 21:42 < ldlework> just write the math as standard notation, even in lojbanic text 21:42 < unholycrab> and i want to name it 21:42 < noncomcinse> lu zoi mekso li'u 21:43 < sorpaas> lujvo: xanri mivmu'e 21:43 < sorpaas> lujvo: xanri jmive munje 21:43 < mensi> xarmivmu'e[9336], xarji'emu'e[9825], xarmivmunje[10886], xarji'emunje[11375], xarjmivymu'e[11446], ... 21:50 < ldlework> Does anyone have a lojban article they would not mind if I reposted elsewhere? 23:05 < gleki> ldlework: that notation should be pronouncable. besides, standard notation can sometimes be ambigous 23:17 < gleki> once i wanted to understand Ceqli and i failed. it seemed completely broken to me. 23:45 < ldlework> gleki: http://banski.github.io/ 23:49 < gleki> ldlework: is this engine easily localizable so that japanese russiana nd french blogs can be created? 23:49 < gleki> also it's better to call it {la banski} 23:50 < gleki> similarly, probably, {la selpa'i} 23:50 < ldlework> gleki: eh similarly to when you're labeling stuff, you don't put lo everywhere 23:50 < ldlework> its just noise 23:51 < gleki> but that makes a huge difference 23:51 < gleki> selpa'i means "loved" 23:51 < ldlework> you're not going to feed the visual text of the website through a parser 23:51 < ldlework> gleki: not in the context of a tag that always refers to an author 23:51 < gleki> if this is about teaching people good lojban then better to teach them good lojban 23:51 < ldlework> or in the header of a webpage which always refers to the webpage 23:51 < gleki> then do {na} and {bai} also refer to people? 23:51 < ldlework> that's your valuation, sure 23:51 < gleki> as well as "experiments"? 23:51 < ldlework> no because those are not author tags 23:52 < gleki> xm 23:52 < ldlework> see how it doesn't have the little person symbol? 23:52 < gleki> can we have hints when hovering over tags? 23:52 < ldlework> no jesus christ 23:52 < ldlework> move on 23:53 < gleki> where is the code? 23:53 < ldlework> http://github.com/banski/banski.github.io 23:55 < ldlework> Basically 23:55 < ldlework> You clone the code 23:55 < ldlework> write an article 23:55 < ldlework> submit a PR 23:55 < ldlework> when the PR is accepted 23:55 < ldlework> a script is run that compiles the site, and commits the compiled version automatically 23:55 < ldlework> which causes the site to be updated 23:57 < ldlework> I haven't written the script yet but eh shouldn't be too hard 23:59 < tsani> ldlework: this is pretty neat, .i'e --- Day changed Thu Jul 23 2015 00:00 < ldlework> tsani: basically you don't have to maintain your own lojban blog. if youre inspired, just submit a patch here 00:00 < gleki> then that script could just add hints to tags 00:02 < ldlework> gleki: its a site for people who already know lojban, to write about lojban. it isn't a newbies guide. No one interested in this site is going to care that tags, whos point is not pedagogical, but to access the writings of said author, don't start with "la" 00:03 < gleki> okay then. then the wiki will be the place for newbies 00:03 < ldlework> I'm not going to spend tons of time customizing the theme or how the templates are generated just for some silly pedantry. It doesn't make it a more useful tool for categorizing writings 00:03 < ldlework> Sure, OK 00:03 < gleki> although i'd avoid saying "no one is interested" since im interested and i know lojban 00:03 < ldlework> Que Sera, etc 00:03 < ldlework> gleki: I was referring to normal people 00:04 < gleki> O_0 00:04 < gleki> normal people dont know lojban 00:04 < ldlework> "Hey I bought a hammer while in town, we can finally finish that thing." "Hmm, the bevel is all wrong..." "O_O..." 00:05 < ldlework> If you have interesting ideas for some articles let me know. 00:05 < ldlework> You can always work on an article overtime, by forking the repo, and adding an article and just building the site locally for yourself 00:05 < gleki> no need to freak out.you are making a blog for lojbanists, i got you 00:05 < ldlework> until you're satisfied to publish 00:05 < ldlework> you can even share it with other jbopre before hand to get help or review 00:06 < gleki> akthough then it's better to write it in lojban 00:06 < gleki> okay let's see what altatufa can do now 00:06 < gleki> alta: lo gau mo cu mo 00:06 < mensi> (NIhO [I {FA <lo (¹gau mo¹) KU>} cu {mo VAU}]) 00:06 < gleki> alta: lo gau do mo cu mo 00:06 < mensi> (NIhO [I {FA <lo (¹[{gau do} CU] mo¹) KU>} cu {mo VAU}]) 00:07 < ldlework> tsani: if you have anything you want to submit, at any time, you know what to do 00:07 < gleki> so basically sumti tail now is the same as bridi tail in {sei} clause 00:07 < gleki> let's reply to guskant then 00:35 * nuzba @lojban_org: In totally unrelated West #Frisian language and #Lojban the word {do} means "you". [http://bit.ly/1CRtdEc] 00:37 <@Broca> I think this is not coincidental :-) 00:39 < gleki> In Bezhta language "mi" means "yuo" and "do" means "I" 00:39 < gleki> In Bezhta language "mi" means "you" and "do" means "I" 00:58 < ldlework> There's an about page now, http://banski.github.io/pages/about.html 01:08 < ldlework> I updated the wording 01:10 < durka42> coi 01:12 < ldlework> hi durka42 01:15 < ldlework> durka42: seen, http://banski.github.io/ ? 01:15 < durka42> is that the rebranding of "roltadni"? 01:15 < ldlework> indeed 01:15 < durka42> nice, looks good 01:16 < durka42> I see you got selpa'i to participate already 01:16 < ldlework> nah, I just stole his article for filler for now 01:16 < ldlework> I plan to replace it with some lame welcome article 01:16 < ldlework> or anything anyone is willing to contribute to set a good tone 01:17 < durka42> I should write about my theories of BAI or something 01:17 < durka42> really need to get around to reading ilmen's "lo sumtcita cu mo" before writing that though :) 01:17 < gleki> is gaunfake intentionally made recursive? 01:17 < gleki> jbo: gaunfake 01:17 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 01:17 < gleki> mensi: ko cnino 01:18 < mensi> sei ca ca'o jai gau cnino be fai lo pe mi sorcu 01:18 < gleki> durka42: lo sumtcita cu mo? you mean that piratepad article? 01:18 < durka42> yes 01:18 < mensi> i ba'o jai gau cnino 01:18 < ldlework> durka42: mind linking it? 01:18 < gleki> i suggest moving it to the wiki as a backup 01:19 < ldlework> Or cleaned up and published on banski! 01:19 * durka42 notes that pelican also supports markdown :p 01:19 < ldlework> durka42: shhhhhh 01:19 < gleki> e.g. i already lost the link to it 01:20 < durka42> hm 01:20 < durka42> it's in my firefox bookmarks 01:20 < durka42> but firefox isn't running right now because it's slower than LMW during a full DB backup 01:20 < ldlework> I'm hoping tsani will contribute some of his writings 01:20 < durka42> that'd be nice 01:21 < durka42> is selpa'i okay with you reposting his article on another site? 01:21 < ldlework> I didn't ask. I could remove it, it was just a test. 01:21 < ldlework> Get some lorem ipsum or something. 01:22 < gleki> actually robin told me that lmw was slow because it tried to backup al tables even those that are for cache. he disabled some 01:22 < gleki> we need to try to delete more of them 01:23 < ldlework> durka42: I'm trying to think of some main categories 01:23 < ldlework> some really high level ones that'd be hard to argue against 01:24 < ldlework> like... "Formalism", "Culture", "Translation" 01:24 < ldlework> eh 01:25 < ldlework> "Logic" "Lingustics" "Philosophy" "Culture" ? 01:25 < ldlework> Where, most "tinkering" would fall under linguistics? 01:25 < durka42> .title http://pastebin.com/pbM1sdxQ 01:25 < durka42> just make Tinkering a top level category :p 01:25 < ldlework> haha 01:25 < fenki> (NO <title> found) 01:25 < ldlework> Wow its in lojban io 01:26 < durka42> Ilmen wrote it 01:26 < ldlework> uanai 01:26 < durka42> yeah that's why I haven't read it all the way through yet o'anai 01:26 < ldlework> :) 01:27 < durka42> mensi: doi ilmen|menli is this the latest version of "lo sumtcita cu mo" and is it posted anywhere more permanent/discoverable? http://pastebin.com/pbM1sdxQ 01:27 < mensi> durka42: mi ba benji di'u ba lo nu la'o gy.ilmen|menli.gy. di'a cusku da 01:31 < gleki> http://piratepad.net/tPI3w3ucDu 01:31 < gleki> i have this instead 01:32 < gleki> i grepped all piratepad links from our channels and put them here: 01:32 < gleki> http://mw.lojban.org/extensions/ilmentufa/piratepad.html 01:32 < gleki> it'd be nice to transfer them to the wiki so that they become indexed 01:36 < gleki> mensi: doi lmen it'd be nice to transfer them to the wiki over time so that they become indexed http://mw.lojban.org/extensions/ilmentufa/piratepad.html by search engines 01:36 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.lmen.gy. di'a cusku da 01:37 < gleki> ta'o ja'o la selpa'i zei'a bilma 02:41 < nilmenre> . 02:41 < mensi> nilmenre: cu'u la'o gy.durka42.gy.: is this the latest version of "lo sumtcita cu mo" and is it posted anywhere more 02:41 < mensi> permanent/discoverable? http://pastebin.com/pbM1sdxQ | 2015-07-23T08:27:18.826Z 02:41 < mensi> nilmenre: cu'u la'o gy.gleki.gy.: it'd be nice to transfer them to the wiki over time so that they become indexed http:/ 02:41 < mensi> /mw.lojban.org/extensions/ilmentufa/piratepad.html by search engines | 2015-07-23T08:36:27. 02:41 < mensi> 167Z 02:43 < gleki> ue 02:43 < gleki> mensi: doi Ilmen it'd be nice to transfer them to the wiki over time so that they become indexed http://mw.lojban.org/extensions/ilmentufa/piratepad.html by search engines 02:43 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.Ilmen.gy. di'a cusku da 02:47 * nuzba @frecafloros: @fukadumena つesperanto つlojban 好きなの選べ [http://bit.ly/1Ify39Z] 04:13 * nuzba @lojban_org: #Lojban keyboard layout for Mac http://mw.lojban.org/papri/Lojban_keyboard_for_Mac [http://bit.ly/1eiQF15] 04:19 < gleki> en:coi 04:20 < mensi> coi = [COI] vocative: greetings/hello. 04:24 < gleki> alta: li mo 04:24 < mensi> (NIhO [I {FA <li LOhO>} CU {mo VAU}]) 04:25 < durka42> alta: li li mo 04:25 < mensi> (NIhO [I {<FA (¹li LOhO¹)> <FE (¹li LOhO¹)>} CU {mo VAU}]) 04:25 < durka42> alta: li lo'o li mo 04:25 < mensi> (NIhO [I {<FA (¹li lo'o¹)> <FE (¹li LOhO¹)>} CU {mo VAU}]) 04:25 < durka42> I thought you put {li} in CU for some crazy reason 04:26 < gleki> yes, {culis} dialect removed 04:26 < gleki> not crazy but bsfk 04:26 < gleki> alta: mi pe na ku je do cu mo 04:26 < mensi> (NIhO [I {FA <mi (¹pe [na ku] GEhU¹)> <je do>} cu {mo VAU}]) 04:27 < gleki> i could retain culis but that would require {mi li cu mo} and {mi li mo} to mean different things 04:28 < gleki> alta: mi pe na je do cu mo 04:28 < mensi> (NIhO [I {FA <mi (¹pe GEhU¹)> <na je> do} cu {mo VAU}]) 04:28 < gleki> ^ and this is the unavoidable craziness of lojban 04:34 < durka42> ro terjonma'o ciste cu cizra ie 04:40 * nuzba @L16777216: 車内でやること ・twitter ・小説書く ・lojban勉強する できないこと ・積ん読ラノベ消化 [http://bit.ly/1CRRcmG] 04:52 * nuzba @oka_iu_tcan: ちなみにロジバン自体の最新事項はwitkiをチェケラッチョ http://mw.lojban.org/papri/Lo_Lojbo_Nuzba [http://bit.ly/1CRSvC9] 05:32 * nuzba @oka_iu_tcan: http://jbovlaste.lojban.org/dict/ianlaua とうとうトキポナフヒヴラが [http://bit.ly/1JAif2M] 05:52 * nuzba @fotono: ui la #lojban zantufa tricu cu banro i ku'i .uu za'a lo pa jimca cu bilma jecu maltufa http://guskant.github.io/gerna_cipra/zantufa_farvi.svg [http://bit.ly/1JAkAuz] 09:13 < gleki> it should have happened sooner or later. i gonna rewrite terms completely. it's around 50% of the current altatufa grammar 09:40 < zipcpi> ue la .guskant. cu jmina zo maltufa la jbovlaste 09:41 < zipcpi> gy pilno my ma 09:42 < zipcpi> (Oh crap just realized there's a category missing from my list of categorical pronouns. Text.) 09:42 < zipcpi> That's sort of the problem with categorical pronouns... 09:50 < zipcpi> lo kutydza cu klama lo kutyzda lo kutydza lo kutydza lo kutydza 09:51 < zipcpi> -srera- 09:57 < zipcpi> u'i zo upsi 10:01 < zipcpi> coi durka42 10:02 < durka42> coi 10:02 < zipcpi> do kanpe ma ne le nu la .guskant. cu jmina zo maltufa la jbovlaste 10:02 < durka42> mi zvati lo dotco ke ckafi vecnu 10:02 < zipcpi> ua 10:02 < durka42> xu puzi jmina 10:02 < zipcpi> ja'a 10:03 < durka42> la'a srana ti fa'o http://guskant.github.io/gerna_cipra/zantufa_farvi.svg 10:03 < zipcpi> u'i 10:04 < zipcpi> ua implemente zo iau 10:04 < durka42> ma smuni sei mi kucli 10:06 < zipcpi> xy'y ju'onai lo versiio poi jai se srera vau jikau lo versiio poi se cikre ma'i lo pante vau jikau lo drata 10:10 < zipcpi> zantufa: do sijdu mi lo nu zenba lo ka ricfu iau ui 10:10 < zantufa> ([do CU {sijdu <mi (¹lo [nu {CU <zenba (²lo [ka {CU <ricfu VAU>} KEI] KU²) VAU>} KEI] KU¹)> VAU}] [iau ui]) 10:11 < zipcpi> zantufa: do sijdu mi lo nu zenba lo ka ricfu iau lo co'e 10:11 < zantufa> ([do CU {sijdu <mi (¹lo [nu {CU <zenba (²lo [ka {CU <ricfu VAU>} KEI] KU²) VAU>} KEI] KU¹)> VAU}] [iau {lo co'e KU} VAU]) 10:11 < zipcpi> .y. makfa 10:11 < durka42> full implemention 10:11 < durka42> io 10:12 < zipcpi> So she made it also close the last {vau}, but allow starting a new bridi-tail 10:12 < durka42> a new tail or a new bridi 10:12 < zipcpi> do sijdu mi lo nu zenba lo ka ricfu iau lo co'e cu broda 10:12 < durka42> zantufa: ka ka ka ka ka ka co'e iau do ma lifri 10:12 < zantufa> ([CU {ka <CU (¹ka [CU {ka <CU (²ka [CU {ka <CU (³ka [CU {co'e VAU}] KEI³) VAU> KEI} VAU] KEI²) VAU> KEI} VAU] KEI¹) VAU> KEI} VAU] [iau {do ma} CU {lifri VAU}]) 10:12 < zipcpi> zantufa: do sijdu mi lo nu zenba lo ka ricfu iau lo co'e cu broda 10:12 < zantufa> ([do CU {sijdu <mi (¹lo [nu {CU <zenba (²lo [ka {CU <ricfu VAU>} KEI] KU²) VAU>} KEI] KU¹)> VAU}] [iau {lo co'e KU} cu {broda VAU}]) 10:13 < durka42> zantufa: ka ka co'e iau ka ka co'e iau ka ka makfa iau oi 10:13 < zantufa> ([CU {ka <CU (¹ka [CU {co'e VAU}] KEI¹) VAU> KEI} VAU] [iau {CU <ka (¹CU [ka {CU <co'e VAU>} KEI] VAU¹) KEI> VAU}] [iau {CU <ka (¹CU [ka {CU <makfa VAU>} KEI] VAU¹) KEI> VAU}] [iau oi]) 10:13 < zipcpi> Hm not much different than Gleki's {ibo} apparently 10:13 < durka42> zantufa: co'e iau tu'e broda .i brode 10:13 < zantufa> ([CU {co'e VAU}] [iau {tu'e <(¹CU [broda VAU]¹) (¹i [CU {brode VAU}]¹)> TUhU}]) 10:13 < durka42> so it's "iau <statements>" 10:13 < zipcpi> Ah 10:14 < durka42> zantufa: co'e iau broda .i brode 10:14 < zantufa> ([CU {co'e VAU}] [iau {CU <broda VAU>} {i <CU (¹brode VAU¹)>}]) 10:14 < durka42> ue 10:14 < durka42> how do you escape from a {iau}?? 10:14 < zipcpi> OK that scope is weird 10:14 < durka42> zantufa: tu'e co'e iau broda tu'u .i brode 10:14 < zantufa> ([tu'e {<CU (¹co'e VAU¹)> <iau (¹CU [broda VAU]¹)>} tu'u] [i {CU <brode VAU>}]) 10:14 < zipcpi> There is no escape bu'a'a'a'a'a'a 10:14 < durka42> escape requires forethought 10:14 < durka42> hmm somehow profound :p 10:15 < zipcpi> lo kutydza cu klama lo kutyzda lo kutydza lo kutydza lo kutydza 10:16 < zipcpi> .y. zo kutyzda 10:17 < durka42> kurtyzda xu 10:18 < zipcpi> mi finti su'o valsi poi xo'o se cmene be'ei la kurtis 10:18 < zipcpi> http://jbovlaste.lojban.org/dict/kutpoi 10:20 < zipcpi> finti mu'i no da mapti valsi lo si'o kunti co'e 10:20 < zipcpi> .y. mu'i lo nu 10:21 < zipcpi> aunai (to uedai zo'o toi) finti lo malma'o 10:23 < durka42> to .u'i 10:25 < zipcpi> da'i zo zai'e'oi zo zai'o'oi zo zai'e'u zo zai'a'a 10:34 < gleki> altatufa wont develop for several days because im basically rewriting the whole lojban grammar 10:37 < zipcpi> la kutydza cu cusku lo kutyuenzi jecu ciska lo kutlerpoi 10:40 < zipcpi> su nu cusku lo kutyuenzi ka'e nu smusku so'i da 10:41 < zipcpi> exp: su'o nu cusku lo kutyuenzi ka'e nu smusku so'i da 10:41 < mensi> ([{su'o BOI} {nu <CU (¹cusku [lo kutyuenzi KU] VAU¹)> KEI} KU] [CU {ka'e <nu (¹CU [smusku {so'i BOI} da] VAU¹) KEI>} VAU]) 10:42 < gleki> spagetufa: kutyuenzi 10:42 < spagetufa> (CU [kuty/ŭènzi VAU]) 10:42 < zipcpi> spagetufa: briii 10:42 < spagetufa> (CU [brìĭi VAU]) 10:42 < durka42> zantufa: kutyuenzi 10:42 < zantufa> (CU [kutyuenzi VAU]) 10:42 < durka42> ue 10:43 < zipcpi> uenmo ma .i xu la .guskant. pu smusku lo du gy na nelci lo'e zevlyjvo 10:43 < mensi> sei mi stace mi bi'u na mutce nelci 10:43 < gleki> ^correct usage of bi'u 10:44 < gleki> much better than ji'a here 10:44 < zipcpi> I dunno, to me bi'u means that {mi} is a new object; sort of like {ko'a bi'u} 10:45 < gleki> and this is where one can see that mi'unai can't go there 10:45 < gleki> {bi'u} means new information 10:45 < gleki> ~= what is new here is that it's me who broda 10:45 < zipcpi> Yeah but when it tags a sumti I believe that means that the sumti represents a new object 10:46 < zipcpi> Sort of like the difference between {lo bi'unai lorxu} and {lo bi'u lorxu} 10:46 < gleki> not object. the bject can be described zillions times before but situation in which it appears has him never described before 10:46 < zipcpi> Doesn't seem to match usage of {bi'unai} ga'a mi 10:47 < gleki> ... in this ver situation 10:47 < gleki> *very 10:47 < gleki> jb: bi'u 10:47 < mensi> bi'u = bi'u [interjection] — newly introduced information, bi'u nai — previously introduced 10:47 < mensi> information 10:47 < mensi> :lo bi'u nai nixli pu klama lo panka i bi'u le nixli ca tavla lo nanla bu'u le panka — That girl (mentioned earlier), 10:47 < mensi> she went to a park. What is new is that she is talking to a boy there. 10:47 < mensi> :Related words: cnino 10:47 < gleki> another example but mensi's is better 10:47 < zipcpi> Yes, but here bi'u tags an entire sentence, so I have no problem. It's when it tags a sumti I worry about the semantics 10:48 < gleki> similar to {ui} tagging sentences 10:48 < gleki> and sumti 10:48 < gleki> what is colloquially called "emphasis". 10:48 < zipcpi> (lo bi'u nu) le nixli li'o... 10:48 < zipcpi> I don't have a problem there 10:48 < gleki> the situation is not new but that mensi appeared in it hasn't been mentioned earlier 10:48 < zipcpi> It's when you use {le bi'u nixli} to refer to a girl mentioned earlier, then I have a problem 10:49 < gleki> no, bi'u is new 10:50 < gleki> it's a bit similar to "NB" in that it introduces new entities, new facts, not just rephrases earlier facts. 10:50 < zipcpi> But that's the thing... to me there's no difference in what kind of thing {le bi'u nixli}, {ko'a bi'u}, and {mi bi'u} tags 10:51 < gleki> what is {le}? 10:51 < zipcpi> I dunno, you used it :p 10:53 < zipcpi> Hm... does make me wonder about how UI affects descriptions 10:53 < zipcpi> {lo nixli ku ji'a} != {lo ji'a nixli}? 10:53 < zipcpi> But what about {ko'abi'u} then 10:53 < durka42> that's supposed to be the same 10:54 < gleki> i used {le} anaphorically which isn't related to {bi'u} 10:54 < zipcpi> Then do we need {lo ke ji'a nixli} for "that which is also a girl"? 10:54 < gleki> that's = nixli ji'a 10:54 < zipcpi> Crap 10:55 < gleki> but not nixli ku ji'a 10:55 < zipcpi> lo poi'i ke'a ji'a nixli 10:55 < zipcpi> Yuck 10:55 < zipcpi> Well in CKTJ you'd save two syllables there :p 10:56 < gleki> zo'e noi ke'a ji'a nixli 10:56 < zipcpi> Not much better 10:56 < gleki> -1 syllable 10:57 < gleki> and -1 syllable in la alta dialect 10:57 < gleki> jia 10:57 < zipcpi> lo poi'i = 3 syllables, zo'e noi = 3 syllables 10:57 < gleki> lo poi'i = 4 syllables 10:57 < zipcpi> No 10:57 < gleki> [jh] ? 10:57 < zipcpi> Oh, by that count... I don't use that count 10:57 < gleki> [h] can happen only between vowels, not semivowels 10:58 < gleki> poh-eeh-heeh 10:58 < zipcpi> No... I consider /oi/ [oi], not [oj] 10:58 < zipcpi> But I don't know how I pronounce it 10:59 < latro`a> I think the official morphology agrees with zipcpi on this one 10:59 < latro`a> the diphthongs are actual diphthongs, except arguably for iV and uV 10:59 < zipcpi> k: poii'i 10:59 < mensi> SyntaxError: Expected [yY] but "o" found. 11:00 < zipcpi> Probably breaks to {po ii'i} 11:00 < zipcpi> But {ii'i} isn't a word, so... 11:00 < latro`a> no, I don't think so 11:00 < latro`a> *maybe* po i'i, though 11:00 < latro`a> but not ii'i 11:00 < zipcpi> k: lo broda poii ko'a 11:00 < mensi> ([LE:lo {G:broda <GOI:po UI:ii> KOhA:ko'a GEhU} KU] VAU) 11:00 < latro`a> unfortunately po i'i actually parses 11:00 < gleki> i dont understand, each vowel is one syllable 11:01 < latro`a> nope 11:01 < zipcpi> No, that's not how Lojban diphthongs work 11:01 < gleki> and it can't be [pojhi] 11:01 < gleki> [oi] is not a diphthong 11:01 < gleki> [oj] is. 11:01 < latro`a> but {oi} is a diphthong 11:01 < gleki> [poi] but [poihi] 11:01 < gleki> {.oi} is one syllable 11:01 < gleki> {poi} is one syllable 11:02 < latro`a> why is poi'i then three syllables? 11:02 < zipcpi> I dunno, if you want to see oi as VG instead of VV, I think VGC is allowed, even though CGV isn't 11:02 < zipcpi> I mean 11:02 < gleki> {poi'i} is three syllables because [poihi] 11:02 < zipcpi> Lujvo have VGC everywhere for crying out loud 11:02 < latro`a> do you mean lujvo or zi'evla 11:02 < latro`a> ? 11:02 < zipcpi> Both 11:02 < latro`a> what's a lujvo with that? 11:03 < zipcpi> vlaste: *aut* 11:03 < vlaste> no results. http://vlasisku.alexburka.com/%2Aaut%2A 11:03 < zipcpi> Er 11:03 < zipcpi> CVVCCV 11:03 < latro`a> ah, ok 11:03 < zipcpi> kaitcu 11:04 < latro`a> but it seems like reading it as VG is forcing a buffer vowel 11:04 < latro`a> in other words, you get po iy hi 11:04 < latro`a> for a suppressed {y} 11:04 < latro`a> rather than poi hi 11:05 < latro`a> which incidentally is how I think I say it 11:05 < latro`a> at any rate, if oi is not a diphthong then nothing is, no? 11:05 < latro`a> oh nvm you meant IPA 11:05 < latro`a> derp 11:06 < zipcpi> Not to mention 11:06 < zipcpi> k: atloiua 11:06 < mensi> (CU [Z:atloiua VAU]) 11:06 < zipcpi> (Can't think of a real word with the pattern I want to demonstate) 11:06 < zipcpi> But notice the diphthong before the semivowel u 11:06 < zipcpi> spagetufa: atloiua 11:06 < spagetufa> (CU [.atlòiŭa VAU]) 11:08 < zipcpi> Which is a different word from 11:08 < zipcpi> spagetufa: atloiiua 11:08 < spagetufa> (CU [.atloĭìŭa VAU]) 11:08 < zipcpi> Same with 11:08 < zipcpi> spagetufa: atloiva 11:08 < spagetufa> (CU [.atlòiva VAU]) 11:08 < zipcpi> spagetufa: atloiiva 11:08 < spagetufa> (CU [.atloĭìva VAU]) 11:08 < zipcpi> All different words 11:08 < durka42> manabmi 11:09 < zipcpi> So pronouncing poi'i as po-ii-'i is just wrong 11:09 < durka42> jetnu 11:09 < zipcpi> Gleki asserted that poi'i is three syllables 11:09 < durka42> po ii'i is two words :) 11:09 < durka42> er no 11:09 < durka42> it's two 11:09 < gleki> poiki is two syllables, poi'i is three 11:10 < gleki> syllable is what vowel creates. 11:10 < gleki> semivowels dont count 11:10 < durka42> what are you talking about 11:10 < durka42> poi'i has no semivowels 11:10 < durka42> poi - 'i 11:10 < gleki> and thus it's three syllables 11:10 < durka42> if it has three syllables where are the syllable boundaries? 11:10 < gleki> im talking about conciseness of speech 11:10 < durka42> I think you're talking about misunderstanding of "syllable" 11:10 < gleki> you can sing vowels but cant sing semivowes and consonants 11:11 < durka42> but I'm not sure 11:11 < durka42> hence the question 11:11 < gleki> pooooooo iiiiiiii hiiiiiiiiiiiiii 11:11 < durka42> ah 11:11 < durka42> but {oi} is a diphthong 11:11 < durka42> so it combines into one syllable 11:11 < zipcpi> That's poii'i -> po ii'i 11:11 < gleki> ooooooooooooooooooooi 11:11 < durka42> yes 11:12 < durka42> diphthong 11:12 < durka42> .w diphthong 11:12 < gleki> {.oi} is [oj] 11:12 < zipcpi> pooooooi hiiiiii 11:12 < fenki> "A diphthong (/ˈdɪfθɒːŋ/ DIF-thong or /ˈdɪpθɒːŋ/ DIP-thong[1]|) (Greek: δίφθογγος, diphthongos, literally 'two sounds' or 'two tones'), also known as a gliding vowel, refers to two adjacent vowel sounds occurring within the same syllable." - https://en.wikipedia.org/wiki/Diphthong 11:12 < zipcpi> If you want to think of oi as [oj], then yes, [jh] is allowed 11:12 < gleki> 11:12 < zipcpi> pooooooi hiiiiii <-- impossible. h can be only between vowels 11:12 < zipcpi> [hj] is not allowed, but [jh] is allowed 11:12 < durka42> "occurring within the same syllable" 11:13 < zipcpi> between one particular definition of vowels 11:13 < zipcpi> That considers diphthongs as VV, not VG 11:13 < durka42> a diphthong counts as a vowel for the purposes of ' being between two vowels 11:13 < durka42> if you want to think about it that way 11:15 < zipcpi> spagetufa: atloi'i 11:15 < spagetufa> (CU [.atlòi'i VAU]) 11:15 < zipcpi> spagetufa: atloii'i 11:15 < spagetufa> (CU [.atloĭì'i VAU]) 11:15 < zipcpi> Again, different words 11:16 < zipcpi> spagetufa: mau'ucmo 11:16 < spagetufa> (CU [mau'ùcmo VAU]) 11:16 < zipcpi> spagetufa: mauu'ucmo 11:16 < spagetufa> ([FA ma] [CU {ŭu'ùcmo VAU}]) 11:23 < zipcpi> Oh and about {le} and {bi'u}; you told me {le} can be exophorical too, which can be {bi'u} in certain contexts, at least IMO 11:25 < gleki> mensi: ko cnino 11:25 < gleki> exophora means context, i.e. what you have 11:26 < zipcpi> So {le bi'u nixli} = The girl that also exists in the context, but which is not anything mentioned earlier 11:26 < mensi> sei ca ca'o jai gau cnino be fai lo pe mi sorcu 11:26 < gleki> it can be mentioned earlier 11:26 < durka42> did we convince you about the syllables? :) 11:26 < gleki> no 11:27 < durka42> :( 11:27 < gleki> durka42: do you pronounce {poi'i} as [pojhi]? 11:27 < durka42> based on my nonexistent knowledge of IPA, yes 11:27 < zipcpi> I don't know whether I pronounce diphthongs as [oi] or [oj] 11:27 < zipcpi> But it's the same to me either way 11:27 < gleki> how can h be after a semivowel? 11:27 < durka42> ah j is a semivowel? 11:27 < durka42> all I know is that oi is a diphthong 11:28 < durka42> two vowels combined into one 11:28 < zipcpi> ca'e [jh] is allowed. It has to be if we don't want to break morphology yet again 11:28 < gleki> [j] is like "y" in "yes" 11:29 < gleki> so {.ie} is [je] 11:29 < durka42> oh then no 11:29 < durka42> oi is just oi 11:29 < durka42> why are we overthinking this :) 11:29 < gleki> {.oi} is [oj] 11:30 * durka42 looks up IPA diphthongs 11:30 < zipcpi> durka42: Good luck. I have no idea what's the difference between [oi] and [oj] 11:30 < gleki> http://mw.lojban.org/extensions/ilmentufa/cirkotci.html 11:30 < gleki> see how {oi} is pronounced 11:30 < zipcpi> Other than something about semivowels that confuses Gleki 11:31 < durka42> this sound is what I use ({oi} as in "boy") http://cmed.faculty.ku.edu/ipafolder/vowels.html 11:31 < gleki> [oi] is aw-eeh 11:31 < gleki> [oj] is like in "boy", yes 11:31 < gleki> and [j] is a semivowel 11:31 < zipcpi> aw-eeh is [oji], at least according to Lojban morphology 11:31 < durka42> which is the same thing cirko is saying the video 11:31 < durka42> ok then {poi'i} is [pojhi] 11:32 < zipcpi> Like I said, ca'e [jh] is allowed if you insist on seeing /oi/ as [oj], cause that's the only way to not break morphology again 11:32 < gleki> [oi] is like in "Awe! East!" 11:32 < gleki> [oj] is like in "boy" 11:33 < zipcpi> I've never seen an IPA source that agrees with that 11:33 < durka42> the link I postd had a different IPA 11:33 < durka42> not one I can type though 11:33 < gleki> east is [iːst] in most dictionaries 11:33 < durka42> anyway I guess I agree with zipcpi then 11:34 < zipcpi> durka42: Same principle. That small-capital-i is a vowel, but it's considered one syllable 11:34 < durka42> uanai 11:34 < zipcpi> IPA considers two vowels put together a dipththong 11:34 < zipcpi> *diphthong 11:35 < durka42> well good that's what we need 11:35 < zipcpi> For something like "awe-east" it would be spelled [ɔ.i:st] 11:36 < zipcpi> Which is different from [ɔist] 11:36 < zipcpi> Or even [ɔ.ist] 11:36 < zipcpi> IPA . = old-Lojban , 11:37 < gleki> http://lojban.github.io/cll/3/10/ 11:37 < gleki> [i] or [i:], doesnt matter 11:39 < durka42> ...okay? 11:40 < zipcpi> [ɔist] != [ɔ.ist] 11:40 < durka42> true 11:40 < gleki> li'a 11:40 < durka42> in that the former is not transcribable into lojban :) 11:41 < gleki> because you put [.] in between 11:41 < zipcpi> IPA [.] is syllable break 11:41 < gleki> so in {poi'a} you pronounce {i} as [i] but in {poi} as [j]? 11:41 < zipcpi> I don't know how I pronounce it 11:42 < zipcpi> I don't think [ojhi] is impossible to pronounce 11:43 < zipcpi> It just doesn't fit with certain dogma about what Lojban considers vowels and where the apostrophe goes, but that's how it works 11:44 < zipcpi> Anything else would just break the morphology yet again 11:44 < zipcpi> And this time we won't have just zi'evla or cmevla to replace 11:44 < zipcpi> But cmavo and lujvo as well 11:45 < durka42> the CLL has nothing to say about this since it doesn't define any cmavo with >2 vowels 11:45 < durka42> but, pragmatically, it works fine 11:45 < durka42> so no need to break anything 11:45 < durka42> why can't {poi} be {poj}? 11:45 < durka42> er, [poj] 11:46 < zipcpi> Nor does it define zi'evla much beyond "What doesn't contain y and isn't a gismu, a lujvo, a tosmabru, or a slinku'i 11:46 < durka42> well that's what we have the modern morphology for 11:47 * zipcpi nods 11:50 < BenLubar> {mi vi'ecpe lo tafyvi'ujibri ku jo'u la jyfykys. jo'u la stalin.} 11:50 < BenLubar> how would I say the second sentence in http://i.imgur.com/fycHx.jpg ? 11:51 < zipcpi> - no'u la .jyfykys. jo'u la .stalin. 11:52 < BenLubar> thanks 11:55 < gleki> zipcpi: in JVS pls be careful with latex, especially such things as "10^30" which should be $10^{30}$ 11:55 < zipcpi> Erm, I tried, but it broke 11:56 < gleki> this is what works and produces non broken pdf 11:56 < gleki> i corrected all your broken definitions 11:56 < zipcpi> Well, thanks. Before though I distinctly remembered getting an error 11:57 < zipcpi> When I tried to include those expressions as Latex 11:57 < gleki> mi di'a morji lo du'u mi tolcau ro jai sarcu be lo nu finti JVS poi cnino 12:01 < gleki> i xu ro da pu'i tcidu la'e ju'oi http://mw.lojban.org/papri/Proposal:_distinguishing_digits_from_numbers 12:01 < durka42> ua nu'o 12:01 < durka42> xu cnino 12:02 < gleki> li'a 12:02 < gleki> i kurtypapri 12:02 < durka42> whoa 12:02 < zipcpi> mi ja'a pu'i tcidu 12:02 < durka42> mi nu'o tcidu lo finti be K jenai clavelcki 12:02 < gleki> i xo'ei zei saske vau zo'oru'e 12:03 < zipcpi> saske lo si'o cusku lo kutyuenzi 12:04 < zipcpi> mutce lo ka za'e zenkai 12:05 < zipcpi> vlaste: zendo (fapro) 12:05 < vlaste> zo zendo gimkamsmikezypro no catni gismu .i zo zendo gimkamsmikezypro no cipra gismu 12:06 < zipcpi> zengoikai u'i 12:07 < zipcpi> sepi'o ra'oi -goi- noi selckikynomrafsi 12:15 < zipcpi> ra'oi -goi- galfi lo li'erla'i rafsi lo kulnu rafsi .i na vajni fa lo du le rafsi pu rafsi makau 12:18 < zipcpi> le kurtynompapri zo'u: I don't know. I can see the problem, but we definitely need a way for {xo'e} to represent several digits as well 12:19 < zipcpi> And uenaisai that he doesn't like xornunsei :p 12:24 < zipcpi> But ugh, I don't want *more* mekso grammar 12:25 < durka42> I was surprised at how few cmavo he introduced, .u'i 12:25 < zipcpi> u'i 12:25 < durka42> but yeah, I'm 100% sure those rules can be simplified 12:26 < zipcpi> Also his {xo'e'o'e}... I don't know how that would work. From what I can tell it's meant to be able to replace both BY and PA as needed 12:27 < zipcpi> But between the ci'i-ba'u proposals about what to do with BY and PA, I don't know how to make it work :p 12:27 < zipcpi> I can see it's utility, but I don't know how it would work 12:27 < zipcpi> *its 12:28 < zipcpi> Also I'd probably want it to be {xo'e'e} 12:30 * nuzba @Saffy: Watching language vids about learning multiple languages @MariiKusha frowns @ the Spanish & says 'thats not #lojban theyre talking it wrong' [http://bit.ly/1TUtYQS] 12:30 < zipcpi> I'd probably just use {bu'o'e} or even {zo'ebu} for "illegible/unknown symbol" 12:32 < durka42> zo'ebu :p 12:32 < durka42> compositionality ftw 12:32 < zipcpi> Yep 12:35 < bigcentaur> coi ro do 12:38 < durka42> coi 12:38 < durka42> mi zvati lo dotygu'e 12:42 < bigcentaur> .ue ma mukti la'e di'u 12:43 < durka42> skami cilre nunpe'i 12:49 < bigcentaur> .i xu plixau do 12:50 < durka42> go'i 12:53 < bigcentaur> .i xu do di'i zvati lo dotygu'e 12:56 < durka42> na go'i 12:56 < durka42> pu noroi 12:58 * nuzba @Saffy: Mr @alaric what is Musee/Musi/Mooosey or possibly Lucy in #lojban @MariiKusha thinks it's a farm animal but it's not the one she thought [http://bit.ly/1TUyIpE] 13:01 < zipcpi> I kinda wonder if the annoying vagueness that some gismu have is motivated by the desire to avoid metaphor in lujvo 13:01 < zipcpi> Like {kafke} 13:02 < zipcpi> It *could* have been just defined as "cough", yet allow lujvo built with it to mean different things like flatulence 13:02 < durka42> the gismu were intended to be quite broad 13:02 < durka42> and specificity achieved with lujvo 13:02 < zipcpi> Hm 13:06 < zipcpi> Hm the Champernowne's constant thing... that's what fancylojban is for. Define a new brivla, and either explain it with several examples if it's meant for human usage, or define it programatically if it's a technical document that leavs no stone unturned 13:24 < zipcpi> x1 is the Champernowne's constant in base x2 13:26 < zipcpi> The question is how do you define an infinite series? Maybe via infinite recursion 13:26 < zipcpi> Of course, that will make any computer program just get stuck :p 13:32 < zipcpi> How many downvotes will I get if I made {cmuna'u} "prime number"? :p 13:38 < zipcpi> x1 is a prime number -> bo'a mulna'u gi'e dubmau li pa .ije no da poi ke'a mulna'u gi'e dubmau li pa gi'e dubme'a bo'a zo'u: bo'a dilcymu'o da 13:39 < zipcpi> Maybe {dilcycmu} 13:51 < rlpowell> gleki: THe backup now only takes a couple of minutes; I don't think any further work is needed there. 14:10 < noncomcinse> selma'o: tau 14:11 < noncomcinse> mensi: selma'o: tau 14:11 < noncomcinse> jbo: tau 14:11 < zipcpi> LAU 14:11 < noncomcinse> jbo: dau 14:11 < noncomcinse> Oh, I thought that worked doi la zipcpi 14:11 < noncomcinse> .u'i zy na sampre 14:12 < zipcpi> u'i 14:12 < noncomcinse> What I really was looking for, ki'e la vlasisku, was PA2. 14:12 < noncomcinse> en: xexsi 14:13 < zipcpi> xexso 14:13 < zipcpi> 10^18 14:13 < zipcpi> "exa" 14:13 < noncomcinse> la mensi cu mrobi'o xu 14:13 < zipcpi> si'au go'i 14:13 < zipcpi> za'a na zvati 14:14 < zipcpi> ba'o oicli'a li'a 14:14 < noncomcinse> vlaste: xexso 14:14 < vlaste> xexso = x1 is 10^18 of x2 in dimension/aspect x3 (default is units). 14:14 < noncomcinse> vlaste: hexadecimal 14:14 < vlaste> no results. http://vlasisku.alexburka.com/hexadecimal 14:14 < noncomcinse> What's the base-N switch word? 14:14 < zipcpi> dau fei gai jau xei vai 14:14 < zipcpi> ju'u I think 14:15 < noncomcinse> vlaste: ju'u 14:15 < vlaste> ju'u = binary mathematical operator: number base; [a interpreted in the base b]. 14:15 < noncomcinse> Is there a word for x1 is the color represented by RGB hex code x2? 14:16 < zipcpi> I don't think so... 14:17 < noncomcinse> ji'a cinri: HSV and CYM 14:17 < noncomcinse> vlaste: cyan 14:17 < vlaste> 3 results: cicna, cicnybla, cicnyri'o 14:17 < noncomcinse> vlaste: cicna 14:17 < vlaste> cicna = x1 is cyan/turquoise/greenish-blue [color adjective]. 14:18 < zipcpi> Hmm... need words for "decimal" and "hexadecimal" 14:18 < zipcpi> ju'urdau , ju'urpaxa 14:18 < zipcpi> Dunno what the place structure should be though 14:19 < noncomcinse> x1 is the color represented by hex code x2 in system x3 14:20 < noncomcinse> cei broda 14:20 < zipcpi> skarnxunri'obla ? 14:21 < zipcpi> skarnju'upaxa ? 14:21 < noncomcinse> {lo tsani cu skari lo broda li rei gai no rei rei} 14:22 < noncomcinse> broda fi la xunre ce crino ce blanu 14:22 < zipcpi> #0EC0EE 14:23 < zipcpi> I don't know what goes in x3 though 14:23 < noncomcinse> I screwed it up, aimed for 7ec0ee 14:23 < noncomcinse> sa'ai li ze le'ai 14:24 < noncomcinse> http://www.color-hex.com/color/7ec0ee 14:24 < noncomcinse> .title 14:24 < fenki> [ #7ec0ee Color Hex Sky Blue 2 ] 14:24 < noncomcinse> the sky is colored by the color labeled with 7ec0ee 14:25 < noncomcinse> x3 would be la xunre ce crino ce blanu 14:26 < noncomcinse> Also, the selstidi valsi you said, mind explaining the rafsi/parts in them? 14:27 < zipcpi> xunri'obla -> xunre crino blanu 14:27 < zipcpi> ju'upaxa is just ju'u pa xa 14:28 < noncomcinse> zo ju'u se xi pajeno rafsi naku 14:28 < noncomcinse> s/no/re/ 14:28 < fenki> noncomcinse meant to say: zo ju'u se xi pajere rafsi naku 14:28 < zipcpi> Yeah... 14:28 < noncomcinse> exp: se xi pajere rafsi 14:29 < noncomcinse> xu la .camxes. ji'a cu mrobi'o paunai 14:29 < zipcpi> se xi veipajere rafsi 14:29 < zipcpi> http://mw.lojban.org/extensions/ilmentufa/camxes-exp.html 14:29 < noncomcinse> ki'e 14:29 < zipcpi> It was mensi that answers those queries 14:29 < noncomcinse> Ah. 14:30 < noncomcinse> lonu la mensi na .irci cu te nandu ma'a 14:30 < zipcpi> vei is funny lol 14:30 < durka42> coi la camxes 14:30 < camxes> coi 14:31 < noncomcinse> I'm getting sidetracked, but what would you think if there was a cmavo X that quoted like {sa} and terminated the most recently constructed construct? 14:31 < noncomcinse> Like {X lo} would ter-- Right. 14:31 < durka42> didn't zipcpi invent that yesterday or am I mixing people up 14:31 < noncomcinse> that was me .u'i 14:31 < durka42> whoops 14:31 < noncomcinse> I thought of it while thinking of {X .i} 14:32 < noncomcinse> Being able to terminate the most recent .i without starting a new one without doing {.i'au vau} 14:32 < zipcpi> Or just {mifyzilska}: x1 is a color represented by code x2 in system x3 14:33 < noncomcinse> Also, I've been thinking about a meta-delete-undo character. 14:33 < noncomcinse> For accidental {su}, mainly. 14:33 < ldlework> garrr, bach is so hard 14:34 < noncomcinse> Parser already has enough problems with delete words. 14:34 < durka42> that gets circular real fast 14:34 < noncomcinse> It doesn't cancel itself. 14:35 < noncomcinse> And can't be deleted by si 14:35 < noncomcinse> So it would have more magic than si sa su 14:36 < zipcpi> Or hey, I have {bitni} and {baitni} 14:36 < noncomcinse> .i mi klama si tavla -> .i mi /**klama si**/ tavla 14:36 < zipcpi> {cibybaitnyska} :p 14:37 < zipcpi> I forgot what the zi'evla for HTML was 14:37 < noncomcinse> .i mi sutra si X cadzu -> .i mi sutra /**si X**/ cadzu -> .i mi sutra cadzu 14:38 < noncomcinse> .i mi sutra si X X cadzu -> .i mi sutra /**si X**/ /**X**/ (second one is useless, not after a delete word) cadzu 14:38 < noncomcinse> ua pei 14:40 < selpahi> I think {xetmele}. Somehow it's not in the corpus. 14:40 < selpahi> But wtf is this: http://getanswerz-hugo.s3-website-us-west-2.amazonaws.com/lojban/2015/03/01/Would-it-not-be-good-to-post-it-to-the-main.html 14:41 < noncomcinse> vlaste: xetmele 14:41 < vlaste> no results. http://vlasisku.alexburka.com/xetmele 14:41 < durka42> wat 14:43 < selpahi> Is this relevant: http://selpahi.weebly.com/archive-pre-2014/systematic-color-fuivla 14:44 < zipcpi> Somewhat 14:44 < zipcpi> I dunno though. I think two zi'evla might be useful 14:44 < durka42> why use one zi'evla when you could use two? 14:45 < zipcpi> One primitive that's like x1 is the color with red level x2, green level x3, blue level x4 14:45 < selpahi> datru style 14:45 < zipcpi> With x2~x4 being 0-1, or subjective numbers 14:45 < zipcpi> Yep 14:45 < zipcpi> zo skatro zo'o 14:46 < zipcpi> Then another that's just the hexcode 14:47 < selpahi> Eventually I'll make words for "x1 is a color", and "x1 is the color red", "x1 is the color green" etc etc. 14:47 < zipcpi> There is already zilska, but hm 14:47 < selpahi> I can't have to run to Toaq Dzu everytime I want to talk about colors. 14:50 < selpahi> mi tolmo'i lo du'u mu'i ma kau mi pu zi co'a sazri lo skami 14:51 < zipcpi> No {sambau}? 14:53 < durka42> xu na ku lo nu tavla mi'o iau .u'i 14:53 < durka42> selpahi: ui nai la'a mi na klama la berlin 14:54 < selpahi> je'e 14:54 < durka42> xu do ba zvati la franfirt 14:54 < selpahi> .i .a'o la cirko ba jdice lo du'u xu kau ri klama la .franfurt. 14:56 < zipcpi> http://jbovlaste.lojban.org/dict/skarxuciba 14:58 < durka42> selpahi: xu lo se jdice be do cu fancu lo se jdice be la cirko 14:58 < zipcpi> http://jbovlaste.lojban.org/dict/skarxetmele 14:59 < durka42> zipcpi: ugh 14:59 < durka42> please pick one 0-1 or some other scale 14:59 < durka42> ambiguity between 0-1 or 0-255 is a neverending headache 14:59 < zipcpi> skarxuciba is a 0-1 scale 15:00 < durka42> that's not what you wrote in the notes 15:00 < zipcpi> "or subjective numbers" 15:00 < zipcpi> Like so'u / so'i 15:00 < selpahi> durka42: milxe lo ka fancu 15:00 < zipcpi> Which can stand for a 0-1 scale just fine 15:00 < durka42> zipcpi: ah. that wasn't clear to me from what you wrote 15:04 < selpahi> One of the points of {raunza} (apart from replacing banzuka) is to finally get a sort of RDA based off it. Before we would always use {banzu} as a seltau. 15:04 < selpahi> {raunzu} * 15:05 < durka42> vlaste: affix:rau 15:05 < vlaste> gradu = x1 [magnitude] is a unit/degree of/on scale/reference standard x2 (si'o) measuring property x3. 15:05 < durka42> xm 15:06 < durka42> I guess you don't see that as a problem :) 15:06 < zipcpi> I dunno, in my idiolect (which might not correspond with standard Lojban), {lo ka xunre} is a {zilska} 15:07 < selpahi> durka42: A sacrifice, maybe? :) 15:07 * zipcpi hides from the downvotes 15:07 < durka42> sacrifice is fine 15:08 < ldlework> selpahi: I made a website and used an article of yours as filler. If you mind, I can take it down. 15:08 < durka42> I wasn't sure if you were proposing to _replace_ -rau or just leave it with two selrafsi 15:08 < ldlework> http://banski.github.io 15:08 < selpahi> ldlework: I know. Maybe pick a better article :P 15:08 < ldlework> Beyond that, I'd love for any articles you'd like to submit to the site 15:09 < selpahi> Yeah, but I already have a blog, with comment section! 15:09 < durka42> I have a mental lojban blog todo list but also a rust one 15:09 < durka42> I'm not sure which one is higher priority :p 15:09 < ldlework> selpahi: Yeah I understand 15:09 < zipcpi> So {lo xunre} is "red", {lo ka xunre} is "the color red"... or I dunno 15:09 < ldlework> Might not make sense to port existing articles 15:09 < ldlework> lo ka xunre is a proposition 15:09 < zipcpi> If you don't like it you can make lujvo {kamxu'e} or something :p 15:09 < zipcpi> That's not how {ka} works 15:10 < ldlework> a very vague one 15:10 < selpahi> {lo xunre} is e.g. a red ball. 15:10 < ldlework> that's how all the abstractors work 15:10 < zipcpi> I meant "something red" 15:10 < zipcpi> I must be tired 15:10 < zipcpi> {lo xunre} is "something red" 15:11 < durka42> seems like {lo si'o xunre} would matmau but nobody seems to like {si'o} except me 15:11 < zipcpi> u'i 15:11 < zipcpi> Well cause I don't use {si'o} with a {ce'u} 15:11 < selpahi> I know some people disagree with this, but to me "the property of being a red object" is not the same as "the color red". 15:12 < ldlework> Especially when you consider that {lo ka da xunre de} doesn't seem anything like the same kind of thing as "the color red" or "redness" 15:12 < durka42> ieru'eju'ocu'i 15:12 < durka42> that's why I like {si'o} 15:12 < zipcpi> To me {si'o} is sort of like elliptical {ce'u} place, or just the idea of something being true 15:12 < ldlework> I mean it doesn't matter what abstractor is used 15:13 < ldlework> Each returns a thing which is a proposition 15:13 < selpahi> Yes, but to me {si'o} doesn't change anything here. "The concept of being a red object". 15:13 < selpahi> It's not the color red. 15:13 < ldlework> ie 15:14 < durka42> oh I see 15:14 < ldlework> "the color red" would be a singular reference "thing" 15:14 < ldlework> but the abstractors give you whole propositions 15:14 < zipcpi> Then just define {kamxu'e}. It's not like you care what -kam- means anyway :p 15:14 < durka42> lo selkai be lo'e xunre 15:16 < ldlework> I wonder though 15:16 < ldlework> is "red" the name for the property? 15:16 < ldlework> IE, in english, when we say "I love the color red" do we really mean "I love the property who's central proposition is 'red'" 15:16 < zipcpi> And that just reminded me that I defined my color-zi'evla somewhat ambiguously >.< 15:17 < ldlework> With the subtext, "And the things to which it applies" 15:17 < zipcpi> lo tsani cu skarxetmele zo'e ji cu skari lo skarxetmele be zo'e 15:17 < ldlework> I love the set of propositions for which the predicate is red. 15:18 < zipcpi> I dunno, I like the first one better somewhat, but if we're gonna be picky over the two, then we're gonna need even more brivla 15:19 < zipcpi> That's kinda why I abuse {ka} 15:19 < ldlework> I'm not even sure its abuse 15:20 < ldlework> I wonder if its just how lojban refers to properties 15:20 < ldlework> And instead of needing brivla to denote every possible proerpties 15:20 < ldlework> such as in other languages 15:20 < ldlework> we just generate those predicates on demand 15:20 < ldlework> like "red" as in the noun, is really just "lo ka xunre" 15:20 < ldlework> Its like "I love wet." 15:21 < ldlework> Well in English, there's just the trick were you don't have to add "ness" to "red" to talk about it 15:21 < ldlework> But really anywhere you see "I love the color red" can be expanded to "I love the property of redness." 15:21 < zipcpi> Hmm... I do see a potential problem though 15:21 < zipcpi> In that this clashes with the {lo ka} = {lo nu ckaji lo ka} polymorphism usage 15:22 < zipcpi> {lo nu vo'a ckaji lo ka} 15:22 < ldlework> Clashes how? 15:22 < zipcpi> mi nelci lo ka xunre 15:22 < zipcpi> Does that mean 15:22 < zipcpi> "I like the color red" 15:22 < zipcpi> Or 15:22 < zipcpi> "I like being red" 15:22 < zipcpi> ? 15:22 < ldlework> It means I love the events where something is charactized as having the property red 15:22 < durka42> exactly 15:23 < durka42> it depends on the brivla 15:23 < ldlework> Which is kind of like saying "I love the general proposition of "zo'e xunre zo'e zo'e zo'e zo'e" 15:23 < durka42> because some of them are defined to "apply" the ka 15:23 < durka42> it does clash 15:23 < ldlework> I don't understand 15:23 < durka42> where's that comic 15:24 < zipcpi> http://imgur.com/Uh7Nm44 15:24 < durka42> hmph you guys are terrible mind readers :p 15:24 < durka42> ue 15:26 < ldlework> Well its fine 15:26 < durka42> ldlework: like if you give a ka to djica, it fills in x1 to the ka's ce'u 15:26 < ldlework> just don't use the property 15:26 < ldlework> just use si'o or whatever 15:26 < durka42> if nelci does that, then {mi nelci lo ka xunre} means "I like being red" 15:26 < ldlework> "I love the propositions in which things are asserted to have redness about them" 15:26 < ldlework> mi nelci lo si'o xunre 15:26 < durka42> well yes that was my suggestion 15:26 < durka42> it didn't seem to be that well received :) 15:26 < selpahi> If you replace {ka} with {si'o} you lose all the {zo'e} and have them replaced by {ce'u} 15:26 < zipcpi> Well, my main objection to si'o is, does it take with ce'u? 15:26 < selpahi> It's not equivalent 15:26 < ldlework> Is that a problem? 15:26 < durka42> to me, {lo si'o xunre} is not just "the property of something being red" 15:27 < ldlework> I would say it is a very vague proposition 15:27 < zipcpi> selpahi: Now that's me'ei-si'o, which is yet another kettle of fish altogether T.T 15:27 < ldlework> And the only information encoded into it, is the predicate 'red' without any arguments 15:27 < zipcpi> No one seems to know what si'o means 15:27 < durka42> I highly doubt we're going to see selpa'i using {me'ei}! 15:28 < ldlework> Think of it like a pattern 15:28 < ldlework> It represents all possible propositions of that form 15:28 < selpahi> The predicate is "to be a red object", "red" alone is ambiguous and can deceive one's intuition here. 15:28 < ldlework> for example 15:28 < ldlework> mi pensi lo si'o do dunda 15:28 < zipcpi> Except that clashes with everything else 15:29 < ldlework> I'm thinking about the domain that contains all the other propositions where you give various things 15:29 < zipcpi> That's how I've been using si'o 15:29 < zipcpi> But it clashes with everything else 15:29 < ldlework> I don't know what you mean 15:29 < zipcpi> The big question is does si'o take one or more {ce'u} 15:30 < ldlework> ce'u is just like ke'a right? 15:30 < ldlework> its an externally bound variable? 15:30 < selpahi> All places are filled with {ce'u}, is the most common view. 15:30 < ldlework> ce'u bound to what though? 15:30 < selpahi> na'i 15:30 < ldlework> what does it mean for ce'u to appear unbound? 15:30 < durka42> yeah I don't really understand that view 15:30 < selpahi> There is no such thing as a bound ce'u 15:30 < durka42> like if you have a si'o that's chock full of ce'u 15:30 < ldlework> What 15:30 < durka42> and you pass it to a brivla that "applies" 15:30 < durka42> then what happens 15:31 < durka42> it explodes 15:31 < durka42> it seems like si'o is something different 15:31 < selpahi> si'o isn't really used with apply-selbri 15:31 < ldlework> Do you guys know the concept of "axiom schemas" ? 15:31 < zipcpi> If you have a si'o that's chock full of ce'u then there is no difference betwen {lo si'o vecnu} and {lo si'o te vecnu}, except for what place structure you get if you {me'au} it 15:31 < selpahi> Although Ilmen does use {si'o} like that now and then. 15:31 < ldlework> Anyone? 15:31 < selpahi> zipcpi: Yes. 15:32 < durka42> so you would say it's bad practice to give a {si'o} to any place such as djica2 or zukte2 15:32 < selpahi> Plus focus/emphasis considerations arguably 15:32 < durka42> ldlework: no 15:32 < selpahi> Yes, it's bad practice, because zukte2 is unary. 15:32 < ldlework> durka42: basically, in a formal system, the axiom schema is the "form" that theorems must take in order to be valid 15:33 < durka42> what if {si'o} were chock full of {zi'o}, instead of chock full of {ce'u} 15:33 < ldlework> Like you an say, statements must be of the form: xpxqx, where x is a string of hypens 15:33 < durka42> zu'u'u'ai 15:33 < selpahi> durka42: You mean zero ce'u? Then it still can't be applied 15:33 < ldlework> I'll try again later 15:33 < durka42> I mean lo si'o xunre => lo su'u zi'o xunre 15:33 < selpahi> I know. 15:33 < durka42> ldlework: sounds like a formal language 15:34 < selpahi> Wait, su'u? 15:34 < ldlework> durka42: so I suggest that si'o abstractions denote a sort of 'axiom schema' 15:34 < ldlework> They refer to the domain of matching propositions 15:34 < durka42> selpahi: well I didn't want to just repeat si'o, and I sure didn't want to say ka, so... :) 15:34 < ldlework> They don't get 'applied' as they are no functions persay 15:35 < durka42> I guess ka is safe since I removed all the places 15:35 < ldlework> they are function "prototypes" 15:35 < selpahi> A {ka} without {ce'u} doesn't exist, or is a {du'u}. 15:35 < durka42> I'm not all that interested in redefining a lojban concept in terms of some obscure english concept 15:35 < durka42> I'd rather know what it means in lojban :) 15:35 < ldlework> That's what I'm trying to illuminate but ok 15:35 < durka42> I agree they aren't functions 15:35 < durka42> and can't be applied 15:36 < durka42> that's why it seems strange to me to fill them up with ce'u 15:36 < ldlework> Its like a du'u with ambiguous places 15:36 < durka42> so fine, lo si'o xunre => lo du'u zi'o xunre 15:36 < ldlework> but instead of being a solid proposition 15:36 < ldlework> with simply, some places being unfilled 15:36 < ldlework> the unfilled places are like asterisks 15:36 < durka42> lo si'o klama => lo du'u zi'o zi'o zi'o zi'o zi'o klama 15:36 < ldlework> and it expands to all propositions of that form 15:36 < zipcpi> Maybe just redefine {za'i}. No one uses {za'i} :p 15:36 < selpahi> Is {zi'o xunre} easier to interpret for you than {ce'u xunre} ? :O 15:36 < durka42> I use za'i! 15:36 < durka42> it's about as easy to interpret as "the abstract property of redness" :/ 15:37 < selpahi> of being a red object. 15:37 < durka42> but that's what you said you don't want 15:37 < durka42> we keep saying it shouldn't be about the objects 15:37 < selpahi> That's what xunre is 15:37 < durka42> so let's just remove that place! 15:37 < ldlework> I think the abstract property of redness, is a pattern that matches against the domain of possible concrete propositions 15:37 < durka42> I remain firm in my stance of refusing to fear zi'o 15:37 < ldlework> As is the property of drinking milk 15:37 < selpahi> No, xunre is inherently coupled to objects that have that color. I don't want any of that. 15:38 < ldlework> I agree you can't remove the places 15:38 < ldlework> It becomes a different predicate then 15:38 < durka42> but it sounds like we need a different predicate! 15:38 < durka42> since xunre is inherently coupled to objects, ti'e 15:38 < selpahi> Yes, new predicates please. 15:38 < zipcpi> But obviously we want it to be based on xunre 15:38 < durka42> zilxunre :) 15:38 < zipcpi> So what should it look like? 15:39 < selpahi> Don't make me throw up zo'o 15:39 < ldlework> lo si'o citka => lo du'u do citka, lo du'u do citka da, lo du'u do citka de, lo du'u mi citka, lo du'u mi citka da, lo du'u mi citka de, etc ad infinitum 15:39 < durka42> well, would this problem be solved if there were predicates meaning "redness", "blueness", etc and xunre/blanu/... were defined in terms of them instead 15:39 < durka42> 'cause we can do that :) 15:39 < zipcpi> Or we make a new NU 15:39 < selpahi> It would solve it while breaking half the corpus. 15:39 < zipcpi> But I won't know how to define it 15:39 < ldlework> the only things that would be valid for the predicate "is-redness" would be the property of redness 15:39 < durka42> selpahi: where did I suggest breakage? 15:40 < selpahi> Because {lo zdani cu xunre} stops being valid 15:40 < selpahi> Let me reread 15:40 < durka42> no I didn't say to redefine xunre 15:40 < durka42> I mean like the way Ilmen pulled {zilzena} out of thin air and redefined {zenba} based on it 15:40 < ldlework> a predicate that denotes redness itself, rather than things that have the property of redness 15:40 < selpahi> I agree 15:41 < ldlework> I argue you don't need it 15:41 < ldlework> the one thing for which that predicate would be true of, is the singular property of redness 15:41 < ldlework> and we have a way to generate a description to that thing, easily 15:41 < zipcpi> Ilmen's magic predicates ^^ 15:42 < durka42> like if broda means redness, then xunre = skari lo broa 15:42 < ldlework> si'o is like a proposition query 15:42 < ldlework> it queries for all possible propositions of the same form 15:42 < selpahi> durka42: Yes, that's the idea. 15:42 < selpahi> Though I'd prefer if skari1 and skari2 were switched. 15:42 < ldlework> I'd say its the oppostie 15:43 < ldlework> redness = skari lo ka xunre 15:43 < ctefaho> all you need is to make 1 new gismu to solve that color mess 15:43 * selpahi won't make any gismu 15:43 < ctefaho> if you are talking about what I think you are this is easy to solve 15:43 < ldlework> Then we need words for wetness 15:43 < ldlework> and dryness 15:43 < ldlework> and happiness 15:43 < ldlework> and so on 15:43 < ldlework> this is the nice thing about lojban 15:43 < ldlework> you don't need adverbs, or adjectives 15:43 < ldlework> or nouns 15:44 < ldlework> just verbs 15:44 < zipcpi> That's why I wonder if we need a new NU 15:44 < durka42> what's wrong with si'o 15:44 < ctefaho> verbs still have to mean different things 15:44 < ldlework> its si'o! 15:44 < durka42> nobody uses it anyway, we can use it for this 15:44 < ctefaho> you can't just cram shit together like yall do right now 15:44 < zipcpi> Because everyone uses si'o differently T.T 15:44 < ldlework> ctefaho: xunre != ka xunre 15:44 < ctefaho> agreed 15:44 < durka42> making a new NU isn't the solution 15:44 < selpahi> From your last lines it looks like you are still not seeing the nuance between redness and red-object-ness 15:44 < ctefaho> or any broda = ke ce broda 15:45 < selpahi> Because dryness is not analogous to ka xunre 15:45 < durka42> indeed I do not see that nuance 15:45 < ctefaho> i take no part in your people's ka abuse 15:45 < ldlework> I'd say that aiming at the set of propositions that adheres to some propositional form is a very good mapping of the concept of concept 15:45 < zipcpi> durka42: Does si'o take a ce'u? Cause to me if it's gonna represent "redness", it should. 15:45 < ldlework> it duly serves si'o 15:45 < durka42> zipcpi: why? what would you do with it that requires a ce'u? 15:45 < ctefaho> and now I am gonna brb 15:45 < ctefaho> or bbiab, more likely 15:45 < zipcpi> To distinguish between {lo si'o vecnu} and {lo si'o te vecnu} 15:45 < durka42> another thing we could try is {lo'e ka}... 15:46 < zipcpi> Ahahahaha 15:46 < selpahi> That's so 2003. 15:46 < ldlework> I don't understand how lo si'o vecnu is not different than lo si'o te vecnu 15:46 < ldlework> Especially if you think of it interms of structure or schema 15:46 < durka42> it's only really different in emphasis 15:47 < ldlework> I would say that, attempting to fill the empty places of a si'o proposition is to misunderstand what si'o does 15:47 < ldlework> with anything, zo'e, ce'u, zi'o anything 15:47 < durka42> right? what's the difference between buying and selling anyway? both have a buyer and a seller... 15:47 < ldlework> We /want/ the proposition that doesn't fill places with anything 15:47 < durka42> I can get behind that maybe 15:47 < ldlework> As to select all the propositions that have the same form that fill it with different things 15:47 < ldlework> like 15:47 < ldlework> lo si'o citka lo ladru 15:48 < ldlework> this is the concept of drinking milk 15:48 < durka42> it's basically the concept of these places being able to be filled 15:48 < ldlework> you drinking milk 15:48 < ldlework> me drinking milk 15:48 < ldlework> no one drinking milk 15:48 < ldlework> etc 15:48 < zipcpi> "The color red is beautiful." "Buying is beautiful." "Selling is beautiful." ? 15:48 < ldlework> with all possible prepositions and tenses and so on 15:48 < zipcpi> I dunno, I admit I'm straining for examples here 15:48 < zipcpi> But I'm not good with them 15:48 < durka42> yeah. I don't see why color-concepts are so special tbh 15:48 < ldlework> I gave one 15:49 < ldlework> lo si'o xunre cu melbi 15:49 < ldlework> all the possible propositions involving something being red, is beautiful 15:49 < selpahi> ldlework: I don't think your interpretation is in conflict with usage. 15:49 < ldlework> the concept of red is beautiful 15:49 < zipcpi> I meant to explain why I'd want to distinguish {lo si'o vecnu} from {lo si'o te vecnu} 15:49 < ldlework> since si'o is never a function, it doesn't matter 15:49 < ldlework> since it never has variables applied in some order 15:50 < ldlework> it is a "mask" or "schema" to other possible concrete propositions 15:50 < ldlework> where as "lo du'u citka" is a proposition in which something is eating something, just unspecified 15:50 < ldlework> "lo si'o citka" is not 15:50 < selpahi> You can argue that buying is different from selling, but {te} shouldn't really be enough to convey that difference, since by definition {te} does nothing but rearrange places, not add meaning related to agency etc. 15:50 < zipcpi> I actually kinda like the {lo'e ka} idea, but others might not agree :p 15:51 < ldlework> Yea I think you'd use a lujvo for that 15:51 < durka42> zipcpi: yay :) 15:51 < ldlework> I dunno why my idea doesn't resonate zipcpi 15:51 < selpahi> ro bu'a zo'u pa da ka ce'u bu'a .i ja'o lu lo'e ka broda li'u lu lo ka broda li'u ma frica 15:52 < durka42> ro ckaji cu pamei xu 15:52 < selpahi> ro da pa mei 15:52 < ldlework> Also 15:52 < zipcpi> selpahi: Yeah that's the whole problem with defining {erve} or {vujnu} isn't it? 15:52 < ldlework> by referring to {lo si'o te vecnu} you explicitly refer to all the propositions where the buyer is emphasized 15:52 < selpahi> That's the motivation for {erve}. 15:52 < ldlework> I think it still works 15:52 < ldlework> Even though you denote all the propositions that have the same truth values of those that use {vecnu} instead 15:53 < zipcpi> Cause propositionally {erve} = {te vecnu} and {vujnu} = {se sumji}, but the agentive semantic and li'o li'o li'o 15:54 < ldlework> In fact, by referring to {lo si'o vecnu} you refer to *none* of the predicates that emphasize buying, and vice versa 15:54 < ldlework> none of the propositions* 15:54 < zipcpi> I understand why {erve} is created. I just don't know how to define it in Lojban other than through {te vecnu} 15:54 < ldlework> I would say the only possible alternative is bu'u 15:55 < ldlework> Like how do you define te vecnu beyond vecnu outside of emphasis? 15:55 < zipcpi> Hm 15:55 < ldlework> all the same referrent sets have the same exact truth values for propositions of either form 15:55 < ldlework> they _are_ the same proposition 15:55 < ldlework> conceptually 15:56 < ldlework> It seems strange to think that predicates in the abstract sense have order outside locution 15:56 < durka42> oh no... someone used the word "locution" 15:56 < durka42> time to go zo'o 15:57 < ldlework> Does anyone else like my model of si'o? 15:57 < ldlework> du'u then becomes easily explainable 15:57 < ldlework> instead of remaining as unfilled propositions 15:57 < ldlework> or "degenerate propositions" 15:57 < ldlework> they get filled with zo'es 15:58 < ldlework> to become concrete 15:58 < ldlework> and have truth 15:58 < ldlework> nu becomes temporally oriented du'u 15:58 < zipcpi> So {si'o} means "concept", so it's the abstract concept of all propositions that have that form? 15:58 < ldlework> zipcpi: yeah basically 15:58 < ldlework> ka becomes parametric nu 15:59 < zipcpi> So then empty places in si'o aren't exactly {zo'e} then... not sure what they are. Maybe the kurtyvla {mai'i} 16:00 < zipcpi> I know more about kurtyvla than any human has a right to; I need to wash my brain T.T :p 16:01 < durka42> en:mai'i 16:01 < durka42> vlaste: mai'i 16:01 < vlaste> mai'i = pro-sumti: the universal argument/value; syntactically-contextually and type-permitted maximally generic in its typing 16:01 < durka42> we need a kurtyvla for "mark next word as kurtyvla" 16:01 < zipcpi> {ci'ai}? :p 16:01 < selpahi> There are already two ci'ai 16:02 < zipcpi> Not in JVS 16:02 < selpahi> http://jbovlaste.lojban.org/dict/ci%27ai 16:02 < selpahi> The other one was mine, but it was later added as mei'e 16:02 < zipcpi> Yeah I only see one 16:03 < ldlework> I mean you can easily just understand them as unfilled 16:03 < zipcpi> But see I proved my point. As soon as durka42 said that I immediately thought of the right kurtyvla 16:04 < selpahi> Why not {ce'u} 16:04 < ldlework> There's no inherent need to fill the places with an arbitrary thing. The grammar wouldn't require it. 16:04 < ldlework> The idea that they are degenerate incomplete proposition helps with the conceptualization as well 16:04 < ldlework> And the way in which they become du'u 16:05 < zipcpi> selpahi: Because if we fill all the places with {ce'u} then we can't use them for "the color red" 16:05 < ldlework> du'u should be the first truth-bearing level of abstraction 16:05 < zipcpi> Or "buying"-ness 16:05 < selpahi> Why not? 16:06 < zipcpi> Well xunre has only one place, but what of other predicats we might want to convert? 16:06 < selpahi> What of them? 16:06 < ldlework> selpahi: what's the functional motivator needing to fill them with a thing that wont/doesn't actualy do anything 16:06 < durka42> you were the one who said it doesn't work! 16:06 < ldlework> It doesn't change the structure, semantics or anything 16:06 < ldlework> It seems like an arbitrary requirement 16:07 < selpahi> The reason it doesn't work is not because of {ce'u}, but because of {xunre}. If you argue that it works, I wonder why you think {ce'u} can't fill the places for you 16:07 < ldlework> That the si'o places can be left unfilled, sort of justifies the way in which they suck in all possible values 16:07 < zipcpi> To me there's a big difference between any definition of {si'o} that would be useful, and {me'ei} 16:08 < ldlework> I don't even understand what the role of ce'u is in a non-function? 16:08 < durka42> well, it'd be nice if I didn't have to worry when I'm carrying around a si'o, that some brivla is going to come out of nowhere and shove sumti into the ce'us 16:08 < ldlework> I think I missed that, if it was explained above 16:08 < ldlework> okay so they can be filled 16:08 < ldlework> if they can be filled, it seems to break the utility of the thing 16:08 < ldlework> if they can't, they seem superflous and noisy 16:09 < zipcpi> But what I don't get is why we need new brivla that couldn't be algorithmically formed from {xunre}, because it is unsustainable otherwise 16:09 < selpahi> Lojban simply requires that there be something in each sumti place. 16:09 < zipcpi> It's not just colors 16:09 < ldlework> Since when? 16:09 < zipcpi> But "wetness", "dryness" 16:09 < selpahi> Since 1987 16:09 < zipcpi> anything-ness 16:09 < ldlework> selpahi: where is that described? 16:09 < ldlework> selpahi: also doesn't it seem arbitrary to you? 16:10 < zipcpi> So either use {si'o}, or some new NU 16:10 < ldlework> unframed from some prescription 16:10 < selpahi> I didn't make the language. 16:10 < ldlework> selpahi: sure I'm asking for your own reasoning 16:10 < selpahi> {zo'e} and {zi'o} exist because of this 16:10 < ldlework> But they do things that we don't want here 16:10 < ldlework> We certainly don't want to modify our predicates 16:10 < ldlework> And we don't want to denote anything in particular and just be silent about what is denoted. 16:11 < durka42> I think zi'o does exactly what we want, but what do I know 16:11 < ldlework> I'd say zi'o creates a new predicate 16:11 < zipcpi> durka42: What does a predicate with zero places mean? o.o 16:11 < ldlework> A predicate that is useful and standalone 16:11 < ldlework> IE, an entirely different distinct predicate 16:11 < durka42> a predicate stripped down to its essential concept and nothing more? 16:11 < selpahi> zipcpi: {lo ka ce'u broda} gives you "broda1-ness". As long as you have a suitable broda1 then you're good. 16:11 < ldlework> I'd say a predicate with no places is no predicate at all 16:11 < ldlework> It defies the very nature of predication 16:12 < durka42> selpahi: so what's the difference between redness and dryness 16:12 < durka42> ldlework: so does a predicate with unfilled places :) 16:12 < zipcpi> en: sudga 16:12 < zipcpi> vlaste: sudga 16:12 < ldlework> durka42: But si'o's are not predicates 16:12 < vlaste> sudga = x1 is dry of liquid x2; (adjective:) x1 is arid. 16:12 < ldlework> They're degenerate 16:12 < selpahi> Well, what I want is the color red. 16:12 < ldlework> err propositions 16:12 < ldlework> selpahi: malgli 16:13 < ldlework> "the color red" in English denotes a relation 16:13 < durka42> the color red ‽= redness 16:13 < ldlework> just like "dryness" 16:13 < ctefaho> you can't get the color red from just xunre 16:13 < ctefaho> you need something else 16:14 < ldlework> It depends on what it means to "get at a relation" 16:14 < ldlework> If you want to refer to a relation, that's pretty easy in lojban 16:14 < selpahi> ctefaho: You get me :P 16:14 < ldlework> If you want to refer to a relation without being forced to denote what is *being related* I argue you need si'o 16:14 < ctefaho> do reflect on the difference between having a color and being a color 16:14 < ctefaho> selpahi: I solved this ages ago 16:14 < ldlework> do reflect on the difference between being wet and being wetness 16:15 < durka42> O enlighten us then, dear ctefaho 16:15 < selpahi> ctefaho: Good for you. I did, too. But whenever I bring it up I am told to use si'o. 16:15 < ctefaho> none of you are going to like my solution 16:15 < ctefaho> except perhaps selpai 16:15 < durka42> wait, so is the color red the same as redness or not 16:15 < durka42> oh I misread the above 16:15 < ldlework> which part 16:16 < durka42> so apparently there are two people who have solutions, but they are both secret 16:16 < durka42> and all we are sure about is that {si'o} doesn't work, even though we can't explain why 16:16 < durka42> great!! 16:16 < ldlework> what 16:16 < ctefaho> I am going to give you a super-hint and just point out that your skari tries to do both a unary and a binary relation 16:16 < ldlework> Why doesn't si'o work? 16:16 < durka42> I don't know 16:16 < durka42> apparently it doesn't describe the color red 16:16 < durka42> because colors are more special than other properties for some reason 16:17 < durka42> why isn't there "the color dry"? 16:17 < ldlework> Yeah I think it is the case of English using one part of speech's form for another. 16:17 < ldlework> IE the definite reference to refer to a predicate 16:17 < selpahi> durka42: My solution is anything but a secret. I talk about it anytime colors come up. Like 20 minutes ago. 16:17 < durka42> okay, sorry 16:17 < ldlework> IE, in English, the predicates which ascribe color have been given proper names 16:17 < selpahi> Make a predicate for each color, and one for "x1 has color x2" 16:17 < durka42> I must have read too fast 16:17 < durka42> yeah, I'd get behind that solution 16:18 < durka42> I still don't see why colors are special though 16:18 < ldlework> But 16:18 < ldlework> What is true of that predicate?! 16:18 < zipcpi> But it's not sustainable. Case in point. I just made the words {skarxuciba} and {skarxetmele} 16:18 < ldlework> No there's a bigger reason why that's bad 16:18 < zipcpi> Are we gonna have abstract-color words for those too? 16:18 < ldlework> What is true of that red predicate? 16:18 < durka42> it's not sustainable if we need a new brivla for every {kambroda} 16:18 < ldlework> x1 is-red-itself 16:18 < ctefaho> ldlework: I am very interested in your take on ka and si'o, though 16:18 < ldlework> what is true of this predicate? 16:18 < selpahi> Juss that. 16:18 < ldlework> no 16:18 < ldlework> you know what I mean 16:19 < ldlework> what could ever logically be quantified and be true of this predicate 16:19 < durka42> seems like it can only be used as {lo broda} 16:19 < ldlework> other than a tautology 16:19 < ctefaho> Something being a color and something having a color is not the same thing 16:19 < durka42> nobody said it was 16:19 < ldlework> right I'm asking, what Thing is true of the predicate "is-red-itself" 16:19 < selpahi> The color red. 16:19 < durka42> it seems like there is only one such Thing, the same way selpa'i was saying earlier that there is only one {lo ka broda} 16:20 < ldlework> selpahi: so you realize you've limited your expansion of the predicate to a tautology 16:20 < durka42> but if we can't fill the place in lojban... 16:20 < ldlework> there is no way to refer to anything in the world, that is true of the predicate, other than using the predicate itself 16:20 < ldlework> In GEB, a very nice insight is made 16:20 < ldlework> That predicates are defined in terms of what other predicates, are true of the things that the predicate in question is true of. 16:21 < ldlework> Beyond the utilitarian existence of dictionaries that is 16:21 < ldlework> If you want to understand what a predicate means 16:21 < ldlework> You look at all the things that are true of that predicate 16:21 < ldlework> And the properties they share, IE, third-party predicates that are also true of that set 16:21 < ldlework> Defining red like this is utterly uselss. 16:21 < selpahi> lo [se] skari be ta cu *red. lo *red cu melbi [se] skari 16:22 < ldlework> all you have is 16:22 < ldlework> lo *red cu *red 16:22 < ldlework> that's the best you can ever do 16:22 < ldlework> "Red" denotes a predicate, not a noun 16:22 < ldlework> You've been fooled. 16:22 < durka42> I'm not gonna go for hyperbole like "utterly useless"... but it does seem like a predicate with very limited application, and that language constructs are at least _supposed_ to cover 16:23 < durka42> it solves the one problem of referring to "the color red", but I don't know what that is anyway 16:23 < selpahi> The predicate is "to be the color red". 16:23 < selpahi> That's not a noun. 16:23 < durka42> or why we wouldn't end up needing a predicate for "the color dry" 16:23 < ldlework> selpahi: right but what you're doing is 16:23 < ldlework> making a predicate that attributes predicateness 16:23 < ldlework> because you have to expand "the color red" anytime it appears 16:23 < ldlework> even in your definitions 16:23 < ldlework> to a relational phrasing 16:24 < ldlework> because that's what it really means in the English, (and elsewhere I'd argue) 16:24 < ldlework> The predicate is "to be the predicate is red" 16:24 < ldlework> rednessness 16:24 < ldlework> Its like 16:24 < ldlework> "He left me in the lurch." 16:24 < ldlework> English abuses the definite reference for things that are not definite references all the time 16:25 < ctefaho> Red is red 16:25 < selpahi> This is a good time to stop the discussion. 16:25 < ctefaho> Redness is redness 16:25 < ldlework> If you replace red with wet, you'll see. 16:25 < ctefaho> Red is not redness 16:26 < ldlework> You haven't addressed the assertion that Red isn't really a definite reference in English 16:26 < ldlework> You keep repeating an arguement that is addressed by that assertion 16:26 < ctefaho> I know what red is and isn't in English 16:27 < ctefaho> also I am going to sleep 16:27 < ctefaho> ciao 16:28 < zipcpi> I don't know what it means to be red independent of {ko'a ckaji lo ka xunre} 16:28 < e`ogan> Welp all natlangs are guilty of that 16:28 < ldlework> en: skari 16:29 < e`ogan> Even Esperanto 16:29 < ldlework> ... 16:29 < zipcpi> I don't know what {skari} is good for other than being a specialized {ckaji} either. 16:29 < e`ogan> red is an approximation of a lot of colours 16:29 < ldlework> zipcpi: the bigger point I think is, notice that skari doesn't denote colors 16:29 < zipcpi> vlaste: skari 16:29 < ldlework> it ascribes colors to other things 16:29 < vlaste> skari = x1 is/appears to be of color/hue x2 as perceived/seen by x3 under conditions x4. 16:29 < ldlework> that implies that skari denotes properties 16:31 * nuzba @masterq_mogumog: 後置か中置で述語を適用するのか。すごい聞いてる方が脳味噌がこんがらかりそう。。。と思うのは今の自然言語に慣れた脳味噌だからなんだろうか。 #ロジバン #述語論理 [http://bit.ly/1GIfiuY] 16:32 < durka42> selpahi: thanks for discussion. need to think more 16:33 < durka42> also need to sleep 16:37 < zipcpi> I also don't know how "Red is a beautiful color" is different from {lo'e xunre cu melbi} lol 16:38 < selpahi> Because it means "Red things are beautiful". 16:38 < selpahi> The first sentence makes no claims about any objects at all. 16:39 < selpahi> And it shouldn't, either 16:39 < zipcpi> Except you consider the "color red" an object 16:39 < zipcpi> Otherwise nothing can go into x1 of this new predicate 16:39 < selpahi> Imagine you like the color red, but happen to hate all things that are red. The two are not really related. 16:40 < selpahi> I like the color blue, but I hate all my blue toys, because they are old and broken. 16:40 < selpahi> Not "nothing", but any color that qualifies as red. 16:40 < selpahi> What's the problem with that... 16:40 < zipcpi> But it's like durka42 said, I don't see how "extracting" a color from "objects that are a color" is different from "extracting" dryness from "objects that are dry" 16:41 < zipcpi> I meant 'object' in terms of {sumdai} 16:41 < zipcpi> Or {sumdza}... whatever 16:42 < selpahi> Or {pavmei} 16:42 < zipcpi> lo mu'oi gy the color red gy : lo xunre = ??? : lo sudga 16:45 < zipcpi> This is the problem I'm referring to... what is the relation described here? 16:45 < zipcpi> And if we need new brivla to solve this problem for colors, why don't we need brivla for everything else? 16:46 < ldlework> If you like the color blue but hate all your toys, nothing is in violation because its not the toys that you like as an entail of liking blue, its liking _that_ your toys are blue that is an entailment of liking blue 16:47 < selpahi> I don't care that they are blue. 16:48 < selpahi> Their color has nothing to do with me disliking them, nor with me liking the color they happen to have. 16:48 < selpahi> Objects are *irrelevant* to the color. 16:49 < ldlework> I was trying to also exhibit that the quality of the toys is divorced from your liking of blue/blueness 16:52 < durka42> not exactly irrelevant 16:52 < durka42> since they satisfy the property of having that color 16:52 < ldlework> Basically my problem with your solution is that it entail creating N predicates which are special in that they entail no other conjunctive relations to any other predicates. 16:53 < ldlework> Or rather, conceptually, such conjunctive relations are impossible 16:55 < ldlework> Its a strange form of raising. To create predicates, which are only true of an individual single other predicate (isxunre is only true of xunre). 16:56 < ldlework> We have a grammar for creating such nonce predicates which are true of specific predicates 16:57 < ldlework> Its like wanting water. 16:57 < ldlework> You want something to do with the water. 16:58 < ldlework> You can't like red. 17:00 < zipcpi> srana fa lo dotygu'e sei me'oi http://www.itchyfeetcomic.com/2013/11/second-mother-tongue.html#.VbF_sPmqpBc 17:02 < ldlework> I think one of the interesting thing about Lojban is that it proves that all there is is predicates 17:02 < ldlework> Any noun you can create in any language is really just the presupposition of a true attribution of some predicate 17:03 < ldlework> This is why there is no such existential referent for the predicate is-the-color-red beyond the property, or predicate relation of some other thing exhibiting redness. 17:05 < durka42> co'o 17:06 < e`ogan> I wonder if there could be devised an artificial colour system naming that would allow to describe pretty much all colours accurately with just a few names... 17:06 < e`ogan> After all we have luminosity of colours too 17:07 < zipcpi> I just made {skarxuciba} and {skarxetmele} lol 17:07 < zipcpi> Which theoretically can describe any color in RGB space 17:07 < zipcpi> But... 17:09 < e`ogan> YCbCr would be the best way to describe colours 17:09 < zipcpi> I don't know how to define that 17:09 < e`ogan> Basically it is an xyz 3D definition of colour 17:10 < ldlework> Even if you wanted to talk about "The wavelength of red" there is no such thing as "a wavelength", its a property too. 17:10 < e`ogan> where xy are the colours and Z (named Y) is the luminosity 17:10 < ldlework> e`ogan: are you into color theory? 17:11 < e`ogan> I'd go one step further and add another dimension 17:13 < ldlework> e`ogan: you might really enjoy, https://www.youtube.com/watch?v=xAoljeRJ3lU 17:13 < e`ogan> Basically have a colour wheel that is extruded into a cylinder where a degree° describes colour, the Z axis value describes the luminosity, and the distance from the center of the wheel describes the degree of reflection 17:14 < deltab> sounds like HSL, except I don't know what you mean by reflection 17:14 < deltab> https://en.wikipedia.org/wiki/HSL_and_HSV 17:14 < e`ogan> CMYK-RGB-mirror 17:15 < e`ogan> Basically CMYK is the format of colours that bounce off the ambient light, RGB is the colour of the light emitter 17:16 < e`ogan> Basically with same three numbers (255;255;255) you could describe more parameters than Y Cb Cr 17:17 < zipcpi> skarxuciba, skarcinupe, skar... dunno 17:17 < zipcpi> skarje'ike skarjejbujcru 17:17 < zipcpi> ... darnit 17:18 < e`ogan> by making colour a degree of a colour wheel I'd have a distance from the center and the Z axis as another two parameters 17:18 < zipcpi> skarpantone 17:18 < ldlework> Oh there is one predicate true of color predicates, in that they are true of the predicate 'is-a-color' 17:19 < zipcpi> Basically the problem here is what goes into skari2 17:19 < ldlework> So basically, creating predicates for color predicates would allow you to avoid using abstractions in the cases where the predicate {skami} appears 17:19 < ldlework> oops skari 17:19 < ldlework> I think that would be the less lojbanic thing to do 17:19 < ldlework> I think we should be proud of lojban and use lo NU xunre to refer to redness, just as we would if we were talking about wetness 17:20 < ldlework> But that argument seems baseless since there's nothing inherently wrong with creating predicates to refer to predicates 17:20 < ldlework> just seems like a slippery slope 17:20 < ldlework> but maybe colors are just naturally suited to being specially denoted 17:21 < ldlework> I'm not really sure tbh, I think its a reasonable debate 17:22 < ldlework> "more words" probably wins out 17:25 < e`ogan> I'd vote for less words that can domore when combined 17:25 < e`ogan> do more* 17:26 < ldlework> "There is almost universal consensus, however, on the idea that only properties can be predicated and exemplified. For example, ordinary objects like apples and chairs cannot be predicated of, and are not exemplified by, anything." 17:27 < ldlework> Since things other than redness itself can "be red", this sort of reinforces the idea that the kind of thing that Red is, is a property. 17:27 < ldlework> If we're to objectify it, my making a predicate that denotes it, surely, that predicate denotes the predicate "is-red" 17:28 < e`ogan> I agree 17:28 < e`ogan> red is not an object on itself 17:29 < ldlework> e`ogan: or rather, in english, the noun-phrase "a red", or "that red" denotes a property not an object (beyond the sense in which properties /are/ objects) 17:30 < e`ogan> I am sure we could get to the basic root of it: predicates are things we perceive and objects are more the things that are true regardless of our perception of them 17:31 < ldlework> well, more plainly, predicates ascribe properties, and objects are the things they ascribe them to. 17:31 < ldlework> Its a matter beyond perception 17:31 < e`ogan> Chair is a chair for everyone: it is a contraption to have a seat; while the colour, softness, comfort are the things we perceive about the said chair 17:32 < ldlework> I agree, but I think that's orthogonal to what kind of thing "the color red" is 17:32 < e`ogan> Well language relates strongly to perception, so i don't think it can really deal in absolute terms 17:33 < ldlework> But what red is, isn't a matter of who sees it. 17:33 < ldlework> No more than what "wet" is. 17:33 < e`ogan> Yes "the red" is a bit of a misnomer 17:33 < ldlework> Exactly. 17:34 < e`ogan> Welp i'd disagree: we'd perceive both wet and red differently because we have different number of receptors for them and we taught ourselves to perceive things in our own unique ways 17:34 < ldlework> It has nothing to do with whether statemetns are truth, or isomorphic with the world 17:34 < ldlework> Its a more primitive matter 17:35 < ldlework> Its a question of what part of speech "red" really is. 17:36 < ldlework> Color names in general. 17:36 < ldlework> If you had a predicate that meant, "exudes-redness" you could point to a firetruck and say that it is true of this predicate. 17:37 < ldlework> If you had a predicate that meant, "is-literally-redness-itself" what could you point to and say that it is true of? 17:37 < ldlework> Even in the abstract 17:37 < ldlework> It is basically only true of itself, and of its determinable, "color" 17:38 < ldlework> (on determinates and determinables, http://plato.stanford.edu/entries/determinate-determinables/) 17:40 < ldlework> In otherwords, if this "is-literally-redness-itself" predicate was {broda} the only way to tell me what {broda}'s is by saying {lo broda cu broda} 17:40 < ldlework> Because there would be no other predicate that you could use to refer to whatever we're referring to in this sentence. 17:41 < ldlework> And the only other substantial statement you could make before going into the utterly abstract is 17:41 < ldlework> {lo broda cu se skari} 17:41 < ldlework> But that's it 17:42 < ldlework> You could of course make other non-entailing statemetns like {mi nelci lo broda} and so on but we'd never be able to learn anything about what a {lo broda} is from statements such as those 17:42 < e`ogan> I see the issue now 17:42 < e`ogan> For all intents and purposes in lojban red needs to be a predicate 17:43 < ldlework> right and we have a perfectly awesome way to refer to predicates 17:43 < ldlework> a whole family of ways, called NU 17:44 < ldlework> Like I said before, there's nothing inherently invalid about creating a predicate which refers to some other predicate. Not really. But it seems exclusionary and malgli. 17:45 < ldlework> And I think the whole business about which NU you should use to say something like "I like Red." is a non-sequitur 17:45 < ldlework> Any NU you use, will be technically correct, but will have different senses just like always 17:45 < ldlework> mi nelci lo ka xunre = I like my redness 17:46 < ldlework> mi nelci lo nu xunre = I like when things are red 17:46 < ldlework> mi nelci lo du'u xunre = I like that things (whatever they are) are red 17:46 < ldlework> mi nelci lo si'o xunre = I like redness 17:48 < ldlework> The du'u one can also be interpreted "I like that there exists the potential for redness and non-redness." 17:49 < ldlework> Since the du'u may contain a zo'e which points to "no da" or "su'o da" which implicates the posibilities of both redness and non-redness 17:49 < ldlework> Or even just a hidden negation of xunre itself. 17:49 < zipcpi> I'm not sure du'u works that way 17:49 < ldlework> du'u encodes propositions 17:50 < ldlework> everything I said, is an entailment of what propositions are 17:50 < zipcpi> Yes but it then has a truth value; how can it represent both X and na X? 17:50 < ldlework> the du'u here is vague 17:50 < ldlework> it represents any possible tense, or set of arguments, prepositional tags, and so on 17:50 < zipcpi> No, du'u is what goes into the x1 of jetnu or jitfa 17:50 < ldlework> until I whittle that space down by providing more explicit things in the du'u construction 17:51 < ldlework> right...? 17:51 < ldlework> or djuno, or jinvi 17:51 < ldlework> and so on 17:51 < zipcpi> But {ja'a} is the default 17:51 < ldlework> So? 17:51 < zipcpi> So how can you say that {lo du'u xunre} can represent {lo du'u na xunre}? 17:52 < ldlework> That a predicate exists and therefore can be used in propositional construction implies the possibility of the predication of descriptions 17:52 < ldlework> It doesn't imply anything about the world... 17:52 < ldlework> That anything actually does or does broda or xunre or whatever.. 17:52 < ldlework> a possible interpretation of {lo du'u xunre} is {lo du'u na xunre} and {lo du'u pu xunre} and {lo du'u le mamta be do cu xunre} and so on 17:53 < zipcpi> I don't think so 17:53 < zipcpi> Where does the {na} come from if {ja'a} is the default? 17:53 < ldlework> It really is. The ja'a is merely a default of pragmatics 17:53 < zipcpi> There's a reason I invented {cau'a} 17:53 < ldlework> Its the default illocutionary force in speech, but not grammatical construction 17:54 < ldlework> Its mechanical observation that things unspecified in a du'u are unspecified. 17:55 < ldlework> How you are to be taken in the case of unspecified speech is a different question all together. 17:56 < ldlework> Anyway, supposing that logical affirmation is "built-into" lojban propositions, unlike tense, or other parts of speech, the points still remain. 17:56 < ldlework> If I was an inference engine 17:57 < ldlework> From the mere apperance of the statement {lo du'u broda} I can infer the possibility of {lo du'u na broda} 17:57 < ldlework> That's all I mean 17:58 < ldlework> Not that the statement means, or should be interpreted as the negation, but that the conceptual boundries of what is possible is expanded both ways by the existence of a predicate. 17:58 < ldlework> Hope that makes it more clear 17:59 < zipcpi> Hm 17:59 < ldlework> Just as the statement {lo du'u na broda} suggests a conceptual potentiality of {lo du'u broda} 18:00 < ldlework> If one thing isn't broda'ing, it makes sense, logically, that other things might broda. 18:00 < zipcpi> I'm not sure... what about {lo du'u li re sumji li pa li pa}? 18:00 < ldlework> Even if it turns out that nothing broda's real or imagined. 18:00 < ldlework> Is there a predicate there? 18:00 < zipcpi> {sumji}? 18:01 < ldlework> you don't need to close the li? 18:01 < ldlework> oh dur 18:01 < ldlework> Sure, I don't see why this doesn't also imply the existence of the statement {lo du'u li re na sumji li pa li pa} 18:02 < ldlework> Which could be true under a different interpretation of mathematics or simply exist as a false statement. 18:02 < zipcpi> Hm 18:02 < ldlework> But it allows for other true statements with other arguments. 18:02 * zipcpi nods 18:02 < ldlework> Even if it didn't.. 18:04 < ldlework> The nice thing about formal systems is being able to derive new statements typographically 18:08 < e`ogan> I found an interesting combo of english words: [footage a rest ore] or [foot age are store] 18:09 < e`ogan> spaces, they matter all too much in english 18:09 < zipcpi> Technically spaces are phonemic in Lojban writing too, unless you mark stress 18:09 < zipcpi> LOjbo bau vs lo JBObau 18:10 < zipcpi> But if you mark stress, scriptum-continuum is possible 18:10 < zipcpi> Oh yeah and mandatory pauses 18:17 < zipcpi> More pertinent though is phonetic word-boundary ambiguity, like "propagate" vs "prop a gate" :p 18:27 < dmiles_afk> in coding an infenerce engine negation (actualyl MOdus Tolens logic) is alwasy the funnest 18:29 < dmiles_afk> languages that purposely remove the concept of negation by logic (such as RDF) end up becoming unusable to a logicain 18:30 < dmiles_afk> all i am saying is negation *must* be said to invert quantifiers and even nescisty/possibility 18:31 < zipcpi> That can be achieved by fronting {na} 18:31 < dmiles_afk> (regardless of the speakers intent to recognise their logic as modal) 18:32 < dmiles_afk> (regardless of the speakers intent to not recognise their logic as modal) 18:57 < deltab> e`ogan: foot a gear e-store 18:58 < e`ogan> Indeed XD 20:00 < noncomcinse> coi .irci prenu noi jbobau ki 20:17 < bigcentaur> ki ki'a 20:18 < noncomcinse> .i tcekitaujau 20:18 < noncomcinse> .i zo ki zo ke'a simxu loka ce ce binxo 20:19 < noncomcinse> .i zo ce zo ce'u go 20:19 < noncomcinse> si go'i 20:19 < noncomcinse> ua pei 20:20 < noncomcinse> do'oi bigcentaur 20:21 < bigcentaur> uanai 20:21 < noncomcinse> ce ki tau jau dialect 20:22 < bigcentaur> i'm passing some of your sentences through camxes, but they're not parsing 20:22 < noncomcinse> It swaps the following pairs of words: ce/ce'u ki/ke'a tau/tu'a 20:22 < bigcentaur> what does? 20:22 < noncomcinse> Hang on. 20:24 < noncomcinse> tcekitaujau 20:24 < noncomcinse> Forgot to copy the link. 20:24 < noncomcinse> bigcentaur: http://mw.lojban.org/papri/ce_ki_tau_jau 20:26 < noncomcinse> vlaste: jo'u 20:26 < vlaste> jo'u = non-logical connective: in common with; along with (unmixed). 20:27 < noncomcinse> camxes: jo'au tcekitaujau loka ce ce klama 20:27 < camxes> SyntaxError: Expected [,] but "k" found. 20:28 < noncomcinse> camxes: +exp jo'au tcekitaujau loka ce ce klama 20:28 < camxes> SyntaxError: Expected [,] but "k" found. 20:28 < noncomcinse> Man, and I thought English dialects were hard. 20:32 < noncomcinse> bigcentaur: Once you get used to saying {djuno lodu} instead of {djuno lodu'u}, you don't want to go back. 20:48 < bigcentaur> i'll take whatever camxes says is legal 20:48 < noncomcinse> bigcentaur: so just swap ce/ce'u du/du'u ki/ke'a 20:49 < noncomcinse> 22:00:26 <noncomcinse> coi .irci prenu noi jbobau ki -> coi .irci prenu noi jbobau ke'a 20:50 < bigcentaur> don't get me wrong, i'm all for having the most useful cmavo being as short as they can 20:51 < noncomcinse> I strongly disagree with su/su'o swap tho 20:51 < noncomcinse> Because it breaks an IAU patter 20:51 < noncomcinse> zi za zu si sa su etc 22:33 * nuzba @hinoki__: 寝起きで口が回らない場合や、思い出し思い出し言葉を選ぶときについ言葉尻が伸びてしまうとき、あらぬ誤解が無くなる。lojban言語(logになってた:P) https://twitter.com/hinoki__/status/624451320397737985 [http://bit.ly/1VDA8H6] 23:27 < gleki> ugly 23:27 < gleki> fegli 23:27 < gleki> i cant edit russian definitions 23:27 < gleki> looks like the only way is not to use jvs 23:27 < gleki> oh, si, sa,su 23:28 < gleki> i wonder how we can can {sa} powerful but not make the grammar swell 23:31 < gleki> en:coi 23:39 < zipcpi> Maybe what things like "the color red" or "dryness (in general)" is really something like {si'o} + {ka} 23:40 < zipcpi> "x1 is the concept of the property shared by all things that fit this description; ce'u binds to the subject of description" 23:40 < gleki> i dont understand ur color problems at all 23:40 < zipcpi> Or of course there is also {lo'e}; but... hah try convincing selpahi to use {lo'e} 23:41 < zipcpi> I don't fully understand it either 23:41 < zipcpi> But what do you use? 23:41 < gleki> xunre1 is any red object, yes. read Uqbar, it can cover half of the universe, this xunre1 23:41 < zipcpi> OK. Then what goes in skari2? 23:42 < gleki> either {lo xunre} or {lo ka xunre} or {lo su'u'u'u'u'u xunre} depending on what decide in bpfk 23:42 < zipcpi> lol 23:42 < gleki> *what you 23:43 < gleki> but i'd suggest {lo xunre} because Zipf 23:43 < zipcpi> What *I* decide? Won't that be the BSFK then? :p 23:43 < gleki> or bsfk 23:43 < gleki> so many committees 23:43 < zipcpi> lol 23:43 < gleki> i need to create "Academy of Lojban" then 23:44 < gleki> http://mw.lojban.org/papri/Academy_de_Lojban 23:45 < zipcpi> Hmm... how about "I see the color red" when there is actually nothing with the color red. Maybe you're hallucinating 23:45 < zipcpi> And know you're hallucinating 23:46 < zipcpi> Maybe it's really {se xunre} lol 23:46 < gleki> it's about external predicate 23:46 < gleki> something like {menlyganse} 23:46 < gleki> lujvo: menlyganse 23:47 < gleki> lujvo: menli ganse 23:47 < mensi> menga'e[6387], menganse[7937], menlyga'e[8517], menlyganse[10067] 23:47 < gleki> i need to fix this lujvo tool 23:48 < zipcpi> Or maybe something that's just "the sensation of"... don't think {li'i} works though 23:48 < zipcpi> "the sensation of red" 23:48 < zipcpi> "the sensation of dryness" 23:49 < gleki> no, not li'i, a new brivla is needed 23:49 < zipcpi> And maybe a new NU 23:49 < gleki> "x1 mentally perceives x2 as if x1 used sensation x3 for real" 23:50 < gleki> lujvo will work here 23:50 < gleki> since it's not a borrowed concept 23:50 < gleki> btw, what is the term for purely apriori words like {kreivi}? 23:50 < gleki> not jbofuvi 23:51 < gleki> zi'evla as opposed to fu'ivla 23:51 < gleki> but zi'evla is a bit misused 23:51 < zipcpi> Hm we might need a word for "a priori" 23:51 < gleki> *but {zi'evla} is a bit misused 23:51 < gleki> en: jbofuvi 23:51 < mensi> jbofuvi = x1 is a Lojban-derived zi'evla with meaning x2 derived from word(s) x3 | From lojbo + fu'ivla, shortening of 23:51 < mensi> jbojbofu'ivla. See zi'evla. Examples of jbofuvi: crenzuue, gentufa, aiste, and of course, jbofuvi itself. | 23:51 < mensi> spheniscine 23:51 < gleki> en: mu'umgu 23:51 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 23:51 < gleki> jbo;' mu'umgu 23:51 < gleki> jbo: mu'umgu 23:51 < mensi> mu'umgu = x1 finti x2 lo nu x3 tcaci x4 | lo'e se mu'umgu cu valsi .i lo'e mu'umgu cu krici lo du'u lo da'i nu lo se 23:51 < mensi> mu'umgu cu zasti cu zmadu lo da'i nu lo se mu'umgu cu na'e zasti kei lo ka lo noi lo mu'umgu cu djica ke'a ku'o nu lo te 23:51 < mensi> mu'umgu cu tcaci lo ve mu'umgu cu lakne ce'u .i zo mu'umgu cu se'o sezysarxe se mu'umgu la .telselkik. lo nu mu'umgu kei 23:52 < mensi> loi jbopre .i ra'oi -mug- rafsi ca'e zo mu'umgu | selckiku 23:52 < zipcpi> Is that it? 23:52 < gleki> no 23:52 < zipcpi> Right 23:52 < gleki> x1 is a zi'evla that has the only etymology in Lojbanist's head 23:52 < zipcpi> lol 23:52 < gleki> zo'onai 23:53 < gleki> lo jbofuvi have etymology in other lojban words 23:53 < gleki> fu'ivla have etymology in borrowed words 23:53 < gleki> gismu mostly too 23:53 < gleki> lujvo are like jbofuvi 23:54 < gleki> but what is {kreivi}? 23:55 < gleki> i failed to find anything like that in any natlangs so i used my imagination or maybe arbitrary letter generator in my head and here it is 23:55 < zipcpi> lol 23:55 < zipcpi> Like I wonder how selckiku makes his words too 23:57 < gleki> en: etymlo 23:57 < gleki> en: etymol 23:57 < gleki> mensi, search! 23:58 < gleki> dont die 23:58 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 23:58 < mensi> 6 da se tolcri: vlakra, farka, pamniarkovni, u'imla, u'ivla, zgibite 23:58 < gleki> camxes: pamniakovni 23:58 < camxes> SyntaxError: Expected [.\t\n\r?! ] but "p" found. 23:58 < zipcpi> Maybe it's related to {ca'engau} or {autca} :p 23:58 < gleki> camxes: pamniarkovni 23:58 < camxes> SyntaxError: Expected [.\t\n\r?! ] but "p" found. 23:59 < gleki> that's basically it. mukti said he wouldnt fix JVS --- Day changed Fri Jul 24 2015 00:00 < gleki> why do we even need JVS. CLL says Lojban is complete :P 00:00 < zipcpi> lol 00:01 < gleki> why not call {kreivi} words blalalavla? 00:01 < zipcpi> lol 00:01 < gleki> any other suggestions? 00:01 <@Broca> Lojban _is_ complete, just poorly documented! (If we are to believe the 2002 baseline statement ...) 00:09 < zipcpi> $x_1$ valsi gi'e se finti fo lo poi'i se djuno le finti be $x_1$ ku po'o 00:10 < zipcpi> $x_1$ valsi gi'e se finti fo lo narju'odza be lo na'e finti be $x_1$ 00:12 < gleki> mensi: doi akmnl xu do nitcu lo fukpi be lo JVS datni sorcu i mi ka'e benji fi do i ma mrilu judri 00:12 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.akmnl.gy. di'a cusku da 00:20 < zipcpi> Hm that little discussion about {ganai/ko} makes me wonder if {du'u} is in fact the most natural type for minde3 00:21 < gleki> those nu/du'u rants ... 00:21 < zipcpi> lol 00:22 < zipcpi> Or just koinde3 00:23 < zipcpi> koinde is a word made to cover all the illocutionary "imperative" modes without implying "order one to do something" 00:23 < zipcpi> But could also cover "weaker" imperative modes 00:23 < gleki> koinde != minde 00:23 * zipcpi nods 00:23 < gleki> but as for e'inde im not sure 00:34 < zipcpi> Hm {joi'i} might be the only VUhU I'd use... and I'd use it more like JOI than VUhU o.o 00:35 < zipcpi> ke zo ko'oi jonai zo e'i ke'e jo'i ... 00:35 < zipcpi> *joi'i 00:42 < zipcpi> Huh, {poi'i} could actually be at least one possible solution to the {gau/gi'e} problem 00:42 < zipcpi> la djan poi'i gau ke'a lo broda cu brode kei gi'e brodi 00:43 < zipcpi> Hm... might not be very economic with the syllables though 00:43 < zipcpi> Not even with CKTJ 00:54 < gleki> *cu 00:55 < zipcpi> Oh right 00:55 < zipcpi> exp: la djan cu poi'i gau ke'a lo broda cu brode kei gi'e brodi 00:55 < mensi> ([la djan KU] [cu {poi'i <(¹[gau ke'a] [lo broda KU]¹) (¹cu [brode VAU]¹)> kei} VAU] [gi'e {CU <brodi VAU>} VAU]) 00:59 < zipcpi> lo kutydza cu vlakra :p 01:00 < zipcpi> I don't even know what {kutydza} actually means. It means whatever NULL means lol 01:02 < zipcpi> I'm pretty sure {nu} is from {fasnu} 01:03 < zipcpi> And I don't think it needs to be restricted to being a Lojban word, it's like {fu'ivla} 01:03 < zipcpi> {jbofuvi} is restricted to being a Lojban word, because it's definition depends on that of {zi'evla} 01:07 < zahlman> en: kutydza 01:07 < mensi> kutydza [< kunti da ≈ Empty X] = x1 is the null value/argument | See kunti, goilka'i, narju'odza, kutpoi, kutselcmi, 01:07 < mensi> kutlerpoi, kutyuenzi, zai'o | spheniscine 01:18 < zipcpi> Oh, I was talking about {blalalavla} btw 01:23 < gleki> NULL like in "the value hasn't been assigned yet"? 01:24 < zipcpi> Maybe. But you can also assign something *to* NULL 01:26 < zipcpi> I think, like NULL, it doesn't actually mean anything other than something to compare itself with 01:26 < zipcpi> lo kutydza cu kutydza 01:26 < zipcpi> lo na'e kutydza na kutydza 01:26 < gleki> okay, its default unassigned value but you are free to assign it back to it 01:28 < zipcpi> So {lo kutydza} = NULL. {cu kutydza} = isNull() 01:28 * zipcpi shrugs 01:29 < gleki> oh you just discovered nouns and verbs in programming languages :D 01:29 < zipcpi> Haha 01:38 < gleki> strange that you needed so many new words to emulate proglangs. ilmen ba'anairu'e reused existing words mostly 01:38 < gleki> but if you do create new words for proglangs then we need to start using them outside programming 01:38 < zipcpi> faubganke and ganfauke have use outside programming 01:38 < zipcpi> And goilka'i can be useful for instructional texts 01:39 < zipcpi> "Let X be the amount of mana paid here" 01:39 < zipcpi> Or something 01:39 < zipcpi> Ilmen used the nonce word {zvailu}, which I formalized as {ganfauke} 01:40 < zipcpi> {goilka'i} was invented to fill a shortcoming with using {goi} to assign variables 01:41 < zipcpi> And the kunti-stuff is invented because Lojban doesn't really have a way to not put anything into a {ce}/{ce'o} construct 01:42 < zipcpi> Xalbo used the kurtyvla {zai'o} at my suggestion, but the problem is its type is unspecified 01:42 < zipcpi> Wonder if we need an empty mass too. Not sure how a mass is used in programming 01:43 < zipcpi> But since Lojban has that type, it might as well exist 01:43 < zipcpi> loi no co'e 01:44 < zipcpi> Hm if zero is a problematic number for {ce}/{ce'o} constructs to have, one is problematic as well 01:45 < zipcpi> {ko'a ce zi'o}? {zi'o ce zi'o}? 01:46 < zipcpi> Same goes for {joi} 01:51 < gleki> how to specify type? 01:52 < zipcpi> I'm not sure what Curtis intends with {zai'o} Maybe {zai'o poi porsi} or {zai'o poi setmima} or something 01:56 < zipcpi> There's also {tu'o} but that's a PA 01:59 < gleki> i thought tu'o is {zi'o} http://mw.lojban.org/index.php?title=ELG._Context 01:59 < zipcpi> It's also why I added a section for "informal algorithms" to the fancylojban challenges page 01:59 < zipcpi> Yes, but what does {zi'o ce zi'o} mean? 01:59 < zipcpi> Can that be used for the empty set? 01:59 < gleki> i dont think so 02:00 < zipcpi> Yeah... 02:00 < gleki> zi'o ce zi'o cu simxu lo ka ce'u ce'u tavla lo bergu lo jbobau 02:00 < zipcpi> I think I might actually have to talk to Curtis about zai'o 02:00 < gleki> it means "Bergu is discussed in Lojban" but there can't be anyone to discuss it between each other. 02:01 < zipcpi> Whether zai'o can be used to mean {lo kutpoi ja kutselcmi ja co'e li'o} 02:01 < gleki> so it's a new predicate, and that's what zi'o is fort 02:01 < gleki> *for 02:01 < zipcpi> Or whether he intends it to be used like {zai'o ce zai'o} = {lo kutselcmi} 02:02 < zipcpi> Then it can be also used for sets with only one member 02:02 < zipcpi> {li pa ce zai'o} 02:05 < zipcpi> Hm I see you've already talked to him 02:05 < zipcpi> But I'm not sure then, that {zai'o} can be used the way I want it 02:05 < zipcpi> Ugh I think we might need a new cmavo 02:05 < zipcpi> {tu'oi} or something 02:05 < gleki> en: incita 02:05 < mensi> incita = x1 (text) is an adposition expressing relation x2 (property of nonce place, has places for ce'u referring to 02:05 < mensi> members of x3) and connecting members of the set x3 (ordered group of text) | The first member of x3 is what follows 02:05 < mensi> after the adposition (unless co is used) and is usually called 'complement' or 'argument'. The second member of x3 can 02:05 < mensi> sometimes be called 'noun phrase of adposition'; it can e.g. be the current clause (lo bridi), x1 of it, x2 of it etc. 02:05 < mensi> x3 doesn't need to represent the full specification, only the first member can be specified. To specify both members 02:05 < mensi> ce'o can be used. See also bridi, sumti, sumtcita | gleki 02:05 < gleki> i start talking like Curtis 02:05 < zipcpi> lol 02:06 < zipcpi> lo kurtyvla poi se finti la gleki 02:06 < gleki> si'au 02:08 < gleki> lu sepi'o li'u incita lo ka se te se pilno vau lu lo forca li'u ce'o lu lo ka citka li'u 02:08 < gleki> i mi citka sepi'o lo forca 02:10 < gleki> i va'i lu lo forca li'u cu te incita fi lu sepi'o li'u 02:11 < gleki> i na djuno lo jei jai se kufra 02:11 < zipcpi> exp: la djan joi ko'a 02:11 < mensi> ([{la djan KU} {joi ko'a}] VAU) 02:12 < gleki> exp: pa mlatu cu mo 02:12 < mensi> ([{pa BOI} mlatu KU] [cu {mo VAU}]) 02:12 < gleki> KU! 02:12 < gleki> mensi: gadri 02:12 < mensi> lo [PA] broda = zo'e noi ke'a broda [gi'e zilkancu li PA lo broda] 02:12 < mensi> la [PA] broda = zo'e noi lu [PA] broda li'u cmene ke'a mi 02:12 < mensi> lo PA sumti = lo PA me sumti 02:12 < mensi> la PA sumti = zo'e noi lu PA sumti li'u cmene ke'a mi 02:12 < mensi> loi [PA] broda = lo gunma be lo [PA] broda 02:12 < mensi> lai [PA] broda = lo gunma be la [PA] broda 02:12 < mensi> loi PA sumti = lo gunma be lo PA sumti 02:12 < mensi> lai PA sumti = lo gunma be la PA sumti 02:12 < mensi> lo'i [PA] broda = lo selcmi be lo [PA] broda 02:12 < mensi> la'i [PA] broda = lo selcmi be la [PA] broda 02:12 < mensi> lo'i PA sumti = lo selcmi be lo PA sumti 02:12 < mensi> la'i PA sumti = lo selcmi be la PA sumti 02:12 < mensi> PA sumti = PA da poi ke'a me sumti 02:12 < mensi> PA broda = PA da poi broda 02:12 < mensi> piPA sumti = lo piPA si'e be pa me sumti 02:12 < gleki> pa broda = pa da poi broda 02:12 < gleki> izo ku mo 02:13 < zipcpi> mo'u finti lo kurtyvla fa'o http://jbovlaste.lojban.org/dict/tu'oi 02:14 < gleki> xm i lu lo broda li'u simsa lu pa broda li'u lo ka se stura 02:14 < gleki> i i'a 02:14 < zipcpi> ba'asai la .kurtis. ba ma'udzau 02:34 < zipcpi> vlaste: setmi (fapro) 02:34 < vlaste> zo setmi gimkamsmikezypro no catni gismu .i zo setmi gimkamsmikezypro no cipra gismu 02:36 < zipcpi> Problem with zi'evla with shapes like {setmima} is that you can't elide the last vowel when making zevlyjvo 02:37 < zipcpi> VCCV is also iffy, due to high possibility of clashes 02:37 < akmnlrse> gleki: ja'a nitcu se ju djica 02:37 < mensi> akmnlrse: cu'u la'o gy.gleki.gy.: xu do nitcu lo fukpi be lo JVS datni sorcu i mi ka'e benji fi do i ma mrilu judri | 02:37 < mensi> 2015-07-24T07:12:13.405Z 02:38 < zipcpi> coi nandyselbacru 02:38 < akmnlrse> coi bauspo 02:38 < zipcpi> u'i 02:38 < akmnlrse> spagetufa: akmnlrse 02:38 < spagetufa> (CU [.àkmynlyrse VAU]) 02:39 < zipcpi> ua 02:39 < zipcpi> la spagetufa cu tisnyga lo zunsna slaka me'o ybu 02:40 < akmnlrse> ka'e ku su'o se bangu zo'u lo zunsnaslaka cu bacru jai filmau .i ku'i lo tisna tarmi zo'u frili fa lo ka djuno lo du'u ma kau noi lerfu cu pagbu ma kau noi slaka 02:41 < zipcpi> ie 02:47 < akmnlrse> tersmus: ro da poi ke'a su'o de viska cu se kanla su'o di 02:47 < tersmus> FA x1:(EX x2. viska(_,x2)). EX x2. kanla(x2,x1) 02:51 < akmnlrse> zantufa: ge ti gi'i 02:51 < zantufa> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 02:53 < akmnlrse> gerna: ko'a e ko'e e ke ko'i e ko'o 02:53 < mensi> (0[{<ko'a e ko'e> e <ke (1ko'i e ko'o)1 KE'E>} VAU])0 02:56 < akmnlrse> exp: getigitajetu 02:56 < mensi> ([{ge ti gi ta} {je tu}] VAU) 02:56 < akmnlrse> alta: getigitajetu 02:56 < mensi> (NIhO [I {FA <ge ti gi (¹ta [je tu]¹)>} CU {COhE VAU}]) 02:57 < akmnlrse> zantufa: getigitajetu 02:57 < zantufa> ([ge ti {gi <ta (¹je tu¹)>} GIhI] VAU) 02:57 < akmnlrse> gerna: getigita etu 02:57 < mensi> (0[{<ge ti gi ta> e tu} VAU])0 03:01 < akmnlrse> gerna: getigita ebotu 03:01 < mensi> (0[{<ge ti gi ta> <e bo> tu} VAU])0 03:06 < gleki> la altatufa xi stodi de'a farvi 03:07 < akmnlrse> gerna gu'e mo gi nei je bo go'i 03:07 < akmnlrse> gerna: gu'e mo gi nei je bo go'i 03:07 < mensi> (0[{<gu'e mo gi nei> <je bo> go'i} VAU])0 03:11 < zipcpi> exp: broda xi brode 03:11 < mensi> SyntaxError: Expected [,] but "b" found. 03:14 < gleki> na'u 03:19 < gleki> i ke ji'a la altatufa nu'o kaken tu'a lu mo xi mo li'u 03:19 < gleki> *kakne 03:29 < zipcpi> xy'y simsa lo te finti be zo ke'oi bei mi 03:30 < noncomcinse> ki'a 03:30 < zipcpi> en: ke'oi 03:30 < mensi> ke'oi = [KE] reverses modification order of contained tanru (does not affect lujvo). | Marks a "selskiski"/" 03:30 < mensi> military-style"/"noun-adjective" tanru, i.e. tertau come *before* seltau. This differs from co in that it does not 03:30 < mensi> affect the place structure of the sumti to the right; i.e. in mi ke'oi broda brode do (= mi brode broda do), both mi and 03:30 < mensi> do are arguments of broda. Default grouping is still left: lo ke'oi ckule nixli cmalu melbi resolves to lo ke'oi 03:30 < mensi> (((ckule nixli) cmalu) melbi), which is equivalent to lo melbi ke cmalu ke nixli ckule. co does not (re-)reverse the 03:30 < mensi> order of tanru-modification; it only changes place structure. See ke'e'u. | spheniscine 03:30 < noncomcinse> zo xy'y 03:30 < zipcpi> en: xy'y 03:30 < mensi> xy'y = [UI1b] naturalistic interjection: in thought/contemplation | Differs from y in that it doesn't imply hesitation 03:30 < mensi> or a desire to say anything after. The first syllable may be lengthened as desired. Has a mandatory pause after. See 03:30 < mensi> pensi. | spheniscine 03:31 < noncomcinse> simsa zo .xm. (noi ku'i cmevla) 03:31 < zipcpi> ie 03:32 < noncomcinse> lu sa'ei .xm. li'u drata .i .u'u .i .a'i toi'e .ai 03:32 < noncomcinse> .i ke'oi prenu bangu logji ke fa lo logji bangu prenu xu 03:33 < zipcpi> ke'oi prenu ke bangu logji 03:33 < zipcpi> Still left-groups 03:34 < zipcpi> Only order of operation is reversed 03:35 < zipcpi> Nor does it change the meaning of {ke} or {ke'e} 03:35 < zipcpi> I'll probably need to finish my page sometime 03:36 < zipcpi> Basically translating that pages about tanru in CLL 03:36 < zipcpi> into selskiski form 03:36 < gleki> exp: ca tu'e mi mo i mo tu'u 03:36 < mensi> (ca tu'e [{mi <CU (¹mo VAU¹)>} {i <CU (¹mo VAU¹)>}] tu'u) 03:36 < gleki> exp: caku tu'e mi mo i mo tu'u 03:36 < mensi> SyntaxError: Expected [yY] but "u" found. 03:37 < gleki> exp: fa tu'e mi mo i mo tu'u 03:37 < mensi> (fa tu'e [{mi <CU (¹mo VAU¹)>} {i <CU (¹mo VAU¹)>}] tu'u) 03:37 < gleki> O_0 03:38 < noncomcinse> zipcpi: I can't put three seltau in ke'oi? 03:38 < zipcpi> You can 03:39 < zipcpi> Erm, what exactly do you mean? Give me a skiselski tanru and I'll try to translate it 03:39 < noncomcinse> I did logji bangu prenu 03:39 < zipcpi> Which is ((logji bangu) prenu) 03:39 < zipcpi> Reversed, it's ke'oi (prenu ke( logji bangu)) 03:39 < zipcpi> Er 03:40 < zipcpi> s/logji bangu/bangu logji 03:40 < fenki> zipcpi meant to say: Reversed, it's ke'oi (prenu ke( bangu logji)) 03:40 < gleki> tokipona invertor 03:40 < zipcpi> Ahaha 03:40 < zipcpi> My inspiration was Malay and Latin though 03:41 < noncomcinse> Why isn't ke'oi prenu bangu logji ke valid? 03:41 < zipcpi> Heck, the name {selskiski} comes from a Malay idiom/mnemonic for that rule 03:41 < zipcpi> ke'oi is selma'o KE 03:42 < zipcpi> exp: ke prenu bangu logji ke 03:42 < mensi> SyntaxError: Expected [yY] but "e" found. 03:42 < zipcpi> alta: ke prenu bangu logji ke 03:42 < mensi> SyntaxError: Expected ['h] or ['hiI] but end of input found. 03:42 < noncomcinse> whut 03:43 < noncomcinse> off: ke prenu bangu logji ke'e 03:43 < mensi> ([ke {prenu bangu logji} ke'e] VAU) 03:43 < akmnlrse> spagetufa: ke prenu bangu logji 03:43 < spagetufa> (CU [{ke <(¹prènu bàngu¹) lògji> KE'E} VAU]) 03:43 < noncomcinse> I meant {ke'oi prenu bangu logji ke'e} 03:43 < zipcpi> And all that'll do is close {ke'oi}, which is automatic anyway 03:44 < zipcpi> It won't make it right-group 03:44 < noncomcinse> But why were you using {ke logji bangu}? 03:44 < noncomcinse> Can't we have everything under ke 03:44 < akmnlrse> ke'oi prenu bangu bo logji 03:44 < noncomcinse> s/ke/ke right group?/ 03:44 < fenki> noncomcinse meant to say: Can't we have everything under ke right group? 03:44 < noncomcinse> s/ke/ke'oi/ 03:44 < fenki> noncomcinse meant to say: Can't we have everything under ke'oi right group? 03:44 < zipcpi> If we want right grouping there are already a bunch of cmavo for that 03:45 < zipcpi> en: ke'ai 03:45 < mensi> ke'ai = [KEhAI] Toggles to no grouping | rafsi -ked- The grouping of gismu/rafsi is left ambiguous. So with a statement 03:45 < mensi> like: ke'ai melbi cmalu nixli ckule or kedmlecmaxlicu'e it is unclear whether this is referring to a (((pretty 03:45 < mensi> little)girls)school), (pretty(little(girls(school))), etc. Overridden by ke'ei or ke'ei'a. (Applies only to next selbri. 03:45 < mensi> ) | ues 03:45 < zipcpi> Oops, not that one 03:45 < zipcpi> en: ke'ei 03:45 < mensi> ke'ei = [KEhEI] Toggles to right grouping of tanru/lujvo. | rafsi -keb- "ke'ei'a" toggles back to left grouping. 03:45 < mensi> (Applies only to next selbri.) | ues 03:46 < zipcpi> I don't know how these selma'o work though 03:48 < zipcpi> {ke'oi} is not meant to be "just reverse the words", but a whole new paradigm of making tanru. For example, under ke'oi {co} only changes place structure, not the ordering (though right now parsers don't treat KE and CO the way I want ke'oi to work) 03:48 < noncomcinse> en: xk 03:48 < mensi> 32 da se tolcri: banxe'ixu, banbuxuku, bangaxuku, bankuxuku, banmuxuku, bansuxuku, banxuku'a, banxukubu, banxukucu, 03:48 < mensi> banxukudu, banxuku'e, banxukufu, banxukugu, banxuku'i, banxukuje, banxukuju, banxukuke, banxukuku, banxukulu, banxukunu, 03:48 < mensi> banxuku'o, banxukupu, banxukuru, banxukusu, banxukutu, banxuku'u, banxukuve, banxukuvu, banxukuxe, banxukuxu, banxukuzu, 03:48 < mensi> banxuxuku 03:48 < noncomcinse> ... 03:48 < noncomcinse> en: xo'i 03:48 < mensi> xo'i = [XOhI] Extracts selbri from a tag, inverse of fi'o | xo'i bau is equivalent to bangu, xo'i fi'o broda is 03:48 < mensi> equivalent to broda | selpahi 03:48 < noncomcinse> exp: ma xo'i na 03:48 < mensi> SyntaxError: Expected [yY] but "a" found. 03:49 < noncomcinse> exp: ma xo'i bai 03:49 < mensi> SyntaxError: Expected [,], [mM], [nN], [rR] or [vV] but "b" found. 03:49 < noncomcinse> exp: ma xo'i fi'o zukte 03:49 < mensi> SyntaxError: Expected [,] but "z" found. 03:52 < zipcpi> Left grouping is easier because if we want to force left grouping if right grouping is default, lots of forethought is needed 03:53 < zipcpi> da'i lu ke'oi ke ke ckule nixli ke'e cmalu ke'e melbi 03:53 < zipcpi> li'u 03:53 < zipcpi> This is why {ke'oi} doesn't change grouping to right 03:54 < zipcpi> Or {bo} can be used I suppose, but still 03:55 < noncomcinse> ke'oi is like co but in forethought and without spitting the tail? 03:55 < zipcpi> Yep 03:59 < noncomcinse> Okay, now I see its use. uanaibu'onai 04:00 < zipcpi> And then there is {ke'e'u}, which semi-permanently changes all tanru to selskiski 04:00 < zipcpi> ke'e'unai changes it back 04:01 < zipcpi> It's basically meant to be a "what-if" experiment 04:01 < zipcpi> Given some discussions with... e'ogan I think? about why Lojban chose skiselski over selskiski 04:02 < zipcpi> (hint, I suspect malgli reasons, although I do have one issue with selskiski that I haven't quite solved satisfactorily yet, namely what to do with {be}) 04:03 < noncomcinse> selsmi 04:03 < noncomcinse> selskiski? 04:03 < gleki> {xo'i} isnt supported by any parser 04:03 < zipcpi> lo ke'oi bangu logji be ko'a 04:03 < noncomcinse> .oi cmalu batkyfoi 04:04 < zipcpi> selskiski is what I call the "reversed" tanru 04:04 < zipcpi> selski = "thing described", a metaphor for the tertau 04:04 < zipcpi> skicu = "describer", a metaphor for the seltau 04:06 < zipcpi> gleki: Hm yeah that's annoying. Just put it in GOhOI and call it a day 04:06 < zipcpi> Hm I wonder if zantufa implemented GOhOI 04:06 < gleki> exp: go'oi mo 04:06 < mensi> SyntaxError: Expected [yY] but "o" found. 04:06 < zipcpi> zantufa: zantufa ze'oi fa 04:06 < zantufa> (CU [zantufa {ze'oi fa}] VAU) 04:06 < gleki> neither gohoi is supported 04:06 < zipcpi> It should be 04:06 < gleki> en;ze'oi 04:06 < gleki> en:ze'oi 04:06 < mensi> ze'oi = [GOhOI] converts following word to selbri-unit: "x1 is related to the meaning of this word in aspect x2" 04:07 < zipcpi> zantufa: zantufa xo'i ba'o 04:07 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "a" cu se tolcri 04:08 < zipcpi> {ze'oi} is a very useful word; it basically allows one to avoid making nonce-zei-lujvo 04:08 < zipcpi> And {go'oi} is pretty popular too 04:09 < zipcpi> So yeah it should be implemented someday 04:09 < gleki> "this word" isnt nice. neither "(n)" is nice in official definitions. it's argument 04:10 < zipcpi> Well it isn't x-anything, unless we want to use {goi} in English defs :p 04:10 < zipcpi> converts following word (goi ko'a) to selbri-unit: "x1 is related to ko'a or the meaning of ko'a in aspect x2" 04:11 < zipcpi> Hah 04:18 < zipcpi> zantufa: zantufa go'oi fa 04:18 < zantufa> (CU [zantufa {go'oi fa}] VAU) 04:24 < ctefaho> coị 04:24 < durka42> coi 04:27 < noncomcinse> coi 04:27 < noncomcinse> doi lo rinsa do mo 04:30 < ctefaho> ɂi mi tinkerino 04:30 < zipcpi> gleki: I dunno, if you don't want to change your def for blalalavla I might just make {comkravla} for a-priori-words in general instead. Your word is fun though, I'll at least give you that :p 04:32 < gleki> zipcpi: ok change 04:34 < zipcpi> Done 04:34 < zipcpi> Er oops 04:34 < zipcpi> I accidentally changed the notes 04:34 < zipcpi> There fixed 04:36 < zipcpi> And lol kreivi. I myself, when making jbofuvi, will just mash together phonemes from appropriate root words. For example if I wanted to make a word like {kreivi} I might take {danlu} and {zutse} and mash them into {dau'utse} 04:38 < durka42> ua 04:38 < durka42> la jbonunsla ba stuzi la .prinstn. 04:38 < durka42> mi pu nova'e co jundi lo nuncasnu be jy 04:38 < zipcpi> se stuzi le'aipei 04:39 < noncomcinse> ua 04:39 < durka42> ku'i la .prinstn. ka'e se klama mi 04:39 < noncomcinse> .i .au .ai klama ju frili mi 04:39 < durka42> oi zo stuzi cu dukti zo zvati 04:40 < noncomcinse> ia ie 04:40 < noncomcinse> .i xu ja'ai zvati la .prinstyn. 04:40 < noncomcinse> ba 04:41 < zipcpi> http://jbovlaste.lojban.org/dict/tu'oi <-- xu kurtykai 04:41 < durka42> la .piier. cu mrisku ca lo prulamdei 04:41 < noncomcinse> ua 04:41 < durka42> zipcpi: xu sarcu .i zo po'o sarcu 04:42 < durka42> y si banzu 04:42 < zipcpi> If you use po'o you can't make it a member of a set or mass or anything 04:42 < noncomcinse> .i ru'i nai mriste tcida 04:42 < durka42> la djon joi tu'oi = la djon ku po'o 04:43 < noncomcinse> zo tu'oi ki'a 04:43 < zipcpi> You can use {lo gunma bepo'o la .djan.} but that's longer 04:43 < zipcpi> It's a typing thing 04:43 < durka42> isn't there a LAhE for that 04:44 < durka42> lu'o la .djon. ku po'o 04:44 < zipcpi> exp: lu'o la .djon. ku po'o 04:44 < mensi> ([lu'o {la djon <ku po'o>} LUhU] VAU) 04:44 < durka42> lu'o la po'o .djon. 04:44 < durka42> la .djon. joi tu'oi 04:44 < durka42> it's a few pixels longer on my screen but involves zero new words :p 04:45 < zipcpi> lol 04:45 < durka42> also 04:45 < durka42> lo valsi cu kurtykai soi cumki 04:45 < durka42> lo velcki ba'e na kurtykai 04:45 < zipcpi> exp: lu'o po'o la .djon. 04:45 < mensi> ([{lu'o po'o} {la djon KU} LUhU] VAU) 04:45 < durka42> vasru no me fi'ubu 04:45 < zipcpi> Yes it needs to be a lot more longer and dense to qualify :p 04:45 < durka42> pe'i lu lu'o po'o la djon li'u du lu lu'o la djon lu'u po'o li'u 04:46 < zipcpi> zantufa: lu'o ke po'o la .djon. 04:46 < zantufa> ([lu'o {<ke po'o> <la djon KU> KEhE} LUhU] VAU) 04:46 < durka42> "only [the mass composed of Joan]" vs "the mass composed of only Joan" 04:46 < noncomcinse> en: lu'o 04:46 < durka42> this {ke} trick is neat 04:46 < durka42> never knew about it before 04:46 < durka42> is it new? 04:46 < mensi> lu'o = [LAhE] the mass composed of; converts another description type to a mass composed of the members. | 04:46 < mensi> officialdata 04:47 < durka42> or is it an old idea that was never parseable? 04:47 < durka42> yacc: ke po'o la djon 04:47 < zipcpi> I dunno. Selpahi suggested it 04:47 < gleki> zipcpi: the point is that {zutse} here can be malgli ju malropno 04:47 < durka42> yacc says "unknown cmavo po'o" 04:47 < durka42> wat 04:47 < zipcpi> gleki: Which is why a zi'evla 04:47 < zipcpi> lol 04:48 < zipcpi> I mean I purposely chose a malpevna when making {li'armi} 04:48 < zipcpi> And {li'a} itself coms from {klina}, which is another malpevna :p 04:49 < durka42> hh 04:49 < durka42> heh 04:49 < noncomcinse> Is there a bot command to pull up the vlasisku page for X? 04:50 < zipcpi> I don't know 04:51 < noncomcinse> vlaste: this-should-get-no-results 04:51 < vlaste> no results. http://vlasisku.alexburka.com/this-should-get-no-results 04:51 < noncomcinse> I was sorta hoping that might work but nah. 04:53 < noncomcinse> en: lu'e 04:53 < mensi> lu'e = [LAhE] the symbol for (indirect discourse); uses the symbol/word(s) for a sumti as the desired sumti. | 04:53 < mensi> officialdata 04:53 < noncomcinse> uanai 04:53 < selpahi> zipcpi: I think this particular function of {ke} was all guskant's idea. It seemed like a good fix to the {fa ji'a ko'a} problem, but sadly the {ke'e} doesn't auto-terminate after the sumti (because it tries to make a full termset). 04:54 < gleki> alta: ke ji'a mi do prami 04:54 < mensi> (NIhO [I {<FA (¹ke ji'a¹) mi KEhE> <FE do>} CU {prami VAU}]) 04:54 < gleki> pa drata cu se jinvi mi 04:55 < gleki> i zo ke pe la alta na dunli zo ke pe lo drata 04:57 < zipcpi> selpahi: spagetufa also implements it but doesn't do the termset thing 04:57 < gleki> alta: ke fa ji'a mi fe do prami 04:57 < mensi> (NIhO [I {ke <(¹fa ji'a¹) mi> <fe do> KEhE} CU {prami VAU}]) 04:57 < zipcpi> Oh yeah alta too 04:57 < selpahi> la'a ku'i la .ilmen. cu zanru zo ke pe la zantufa .i ra pu ta'e pante ru'e lo nu zo ge pe lo sumpoi zo'u lo nu lo jorne sumpoi jo'u lo se jorne sumpoi cu dunli lo ka xo kau da cmima .i si'au zo ke pe la zantufa cu jai do'e na'e sarcu 04:57 < zipcpi> spagetufa: ke ji'a mi do prami 04:57 < spagetufa> ([{FA <(¹ke ji'a¹) mi KE'E>} {FE do}] [CU {pràmi VAU}]) 04:58 < selpahi> zantufa: ke mi do ke'e je ke do mi ca lo pavdei cu prami 04:58 < zantufa> ([{ke <mi do> ke'e} {je <ke (¹do mi [ca {lo pavdei KU}]¹) KEhE>}] cu [prami VAU]) 04:58 < selpahi> .i ju'o sai la .ilmen. cu zanru 04:58 < gleki> zantufa: ge mi viska gi do gi'i 04:58 < zantufa> minajimpe: .i kanpe tu'a [oO] .onai [uU] .i ku'i "i" cu se tolcri 04:58 < gleki> zantufa: ge mi viska gi do 04:58 < zantufa> minajimpe: .i kanpe tu'a [yY] .i ku'i "o" cu se tolcri 04:58 < gleki> zantufa: ge mi viska gi do mo 04:58 < zantufa> (ge [mi CU {viska VAU}] [gi {do CU <mo VAU>}] GIhI) 04:59 < gleki> i ju'onai mi ka'e cikre 04:59 < zipcpi> Oh yeah selpahi, I also added to the TMC list some ideas about the colors thing 04:59 < gleki> selpahi: ta'o xu do di'a kanro 04:59 < selpahi> na go'i 04:59 < gleki> vau coi 04:59 < zipcpi> That fundamentally, the "color red" is the sensation produced by things that are red, is isolation of whether there is anything red to produce them. 05:00 < zipcpi> Similarly, the sensation of dryness 05:01 < noncomcinse> lo skari be losi'o tolcinmo gidei 05:01 < noncomcinse> si si 05:03 < zipcpi> durka42, you might want to have a look at the TMC list - I added some fuel to the color-fire there 05:04 < selpahi> I might need some help later or tomorrow or whenever I'm less sick, with making a couple brivla that I need for my novel. 05:04 < zipcpi> https://docs.google.com/spreadsheets/d/1LH_bawafDZBp5h4shGoajnTU7sttOilfpyxqUD4B5BM/edit#gid=0 05:04 < zipcpi> Oh, get well soon 05:04 < selpahi> ki'e 05:05 < durka42> ko dei'a bilma 05:05 < durka42> a'o lo nu casnu zo si'o na bilma rinka 05:06 < noncomcinse> .i mi zanfunpa'a be fi do loka ce tolbilma binxo xoi sutra 05:06 < selpahi> mi pu tu'a zo si'o co'a bilma .i pu za lo djedi be li re ja ci 05:07 < selpahi> ki'e la noncomcinse 05:07 < durka42> ua .ei mi na jbibi'o la .berlin. .i da'i mi co'a bilma 05:07 < durka42> zo'o 05:08 < durka42> do finti ji fanva lo cukta 05:08 < durka42> si se cukta 05:08 < selpahi> finti lo brafi'a 05:08 < noncomcinse> lu li re ci li'u simsa banzu .i na nitcu zo'ei zo je 05:08 < durka42> ba'anairu'e i'e 05:08 < selpahi> reci = 23 05:08 < selpahi> re ja ci = 2 or 3 05:09 < selpahi> .i .e'o doi la durka42 ko za'u re'u mojgau mi lo du'u co'a ma kau je co'u ma kau do zvati la .franfurt. 05:10 < noncomcinse> camxes: li rejaci 05:10 < camxes> SyntaxError: Expected [yY] but "j" found. 05:10 < noncomcinse> camxes: li vei rejaci 05:10 < camxes> SyntaxError: Expected [yY] but "j" found. 05:10 < selpahi> zo .a 05:10 < noncomcinse> exp: li re ja ci 05:10 < mensi> ([li {<re BOI> <ja (¹ci BOI¹)>} LOhO] VAU) 05:10 < noncomcinse> zo'oi .exp. 05:11 < durka42> selpahi: oi za'o na djuno .i la'a lo rebimoi co djedi po'oru'e 05:11 < durka42> mi voikla ca lo cerni be lo resomoi 05:12 < selpahi> uat .i va'i pei zvasta ze'i lo djedi be li so'u sai ba'e pa ?! 05:12 < durka42> .u'i 05:12 < durka42> ti'e la .franfirt. na mutce co cinri lo datygugvi'e .i .ai vitke la .muncen. 05:12 < durka42> pei 05:12 < selpahi> >_< 05:13 < durka42> za'o ka'e platu galfi 05:13 < durka42> ma se stidi do 05:13 < selpahi> .i da'i mi ta'i ma penmi do va'o lo nu do ji'a sai na djuno lo du'u zvati ma kau 05:13 < durka42> mi'e la .karmensandiiegos. 05:13 < zipcpi> u'i 05:13 < selpahi> ... 05:14 < durka42> samkei sitna 05:14 < durka42> .wik Where in the World is Carmen San Diego? 05:14 < selpahi> mi .u'inmo 05:14 < fenki> "Where in the World Is Carmen Sandiego? is the title of several educational computer games in the Carmen Sandiego series that teach geography and reference skills." - https://en.wikipedia.org/wiki/Where_in_the_World_is_Carmen_San_Diego? 05:14 < selpahi> .i .ei pei mi retsku lo drata no'u di'e .i co'u ma do zvati lo dotygu'e 05:15 < noncomcinse> .i ma se zvati la durka po'u kysy 05:16 < noncomcinse> Can I use le'ai to replace su if I accidentally su? 05:16 < durka42> co'u li D 29 C 13 05:17 < selpahi> cacra xu 05:17 < durka42> cacra 05:17 < zipcpi> noncomcinse: If you accidentally su, not even lawyers can save you zo'o 05:17 < selpahi> ne xu la zelmasti 05:17 < zipcpi> Man that was terrible 05:18 < durka42> ne lo jeftu pamoi'o ie 05:18 < noncomcinse> I accidentally su root is that bad. 05:18 < zipcpi> Oops I accidentally su the world 05:18 < zipcpi> I accidentally su jbovlaste 05:19 < selpahi> zipcpi: Not much worse than all the variations of "You forgot you shoes" when people accidentally drop their {cu}. 05:19 < zipcpi> lol 05:19 < selpahi> Though it carried over into Lojban with {ko dasni lo cutci} 05:20 < zipcpi> :p 05:20 < selpahi> durka42: .oi .i suksa fa lo nu co'u zvati .i mi pu kanpe lo drata ki'u lo na se sanji be mi 05:21 < noncomcinse> .i mi tolmo'i lodu lo mi cutci zvati makau 05:21 < durka42> selpahi: ie mi mabla gunrivzu'e lo nu platu .iau .u'u 05:21 < selpahi> #mencre_klamburi 05:22 < noncomcinse> ie 05:22 < selpahi> Yow! Oohoo! lol #lojban 05:22 < durka42> lol 05:31 < gleki> how do we say "case" as for capitalized letters? 05:31 < noncomcinse> tau 05:31 < gleki> en:tau 05:31 < mensi> tau = [LAU] 2-word letteral/shift: change case for next letteral only. 05:31 < gleki> en:ga'e 05:32 < mensi> ga'e = [BY1] upper-case letteral shift. 05:32 < noncomcinse> Mat = tau my .a'y ty 05:32 < gleki> en:en:to'a 05:32 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 05:32 < noncomcinse> MAT = ga'e my .a'y ty 05:32 < gleki> en:to'a 05:32 < mensi> to'a = [BY1] lower-case letteral shift. 05:32 < gleki> i need to translate "match case" 05:32 < gleki> e.g. in regex 05:32 < noncomcinse> MaT = ga'e my to'a .a'y ty 05:33 < noncomcinse> nilbra 05:33 < noncomcinse> nilbravai 05:33 < noncomcinse> nilbranalvai 05:33 < gleki> en: nilbra 05:33 < mensi> nilbra [< ni barda ≈ Amount abstract big] = x1 is the size of x2 in property/dimension x3 on scale x4. | 05:33 < mensi> xorxes 05:33 < gleki> ienai 05:34 < noncomcinse> capital letters are fanva as barda lerfu 05:34 < noncomcinse> rafsi: jei 05:34 < mensi> zo'oi jez rafsi zo jei .i zo jenmi se rafsi zo'oi jei 05:34 < noncomcinse> jezbra 05:34 < noncomcinse> pei 05:35 < durka42> selpahi: xu mi'o ba penmysi'u ca lo za'umoi co nu mi dotygu'e vitke noi ca ke'a mi vitke .ai .au la .berlin. 05:35 < noncomcinse> lujvo: jei barda lerfu kei vajni 05:35 < zipcpi> I dunno. Capital letter is ga'e zei lerfu, lowercase is to'a zei lerfu 05:36 < mensi> jezbralerkezvai[14694], jezbrale'ukezvai[15183], jezbralerkezvajni[16764], jezbralerfykezvai[16824], jezbardylerkezvai[16844], ... 05:36 < selpahi> durka42: xu ku'i djuno lo du'u ba za'u re'u vitke 05:36 < noncomcinse> jezbralerkezvai = case sensitive 05:36 < durka42> nonselju'o 05:37 < zipcpi> rafsi: ja 05:37 < mensi> zo'oi jav rafsi zo ja 05:37 < zipcpi> k: ga'enjafto'a 05:37 < mensi> (CU [Z:ga'enjafto'a VAU]) 05:37 < zipcpi> rafsi: jaf 05:37 < mensi> zo jamfu se rafsi zo'oi jaf 05:37 < zipcpi> u'i 05:37 < noncomcinse> jezbralerkeznalvai = case insensitive 05:37 < zipcpi> ganse jamfu tonga 05:38 < noncomcinse> JA is one selma'o I know all the rafsi for. 05:38 < zipcpi> I'm the king of jbofuvi lol 05:38 < selpahi> I know all cmavo rafsi, but not quite all gismu rafsi. :/ 05:38 < noncomcinse> jbofuvi .u'i 05:39 < noncomcinse> xu su'o da djuno fi lo ro cmavo 05:39 < gleki> li'a na go'i 05:39 < selpahi> xu lo ro cmavo cu se menre lo kurtyvla 05:39 < gleki> la kurti cu du'eva'ei sutra 05:40 < noncomcinse> .u'i 05:40 < zipcpi> la'a mi ji'a to'e sidju fi tau le tcini 05:41 < zipcpi> en: no'ai 05:41 < mensi> no'ai = [PA4] digit/number: absolute zero; nothing; there does not exist; ∄ | Mathematically equivalent to no, but 05:41 < mensi> implies lack of negative values, especially on a scale (va'e/va'ei e.g.). Equivalent to "rone'e no". Cf. no'e'u, su'o, 05:41 < mensi> ro, su'oi, ro'oi | spheniscine 05:42 < gleki> ca kelvo li no'ai 06:04 < zipcpi> tu'oi klama tu'oi tu'oi tu'oi tu'oi 06:04 < zipcpi> This is even worse than zi'o :p 06:15 < durka42> oi mi pu bilga lo ka platu .i ku'i na platu .i bu'u ma mi sipna ca lo bavlamnicte 06:15 < gleki> ue zo platu 06:16 < durka42> mo 06:17 < gleki> so'uroi se pilno 06:17 < noncomcinse> coi ru'e 06:17 < noncomcinse> en: unquantifiable 06:18 < mensi> lo nu mulno sisku zo'u: y no da se tolcri 06:23 < noncomcinse> zo tu'o 06:25 < zipcpi> bu'u tu'oi zo'o 06:25 < zipcpi> lo kutydza cu kutydza 07:30 < durka42> .ai vitke lo badydi'u ca lo bavlamdei 08:12 * nuzba @frecafloros: .i mi tavla lo lojban .iu [http://bit.ly/1GKFBAJ] 08:13 < durka42> za'a cmevla brivla co'e 08:26 < ldlework> The difference between universals and tropes is fascinating 08:28 < ldlework> heh interesting, lojbanic form found in non-lojban literature: 08:28 < ldlework> "Relations are usually taken to have a ‘degree’ (‘adicity’, ‘arity’), which depends on the number of objects that they can relate, or, to put it otherwise (somewhat metaphorically), on the number of ‘places’ they come with." 08:34 < farmer> hello everyone 08:35 < noncomcinse> farmer: Hello. 08:36 < ctefaho> ldlework: where is that citation from? 08:37 < ctefaho> hmm this http://plato.stanford.edu/entries/properties/ ? 08:38 < ldlework> ctefaho: yeah 08:38 < ldlework> fascinating page 08:38 < ldlework> "It seems also possible to have definite descriptions and perhaps even indexicals that refer to properties. If Mary happens to prefer wisdom to any other property ‘Mary's favorite property’ seemingly refers to wisdom. Moreover, though more controversially, if someone points to a red object while saying: ‘that shade of red is a beautiful color,’ then the demonstrative ‘that shade of red’ denotes a 08:39 < ldlework> property (Heal 1997)." 08:39 < ctefaho> I will make sure to read through it when the heatstroke has passed 08:39 < ldlework> "Frege (1892) and Russell (1903) had different opinions regarding the ontological import of nominalization." 08:39 < ldlework> haha no kidding 08:42 < ldlework> ctefaho: I have been reading this site for months and months and have notes from all over that are relevant to thinking about lojban 08:42 <@xalbo> coi te cange 08:43 < ldlework> One of my first banski articles will be just a shotgun of links to various sections of various pages, with a snippet and a comment about how it might be interesting to read if you're a lojbanist 08:43 < ctefaho> ldlework: I am most interested 08:43 < ldlework> Will probably do the same for some of the books I have read recently 08:44 < ctefaho> and I love big texts with big reference lists behind 08:44 < ldlework> xalbo: who are you calling a farm? 08:45 <@xalbo> I saw that "farmer" had joined, but missed that ri had quit. 08:45 < ldlework> oh I have parts and joins hidden but I do see 'farmer' above now 08:46 < ldlework> ctefaho: haven't you alluded to articles you have written about various things? 08:46 < ldlework> You're more than just welcome to have them published to Banski 08:46 < ctefaho> ehm I have no articles written anywhere 08:46 < ldlework> Oh okay. 08:46 < ctefaho> I just have my dialect I am tinkering around with 08:46 < ctefaho> dialect/offshot/whatever 08:48 < ldlework> "In seeking explanations in ontology (as in other disciplines) we must frequently weigh tradeoffs between various desiderata, e.g., between simplicity and comprehensiveness, and even between different kinds of simplicity. But one tradeoff is so pervasive that it deserves a name, and we will call it the fundamental ontological tradeoff. The fundamental ontological tradeoff reflects the perennial tension 08:48 < ldlework> between explanatory power and epistemic risk, between a rich, lavish ontology that promises to explain a great deal and a more modest ontology that promises epistemological security. The more machinery we postulate, the more we might hope to explain—but the harder it is to believe in the existence of all the machinery. As we shall see in the following, the inevitability of this tradeoff keeps playing a 08:48 < ldlework> crucial role in current discussions of properties." 08:49 < ctefaho> (The things I think I may have said that implied I have written something about it are just thoughts I have in my head and in some text files here and there) 08:52 < ctefaho> like what I call the "consolodation of truth" which I have more or less just made up 08:53 < ctefaho> but I am very hungry for good reading sources about this 08:53 < ctefaho> http://plato.stanford.edu/ looks awesome though 08:54 < ldlework> it is 08:54 < ldlework> Its my starting point for various topics 08:54 < ldlework> it doesn't really do anything more than scratch the surface of any given topic 08:55 < ldlework> its more of an overview of the history in thinking of those subjects 08:55 < ldlework> so to really get at the content you have to actually go out of the site and look for what they refer to 08:55 < ctefaho> there is no big wall of text better than a well written and well referenced big wall of text 08:55 < ctefaho> I see 08:55 < ldlework> yeah I'm just noting that beyond review it doesn't offer a lot 08:56 < ldlework> but it does offer excellent review 08:56 < ldlework> the best I've found for sure 08:56 < ctefaho> any other similar sites or sources you can recommend? 08:56 < ldlework> its how I found about the concept of Speech Acts 08:56 < ldlework> Which I've been studying pretty specifically for a while 08:56 < ldlework> I have a list of books and whitepapers you can buy 08:57 < ldlework> But no other sites like plato.stanford 08:57 < ctefaho> this one I suppose http://plato.stanford.edu/entries/speech-acts/ ? 08:57 < ldlework> yeah 08:57 < ldlework> and if it interests you, you should definitely get a copy of Searle's Speech Acts 08:57 < ldlework> priceless book imo 09:00 < ctefaho> will look into it 09:00 < ctefaho> but just that site seems awesome enough for now 09:01 < ctefaho> non nom @ all the topics 09:02 < ctefaho> http://plato.stanford.edu/entries/truth/ ftw 09:03 < ctefaho> many thanks for sharing this 09:22 < gleki> la .lojban. 09:22 < gleki> cu ciplanli lo bangu 09:22 < gleki> .i runbau gi'e se jicmu lo bridi logji 09:23 < gleki> en: ciplanli 09:23 < mensi> ciplanli [< cipra lanli ≈ Test analyze] = l1 experiments l2=c2 by means l3. | Cf. lanli, cipra, tadni, birti. | 09:23 < mensi> tijlan 09:24 < ldlework> Yesterday selpahi alluded to a fact that Lojban was defined to require all places to be filled of any bridi in any context 09:24 < ldlework> Does anyone know what he was referring to and where I could look at it? 09:25 < gleki> CLL 09:26 < ldlework> gleki: sure, but where? 09:26 < gleki> https://lojban.github.io/cll/7/7/ 09:27 < ldlework> woah, the CLL even calls zo'e an indefinite pro-sumti 09:27 < ldlework> holy shit 09:27 < ctefaho> hmm? 09:27 < ldlework> I have this theory of zo'e actually being a logical primitive like da 09:28 < ldlework> Where da is an existential reference, zo'e is an indefinite reference 09:28 < ldlework> Where using zo'e directly means something other than "there is really someting filling this place but I'm not saying what" 09:28 < gleki> looks like correct 09:28 < ldlework> Since only da can refer. So zo'e means something else. 09:29 < ctefaho> ldlework: So would "mi klama zo'e" and "mi klama" be different then? 09:29 < ctefaho> or am I misunderstanding 09:29 < gleki> since then koha7 seems to have become *a bit* bigger http://mw.lojban.org/extensions/ilmentufa/i/en/index.html#sisku/KOhA7 09:29 < ldlework> But I feel like the cll just uses 'indefinite' here vastly different than what the word means 09:29 < ldlework> since it goes on to say 09:29 < ldlework> zo'e KOhA zo'e-series the obvious value 09:29 < ldlework> Where an indefinite description doesn't have an 'obvious value' 09:29 < gleki> ma'oste was written by someone else than CLL 09:30 < ldlework> Well that's the example given in the CLL 09:30 <@xalbo> As far as everything I've seen goes, the "there is really something filling the place" comes from using that selbri at all. All {zo'e} does is add the "I'm not saying what", and even that it only does by omission. 09:30 < gleki> 7/7 even a bit contradicts the def. of zo'e in ma'oste 09:30 < ldlework> xalbo: right 09:30 < ldlework> xalbo: I just got a bit excited at the section title "7. Indefinite pro-sumti and pro-bridi: the zo'e-series and the co'e-series" 09:30 < ldlework> Nothing in the actual section reaffirms zo'e as an indefinite reference 09:31 < selpahi> bilma zei coi 09:31 < ldlework> a'oi 09:31 < gleki> bilma zei co'o zei e'u 09:31 < gleki> i vreta zei e'unmo 09:31 <@xalbo> The docs misuse linguist terms all over the place. If they say "indefinite", it's far more likely to be related to what a layperson would call "indefinite" (ie, "eh, who knows?") than whatever real linguists mean by "indefinite". 09:31 < gleki> i lo kibro se tcana cu infekte lo menli be lo'e remna 09:32 <@xalbo> ba zi .a'o kanro fa ko 09:32 < ldlework> xalbo: right 09:33 < ctefaho> ldlework: zo'e as indefinite - indefinite how? 09:34 < selpahi> The ma'oste was written before the CLL, and in many cases where CLL differs from the ma'oste they simply never bothered to update the ma'oste even before any baseline considerations. 09:34 < selpahi> {me} is one of the biggest examples. 09:34 < ctefaho> and I didn't quite get if you meant an explicit zo'e as being different from an empty argument? 09:34 < ldlework> ctefaho: it doesn't really matter, its a theory no one will ever agree with 09:35 < ldlework> gleki: I don't really see anywhere on that page an assertion about filled bridi 09:35 < ctefaho> Well I am still interested what said theory is 09:36 < ldlework> Its an attempt to deal with scoping problems in the expansion of indefinite speech to logical forms. 09:36 < ldlework> By establishing a secondary primitive with different scoping rules 09:36 < ctefaho> (I am myself very concerned with da vs. zo'e) 09:36 < ldlework> But the literature of indefinite scope is strange and hard to understand 09:36 < ldlework> So its still hard for me to fully understand the implications or how it work exactly 09:37 < ldlework> So I'd rather not embarass myself just yet 09:37 < ctefaho> I see 09:37 < ctefaho> Is it based on http://plato.stanford.edu/entries/descriptions/ ? 09:37 < ctefaho> or more than that? 09:38 < ldlework> Well sure, it is a reflection of the yet-unformalization of indefinite speech which is talked about in that article and basically anywhere else descriptions are addressed in LP 09:39 < ldlework> It also has to do with the ontology of properties and predicates 09:39 < ldlework> If indefinite descriptions are not referring descriptions, what do they denote? 09:39 < ldlework> If they denote predicates, why can't you substitute indefinite descriptions for verbial phrases 09:39 < ctefaho> ehm sorry LP as in? 09:40 < ldlework> language philosophy 09:40 < ctefaho> ua 09:41 < ldlework> If they denote properties, what is the ontology of properties, and so on 09:41 < ldlework> And every supposed ontology creates horrible scoping problems that no one really has figured out 09:41 < ctefaho> hmm well I have my own naive idea of definite and indefinite but it seemed to clash with yours 09:41 < ldlework> I don't really have my own ideas, I'm just aware of some established ones. I don't really know what's right. 09:41 < ldlework> Neither does anyone it seems. 09:42 < ctefaho> but it is maybe not the same definite-indefinite as you mean. have to read that article when I am done tinkering some other stuff 09:42 < ldlework> ctefaho: it would seem that it has been established pretty widely that there are two kinds of things that descriptions can denote 09:42 < ctefaho> mhm 09:42 < ldlework> objects, which can exemplify properties, and properties themselves 09:43 < ldlework> But don't exactly divide the kinds of descriptions we can make into these two categories directly 09:43 < ldlework> We actually treat both of these kinds of referrents exactly the same 09:43 < ldlework> And instead we make a distinction on the border of partiuclarity, or rather, identifiability 09:44 < ldlework> In that, descriptions may refer to objects, or properties, but they may do so in one of two modes 09:44 < ldlework> defintie, where some object or property is denoted specifically 09:44 < ldlework> Or indefinite, where some object or property is not denoted specifically, but ... in some otherway that seems to be hard to define, but intuitive to any speaker 09:45 < gleki> ldlework: selpahi here so maybe he can speak for himself. otherwise we can only guess what he meant 09:45 < ctefaho> At the risk of being way wrong I think of indefinite descriptions where something is denoted "referentially" 09:45 < ldlework> If you take the axiom that indefinite description implies that nothing is referred to, then you can go from there. 09:46 < gleki> ldlework: you meant this? http://selpahi.weebly.com/lojban/discourse-sensitive-zoe-and-zio 09:46 < ctefaho> where this referential-something is a referent of something denoted "specifically" 09:46 < ldlework> It would seem that that definition would violate such an axiom 09:46 < gleki> if nothing then thats {noda} 09:46 < gleki> if not applicable then that's {zi'o} 09:46 < ctefaho> "indefinite description implies that nothing is referred to"? 09:47 < ldlework> Yeah 09:47 < ctefaho> Meaning "The cats ate fish" implies no cats are referred to? 09:47 < ldlework> Or more precisely 09:47 < gleki> that seems correct when applied to zo'e 09:47 < ldlework> ctefaho: it means no fish are referred to 09:47 < ctefaho> huh 09:47 < ctefaho> what 09:48 < ctefaho> How is that supposed to make sense? 09:48 < gleki> "referred to" can confuse a lot of people though 09:48 < ldlework> gleki: indeed 09:48 < gleki> since logicians are a tiny part of population 09:48 < ldlework> Reference is a matter of language not logic. 09:48 < ldlework> There is no reference in logic. 09:49 < ctefaho> Ok well in my model, definite refers to a subset of the indefinite 09:49 < ctefaho> if I worded that correctly 09:49 < ctefaho> probably not 09:49 < ldlework> ctefaho: sure, that's getting closer for sure 09:49 < ldlework> Many people conceptualize indefinite reference as referring to some ideal case, or some platonic object, etc 09:49 < ctefaho> "le cmalu noltru" referring to a specific "lo cmalu noltru" 09:49 < ldlework> For which definite references are 'instances' of, etc 09:49 < ctefaho> or to a subset of "lo cmalu noltru" 09:50 < ldlework> ctefaho: well subset implies that lo cmalu noltru refers to the superset 09:50 < ldlework> or simple "all" 09:50 < ldlework> That'd be incorrect 09:50 < gleki> i'd rather rely on discourse since this is easier described both for ordinary people and for programmers 09:50 < ctefaho> I am not sure which terms to use here to say what I mean 09:50 < ldlework> ctefaho: prototype, blueprint, platonic model, etc 09:51 < ldlework> many just use 'property' 09:51 < ldlework> in that "a cat" is really just an instantiation of the property "to-cat" 09:51 < ldlework> In that you're saying 09:51 < ldlework> both there exists something, and that something is characterized by is "catness" 09:51 < ctefaho> ("A dragon" is what predicates to dragon, while "The dragon" simply refers to some "A dragon") 09:51 < ldlework> That's how descriptions are different than non-descriptions 09:52 < ldlework> Right 09:52 < ctefaho> but I probably need to read more to get further than that 09:52 < ldlework> So indefinite descriptions seem to skip the existential coupling 09:52 < ldlework> They seem to posit existence if nessary but don't require it 09:52 < ldlework> Such as 09:52 < ldlework> "Smith's murder is insane." 09:52 < ctefaho> "lo karce" to me is just "zo'e cu karce" 09:53 < ldlework> If no one, especially the speaker, doesn't know who smith's muderer is, the description doesn't refer. 09:53 < ldlework> Instead it seems to stage an instance of the property, and then uses that to predicate upon, to attribute insantiy. 09:53 < ctefaho> (plus xo'e something if one thinks lo should do default quantification) 09:54 < ldlework> ctefaho: that's kind of the direction I was going by establishing zo'e as a non-existential logical primitive 09:54 < ldlework> It would more closely resemble the x in first-order predicate language 09:54 < ldlework> Where in fopl, the variables don't actually represent some external thing for which the predicate are true about 09:54 < ldlework> instead in the caluclus of fopl, x represents an _entailment_ between predicates 09:55 < ldlework> A way in which predicates are logically related to each other 09:55 < ldlework> But in speech when we expand lojban to "fopl-form" 09:55 < ctefaho> and because someone went and made "mei" to a gismu: "lo karce" -> "zo'e cu mecni be li xo'e je karce" 09:55 < ldlework> we abuse da, the thing that represents our referents, as the "x" thing that we see in fopl statements 09:55 < ldlework> But I don't agree this is perfectly isomorphic 09:55 < ldlework> Because what is happening in fopl is not the attribution of predicates to objects 09:55 < ldlework> But the entailment of predicates to other predicates 09:55 < ldlework> And this *exactly* what you do in language 09:56 < ldlework> When you predicate upon indefinite descriptions 09:56 < ldlework> When I say 09:56 < ldlework> Cats have four legs 09:56 < ldlework> I'm not defining some object 09:56 < ctefaho> "le karce" just has that "lo karce" expansion plus me/menre 09:56 < ldlework> and asserting some conjunction 09:56 < ldlework> of 09:56 < ldlework> there exists an X, and that X is a cat AND that X has four legs 09:56 < ldlework> Nay 09:56 < ldlework> Instead I'm establishing a direct link between the predicate "cat" and the predicate "has-four-legs" 09:56 < ldlework> There is no object involved 09:57 < ldlework> So traditionally when we use lojban and expand it to logical forms 09:57 < ldlework> We have been systmatically conflating logical calclus with the action of language 09:57 < ldlework> Conflating definite speech for logical entailment 09:57 < ldlework> When I say 09:57 < ldlework> "Your dad is cool" in language, I only mean to assert that there is some guy, that is cool. 09:58 < ldlework> It is not nessarcy for that guy to be both a dad, and be cool 09:58 < ldlework> and further, there is no logical entailement between dadness and coolness 09:58 < ldlework> But in lojban this is what we aim to prove with the way in which we all go about expanding lojban 09:58 < ldlework> by conflating speech as isomorphic with the calculus of first-order predicate logic 09:59 < ctefaho> mhm 09:59 * selpahi wonders what version of "we" you'd be using if you were saying all this in Lojban. 10:00 < ldlework> Well, when I say "the way in which we expand lojban" I mean everyone who goes about it expanding lojban as far as instances I've seen 10:00 < ldlework> Since everyone seems to use da to refer to the referents of descriptions 10:01 < ldlework> And since everyone seems to agree that the purpose of expanding lojban this way to reflect the underlying predicate logic underneath 10:01 < ldlework> So the apparent conclusion is that everyone makes this conflation 10:01 < ldlework> Between the referents of descriptions in language, the variables in the calculus of predicate logic which reflect predicate entailments. 10:01 < ldlework> s/, /, and 10:01 < fenki> ldlework meant to say: Between the referents of descriptions in language, andthe variables in the calculus of predicate logic which reflect predicate entailments. 10:02 < ldlework> The only kind of speech that happens to map isomorphically to theorems fopl calclus is speech that predicates indefinate descriptions 10:03 < ldlework> as far as I can tell anyway 10:03 < ldlework> Which explains a lot of why CLL descriptions work the way they work honestly 10:04 < ldlework> And why they were not useful and demanded something wholly magical which is lo/xorlo but actually useful of language 10:06 * nuzba @nuttycom: Fantasizing about the creole that would arise among children raised as native speakers of lojban. [http://bit.ly/1Ij3y31] 10:06 < selpahi> How do CLL description work? 10:06 * ctefaho bails out for now, shall return once big walls of text have been conquered 10:07 < gleki> i'e doi la ctefaho 10:07 < ldlework> selpahi: they seem to make assertions about the nature of predicates themselves, having consequence for the scope of what we assert 10:08 < ldlework> Which is all fopl does too, is generate theorems describing entailments between predicates themselves 10:09 < ldlework> But when we speak, we're not always trying to establish new entailments between predicates, but rather ascribe predicates to objects which exemplify those predicates 10:09 < ldlework> It is my impression that this is exactly the problem that xorlo solved 10:09 < ldlework> selpahi: does that differ much from your understanding? 10:13 < gleki> oh my, i seem to have rewritten the whole "terms" section of Lojban.PEG manually including primitive te sumti detection 10:14 < gleki> wht is the tag for {snada}? :P Ithkuil seems to have one. 10:15 < ctefaho> fi'o snada si si 10:15 < gleki> zo uoi zo'o 10:16 < gleki> in ithkuil it's interesting that "to find" is ~= snada lo ka mitysisku 10:16 < ldlework> selpahi: interested to hear what you have to say 10:16 < gleki> not successful at searching but at the goal of the search 10:17 < gleki> snada lo te zukte be lo nu mitysisku 10:26 < selpahi> You keep saying things like "[...] da is an existential reference, zo'e is an indefinite reference. [...] only da can refer. So zo'e means something else." 10:26 < gleki> where is la zipcpi and la kurti when you need them 10:26 < selpahi> This is quite an abuse of terminology. 10:26 < selpahi> It is very far removed from everything we know. 10:26 < ldlework> selpahi: You mean of conventional lojban terminology? Of course, we started this by talking about some theory that no one would agree with. 10:26 < selpahi> (recall your recent discussion with latro'a) 10:26 < gleki> someone wants to have "la.gerku." account on LMW. not sure if it will work out 10:27 < ldlework> 09:34:58 ldlework | ctefaho: it doesn't really matter, its a theory no one will ever agree with 10:27 < ldlework> I don't know why people have such a hard problem dealing with contexts, or universe of discourse that diverge from the typical one, even when explicitly addressed 10:27 < gleki> oh huh it worked! http://mw.lojban.org/papri/User:La.gerku. 10:27 < gleki> dots in the end! 10:28 < ldlework> I was asking specifically about my words regarding the traditional CLL descriptions and why they demanded xorlo 10:28 < ldlework> But I guess we can just skip all substance and defer to, "You're talking in a way that allows me to talk about your way of talking." :/ 10:29 < selpahi> If you could be more specific in your critique of CLL descriptions that would be useful. 10:29 < selpahi> I do not agree that they were useless 10:29 < noncomcinse> coi 10:29 < ldlework> useless is obviously an extreme adjective, I was using for emphasis 10:29 < ldlework> In that they have use, but were less useful for ordinary speech 10:30 < ldlework> was my point 10:30 < gleki> cll was a great attempt at criss-crossing loglang and humanlang. seems to be not interesting these days. still a great attempt. 10:30 < selpahi> Would you like to show an example that goes along with your claim? 10:30 < ldlework> But it doesn't seem like you want to cooperatively explore the thought 10:30 < ldlework> But just wait for me to say something crucially wrong 10:31 < gleki> i can only see whorfian effects there in their greatest form: not speaking a language at all! :D 10:31 < ldlework> I guess I recind my iterrogation. 10:32 < gleki> it's interesting though that CLL was to a large degree written by a kinda programmer+resident of CONLANGS-L mailing list trying to describe formal logic 10:32 < selpahi> He had a completionist attitude, that's why. 10:33 < gleki> xu la kouan cu progypre gi'i ciste admine 10:34 < gleki> i gi'eru'e ba'a 10:34 < selpahi> na .admine ba'e ca ku .i na djuno tu'a lo purci .i ca ku la .aleks. je la .and. cu .admine 10:34 < selpahi> la runbau liste ku sa'e 10:35 < gleki> ti'e la kouan cu se jibri lo ka ciste admine 10:36 < selpahi> mi djuno no da lo srana 10:36 < gleki> ma xe fanva zoi gy. Is Alice more a programmer or a sysadmin.gy. 10:36 < selpahi> .i za'a dai mi fliba lo ka drani spuda 10:36 < gleki> zo'o la alis ku noi progypre cu zmadu la alsi ku noi ciste admine li'o 10:36 < gleki> s/alsi/alis/ 10:36 < fenki> gleki meant to say: zo'o la alis ku noi progypre cu zmadu la alis ku noi ciste admine li'o 10:37 < selpahi> me'ei sampla ji me'ei ciste .admine cu traji lo ka la .alis. cu ckaji ke'a 10:37 < gleki> oi oi i clani 10:37 < gleki> vau uenai 10:38 < selpahi> la .xorban. cu se pagbu su sidju valsi .i ku'i na mapti lo jbobau 10:40 < noncomcinse> .i la .xorban. cu mo 10:40 < selpahi> lojbau 10:40 < noncomcinse> camxes: lojbau 10:40 < camxes> (lojbau VAU) 10:40 < gleki> di'a 10:40 < gleki> s/di'a/de'a/ 10:40 < fenki> la'oigleki pu djica lo ka cusku zoi zoi de'a zoi 10:41 < gleki> y 10:41 < selpahi> zo'o la'oigleki cu terpa tu'a zo skudji 10:41 < noncomcinse> je'e .camxes. .i zo lojbau na'e valsrtosmabru ki'u tau re slaka 10:42 < gleki> terpa 10:42 < gleki> s/terpa/ja'a terpa/ 10:42 < fenki> la'oi gleki pu djica lo ka cusku zoi zoi ja'a terpa zoi 10:43 < selpahi> mi zanru ku'i 10:43 < selpahi> .i go'i lo nu lojbo ca ku 10:44 < noncomcinse> .i ma mukti lonu cusku zo ku .i xu famyma'o djedi zo'o 11:09 < ciuak> hello. 11:10 < gleki> coi 11:11 < ciuak> coi 12:17 < keidji> coi 12:17 < ldlework> coi keidji 12:17 < keidji> doi Idlework do mo 13:17 < selpahi> Where is everyone? 13:19 < selpahi> I was writing a wiki page and then my computer just turned itself off. 13:22 < selpahi> .seen Ilmen 13:22 < fenki> Sorry, I haven't seen ilmen around. 13:23 < Zluglu> menli leaves about 5 hours ago 13:23 < Zluglu> … “left” 13:43 * selpahi starts rewriting everything 18:29 < zipcpi> http://jbovlaste.lojban.org/dict/ganseti 18:54 < iloveass> mi prami lo do zargu 18:55 < iloveass> xu do prami le mi zargu 18:57 < iloveass> co'o rodo 19:28 < sorpaas> How to say "notation" in Lojban? 20:09 < noncomcinse> coi 20:12 < sorpaas> coi 20:12 < noncomcinse> coi sy. do mo 21:28 < gleki> .dict notation 21:28 < fenki> notation — noun: 1. (uncountable) The act, process, method, or an instance of representing by a system or set of marks, signs, figures, or characters, 2. (uncountable) A system of characters, symbols, or abbreviated expressions used in an art or science or in mathematics or logic to express t[...] 21:32 < gleki> lo ciste be lo sinxa - system of symbols/signs/signals 21:33 < gleki> en: kacmyxra 21:33 < mensi> kacmyxra [< kacma pixra ≈ Camera picture] = p1 is a photograph of p2 taken by photographer p3 in medium p4=k3. | 21:33 < mensi> phma 21:34 < gleki> zo'o pixra lo kacma 21:35 < gleki> i pixra lo se kacma 23:20 < gleki> zantufa: ke fami ke fe do mo 23:20 < zantufa> ([ke {<fa mi> <ke (¹fe do¹) KEhE>} KEhE] CU [mo VAU]) 23:21 < gleki> zantufa: ke fami ke'e ke fe do mo 23:21 < zantufa> ([{ke <fa mi> ke'e} {ke <fe do> KEhE}] CU [mo VAU]) 23:21 < gleki> zantufa: ke fa ke mi ke'e ke fe do mo 23:21 < zantufa> ([ke {<fa (¹ke mi ke'e¹)> <ke (¹fe do¹) KEhE>} KEhE] CU [mo VAU]) 23:21 < gleki> zantufa: ke fa ke mi mi ke'e ke fe do mo 23:21 < zantufa> ([{ke <(¹fa [ke mi KEhE]¹) mi> ke'e} {ke <fe do> KEhE}] CU [mo VAU]) 23:28 < gleki> zantufa: ke fa mi do ke fi do mo 23:28 < zantufa> ([ke {<fa mi> do <ke (¹fi do¹) KEhE>} KEhE] CU [mo VAU]) 23:30 < gleki> spagetufa: ke fa mi do ke fi do mo 23:30 < spagetufa> SyntaxError: Expected [,] but "m" found. 23:31 < gleki> spagetufa: ke fa mi do fi do mo 23:31 < spagetufa> (CU [ke {CU <(¹[{fa mi} {FE do} {fi do}] CU¹) mo VAU>} KE'E]) 23:31 < gleki> yyyyyyyyyyyyyyyyyy 23:31 < gleki> akmnlrse: xu te zu'e se gentufa 23:41 < gleki> exp: na ge mo gi mo 23:41 < mensi> (CU [na {ge <CU (¹mo VAU¹)> gi <CU (¹mo VAU¹)> VAU}]) 23:41 < gleki> xm i ma smuni --- Day changed Sat Jul 25 2015 00:30 < cliva> exp: na je gi mo gi mo 00:30 < mensi> (CU [{na je} gi] [CU {mo VAU}] gi [CU {mo VAU}] VAU) 00:30 < cliva> gleki: te zu'e ku ru'e 00:30 < durka42> coi 00:30 < cliva> coi 00:30 < durka42> ca lo bavlamdei mi renkla la muncen 00:30 < cliva> spagetufa: cu so'u nu'i do remna vau mi pavyseljirna vau na nei 00:30 < spagetufa> (cu [so'u BOI] nu'i [{CU <(¹[FI'A do] CU¹) rèmna vau>} {CU <(¹[FI'A mi] CU¹) pavy/sel/jìrna vau>} {CU <(¹na nei¹) VAU>}] VAU NU'U) 00:31 < cliva> xu ca lo pavdei cu zvasto la .franfurt. 00:31 < cliva> .i va'o ku mi klama 00:31 < gleki> spagetufa: ge ba gi ca ma 00:31 < spagetufa> SyntaxError: Expected [,], [nN] or [rR] but "m" found. 00:31 < gleki> i wanted to make {ge ca gi ba ko'a} gendra. 00:31 < gleki> but all these {[ca gi ca ba] [gi ba]} make it ambiguous. 00:32 < gleki> and selbri autorestorations want to inject {co'e} into {ge ... gi} all the time :) 00:32 < cliva> alta: na ge ti gi ta 00:32 < mensi> (NIhO [I {na <ge (¹FA ti¹) CU (¹COhE VAU¹)> gi <(¹FA ta¹) CU (¹COhE VAU¹)> VAU} VAU]) 00:32 < cliva> spagetufa: na ge ti gi ta 00:32 < spagetufa> (CU [na {ge <(¹FA ti¹) VAU> gi <(¹FA ta¹) VAU> VAU}]) 00:33 < cliva> .ei ja'ai cikre 00:33 < gleki> cikre ma 00:33 < cliva> lo bi'u nai .gek. lo ka cupra lo sumti je nai lo brirebla 00:34 < gleki> cikre la spagetufa vau xu 00:34 < gleki> i je'e 00:34 < cliva> go'i fa je lo drata ku ru'a 00:34 < cliva> s/fa/fe/ 00:34 < fenki> la'oi cliva pu djica lo ka cusku zoi zoi go'i fe je lo drata ku ru'a zoi 00:35 < durka42> cliva: la'a no'e zvasto la franfurt 00:35 < cliva> .aidji co klama ma 00:35 < durka42> mi renkla fi la stutygart fe lo voizda 00:35 < gleki> lo fa la altatufa cu cnino cu cupra lo sumti 00:36 < cliva> ua 00:36 < gleki> i ku'i le cnino nu'o bredi 00:36 < durka42> mi vitke bagi la muncen ca lo djedi be li re gi la stutygart ca lo djedi be li pa 00:37 < cliva> je'e .i ka'e penmi zai la .ctutkart. ku si'a 00:37 < cliva> .i lo zei sei za'a 00:38 < cliva> lo fa je ke'a ja bo ce'u .altatufa cu cnino sa .i 00:38 < durka42> cliva: ui da'i 00:39 < cliva> cu'e sa'e 00:39 < cliva> mh: test 00:39 < durka42> la'a .irci slilu .i ka'e mrilu fi'o judrysni durka42@gmail.com 00:39 < cliva> je'e 00:53 < gleki> lu lo fa je ke'a ja bo ce'u .altatufa cu cnino li'u ca na gendra 00:54 < gleki> i ku'i 00:54 < gleki> (NIhO [I {FA <lo (¹[{fa <la altatufa KU>} cu] cnino¹) KU>} CU {COhE VAU}]) 00:55 < gleki> alta: lo fa la .altatufa cu cnino 00:55 < mensi> (NIhO [I {FA <lo (¹[{fa <la altatufa KU>} cu] cnino¹) KU>} CU {COhE VAU}]) 00:55 < gleki> ie i ze'u zo'o gendra 00:57 < akmnlrse> alta: lo cu broda cu broda 00:57 < mensi> SyntaxError: Expected [,] but "b" found. 00:57 < akmnlrse> alta: lo fa cu broda 00:57 < mensi> (NIhO [I {FA <lo (¹[{fa ZOhE} cu] broda¹) KU>} CU {COhE VAU}]) 00:57 < akmnlrse> alta: lo mi fa cu broda 00:57 < mensi> (NIhO [I {FA <lo (¹[NE {DOhE mi} GEhU] [{fa ZOhE} cu] broda¹) KU>} CU {COhE VAU}]) 00:59 < gleki> lo'u lo cu broda le'u ge'e i na djuno lo du'u xukau plixau 00:59 < akmnlrse> ie 00:59 < gleki> alta: mo 00:59 < mensi> (NIhO [I {mo VAU}]) 00:59 < gleki> alta: cu mo 00:59 < mensi> SyntaxError: Expected [yY] but "o" found. 00:59 < gleki> oi 00:59 < akmnlrse> alta: ni'o i mo 00:59 < mensi> (ni'o [i {mo VAU}]) 00:59 < akmnlrse> alta: ni'o i cu mo 00:59 < mensi> SyntaxError: Expected [yY] but "o" found. 01:00 < akmnlrse> alta: ni'o i sei brodacu mo 01:00 < mensi> (ni'o [i {sei <brodacu mo> SEhU}] [COhE VAU]) 01:00 < akmnlrse> alta: ni'o i sei broda cu mo 01:00 < mensi> SyntaxError: Expected [yY] but "o" found. 01:00 < akmnlrse> alta: mo .ijebo cu mo 01:00 < mensi> SyntaxError: Expected [yY] but "o" found. 01:01 < gleki> ie ie i lu cu mo li'u pu jai nabmi 01:02 < akmnlrse> alta: mo vau je mo 01:02 < mensi> (NIhO [I {mo vau} {je CU <mo VAU> VAU}]) 01:02 < akmnlrse> za"a zantufa simsa ca 01:02 < gleki> alta: mi prami 01:02 < mensi> (NIhO [I {FA mi} CU {prami VAU}]) 01:02 < gleki> exp: mi prami 01:02 < mensi> (mi [CU {prami VAU}]) 01:03 < gleki> zo cu di'a te sepli be lo rebla bei lo stedu be'o cmavo 02:58 < f8l> Tachyon`: Small world, eh? 03:05 < Tachyon`> ahh, hello f8l 03:27 < f8l> Do you speak Lojban already? 04:00 < ctefaho> coị 04:27 * nuzba @shukil026: Lojban [http://bit.ly/1SHi2iY] 04:42 < noncomcinse> coi 04:42 < noncomcinse> coi 04:43 < ctefaho> coị coị 04:43 < noncomcinse> coi la ctefa'o do mo 04:46 < ctefaho> .i mi pinxe lo ckafi gi'e co'a cikna 04:47 < noncomcinse> .i ji'a co'a co'e .i ku'i lo tcati lo ckafi tu'a mi zmadu 04:49 < noncomcinse> .i ku'i mi na ponse 05:01 * nuzba @Naikodemus: Retomamos: http://mw.lojban.org/papri/Lojban#Recursos ; #lojban [http://bit.ly/1D10Vaa] 05:10 * mumble jorne co'a fa la'o ce doi darca be ti ko se ba'i klama la jinme bunkere ne zoi fa mumble.metal-bunker.info:64738 fa ce 05:19 < gleki> parsing {ke fa mi fe do mi} etc. in altatufa is as problematic as parsing {fu'e} in standard camxes 05:26 * mumble jorne co'a fa la'o fa akti fa 05:26 * mumble jorne co'u fa la'o fa akti fa 05:28 < akmnlrse> test 05:28 < akmnlrse> s/test/zoi zoi/ 05:28 < fenki> la'oi akmnlrse pu djica lo ka cusku zoi zoi zoi zoi zoi 05:28 < noncomcinse> .u'i 05:28 < noncomcinse> There's a reason the delimiter word has a denpa bu after it. 05:32 < gleki> co'o 05:32 < gleki> s/co'o/coi/ 05:32 < fenki> la'oi gleki pu djica lo ka cusku zoi zoi.coi.zoi 05:33 < gleki> {zmakai} doesnt solve the problem of "Is Cowan more a sysadmin or a programmer or vice versa?" 05:33 < gleki> oops 05:33 < gleki> {zmakai} doesnt solve the problem of "Is Cowan more a sysadmin than a programmer or vice versa?" 05:35 < akmnlrse> la .kouan. ma trajykai lo ka tsisadmine jo'u lo ka sampla 05:35 < noncomcinse> s 05:35 < akmnlrse> .i ji di'u si'a jai se srera 05:36 < noncomcinse> s/s/WQ.zoi coi lo jbopre zoi zoi.WQ/ 05:36 < fenki> la'oi noncomcinse pu djica lo ka cusku zoi zoi.WQ.zoi coi lo jbopre zoi zoi.WQ.zoi 05:36 < noncomcinse> .u'i 05:36 < akmnlrse> .u'i 05:37 < noncomcinse> s 05:37 < noncomcinse> 7:36:35 <noncomcinse> s/s/WQ.zoi si coi lo jbopre fa'o zoi zoi.WQ/ 05:37 < noncomcinse> s 05:37 < noncomcinse> s/s/WQ.zoi si coi lo jbopre fa'o zoi zoi.WQ/ 05:37 < fenki> la'oi noncomcinse pu djica lo ka cusku zoi zoi.WQ.zoi si coi lo jbopre fa'o zoi zoi.WQ.zoi 05:37 < Ilmen> {la .kouan. cu certu traji lo ka tsisadmine ku je'i lo ka sampla} 05:37 < Ilmen> .e'u ru'e 05:38 < gleki> lo ka ciste admine vau lo ka progypre cu se xi xo zmadu lo ka la kouan cu ckaji ce'u 05:39 < Ilmen> {ma traji lo ka la .kouan. ce'u ckaji ku lo ka ciste .admine ku jo'u lo ka sampla 05:39 < Ilmen> } 05:40 < noncomcinse> s 05:40 < noncomcinse> s/s/WQ.zoi si lu coi lo jbopre li'u fa'o zoi zoi.WQ/ 05:40 < fenki> la'oi noncomcinse pu djica lo ka cusku zoi zoi.WQ.zoi si lu coi lo jbopre li'u fa'o zoi zoi.WQ.zoi 05:40 < gleki> i can shut down te sumti autodetection using peg thus allowing any cases of {ke fa mi ke fi do mi ke ma cu broda} (te sumti autodetection can be done using cirko's or mukti's tricks) but the problem of {fu'e} still stands 06:01 < noncomcinse> lo xe ckule na djuno fi lo glare djacu gasnu minji 06:01 < noncomcinse> vau .u'i 06:18 < noncomcinse> .i ma fasnu 06:22 * nuzba @blubberquark: We need to agree on common language, like Esperanto or Lojban or JSON # singularity [http://bit.ly/1D1bvxD] 06:22 * nuzba @blubberquark: We need to agree on a common language, like Esperanto or Lojban or JSON #singularity [http://bit.ly/1I0pKji] 06:22 < zipcpi> lol JSON 06:22 < Zearen> .u'i 06:33 < zipcpi> Hm... tempted to make {zvaksu} just to give an alternative to selpahi's zi'ejvo lol 06:33 < ctefaho> doi gleki 06:34 < zipcpi> Wait that's a slinku'i 06:34 < zipcpi> zvasiiu 06:35 < zipcpi> zipcpi, abuser of morphology 06:35 < zipcpi> la zipcpi voi'i malpli lo vlaturge'a 06:37 < zipcpi> spagetufa: zvasiiu 06:38 < zipcpi> za'a na ca zvati 06:38 < cliva> exp: zvasiiu 06:38 < mensi> (CU [zvasiiu VAU]) 06:45 < cliva> en: zvasisku 06:46 < mensi> zvasisku [< zvati sisku ≈ At seek] = x1=s1 looks for the location of x2=z1 among location(s) x3=s3=z2 | s2 = "lo ka z1 06:46 < mensi> zvati ce'u", so is eliminated. | Eimi 06:47 < zipcpi> lo pendo be mi cu e'onde mi lo ka ciska lo se jibri cpedu xatra bau lo mejbau tezu'e lo nu xajmi 06:48 < zipcpi> .i mi ciska keji'a lo jbobau xatra 06:48 < zipcpi> coi .i au mi se jibri do .i mi mutce lo ka certu tu'a lo'e skami .i mi kulmu'o fi lo balcu'e lo samske jecu vinta lo li'i jitro lo datnyci'e je lo skami te lanme ja ke xagji sofybakni kei ze'a lo nanca be li mu .i e'u ko spuda fi tu'a lo penmi temjudri poi jai cumki .i ki'esai 06:49 < cliva> la'e di'e ka'e cinri fa'o https://xsznix.github.io/me/jb/ralju.html 06:49 < zipcpi> mi tolsnuti cuxna lo valsi poi cizra 06:49 < zipcpi> lu vinta lo li'i jitro lo datnyci'e je lo skami te lanme ja ke xagji sofybakni li'u 06:51 < zipcpi> ua a'u 06:52 < zipcpi> .iku'i mi na djuno lo du makau mapti valsi lo si'o mu'oi gy job experience gy 06:53 < zipcpi> .iseki'ubo mi malpli zo vinta .i xamzu'e po'o 06:54 < gleki> xu citri lo nu se jibri 06:54 < zipcpi> xy'y ka'e mapti 06:55 < zipcpi> te lanme ja ke xagji sofybakni 08:18 < gleki> why was it even necessary to merge tags when they were so nice splitted in the original camxes 08:18 < gleki> split 09:01 < gleki> i found seomthing strange in masato's parser 09:01 -!- Irssi: #lojban: Total of 152 nicks [3 ops, 0 halfops, 0 voices, 149 normal] 09:02 < gleki> anyone listening here? 09:04 < zipcpi> Hi 09:06 < gleki> termset = termset:(gek_termset / NUhI_clause free* gek terms NUhU_clause? free* gik terms NUhU_clause? free* / NUhI_clause free* terms NUhU_clause? free*) {return _node("termset", expr);} 09:06 < gleki> termset85 = NUhI # gek terms /NUhU#/ gik terms /NUhU#/ | NUhI # terms /NUhU#/ 09:06 < gleki> compare the two strings 09:06 < gleki> the second is from cll 09:09 < zipcpi> mi na jimpe 09:09 < gleki> fi lo re xu moi 09:10 < zipcpi> mi na se bangu la .peg. 09:10 < gleki> je'e 09:10 < gleki> i ei denpa tu'a loi drata 09:26 < gleki> gerna: ge ba gi ca ma 09:26 < mensi> O_0 09:26 < gleki> gerna: ge ba gi ca ma mo 09:26 < mensi> O_0 09:27 < gleki> exp: ge ba gi ca ma mo 09:27 < mensi> SyntaxError: Expected [yY] but "o" found. 09:27 < gleki> exp: ge ba ku gi ca kku ma mo 09:27 < mensi> SyntaxError: Expected [,] but "k" found. 09:27 < gleki> exp: ge ba ku gi ca ku ma mo 09:27 < mensi> ([{ge <ba ku> gi <ca ku>} ma] [CU {mo VAU}]) 09:27 < gleki> gerna: ge ba ku gi ca ku ma mo 09:27 < mensi> O_0 09:57 < akmnlrse> gleki: xu tavla fi la gektermsete 09:59 < akmnlrse> .i pagbu la .camxes. ku pe la camgusmi 10:05 * nuzba @oka_iu_tcan: lojbanはサーバーが貧弱すぎる [http://bit.ly/1IkNwpx] 10:08 < akmnlrse> si'au la xorxo cu finti http://webcache.googleusercontent.com/search?q=cache:GCbXv972BnYJ:tiki.lojban.org/tiki/tiki-view_forum_thread.php%3Ftopics_offset%3D1%26topics_sort_mode%3Dhits_asc%26forumId%3D5%26comments_parentId%3D3687+&cd=2&hl=en&ct=clnk&gl=ch&client=ubuntu&strip=1 10:09 < akmnlrse> ku'i lo mupli ne lu "ge zo by. .a zo beiste gi zo zy. .a zo zgana" cu simlu lo ka toldra 10:09 < akmnlrse> exp: ge zo by. .a zo beiste gi zo zy. .a zo zgana 10:09 < mensi> ([{ge <(¹zo by¹) (¹a [zo beiste]¹)> gi <zo zy>} {a <zo zgana>}] VAU) 10:10 < akmnlrse> zantufa: ge zo by. .a zo beiste gi zo zy. .a zo zgana 10:10 < zantufa> ([ge {<zo by> <a (¹zo beiste¹)>} {gi <zo zy> <a (¹zo zgana¹)>} GIhI] VAU) 10:10 < akmnlrse> spagetufa: ge zo by. .a zo beiste gi ke zo zy. .a zo zgana 10:10 < spagetufa> ([ge {FA <(¹zo by.¹) .a (¹zo bèi/ste¹)>} gi {FA <ke (¹[zo zy.] .a [zo zgàna]¹) KE'E>}] VAU) 10:14 < akmnlrse> s/" /" li'u/ 10:14 < fenki> la'oi akmnlrse pu djica lo ka cusku zoi zoi.ku'i lo mupli ne lu "ge zo by. .a zo beiste gi zo zy. .a zo zgana" li'ucu simlu lo ka toldra.zoi 10:24 < akmnlrse> rlpowell: lo kibystu cu simlu lo ka spofu (to denpa tu'a lo selfu co'u lo nu lo jimte cu ditcu toi) 10:29 < gleki> akmnlrse: spofu zea lo mentu be li su'o xo 10:29 < akmnlrse> li su'o 20 10:31 < gleki> mo'u benji SMS 10:31 < akmnlrse> .i'o 10:49 * mumble jorne co'u fa la'o ce doi darca be ti ko se ba'i klama la jinme bunkere ne zoi fa mumble.metal-bunker.info:64738 fa ce 11:13 < noncomcinse> coi lo jbopre do'u mi ve'u klama 11:37 < noncomcinse> eo: tanru 11:38 < mensi> tanru = x1 estas binara metaforo formita per x2 modifanta x3-on, kun signifo x4 en uzo/instanco x5 | (x2 kaj x3 ambaŭ 11:38 < mensi> estas tekstoj aŭ ambaŭ konceptoj si'o) Vidu ankaŭ gismu, smuni. | xorxes 11:38 < noncomcinse> eo: klama 11:38 < mensi> klama = x1 venas/iras al celita loko x2 el deirpunkto x3 tra vojo x4 per rimedo/veturilo x5 | Ankaŭ: vojaĝas, moviĝas al 11:38 < mensi> ... de ...; x1 estas vojaĝanto; (x4 kiel aro inkluzivas punktojn almenaŭ sufiĉajn por difini la vojon). Vidu ankaŭ 11:38 < mensi> cadzu, bajra, marce, vofli, litru, muvdu, cpare, cmavo ka'a, pluta, bevri, farlu, limna, vitke. | 11:38 < mensi> xorxes 11:44 < gleki> cu'u la robin ri darno lo skami 11:45 < gleki> i la gek term set zo'u mabla i se pagbu su'e je su'o re da 12:01 < noncomcinse> mi pilno la mensi lonu sidju mi lonu tavla fo la spero 12:01 < noncomcinse> .i'au vau .u'i 12:12 < noncomcinse> si'au smaji 12:48 < Ilmen> coi 13:10 < ctefaho> coị doị 13:11 < ctefaho> laɂ Ilmen 13:14 < ctefaho> hey Ilmen how common is /c/ in French? 13:14 < ctefaho> iirc you speak it 13:15 < Ilmen> coi 13:15 < Ilmen> .i mi ja'a se bangu .i .e'o ze'i denpa 13:16 < Ilmen> Do you mean [ʃ]? 13:17 < Ilmen> [c] doesn't exist in French. 13:17 < ctefaho> ehm 13:17 < ctefaho> french "qui"? 13:17 < Ilmen> Or do you mean the letter "c"? 13:17 < ctefaho> this https://en.wikipedia.org/wiki/Voiceless_palatal_stop 13:18 < Ilmen> As far as I know, this consonant doesn't exist in modern French. 13:19 < ctefaho> mhm 13:19 < noncomcinse> Is that lojban tc? 13:19 < Ilmen> "qui" is pronounced /ki/ 13:20 < Ilmen> And phonetically realized [ki] as well 13:20 < ctefaho> then why does wiki say [ci] 13:21 < ctefaho> ok, right, a range 13:21 < noncomcinse> is /c/ the same as [tc]? 13:21 < Ilmen> As for the [tᶴ] affricate, I think it's mostly present in loanwords. 13:21 < ctefaho> noncomcinse, no, it doesn't exist in lojban 13:22 < noncomcinse> Okay. 13:22 < noncomcinse> English example? 13:22 < tsani> Ilmen: I believe parisians pronounce "tu" somewhat like {tcu} 13:22 < tsani> (Montrealers pronounce "tu" more like {tsu}) 13:22 < ctefaho> something like english "keen" 13:23 < ctefaho> coị Sky 13:23 < tsani> coi 13:23 < Ilmen> Possible. I'm not parisian and am not familiar with every French accent. 13:23 < ctefaho> but I got my question answered then I think 13:23 < Ilmen> As for me I pronounce "tu" as [ty] 13:23 < Ilmen> ta'o coi 13:24 < tsani> whoops, yes, those "u" should be "y" in my pronunciations. 13:25 < tsani> Point is, I think that in French, after [t] and before a front vowel, many accents do include a sibilant / use an affricate 13:25 < Ilmen> I don't see any reference to [c] on the relevant Wikipedia articles 13:26 < tsani> .i ta'o mi jinvi lo du'u no roi ba'o ku mi do fasysku tinju'i 13:26 < tsani> .i lo nu do lojbo cusku zo'u mi do tirna 13:26 < Ilmen> ra'o ba'anai .u'i 13:26 < tsani> .u'i 13:29 < noncomcinse> My sister is fluent in French and spent a semester in Paris, she might know. It might be better to find a native, ku'i 13:32 < ctefaho> well after checking french phonology wiki page a bit better it seems [c] is not that common 13:32 < ctefaho> in fact the only mention seems to be "Some speakers pronounce /k/ and /ɡ/ as [c] and [ɟ] before /i, e, ɛ, a, ɛ̃/ and at the end of a word." 13:33 < Ilmen> The Wikipedia article lists [x] as an optional French phoneme in loanwords; however when I hear people pronouncing Spanish personal names, it seems to me they make no pronunciation difference between Spanish "r" and Spanish "j" ( [x] ) 13:37 * nuzba @iswebsitedownow: http://lojban.org is offline again! #down http://www.iswebsitedownnow.com/d/lojban.org [http://bit.ly/1JFEEeX] 13:38 < Ilmen> ctefaho: Interesting. I've never noticed such a phenomenon, but it may have escaped my awareness of course. At least myself I never realize /k/ as [k]. 13:38 < Ilmen> er 13:38 < Ilmen> as [c] 13:38 < Ilmen> ja'o se'i mi tatpi 13:39 < ctefaho> I see 13:40 < Ilmen> The wiki article lists /ɲ/ among the phoneme, and I think it says somewhere that /ɲ/ is realized [nj] in some dialects 13:41 < Ilmen> I think I do pronounce /ɲ/ as [nj], although I'm not 100% sure, as I have still some trouble distinguishing them :p 13:41 < Ilmen> It do seem to me that I have the "magner"/"manier" merger 13:42 < durka42> xu casnu lo fraso bankle 13:42 < Ilmen> I pronounce them the same way 13:42 < Ilmen> go'e 13:43 < noncomcinse> What's the IPA for A in American "cats"? 13:43 < Ilmen> æ 13:43 < Ilmen> kæt 13:43 < noncomcinse> kihe 13:43 < Ilmen> Some people think that "a" is better for at least some of the modern dialects 13:44 < noncomcinse> The O in "dog"? 13:44 < noncomcinse> I should really learn the vowels. 13:45 < Ilmen> I'm not good at determining how to pronounce "o" in the American accents, because there have been a split there, causing a disharmony with the British English o sounds 13:45 < Ilmen> But I think it's [dɑg] 13:46 < Ilmen> In British English it would be [dɒg] 13:46 < Ilmen> (the only difference is the lip rounding of the vowel) 13:47 < noncomcinse> kihe 13:48 < ctefaho> I hope yall familiar with [ɧ] 13:48 < Ilmen> It's present in Swedish, if I'm not mistaken 13:48 < noncomcinse> Uh, not familliar with that one. 13:48 < ctefaho> and Kölsch 13:48 < Ilmen> It's a rare consonant 13:48 < ctefaho> https://en.wikipedia.org/wiki/Sj-sound 13:49 < ctefaho> it is somewhat like [ʃ] but not quite 13:50 < ctefaho> we also got [ʏ] vs "[ʏ̫]" for anyone interested 13:50 < ctefaho> I still have no idea how I distinguish them 13:51 < ctefaho> or well somewhat 13:51 < noncomcinse> I'm amazed at the Unicode glyphs my phone can render. 13:51 < ctefaho> second one being the "near-close near-front protruded vowel" 13:52 < ctefaho> which only swe and norwegian seem to have 13:52 < noncomcinse> I'm not very familliar with all the weird consonants. 13:53 < ctefaho> first vowel in https://en.wikipedia.org/wiki/File:Sv-ylle.ogg vs https://en.wikipedia.org/wiki/File:Sv-ut.ogg 13:53 < fushgrhfhxdz> coi 13:53 < noncomcinse> IPA has the coolest facts of what human speech is capable of, though. 13:53 < ctefaho> yeah 13:54 < noncomcinse> coi la'oi . fushgrhfhxdz . 13:54 < ctefaho> hey let's design a language using IPA as its script 13:54 < ctefaho> all the sounds abound 13:54 < noncomcinse> Just looking at the chart ends up revealing a tribe of clicking people. 13:54 < ctefaho> no one will be able to speak it or hear it but it will be awesome 13:55 < noncomcinse> Oh, dear god. If we actually did make a full IPA language, we'd need five keyboards. 13:55 < durka42> nah just one but it's an APL keyboard 13:55 < fushgrhfhxdz> You could try Ithkuil first, you know, as a low level test. 13:55 < ctefaho> you mean you haven't practiced your [ǂ] today? 13:56 < durka42> my double dagger? 13:56 < noncomcinse> What even, ctefaho? 13:56 < ctefaho> https://en.wikipedia.org/wiki/Tenuis_palatal_click 13:57 < ctefaho> [ʘ] though 13:57 < fushgrhfhxdz> Alright I'll make this language. I already see some potential 13:57 < ctefaho> or your https://en.wikipedia.org/wiki/Tenuis_alveolar_lateral_click 13:58 < noncomcinse> ctefaho: If we did this, just as an experiment, I forsee one thing possibly happening if it became popular enough. 13:59 < fushgrhfhxdz> but does it also have to have tones? 13:59 < noncomcinse> People start using every sound in IPA, this language alone causes sounds to split. 13:59 < ctefaho> oh yes tones too! 13:59 < ctefaho> all of them 13:59 < ctefaho> can we make new ones 14:00 < ctefaho> noncomcinse: and creating new sounds with even finer distinctions? 14:00 < noncomcinse> durka42: APL keyboard? 14:00 < noncomcinse> ctefaho: Yeah. 14:00 < durka42> for the programming language APL 14:00 < ctefaho> then we get new sounds to add! 14:00 < noncomcinse> ctefaho: Which would be added to IPA and then be added to our language. Recurse. 14:00 < durka42> http://www.wickensonline.co.uk/apl/union-large.png 14:00 < durka42> afk 14:01 < noncomcinse> Wow. 14:01 < fushgrhfhxdz> inventing phones sounds like fun 14:03 * ctefaho is happy with his iphone 14:03 < noncomcinse> ctefaho: go'i ra'o naku 14:04 < fushgrhfhxdz> not talking about mobile phones 14:04 < ctefaho> me neither 14:04 < ctefaho> I meant [ʏ̫] 14:04 < noncomcinse> ua je'e 14:04 < ctefaho> zo'osai 14:05 < fushgrhfhxdz> loglang, artlang, engelang? 14:06 < noncomcinse> .i lo fonxa zo'u djica tu'a lo fonxa pe me'e lu pa su'i pa li'u 14:07 < fushgrhfhxdz> lets not make another loglang... id say a mix of artlang and engelang 14:07 < ctefaho> engelang? 14:08 < noncomcinse> .i la'e zoi .urli. https://oneplus.net/one .urli srana 14:08 < noncomcinse> le fonxa 14:10 < ctefaho> "Den riktiga flaggskeppsdödaren" 14:10 < ctefaho> oh my 14:10 < ctefaho> such svenska 14:10 < ctefaho> mycket style 14:10 < ctefaho> waw 14:11 < noncomcinse> sa'ei .uat. 14:11 * ctefaho brb gonna fetch more jordgubbisar 14:14 < ctefaho> more jordgubbisar fetched 14:14 < ctefaho> may the nomness begin 14:17 <@rlpowell> gleki: Forcefully restarting the web server VM. 14:17 < ctefaho> rlpowell: What has been up with the website lately? 14:18 <@rlpowell> I'm going to need a much more specific quesiton. 14:19 < ctefaho> Well why is it down and not up:p 14:19 <@rlpowell> Right now? Web server VM got stuck. Dunno what that happens sometimes. 14:19 < noncomcinse> Because the Party says which way is up and which way is Down. 14:19 < ctefaho> ah well I meant more generally 14:19 < noncomcinse> s/D/d/ 14:19 < fenki> la'oi noncomcinse pu djica lo ka cusku zoi zoi.Because the Party says which way is up and which way is down..zoi 14:20 < ctefaho> iirc when I was hanging around half a year ago site was basically up 24/7 14:20 < noncomcinse> zo'o 14:20 < ctefaho> never these downtimes 14:21 <@rlpowell> Yay mediawiki. 14:23 <@rlpowell> Its' up now. 14:24 < niek> umop apisdn 14:24 < ctefaho> lel 14:26 < ctefaho> also yay 14:37 < _mukti_> en: kakne 14:37 < mensi> kakne = x1 is able to do/be/capable of doing/being x2 (event/state) under conditions x3 (event/state). | Also: has 14:37 < mensi> talent; know how to; know how to use (= plika'e). (cf. stati, certu, gasnu (in the time-free potential sense), cmavo 14:37 < mensi> list ka'e, cmavo list nu'o, cmavo list pu'i, djuno, zifre) | officialdata 14:38 < _mukti_> Is the fact that kakne2 is defined as an event rather than a property an artifact of older lojbanic thinking? 14:38 < durka42> pe'i ie 14:38 < _mukti_> coi ki'e la durkavore 14:39 < durka42> I think I remember discussions about this 14:39 < durka42> like whether {mi kakne lo nu do broda} should be interpreted as {mi kakne lo ka gasnu lo nu do broda} or if it's just nonsensical 14:39 < _mukti_> Is there a shade of difference that people perceive between {mi kakne lo ka limna} and {mi kakne lo nu limna{ 14:39 < _mukti_> s/{/}/ 14:39 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi.Is there a shade of difference that people perceive between }mi kakne lo ka limna} and {mi kakne lo nu limna{.zoi 14:39 < _mukti_> haha 14:40 < durka42> well the second one just doesn't make too much sense to me 14:40 < durka42> (yes the new bot is a little overzealous) 14:40 < _mukti_> Is that your handiwork? It's amazing in any case. 14:41 < _mukti_> So for you, {mi kakne lo nu limna} would be received as meaning what today would be expressed as {mi kakne lo ka limna}? 14:42 < _mukti_> .i ta'o xu do ca'o zvati lo dotygu'e 14:43 < durka42> go'i 14:43 < durka42> ca lo bavlamdei mi renkla la .muncen. 14:43 < _mukti_> ua 14:43 < _mukti_> .i xu do ba vitke fi la .berlin. 14:43 < durka42> {mi kakne lo nu limna} raises the question to me that if x1 isn't you, who is it and how are you controlling their limbs? so given that x1 is you, why not use ka? :) 14:43 < durka42> mi na vitke la berlin 14:44 < durka42> darno dukse 14:44 < _mukti_> .i mi djuno lo du'u darno 14:46 < _mukti_> So structurally, I get the idea that since {nu} doesn't inject a {ce'u}, that {lo nu limna} is dependent on the first {zo'e} using it's magic powers of obviousness to represent {mi} 14:46 < _mukti_> I guess I just wondered if {kakne lo nu} had some residual, useful meaning that {kakne lo ka} does not 14:47 < _mukti_> And in any case, it seems like a good candidate for an updated definition. 14:47 < _mukti_> Since {ka} is not an event. 14:47 < durka42> yes 14:47 < durka42> x1 is capable of having the property x2 14:47 < durka42> cumki fa lo nu x1 ckaji x2 14:47 < _mukti_> This came up when I spent some time today with the 25 most frequent gismu in the ralju korpora 14:47 < durka42> jbo:kakne 14:47 < mensi> kakne = lo nu x1 zukte ja ckaji x2 cu cumki x3 14:48 < durka42> zo zukte ueru'e 14:48 < _mukti_> mi ji'a se spaji tu'a zo zukte 14:49 < _mukti_> So one thing I was doing as I looked over the gismu was considering how useful the places are. 14:49 < _mukti_> A favorite old pastime. 14:50 < _mukti_> What do you think of mutce3? 14:50 < durka42> did you come to the conclusion that under-conditions and by-standard places are mabla and should all be burned? :p 14:50 < durka42> vlaste: mutce 14:50 < vlaste> mutce = x1 is much/extreme in property x2 (ka), towards x3 extreme/direction; x1 is, in x2, very x3. 14:50 < _mukti_> haha, pretty much 14:50 < durka42> huh what do you even put in that x3 14:50 < _mukti_> I have never used mutce3 and would have to look it up if someone else did. 14:50 < durka42> it sounds like if you use a non-obvious x3 you could turn {mutce} into {tolmutce}? 14:50 < _mukti_> I wonder what gleki did with it. 14:50 < durka42> lb:mutce 14:51 < durka42> jb:mutce 14:51 < mensi> mutce = mutce — x1(entity) is extreme in x2(property of x1) 14:51 < mensi> :lo mutce — extreme, very. 14:51 < mensi> :ti mutce lo ka kargu — This is very expensive. 14:51 < mensi> :mi mutce lo ka senpi — I highly doubt that. 14:51 < mensi> :Related words: milxe, traji, banli, carmi, nutli 14:51 < durka42> yep he jettisoned it :) 14:51 < _mukti_> Fantastic. Then let's leave it out of "Lojban: The Good Parts". :) 14:51 < _mukti_> cmalu3 was the next one. 14:51 < durka42> is that an upcoming O'Reilly book? 14:51 < _mukti_> jb:cmalu 14:51 < mensi> cmalu = cmalu — x1(entity) is small in length/width etc. x2(property of x1) 14:51 < mensi> :lo cmalu — small. 14:51 < mensi> :lo nu penmi ba cmalu lo ka ditcu — The meeting will be short in duration. 14:51 < mensi> :lo cukta cu cmalu lo ka jai vajni — The book is of little importance. 14:51 < mensi> :Related words: caxno, cinla, jarki, tordu, barda 14:52 < _mukti_> One can only hope! 14:52 < durka42> cmalu3 is a by-standard place so you know my opinion! 14:52 < _mukti_> Chop. Done. 14:52 < _mukti_> viska3 falls to the "under conditions" rule. 14:53 < _mukti_> drata3: standard. 14:53 < _mukti_> binxo3: conditions 14:53 < durka42> by-standard => ma'i 14:53 < durka42> under-conditions => va'o 14:54 < _mukti_> How do you feel about kumfa3? 14:54 < _mukti_> lb:kumfa 14:54 < _mukti_> lb: kumfa 14:54 < Ilmen> {mutce3} has fallen out of favor, provided it has ever been in favor. Having zo'e as a defaut for mutce3 would be awful. Now using {mutce} for "a lot" and {toltce} for "a litte" is the preferred usage. 14:54 < _mukti_> coi la .ilmen. 14:54 < durka42> I don't know any of these high-numbered places .u'i .oise'i 14:54 < durka42> vlaste: kumfa 14:54 < vlaste> kumfa = x1 is a room of/in structure x2 surrounded by partitions/walls/ceiling/floor x3 (mass/jo'u). 14:55 < durka42> jb:kumfa 14:55 < mensi> kumfa = kumfa — x1(entity) is a room in x2(entity) 14:55 < mensi> :lo kumfa — room. 14:55 < mensi> :lo vi kumfa be lo xotli cu barda — This hotel room is large. 14:55 < mensi> :Related words: bitmu, canlu, zdani 14:55 < Ilmen> co'o ta'o 14:55 < durka42> I dunno kumfa3 seems like it could be useful 14:55 < durka42> bitmu only has places for things it stands between 14:55 < durka42> not things it encoses 14:55 < durka42> encloses* 14:55 < Ilmen> the enclosed part is separated from the outside 14:56 < _mukti_> Ok, I accept that. 14:57 < _mukti_> So in the 10 most frequent gismu in the ralju korpora, that's five places that could be omitted from flashcards. 14:58 < durka42> nice 14:58 < _mukti_> Ok, so here's a different question. 14:58 < _mukti_> en: djica 14:58 < mensi> djica = x1 desires/wants/wishes x2 (event/state) for purpose x3. | If desire is for an object, this is sumti-raising; 14:58 < mensi> use tu'a in x2 (or use lujvo = po'edji). See also taske, xagji, mukti, nitcu, nelci, pacna, prami, rigni, trina, xebni, 14:58 < mensi> xlura. | officialdata 14:59 < durka42> oh no, the djica question :) 14:59 < _mukti_> {djica lo nu} v. {djica lo ka} ... is there a meaningful difference except for the somewhat implied binding of {lo djica} as {ce'u} in the case of {ka} 14:59 < _mukti_> Yep, that one. 15:00 < durka42> this one is sometimes answered with that one comic 15:00 < _mukti_> Is there a difference of meaning, for example, between {mi djica lo ka ce'u limna} and {mi djica lo nu mi limna} ? 15:00 < durka42> I don't think there is any difference 15:01 < durka42> hey, when I google image search for "ma djica lo nu cenba" I get one of my xkcd translations :) 15:01 < _mukti_> ha! 15:02 < _mukti_> So, there's another place that needs an amended definition. Since {ka} is certainly valid, and is not a property. 15:02 < _mukti_> s/property/event/ 15:02 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi.So, there's another place that needs an amended definition. Since {ka} is certainly valid, and is not a event..zoi 15:02 < durka42> anyway the comic is a politician talking to a crowd 15:02 < durka42> {ma djica lo nu cenba} *raucous applause* 15:02 < _mukti_> Right, I know the one you mean. 15:02 < durka42> {ma djica lo ka cenba} *crickets* 15:03 < durka42> ei mi sipna 15:03 < _mukti_> .a'e nai dai 15:03 < _mukti_> .i ko kukte senva 15:03 < durka42> ai mi kukte je senva 15:04 < _mukti_> .u'i 15:28 * nuzba @lojytan_bot: mi ca zvati lo vijytcana / 私は今、空港にいます。 #例文 #ロジバン [http://bit.ly/1gccm4X] 15:40 * nuzba @CapnKronaz: #lojban .i mi masno cilre [http://bit.ly/1I1whKu] 15:42 < sorpaas> What is the difference between {vitke} x2 and {vitke} x3? 15:51 < ldlework> en: vitke 15:51 < mensi> vitke = x1 is a guest/visitor of x2 at place/event x3; x1 visits x2/x3. | See also friti, klama, zasni, xabju, zvati. | 15:51 < mensi> officialdata 15:53 < ctefaho> x3 looks like a bu'u even 15:53 < ctefaho> or a va'o 16:00 < sorpaas> So {lo se vitke} is "host" and {lo te vitke} is a location? 17:27 < shanoxilt> coi ro do 17:53 * nuzba @cizyprijev: .u'i sai https://twitter.com/Saffy/status/624300431787868160 [http://bit.ly/1DFRmIK] 18:17 < _mukti_> coi la'o zoi shanohilt zoi 18:17 < _mukti_> s/shanohilt/shanoxilt 18:17 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi.coi la'o zoi shanoxilt zoi.zoi 18:25 < noncomcinse_> coi 20:39 < _mukti_> jbo: cusku 20:39 < mensi> cusku = x1 benji x2 noi valsi porsi ku'o x3 fu x4 | benji; valsi; porsi; tavla; casnu | 20:39 < mensi> xorxes 20:49 < _mukti_> jbo: zvati 20:49 < mensi> zvati = x1 dunli x2 le ka ce'u fasnu bu'u makau | bu'u | xorxes 22:37 < phma> Does Alberto Fujimori wear a poncho? Or is that ponytco? 23:05 < gleki> {mi kakne lo nu limna} doesnt even say tat we ust assume limna1=ce'u 23:06 < gleki> it can be limna2 as well --- Day changed Sun Jul 26 2015 00:03 < gleki> spagetufa: ge mi gi do mo 00:03 < spagetufa> ([ge {FA mi} gi {FA do}] [CU {mo VAU}]) 00:03 < gleki> spagetufa: ge mi do gi do mi mo 00:03 < spagetufa> ([ge {<FA mi> <FE do>} gi {<FA do> <FE mi>}] [CU {mo VAU}]) 00:17 < gleki> akmnlrse: xu lo ofline versiio be la spagetufa cu zasti 00:25 < akmnlrse> gleki: za'o na 00:26 < akmnlrse> (to .ai mulgau lo uebre versiio ca lo cabdei toi) 00:26 < akmnlrse> .i lo krasi samselpla zo'u ju'oi https://github.com/mezohe/ilmentufa/tree/morfologi 00:30 < gleki> je'e je'e i ta'o mi zei'a jinvi lo du'u PEG mabla 00:30 < gleki> i ku'i pilno ma 00:30 < gleki> i sa'e zo fu'e jai nabmi 00:31 < akmnlrse> je'u nabmi .i ku'i porvanbi funsi gerna 00:31 < akmnlrse> .i pe'i PEG joi lo se jmina proga cu banzuka 00:31 < akmnlrse> si jai banzu 00:32 < gleki> xu do pu'i troci tu'a zo fu'e 00:32 < akmnlrse> na 00:32 < akmnlrse> .i .o'a nai ke ji'a lo gloso za'o na tolspofu 00:33 < akmnlrse> spagetufa: nu 00:33 < spagetufa> (CU [{nu VAU KEI} VAU]) 00:33 < akmnlrse> spagetufa: +gloss nu 00:35 < akmnlrse> zu'u nai lo ba'e formata pe PEGJS ja'a pe'i mabla .i ri'a bilga lo ka so'i roi cusku lo mitsi'u je so'i mei 00:36 < akmnlrse> .i ti'e lo panra pe la since cu kanrymau .i mu'a curmi lo nu pa gerna cu se galfi za'u se jmina proga 00:36 < akmnlrse> vau sa'e nai 00:39 < akmnlrse> yyy .i zo mintu .i zo mintu mo 00:45 < akmnlrse> .i si'a xu ke mu'a lu lo so'i protoni li'u ka'e drani 00:48 < gleki> xu lu protoni be li xo'e li'u 00:49 < akmnlrse> na djuno 00:50 < gleki> ti'e su'e je su'o pa lektoni cu zasti 00:50 < akmnlrse> rarna mupli la iolpre broda 00:51 < akmnlrse> s/pr/p/ 00:51 < fenki> la'oi akmnlrse pu djica lo ka cusku zoi zoi.rarna mupli la iolpe broda.zoi 01:07 < gleki> mi kanpe no ka'e zabna danfu be lo zo fu'e nabmi 01:08 < akmnlrse> ba'a simsa lo me ma'oi zoi moi 01:09 < gleki> lo nu setca me'o vei noi cnino cu mabla i lo gerna tricu se tcana cu cenba 01:10 < akmnlrse> ka'e ku zo fu'e se tcita lo ka ma kau ni zo fu'o pe ce'u cu darno 01:40 < akmnlrse> gleki: http://zbagazasni.ax.lt:8000/camxes-exp.html 01:41 < gleki> uesai 01:44 < gleki> i ma judri lo krasi 01:45 < akmnlrse> https://github.com/mezohe/ilmentufa/tree/morfologi (to ba lo nu kibycpa zo'u minde fi tu'a zoi fa git checkout morfologi fa toi) 02:16 < fushgrhfhxdz> banli doi banlihi 02:24 < fushgrhfhxdz> xu voktaha ca lo prulamdje 03:27 < gleki> spagetufa: ge fami klama gi fe do 03:27 < spagetufa> (CU [ge {<fa mi> <CU (¹klàma VAU¹)>} gi {<fe do> VAU} VAU]) 03:27 < gleki> uanai 03:27 < gleki> spagetufa: ke fa mi ke'e mo 03:27 < spagetufa> SyntaxError: Expected [,] but "m" found. 03:27 < gleki> uanai 03:36 < ctefaho> coị 05:08 * nuzba @frecafloros: 俺、この試験が終わったら…lojbanやるんだ… [http://bit.ly/1fwAkXJ] 05:15 < Ilmen> xu la vreji cu nalgunditcu cliva 05:15 < Ilmen> http://www.lojban.org/resources/irclog/lojban/ 05:16 < Ilmen> za'a dai no da vreji tau la zelmasti pe lo cabna'a 05:17 < Ilmen> .i ji'a so'u sai da vreji tau lo prula'i masti 05:22 < gleki> mi puzi jungau la robin i ka'e lanli lo mi vreji be lo se irci citri 05:43 < Ilmen> je'e 05:43 < gleki> spagetufa: li zo mlatu 05:43 < spagetufa> SyntaxError: Expected [,] but "m" found. 05:43 < gleki> zantufa: li zo mlatu 05:43 < zantufa> minajimpe: .i kanpe tu'a [,] .i ku'i "m" cu se tolcri 05:44 < gleki> xm i nabmi 05:45 < gleki> mensi: doi zipcpi la kurtis cu djica lo nu lu li zo mlatu li'u gendra gi'e sinxa lo pa sumti i ja'o ei lu li cu broda li'u co'u skicu tu'a lo kunti ke namcu ja lerfu porsi 05:45 < mensi> gleki: mi ba benji di'u ba lo nu la'o gy.zipcpi.gy. di'a cusku da 07:35 < gleki> test 07:36 < durka42> no it's broken 07:36 < durka42> can't hear you 07:36 < gleki> test 07:38 < ctefaho> still nothing 07:44 < gleki> u'u 07:45 < durka42> zo'o 07:45 < durka42> cipra ma 07:53 < gleki> na'i i vali po'o 07:53 < gleki> sa 07:53 < gleki> i valsi poho 08:19 * nuzba @fotono: lo #lojban zantufa tricu cu banro milxe i zo iau co'a jufra famyma'o http://guskant.github.io/gerna_cipra/zantufa_farvi.svg [http://bit.ly/1DGUGDf] 08:35 * nuzba @s_i_d_a: lo sefta .e lo nenri .e lo menli cu sanji lo crisa #lojban #haiku [http://bit.ly/1DGWpbK] 08:38 < _mukti_> {troci lo ka co'e} v. {troci lo nu co'e} 08:39 < _mukti_> Is there ever any reason to prefer {troci lo nu}, or is that an artifact of older lojban? 09:00 < _mukti_> I see that {troci lo nu} is used exclusively in .alis., where as .oz. alternates about evenly between {troci lo nu} and {troci lo ka}. 09:01 < _mukti_> So now I'm wondering if there is a difference that people understand between troci + event, and troci + property -- setting aside the obvious fact that the property binds {ce'u} 09:04 < _mukti_> Similarly, .alis. only uses {sutra lo nu}. .oz. also prefers it, but has one instance of {sutra lo ka}. 09:17 < _mukti_> {gleki} is like {sutra}: .alis. uses only {gleki lo nu}, .oz. prefers that but has a few uses of {gleki lo ka} 09:18 < _mukti_> Based on these examples, I'm getting the impression that {lo ka} is being used in .oz. as a specialization of {lo nu} that emphasizes the binding of {ce'u}. But I'm wondering if there are other differences intended. 09:22 < _mukti_> lb: djuno 09:57 < gleki> jb: djuno 09:57 < mensi> djuno = djuno — x1(entity) knows fact x2(proposition) about x3(entity) on grounds x4(proposition) 09:57 < mensi> :lo se djuno — known fact. lo ve djuno — grounds of knowledge. 09:57 < mensi> :mi djuno lo du'u mi zasti vau fo lo du'u mi pensi — I know that I exist since I think. 09:57 < mensi> :mi djuno so'i da la Lojban — I know a lot about Lojban. 09:57 < mensi> :mi djuno mo'a da lo cmaci — I know too little about math. 09:57 < mensi> :Comment: to be familiar with is se slabu, na'e se cnino, na'e se fange. 09:57 < mensi> :Related words: krici, jinvi, cilre, certu, facki, jijnu, jimpe, senpi, smadi, kakne, birti, mipri, morji, saske, 09:57 < mensi> viska 09:57 < gleki> "lb" is Luxembourgish and one visitor here not only spoke it but even tried to teach it. 09:58 < gleki> _mukti_: {troci lonu} is pre-ka Lojban. it has an underformalized feature that {lo nu} doesnt have any ce'u implied so {troci lo nu limna} can mean the same as {troci lo nu se limna} 09:59 < gleki> i mean for them it was fine. 09:59 < gleki> for us it's underformalized 10:00 < gleki> i dont know when ka+ce'u appeared but gimste has hints to it and it was before nitciion's era ended http://mw.lojban.org/papri/ka,_du%27u,_si%27o,_ce%27u,_zo%27e 10:02 < gleki> tiki doesnt have any more info: http://tiki.lojban.org/tiki/tiki-pagehistory.php?page=%27%27ka%27%27+versus+%27%27du%27u%27%27%2C+and+%27%27ce%27u%27%27+and+%27%27zo%27e%27%27 10:02 < gleki> imported in nov 2001 from phpwiki. 10:02 < gleki> phpwiki is lost 10:02 < gleki> at least robin doesnt have it 10:02 < gleki> if you can scrap the wole lojban.org section of archive.org maybe you can find it 10:03 < gleki> *whole 10:04 < Taun> I came in here for one quick question: what's an short, easy way to specify distance traveled? "I walked five miles (long distance units)." Would it be appropriate to say "mi cadzu veka'a le minli be li mu?" 10:04 < gleki> mi cadzu vi'i lo minli be li mu 10:04 < gleki> {veka'a} is also fine 10:05 < Taun> ki'e ki'e 10:09 * nuzba @youxkei: ロジバンでLooks good to meってどう言えば良いんでしょう? #lojban [http://bit.ly/1gdQB4E] 10:19 * nuzba @youxkei: galfiのx1は出来事、cnegauのx1は人、ってことで良いんです・・・? #lojban [http://bit.ly/1I2TM5P] 10:39 < selpahi> _mukti_: ju'i 10:44 < _mukti_> re'i la selpa'i 10:44 < selpahi> I see you're asking about {lo ka}. If you're still reading Oz and looking for errata to add to the spreadsheet, {sutra lo nu} may be added as an erratum. And any other situation where you see two different abstraction types in one sumti place does at least warrant a comment. 10:45 < _mukti_> Yes, still working on errata, though mostly lower concept stuff than that. I'll go ahead and flag those uses though. 10:45 < selpahi> When I started working on Oz I was still using {lo nu} with most predicates, but I made the switch to {ka} before I finished, and then I had to replace everything manually. I missed a lot. 10:46 < _mukti_> Right now, I'm going through the gismu and trying to develop a better set of flash cards than the ones I've been using on Memrise forever. 10:46 < selpahi> Asiz already spotted most of them though 10:46 < selpahi> And I corrected them 10:46 < selpahi> That sounds useful 10:46 < _mukti_> In doing so, I've been looking at some of the "event" places, and trying to reconcile them with what I've seen in contemporary usage. 10:47 < _mukti_> I'll post the results of the first 100 for comments. 10:47 < selpahi> Cool. 10:47 < gleki> first 100 most frequent? 10:47 < _mukti_> Yes, but only looking at ralju korpora. 10:48 < _mukti_> Maybe I should throw the net wider. What do you think? 10:48 < selpahi> Oh you're excluding IRC? IRC is by far the biggest corpus 10:48 < selpahi> it's also the most varied though 10:48 < gleki> what people are missing usually is that they remember gismu as "x1 does x2 to x3 at temperature x4" but when they are asked "how to say temperature?" few of them ar ready to say {lo ve broda} 10:48 < _mukti_> I figure that it's the biggest, but also skewed to discussions of language, etc. 10:48 < gleki> and learning isolating words .... idk ... looks like the worst method of learning languages 10:49 < gleki> even {lo ve broda cu xamgu} is 100 times better 10:49 < _mukti_> That said, visual and speech related gismu seem overrepresented in the korpora ue nai 10:49 < selpahi> mi tugni .i ku'i pe'i na palcau fa lo nu jboske lanli lo gismu 10:50 < selpahi> {cu cusku} is soooo common, it makes you think Lojban is only {c}. 10:50 < gleki> lo since sance 10:51 < _mukti_> Another thing that I've noticed as I've walked through the gismu -- I imagine this must have been commented upon at some time -- is a fairly quiet distinction between gismu with a sense of agency in x1 and those without. 10:51 < _mukti_> {klama} v. {muvdu} for example 10:51 < gleki> la bangu dictionary has examples that are sometimes too difficult for newbies to parse. 10:52 < gleki> i have anther idea but it'd require as much time as i spent for this dictionary 10:52 < _mukti_> {viska} v. {catlu} 10:52 < gleki> catlu limited to vision? 10:53 < _mukti_> My understanding of {catlu} is visual, but then I could well be wrong. 10:53 < selpahi> Lojban has to make certain kinds of distinctions (like agency) at the gismu level, because delegating it to compound words would make the words too long. 10:53 < selpahi> Lojban just doesn't have the luxury that Toaq Dzu has. 10:53 < VladVP> coi rodo 10:54 < gleki> ue coi lo ba'o rusko 10:54 < selpahi> Toaq Dzu's predicates are all way more primitive and minimalistic, and they get combined in all sorts of ways to derive distinctions. 10:54 < selpahi> Which costs only one syllable. But ta'o nai. 10:54 < selpahi> There are some gismu where usage sometimes ignores the agency vs event distinction, like {galfi} 10:55 < selpahi> xorxes seems one of the main proponents of this kind of conflation. 10:55 < selpahi> It's practical, no question. 10:56 < _mukti_> I think I've seen {sutra} used with a sense of agency somewhere, like "hurry" 10:56 < VladVP> la .repak. cu xagrai lo ro ra'apxu 10:56 < VladVP> someone confirm whether that was grammatical 10:56 < gleki> in {catlu} the expression of directiontowards which the perceiver turns their gaze in order to see a phenomenon is obligatory. {zgana} doesnt have such implications. But as for "vision" im not sure such ilimtation should be put onto it. 10:57 < selpahi> la re palku 10:57 < selpahi> Yes, VladVP. 10:57 < VladVP> phew 10:57 < selpahi> You still got it :P 10:57 < VladVP> my Lojban is still in order after 6 months 10:57 < _mukti_> gleki: How would that differ from {jundi} ? 10:58 < _mukti_> (which I also understand as directional attention) 10:58 < _mukti_> I guess I should say "targeted" rather than "directional" -- perhaps that's the distinction. 10:59 < _mukti_> {jundi} specifies a target without presupposing a mediating direction 10:59 < _mukti_> On the other hand, "mediating direction" sounds like a missing slot 11:00 < gleki> _mukti_: {catlu} vs. {jundi}? 11:01 < gleki> anyway i stole that quote from http://www.humnet.unipi.it/linguistica/Docenti/Lenci/papers/johnsonLenci.CF.2011.pdf 11:01 < selpahi> What are your opinions on the concept of "canon" in the context of Lojban non-gismu vocabulary. Obviously, there is no such thing as an official gismu, but maybe a milder level of officialness could be beneficial. 11:02 < selpahi> s/official gismu/official lujvo 11:02 < fenki> la'oi selpahi pu djica lo ka cusku zoi zoi.What are your opinions on the concept of "canon" in the context of Lojban non-gismu vocabulary. Obviously, there is no such thing as an official lujvo, but maybe a milder level of officialness could be beneficial..zoi 11:02 < selpahi> Like that. 11:03 < gleki> o'u 11:03 < gleki> u'i 11:03 < gleki> i tcefrica 11:03 < selpahi> .u'i 11:03 < gleki> pe'i lo jvajvo cu ei zasti 11:04 < _mukti_> I think that the notion of canonical lujvo has never been adequately developed. The failure to produce a dictionary that would have defined canonical lujvo gave birth to BPFK, but BPFK has, until recently, always been consumed with improving cmavo definitions. 11:04 < gleki> develop the sceinc 11:04 < _mukti_> I think the jbovlaste voting system was supposed to be a step in the direction of defining a canon for such words, but we all know the limitations of that. 11:05 < gleki> sorry my connection (or is it vrici) is terrible 11:05 < gleki> develop the science of rafsi slotting 11:06 < _mukti_> Sometimes I wish there were two systems for forming lujvo, such that the resulting lujvo would indicate whether or not it was intended to convey a sense beyond what might be inferred from it's components, ala luvjo. 11:06 < gleki> shut down this terrible esperanto affix system 11:06 < _mukti_> s/luvjo/jvajvo/ 11:06 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi.Sometimes I wish there were two systems for forming lujvo, such that the resulting lujvo would indicate whether or not it was intended to convey a sense beyond what might be inferred from it's components, ala jvajvo..zoi 11:07 < gleki> such system was prposed once 11:07 < _mukti_> Which is to say, for a lojbanist who had some familiarity with rafsi, but perhaps not much lujvo vocabulary, they would be able to know where their vocabularity needed to grow. 11:08 < _mukti_> Failing to understand lujvo was a serious obstacle to my reading for a while. And I'm much better now, but as I read lujvo, I'm largely assuming jvajvo meanings, and not aware of lujvo that have actually been narrowly defined. 11:09 < _mukti_> Perhaps "jvajvo" is too narrow here: I mean I assume the same kind of meanings that could have been derived from analogous tanru, etc. 11:09 < gleki> google rafybri. but since this proposal was ignored it wasnt further developed 11:10 < _mukti_> gleki: When was that? (googling now) 11:11 < gleki> now i can only see that the power of {lo TAG broda} and {i ko'a TAG broda} instead of just {ko'a cu broda} has been underestimated. 11:11 < _mukti_> The Great Rafsi Reallocation in some ways aggravated this problem, since it made lujvo unambiguous relative to their rafsi. 11:11 < gleki> but it made the language much harder to learn 11:11 < _mukti_> The reallocation did? 11:11 < gleki> #zgaga'aza'azganzgana 11:12 < durka42> coicoi 11:12 < _mukti_> You mean by encouraging nonce lujvo? 11:12 < _mukti_> coi la durkavore 11:12 < durka42> before the GRR, lujvo weren't 1-to-1? 11:13 < durka42> rafsi* 11:13 < _mukti_> Oh, I think I do have the history wrong, durka. 11:13 < _mukti_> I guess before GRR there were assignments, which were changed. 11:13 < _mukti_> Do I incorrectly remember that there was a time before assignments? 11:14 < durka42> ba'anairu'e in TLI Loglan times they just kinda made up lujvo and sometimes even gave them gismu shapes 11:14 < gleki> loglan version 1 i.e. in SciAm there were no rafsi 11:14 < selpahi> I don't think there was a time before assignments. 11:14 < gleki> there were only jbofuvi 11:14 < selpahi> They changed them to give the more useful rafsi to the more common gismu. 11:15 < selpahi> They kept a few where the userbase was already too used to the old assignment. 11:16 < gleki> #zgaga'aza'azganzgana problem is explained here mukti: https://groups.google.com/d/topic/lojban/6H7JgQOJTzk/discussion 11:16 < _mukti_> ua 11:18 < gleki> also i consider -gau lujvo shifting places to the right as something unnaturalistic. as for {jungau} i perceive it just like a new fu'ivla since this is how known natlangs work (consider english "to inform") 11:18 < _mukti_> Yes, I'm looking at JL1. There's a list of "primitives" and their "affixes", along with discussion of how to create "complexes" with them. 11:18 < _mukti_> I think part of that discussion may touch on the origin of noralujv 11:18 < durka42> I don't think -gau is unnatural 11:18 < durka42> A does B 11:19 < gleki> i consider bramlatu/cmabroda as something naturalistic but wouldn't make them jvajvo. 11:19 < durka42> A causes B to do C 11:19 < durka42> ^ shifted places 11:19 < durka42> but part of the goal of lojban is to be unnatural isn't it :p 11:19 < gleki> that's not exactly how lojban works 11:19 < gleki> {jai gau fai} is naturalistic to me though 11:20 < gleki> although superlong but that's another issue :D 11:20 < _mukti_> Wow, a lot of jboci'a in that thread. 11:20 < selpahi> durka42: The one difference there is that "to do" is a special FA that let's you avoid counting. 11:21 < durka42> yeah well we don't normally count, fair point :) 11:22 < durka42> ta'osai I was wondering today if there is a "null BAI" 11:22 < selpahi> Learning {jungau} as a standalone word is much easier (though you can note to yourself that it follows the ctuca-pattern) 11:22 < durka42> for example if I asked {do klama fi'a la .bastn.} and the answer should be "no such FA/BAI would make that true" 11:22 < selpahi> Thing is, you don't need to even know {gasnu} to learn {jungau}. 11:22 < durka42> no bu'a zo'u mi klama fi'o bu'a la .bastn. 11:23 < durka42> (ta'o toi'e ta'o is that how {bu'a} works‽) 11:23 < gleki> zo'o fa xi tu'o 11:23 < _mukti_> I agree that learning lujvo is a good idea, and not any more difficult than learning other vocabulary. I think it's the lack of a canon which is the stumbling point, and perhaps what selpahi was at earlier. 11:23 < durka42> gleki: fa xi mo'e noda, I guess... 11:23 < gleki> the difference is that natlangs dont proliferate lujvo that much. they have rules 11:23 < durka42> camxes: faximo'enoda 11:23 < camxes> SyntaxError: Expected [yY] but "x" found. 11:23 < durka42> camxes: +exp faximo'enoda 11:23 < camxes> ([{fa <xi (¹mo'e [{no BOI} da] TEhU¹) BOI>} KU] VAU) 11:23 < gleki> exp: faimo'eda 11:23 < gleki> exp: fxaimo'eda 11:23 < mensi> SyntaxError: Expected [0123456789] but "d" found. 11:23 < mensi> SyntaxError: Expected [.\t\n\r?! ] but "f" found. 11:23 < gleki> darn 11:23 < gleki> ju'inai 11:24 < durka42> german proliferates lujvo a lot :) 11:24 < _mukti_> I'm reflecting now on my childhood experience as I first began to recognize latin and greek particles in words. I didn't assume that I didn't need to learn words and could infer them from the particles. 11:24 * gleki to'o bajra gi'e curmi lo nu lo dotco cu spuda 11:24 < selpahi> I think German is the same as English, except German's lujvo and tanru both look like lujvo, whereas in English both look like tanru. 11:25 < _mukti_> fireman 11:25 < gleki> okay, mukti noticed that -ske is a candidate for jvajvo generator 11:25 < durka42> you mean german often uses compound words where english would use a noun phrase 11:25 < durka42> ? 11:25 < selpahi> I mean the insertion of spaces 11:25 < gleki> but then limit -ske lujvo to two places 11:25 < durka42> yeah 11:25 < durka42> I have a small dictionary and I am often guessing word boundaries :) 11:26 < gleki> for fireman it's -pre usually unless you again limit yourself to two places 11:26 < gleki> btw who will finally write Link Grammar for Lojban? paunai 11:26 < durka42> but today I said a full sentence in German and was sort of understood! 11:26 < durka42> with selpahi's help 11:26 < durka42> four whole words :p 11:26 < gleki> xu lo re do pu'i penmi zu'ai 11:26 < _mukti_> "fireman" though is a good example of a non-javni compound... since it could just as easily mean "arsonist" 11:26 < selpahi> Something about beer? 11:27 < durka42> well similar 11:27 < durka42> mi cusku zoi dy Ich essen keine Brot dy 11:27 < selpahi> ma na'o basti lo nanba lo ka do citka 11:28 < durka42> mi vitke lo .itlo gusta .i dunda lo nanba me'oi Vorspeisen .i mi cusku ri .i lo nanba cu canci jecu se basti lo salta 11:28 < durka42> y 11:29 < durka42> lo rismi mu'a 11:29 < selpahi> ua .i xu lo cersai ta'e se pagbu lo rismi 11:29 < durka42> ja lo nanba be na'e bo lo maxri 11:30 < selpahi> je'e 11:30 < durka42> .u'i .i ta'e cersai lo djoguro jo'u lo grute 11:30 < selpahi> je'e 11:31 < gleki> en: basti 11:31 < mensi> basti = x1 replaces/substitutes for/instead of x2 in circumstance x3; x1 is a replacement/substitute. | See also cmavo 11:31 < mensi> list ba'i, binra. | officialdata 11:31 < gleki> li'a nabmi la vrici 11:31 < _mukti_> doi la durkavore do'u .i xu do citka lo bavmi 11:31 < durka42> lo kukfusra plise xalkai cu basti lo birje .i ku'i na basti .u'i 11:32 < gleki> i ja'o sei zabna mi deha jundi 11:32 < durka42> lo bavmi na ku 11:32 < _mukti_> ua 11:32 < selpahi> do ma jinvi lo dotco birje noi ka'u banlyrai 11:32 < durka42> lo xrilanbi sa'e 11:33 < durka42> ti'esai banlyrai 11:33 < durka42> xu da dotco ke plise co'e 11:33 < selpahi> mi djuno no da lo birje ja lo simsa .i mi na pinxe 11:34 < _mukti_> ue sai .i mi sruma lo du'u do taske doi la selpa'i zo'o 11:35 < selpahi> taske lo djacu 11:35 < _mukti_> je'e .u'o 11:35 < _mukti_> s/u'o/o'u/ 11:35 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi.je'e .o'u.zoi 11:36 < durka42> pinxe xu lo drata xalka 11:36 < selpahi> lo xalka zo'u mi pilno ri ji'i no drata be lo ka lumci 11:36 < selpahi> .i na go'e 11:37 < _mukti_> .i xu do ta'e pinxe lo plise xalka doi la durkavore 11:37 < durka42> je'e selpa'i 11:37 < durka42> go'i doi mukti 11:37 < selpahi> ma te frica lo plise xalka lo plise vanju 11:38 < _mukti_> mi ji'a go'i 11:38 < _mukti_> la'a mi pu smuni lo simsa 11:38 < _mukti_> s/smuni/skudji 11:38 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi.la'a mi pu skudji lo simsa.zoi 11:38 < durka42> y lo plise vanju cu mo 11:39 < selpahi> vanju lo plise 11:39 < durka42> lo kukfusra plise xalka cu simsa lo birje 11:39 < _mukti_> lo vanju ja birje be lo plise 11:40 < durka42> lo vanju ta'e na fonmo 11:40 < _mukti_> mi na tugni 11:40 < _mukti_> .i ku'i mi ta'e pinxe la'o zoi prosecco zoi 11:40 < selpahi> ie pei doi la mensi 11:40 < mensi> ba'e mi na tugni 11:41 < durka42> ie lo proseko je lo carmi pagne cu fonmo 11:41 < _mukti_> .u'i 11:41 < _mukti_> ie mi ta'e pagne gi'u carmi 11:44 < durka42> ku'i ma valsi lo se skicu be mi 11:45 < durka42> lo poi'i xalka se pinxe jecu se zbasu fi lo kukfusra te marxa be lo grute (to mu'a lo plise ja lo perli toi) 11:47 < _mukti_> za'a zo'oi champagne cu se vlaka zo'oi campania noi latmo gi'e se smuni lo du'u foldi 11:47 < durka42> smuni lo stuzi xunai 11:49 < _mukti_> What is a smuni anyway. Is it a du'u? 11:49 < _mukti_> I often stumble with du'u and se du'u 11:50 < durka42> I guess it's a du'u 11:50 < durka42> lo sedu'u co'e refers to text 11:50 < _mukti_> That's what I was thinking ... thanks 12:07 < akmnlrse> gleki: za'o na cazgau tu'a lo zo ke termsete .i kanpe lo du'u su'o xagmau cu ka'e danfu 12:07 < akmnlrse> selpahi: na snada lo ka voksnu .i la .ilmen. ku po'o kansa fa noi na zifre lo ka bacru 12:07 < Ilmen> coi mi'e di'a jundi 12:07 < durka42> coi dicra jundi 12:08 < akmnlrse> durka42: la'a mi se cfipu sai tu'a lo nu do ca ma kau ma kau zvati 12:08 < Ilmen> za'a ca lo nu tavla fi mi cu tolcanci fa mi .i mapti nu cabna 12:08 < akmnlrse> .i do cliva ca lo 28 moi vau xu 12:08 < durka42> mi fuzme .i mi mabla platu 12:09 < durka42> mi klama la .ctutkart. lo ca se zvati ca lo 28moi 12:09 < akmnlrse> mi si'a go'i 12:09 < akmnlrse> si go'e 12:09 < akmnlrse> .i xu voikla fi ba'e C 12:09 < durka42> C naje F 12:10 < durka42> ca lo 29moi 12:10 < akmnlrse> ua ru'e 12:10 < akmnlrse> .i na nabmi fa lo nu mi zvati ca lo 28 moi 12:10 < akmnlrse> .i lo trene canlu na tai culno 12:11 < selpahi> doi la .akme ma ralju lo nu mukti be lo nu do benji fi mi fe lo vreji .i xu do kanpe lo nu srana lo zgike 12:11 < selpahi> .i va'o ro da mi gleki lo nu viska la .rnuomer. 12:12 < akmnlrse> na kanpe .i sa'u lo se .irci cu mutce co zenba lo ka tolcando .i je ca lo bi'u nai djedi la sonja cu vitke 12:12 < akmnlrse> soi na sai cafne 12:12 < Ilmen> ŭa .rnu'omer. do'e ma 12:12 < akmnlrse> lo tokpona se .irci 12:13 < selpahi> je'e .i la sonja cu se slabu mi .i mi pu so'o tavla ri 12:13 < Ilmen> ĭe pu za lo djedi be li so'o mi ganse lo nu la tokpona finti cu co'a jundi la .lojban. gi'e cusku zo coi 12:14 < selpahi> s/so'o/so'o roi 12:14 < fenki> la'oi selpahi pu djica lo ka cusku zoi zoi.je'e .i la sonja cu se slabu mi .i mi pu so'o roi tavla ri.zoi 12:14 < Ilmen> coi ro do 12:14 < selpahi> coi 12:14 < Ilmen> .i ma nuzba 12:14 < selpahi> mi ze'a pu bilma 12:14 < durka42> xu mo'u bilma 12:15 < selpahi> na go'i 12:15 < Ilmen> .oi dai .i xu vidru ja jurme bilma 12:15 < selpahi> .i ku'i simlu lo ka zenba lo ka kanro 12:15 < selpahi> .i mi na djuno 12:15 < selpahi> .i la'a .influ'enza 12:15 < Ilmen> .i ku'i lo nu bilma je ba bo cu di'a kanro cu selja'e lo nu zenba co .i'onmo lo nu kanro 12:16 < selpahi> go'i 12:17 < selpahi> .i sa'u doi la akmnlrse mi pu no'e jimpe lo du'u ma kau mukti lo nu benji 12:17 < selpahi> .i ku'i .i'onmo 12:17 < Ilmen> .oi dai .i ju'o cu'i mi no roi bilma la .influ'enza .i la'a lo nu mi darno lo tcadu cu krinu .u'i .i ti'e mabla se bilma 12:17 < selpahi> xu go'i 12:18 < selpahi> .i lo me mi moi na tsali klesi lo .influ'enza .i mu'a na vamtu 12:18 < Ilmen> je'e 12:19 < selpahi> mi xebni lo nu lo nu lo nazbi tubnu cu culno cu rinka lo nu na'e prane tirna 12:21 < selpahi> en:traffic light 12:21 < mensi> 2 da se tolcri: fletrogusmi'i, jezystisni 12:21 < selpahi> .u'i 12:24 < Ilmen> ba'a nai ru'e do pu xusra lodu ba lo nu mo'u bilma cu la'a djucpe fi loka finti su cfika brivo 12:25 < selpahi> go'i .i je'u mi pu zi pensi lo mintu 12:47 < selpahi> xu lo kanvasa poi te ciska no da cu kunti jo'a pei 12:47 < selpahi> .i ma drani valsi 12:47 < durka42> la'a kunti 12:47 < selpahi> .i zoi gy. blank .gy 12:48 < durka42> nu'o co'e 12:48 < selpahi> .i lo kanvasa na vasru .i na sruri lo cinta 12:49 < durka42> relcimde kunti 12:50 < selpahi> .i ku'i .ambigu .i lu lo va kanvasa cu kunti li'u .i se smuni lo du'u no da nenri kei ji lo du'u no da cpana lo sefta 12:50 < selpahi> .i mi na se mansa 12:50 < durka42> ku'i na'i da nenri lo kanvasa 12:51 < Ilmen> pircau 12:52 < Ilmen> nu'o te ciska 12:52 < selpahi> mablyme'a .i mo lo pelji boxfo poi te ciska no da 12:53 < selpahi> .i ma kampu lo ro srana 12:53 < Ilmen> no da su'o de ciska ko'a 12:53 < durka42> lo pelji boxfo cu simsa lo kanvasa lo ka ka'e relcimde kunti pe'i 12:53 < Ilmen> .i me'ei ciska cu kaipti tau lo pixra 12:54 < Ilmen> ba'a nai 12:55 < selpahi> ma melbi brivla 12:55 < durka42> ba'arcau 12:56 < Ilmen> la'a zo nonci'a na'e mapti .i lo lujvo pe zo'oi non ta'e co'e lu broda be no da li'u 12:56 < Ilmen> nonteci? 12:56 < selpahi> mi pu pensi zo barna .i ku'i mi co'a sanji lo nu lo ba'o te ciska pelji ku ji'a ka'e ba'arcau da'i 12:56 < Ilmen> te ciska be no da? 12:57 < durka42> ka'e ba'arcau xu 12:57 < Ilmen> zo skamanfo ji'a ka'e co'e 12:58 < selpahi> "This page of writing is full of stains. Do you have a copy that's stain free?" 12:58 < Ilmen> .i ku'i pe'i zo nonteci na mabla .i ma ni fegli fi do'o 12:58 < selpahi> li du'e ui nai 13:00 < durka42> ua 13:00 < selpahi> .i zo skamanfo zo'u lo boxfo poi morna simsa lo zebra je ku'i nu'o te ciska cu zasti 13:00 < Ilmen> ju'o 13:00 < durka42> mi djica lo boxfo poi barna lo xinmo jenai lo ckafi 13:03 < selpahi> malfu'a fa lo nu zo nu'o ji'a na prane mapti .i ka'e tcica .i mu'a lo rokpisterci'a poi ba'o te ciska ka'e se lumci ja'e lo nu di'a kunti jo'a cu'i 13:03 < selpahi> .i ja'o na nu'o te ciska 13:03 < selpahi> .i ku'i kunti jo'a cu'i 13:03 < Ilmen> ci'anti .i ciska kunti 13:04 < Ilmen> .u'i 13:04 < selpahi> mablyme'a ro lo prula'i 13:04 < durka42> then you need ci'anxra as well 13:04 < Ilmen> ŭa nai 13:04 < Ilmen> ka'e ciska lo pixra ke'u 13:05 < durka42> je'e 13:05 < Ilmen> .i zo ciska na srana fe po'o lo lerpoi 13:05 < Ilmen> ba'a nai 13:05 < Ilmen> en: ciska 13:05 < mensi> ciska = x1 inscribes/writes x2 on display/storage medium x3 with writing implement x4; x1 is a scribe. | Also x3 writing 13:05 < mensi> surface. See also papri, penbi, pinsi, tcidu, xatra, pixra, prina, finti for 'author' or specific authorial works, 13:05 < mensi> barna, pinka. | officialdata 13:06 < Ilmen> xmm do'e ma doi mi mi xusyga'e lo nu su da xusra lodu zo ciska mapti lo si'o xraci'a 13:06 < selpahi> la'a lo lojbo ve skicu pe la .xorxes. 13:06 < durka42> jbo:ciska 13:06 < durka42> la mensi ba spuda ca lo bavlamdei 13:06 < mensi> ciska = x1 gasnu lo nu x2 noi na'o se smuni ja sinxa cu barna ja ba'armo'a x3 13:07 < Ilmen> en: xraci'a 13:07 < mensi> xraci'a [< pixra ciska ≈ Picture write] = c1=p3 draws picture c2=p1 on medium c3=p4 with implement c4 depicting p2. | Cf. 13:07 < mensi> pixra, ciska, xraselci'a, xratai, kacmyxra, kacmyterxra, kacmyxragau, kacmyxrazu'e. | 13:07 < mensi> totus 13:07 < durka42> uasai zo barna 13:09 < Ilmen> pe'i zo ciska joi zo kunti joi zo no joi zo claxu cu kaipti lo ka te jupyzba 13:10 < Ilmen> .i ku'i na'e frili fa lo ka jupyzba lo mletce 13:10 < selpahi> «lu [...] pa'a lo ci'anti pergamena li'u» 13:15 < selpahi> xu mansa .i la'a mansa mi 13:16 < Ilmen> pe'i go'i mi 13:16 < selpahi> .i mi ji do ba jmina zo ci'anti la jbovlaste 13:16 < Ilmen> mi zifre lo ka jmina 13:16 < Ilmen> .i .e'e pei 13:16 < selpahi> go'i ra'o 13:16 < Ilmen> .i ce'u te ciska no da no de 13:16 < Ilmen> si 13:16 < Ilmen> si 13:17 < Ilmen> su'o de 13:17 < selpahi> .i sa'u lo jmina cu .y crepu lo .y ka lo drata cu manci lo banli se finti be ce'u 13:17 < selpahi> :P 13:17 < Ilmen> .u'i 13:17 < selpahi> .i jalge fa lo nu ro da djica lo nu da jmina 13:18 < Ilmen> mi jmina .ai 13:19 < Ilmen> .i xu su da drata brivo je cu jai se djica 13:19 < selpahi> .i ie pei na banzu fa tau lu te ciska no da su'o de li'u .i xu lo vacri je lo xamsi je lo sidbo cu ci'anti ni'i lo du'u no da su'o de ciska 13:19 < mensi> mi tugni i ie mi tugni 13:19 < selpahi> .i .ei lo ci'anti cu se sefta su'o da 13:20 < selpahi> $x_1$ noi ke'a se sefta cu li'o 13:20 < selpahi> .i ja'a ku so'o drata brivla cu jai se djica 13:21 < Ilmen> lu $x_1$ ka'e te ciska su'o da su'o de gi'e ku'i ca'a te ciska na ku su'o da su'o de li'u 13:22 < selpahi> .i'a ru'e .i xu na ku lo ka ciska cu ka punji .i ku'i lo si'o ci'anti cu srana lo ba'e jalge be lo nu ciska 13:23 < selpahi> .i .e'u jmina su pu zei cmavo ja su za'o zei cmavo 13:23 < Ilmen> la'a lu ka'e fa'u ca'a ja'a fa'u na te ciska li'u tordu zmadu 13:23 < Ilmen> .u'i 13:23 < selpahi> .u'i 13:24 < selpahi> mi zo'u zo zi'o simsa sai lu su'o da li'u pe lo do smuvelski 13:25 < selpahi> .i na vajni .i ta'i ma cikre lo do smuvelcki .i zo pu na jai banzu 13:25 < Ilmen> exp: ka'e je na broda 13:25 < mensi> SyntaxError: Expected [,] but "b" found. 13:25 < Ilmen> exp: ka'e ku je na ku 13:25 < mensi> ([{ka'e ku} {je <na ku>}] VAU) 13:25 < Ilmen> mabla 13:27 < selpahi> .i xu zo viska pagbu .i ka'e te ciska gi'e ca'o se cpana no cinta ja co'e poi se viska 13:28 < selpahi> .i .u'i zanfu'a mi fa lo nu do kau je nai mi jmina zo ci'anti 13:30 < Ilmen> mo'u jmina lo lojbo velski 13:31 < selpahi> xu do jimpe lo du'u mi ma kau pante 13:31 < Ilmen> .i xŋŋ mi kucli .i xu lo ve pixra cu ve vreji 13:31 < Ilmen> zo nonvei na na ka'e mapti ju'o cu'i 13:32 < selpahi> la'a .i ju'o su'o ve pixra cu ve vreji 13:35 < _mukti_> The definition and usage of pensi seem to be at odds. 13:35 < _mukti_> {pensi lo du'u} is frequent 13:36 < Ilmen> sa'u va'o lo nu su'o da ciska fo lo na ka'e se viska to mu'a blabi ja skacau xinmo toi zo'u lo dacti cu za'o simlu lo ka na te ciska .i simsa lo nu lo prenu cu viska lo rutni spati gi'e xusra lodu lo spati cu zvati kei fau lo nu fliba loka ganse lonu rutni 13:36 < selpahi> Lots of malgli usage of pensi 13:36 < _mukti_> en: pensi 13:36 < mensi> pensi = x1 thinks/considers/cogitates/reasons/is pensive about/reflects upon subject/concept x2. | Also: x1 is 13:36 < mensi> thoughtful (one sense); x2 is mental (one sense)/intellectual (one sense) (= selpei). See also cmavo list pe'i, jijnu, 13:36 < mensi> menli, morji, sidbo, jinvi, se nabmi, minra, lanli, besna, saske, skami. | officialdata 13:36 < selpahi> je'e la .ilmen. 13:36 < _mukti_> If by malgli you mean "think about", it seems to be encouraged by the official definition. 13:37 < _mukti_> Intuitively, I expect {pensi} to have a similar structure to {jinvi}, at least the first three places 13:37 < _mukti_> But instead, it seems to fuse jinvi2 and jinvi3 into pensi2 13:37 < selpahi> .i ku'i xu do jimpe fi lo drata se xusra be mi .i lo na ci'anti boxfo na bi'ai ca'o te ciska .i ku'i lo do smuvelcki cu curmi lo nu ci'anti 13:37 < selpahi> ua pei 13:38 < selpahi> No, the opposite is what I mean. 13:38 < selpahi> It means "think about X", not "think that X" 13:38 < selpahi> They often use {pensi} when they mean {jinvi} 13:38 < _mukti_> ua 13:39 < _mukti_> I just noticed jinvi also has the gloss "think" 13:42 < Ilmen> ŭa nai go'oi curmi 13:43 < selpahi> lo do smuvelcki zo'u lo nu mi skicu lo na ci'anti boxfo lo ka ci'anti cu drani 13:43 < selpahi> fo lo ka * 13:43 < Ilmen> ŭa 13:44 < selpahi> ki'u lo nu na ba'e ca'o te ciska 13:44 < selpahi> .i ku'i pu te ciska 13:49 < Ilmen> zo ximcau ji'a cinri .i ku'i lo nu curmi lo nu lo linji ja gredile ja mo kau cu zvati cu selja'e lo nu nandu .i da'i ka'e terbri panra zo kunti joi zo jinsa 13:51 < Ilmen> kibro nabmi lifri .i xu lo selsku be mi be'o poi zo nandu pagbu cu mo'u se benji 13:51 < selpahi> go'i 13:51 < Ilmen> si cpacu 13:51 < _mukti_> How do {se xamgu} and {se prali} differ ? 13:51 < _mukti_> Both have a sense of "benefits" 13:53 < _mukti_> xamgu1 can be an event. prali3 must be. 13:53 < _mukti_> and prali1 has no analogue in xamgu 13:53 < _mukti_> oh, and xamgu3 is a "by standard" place, so I'm pretending it's not there 13:54 < Ilmen> X prali Y fi Z is similar to {lo nu Y ckaji X cu se xamgu Z} 13:54 < Ilmen> si'au 13:54 < _mukti_> Really? 13:55 < Ilmen> jbo: prali 13:55 < mensi> prali = x1 xamgu x2 gi'e jalge x3 | xamgu; jalge | xorxes 13:55 < Ilmen> .u'i 13:55 < _mukti_> lo rupnu be li panono cu prali mi lo nu mi gunka 13:56 < Ilmen> loka cpacu lo rupnu 13:56 < Ilmen> ba'a 13:56 < _mukti_> lo nu mi ckaji lo rupnu be li panono cu se xamgu lo nu mi gunka 13:56 < Ilmen> .i lo ka jerna 13:57 < _mukti_> Hmm... Maybe if prali1 were an event ... but the current definition doesn't read that way to me 13:58 < _mukti_> And the lojban definition allows it, since xamgu1 is weird in admitting both events and non-events 13:59 < _mukti_> So keeping with the idea of an event in prali1... and converting it to a property ... lo nu mi ckaji lo ka cpacu lo rupnu be li panono cu se xamgu lo nu mi gunka ... 14:00 < _mukti_> Now we've got surprising (to me at least) semantics for xamgu 14:00 < _mukti_> The idea that the event is the beneficiary? 14:00 < Ilmen> doi la selpa'i .i ma traji loka do nelci ku lo nu da'i zo ci'anti cu simsa zo ximcau to no skari morna cu pagbu .i se ni'i bo no linji ja gredile ja mo kau cu co'e toi kei jo'u lo nu zo ci'anti cu terbri simsa zo kunti .e zo jinsa to mu'a ci'anti lo lerfu morna .enai lo linji toi 14:01 < Ilmen> .i ji'a lo midju pluta cu nu tersumei za'u da 14:01 < _mukti_> Although I think Augustine would assent to the idea that causation is good. :) 14:01 < _mukti_> s/Augustine/Aquinas/ 14:01 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi.Although I think Aquinas would assent to the idea that causation is good. :).zoi 14:02 < _mukti_> Confounding my Doctors. 14:02 < selpahi> doi la .ilmen. .i na birti .i .e'u ma'a facki ca lo nu pilno 14:03 < Ilmen> je'e .i'a 14:05 < Ilmen> lo nu do gunka cu selja'e lo nu do rupnu jerna noi xamgu do 14:05 < Ilmen> doi la mukti 14:09 < Ilmen> ni'o .ai mi spuda la .guskant. tau lo prenexo normale tarmi 14:35 < Ilmen> .i mulfau 14:37 < selpahi> mi viska 14:41 < Ilmen> .a'o mansa la .guskant. 14:43 < Ilmen> za'a mi fliba loka jmina lo ro moi toi zei lerfu kei soi nalvai 14:43 < selpahi> .u'i 14:57 < selpahi> ca'o ku pa pa valsi cu pagbu lo liste 14:59 * Ilmen ca'o tinju'i lo .albumza'o 15:00 * Ilmen zo'u la'a la'au ko cikna cu na'e ke tcika mapti .i ku'i za'a no cmima lu ko sipna li'u se cmene 15:00 < selpahi> .u'i 15:03 < selpahi> la'a ro mi'o ba zi co'a sipna .i ku'i mi ka'e visygau do lo se liste 15:04 < tsani> .i .oi .u'i JVS mrobi'o ba lo nu mi co'a se jaspu ku zi 15:05 < tsani> .i coi rodo 15:07 < selpahi> coi 15:07 < Ilmen> je'e .i .a'u mi liste tcidu 15:07 < Ilmen> da'i 15:08 < Ilmen> coi zdani be lo dilnu 15:08 < Ilmen> jo'u lo tarci 15:10 < selpahi> do ka'e pensi lo cmima be lo liste pu lo nu mi'o di'a casnu .i .a'o ca lo bavlamdei mi mo'u di'a kanro 15:10 < selpahi> .i liste fa di'e 15:10 < selpahi> crossbow threaten to go a few steps further (literally) thorn (e.g. of a plant) goods corrupt (doing dishonest, illegal, or immoral things to gain money or power) log (of wood) ember ban (e.g. a person from a place) apparently strain/burden ("x1 is a burden on x2") 15:14 < Ilmen> je'e 15:16 < Ilmen> zo'oi apparently zo'u ŭa nai ru'e zo si'au joi lu soi faumlu li'u mo .i zo'oi thorn zo'u .o'o xu za'o na pagbu lo jbobau vau .u'i 15:17 < Ilmen> .i .e'u zo spajesni ja zo spajesi 15:17 < selpahi> lo pa moi zo'u ra'u ru'e mi djica lo nu na ku ro srana be zoi gy. apparently .gy se fanva fu zo si'au ja zo simlu po'o .i ku'i ba casnu ca lo drata 15:18 < Ilmen> .i je'e .i lu soi se ganse li'u ji'a ka'e 15:18 < selpahi> ie lo simsa 15:18 < Ilmen> .i sa'u naku ro se ganse cu jetnu 15:18 < Ilmen> ja fasnu 15:20 < Ilmen> zo'oi goods zo'u mi da'i stidi zo xamgu si zo selpo'e 15:20 < Ilmen> .i ku'i ĭe ca lo mapti di'a 15:21 < selpahi> .i'o 15:21 < Ilmen> .i .ai mi ba zi co'u cikna .i .ai mi ba zi co'u cikna 15:21 < selpahi> .u'i 15:21 < selpahi> go'i ra'o 15:21 < selpahi> di'ai 15:22 < Ilmen> co'o :) 19:55 < tsani> .i zukte jo nai zukte .i no da ka troci 19:55 < _mukti_> coi la tsani 19:55 < tsani> coi 19:56 < tsani> .i zoi gy do or do not gy jai nandu co se fanva 19:56 < tsani> .i li'a lu zukte jo zukte li'u smudu'i .i ku'i ... no'e mansa 19:56 < tsani> ni'o do mo doi la mukti .i ma nuzba 19:57 < _mukti_> mi co'a jinvi lo du'u tcika fa lo ny bastygau la jbovlaste 19:57 < _mukti_> .i ba'a nai do pu co'a gunka tu'a lo basti 19:58 < _mukti_> (bastygau fi la jbovlaste) 19:59 < _mukti_> .i xu do co'a gunka 19:59 < _mukti_> s/co'a/ca'o 19:59 < fenki> la'oi _mukti_ pu djica lo ka cusku zoi zoi..i xu do ca'o gunka.zoi 20:03 < tsani> .i je'u mi na co'e .i mi pu troci .i ku'i .u'u binxo co zukcfu dukse 20:04 < _mukti_> je'e .i nandu 20:05 < tsani> .i ma ne lo nu basygau ku do se plate 20:05 < tsani> si platu 20:05 < _mukti_> .i mi co'a ru'e junri platu 20:05 < tsani> .i'e 20:06 < tsani> .i xu ba finti lo je'u cnino .iji favgau la vlasisku te zu'e lo nu kakne dunli JVS 20:07 < _mukti_> .i mi jinvi lo du'u pa mai zbasu zoi zoi API zoi 20:08 < _mukti_> .i ba bo cumki fa lo nu pilno zoi zoi interface zoi pe la vlasisku 20:10 < _mukti_> .i ji'a bo zbasu lo mulno cnino zoi zoi interface zoi poi ke'a se pilno fi lo nu finti jonai jai galfi lo valsi .e lo smuni 20:26 * nuzba @djeikyb: .@bgcarlisle I can imagine myself telling people that the #lojban word for "jet engine" literally means "fart propulsion device". [http://bit.ly/1HTI2F2] 22:30 <@rlpowell> Breaking things for a bit, sorry. 23:47 <@rlpowell> And breaking things again. 23:51 <@rlpowell> Fixed. --- Day changed Mon Jul 27 2015 00:32 * nuzba @yutopp: lojbanの単語catraしか知らない [http://bit.ly/1KsZOzR] 01:04 < zipcpi> gleki: One thing in common with the words you jokingly suggested the other day, which I made Eulojban brivla, is that they all describe sensations/qualia, which is actually a category of words that Lojban might have trouble expressing 01:05 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: la kurtis cu djica lo nu lu li zo mlatu li'u gendra gi'e sinxa lo pa sumti i ja'o ei lu 01:05 < mensi> li cu broda li'u co'u skicu tu'a lo kunti ke namcu ja lerfu porsi | 2015-07-26T12:45:28. 01:05 < mensi> 384Z 01:05 < mensi> zipcpi: cu'u la'o gy.gleki.gy.: la guskant na djica lo nu zo xo'i cmavo zo go'oi iju'oi https://twitter.com/oka_iu_tcan/ 01:05 < mensi> status/614416499105624064 | 2015-07-27T06:54:53.791Z 01:08 < zipcpi> As for Guskant not wanting xo'i in GOhOI... hm. I guess she does want to enforce the "only sumtcita" rule. That is also a valid position. I was only suggesting GOhOI for those either too lazy to program more selma'o, or who wish to not complicate grammar any further than it needs to be :p 01:11 < zipcpi> Yes, I've seen Curtis' suggestion for allowing strings in LI...LOhO. I'm not sure it's needed though... {mo'e} exists, and if we allowed VUhU-joi'i to be used like JOI (or even just moved it to JOI) it'd work great 01:12 < zipcpi> exp: zo gerku joi'i zo zdani 01:12 < mensi> ([{zo gerku} {joi'i <zo zdani>}] VAU) 01:12 < zipcpi> off: zo gerku vu'u zo zdani 01:12 < mensi> SyntaxError: Expected [,] but "z" found. 01:14 * nuzba @yutopp: lojbanで文鳥はなんていうのかな [http://bit.ly/1I47aXu] 01:22 < zipcpi> Heck that's probably how I'm gonna use joi'i the most 01:29 < zipcpi> Guskant also did make RAhOI its own selma'o 01:29 < zipcpi> zantufa: ra'oi tce 01:29 < zantufa> ([ra'oi tce] VAU) 01:29 < zipcpi> zantufa ra'oi bu'a'a'a 01:29 < zipcpi> zantufa: ra'oi bu'a'a'a 01:29 < zantufa> minajimpe: .i kanpe tu'a [oO] .i ku'i "a" cu se tolcri 01:29 < zipcpi> zantufa ra'oi nonlojbanword 01:30 < zipcpi> zantufa: ra'oi nonlojbanword 01:30 < zantufa> minajimpe: .i kanpe tu'a [,] .i ku'i "l" cu se tolcri 01:30 < zipcpi> So yeah, not in ZOhOI, only accepts rafsi 02:03 < gleki> quotes inside {li} will make {lo'o} almost unelidible 02:04 < gleki> ta'osai i think i just won't use neither {volve} nor {xruti} nor any other single brivla. 02:05 < gleki> {volve} is bad for portuguese speakers. others are either clash with existing gismu or even worse than {volve} in being confusing in mnemonics 02:05 < gleki> s/are// 02:05 < fenki> la'oi gleki pu djica lo ka cusku zoi zoi.{volve} is bad for portuguese speakers. others either clash with existing gismu or even worse than {volve} in being confusing in mnemonics.zoi 02:11 <@rlpowell> gleki: Fixed vreji, and taught it to send email if it can't connect. 02:17 < gleki> rlpowell: will you restore lost logs? 02:22 <@rlpowell> Not as easy as it sounds; your logs are in a totally different format. 02:22 <@rlpowell> Well. Not totally, but substantially. 02:22 <@rlpowell> And I don't have #jbosnu 02:22 <@rlpowell> Ah, but that's what the "other" folder is for, I suppose.