Subject: The proposed fourth tense of Lojban Participants: Arthur Protin Bob LeChevalier Carl Burke Doug Landauer Guy Steele Jack Waugh Jeff Prothero Jim Carter Robert Chassell [1] by Jeff Prothero If you grow up in, say, North America, you're likely to get used to seeing the Sun rise once a day in the East, and be tempted to assume that it does this all over the planet. If you're into science or science fiction, you might add a mental caveat that it is actually the horizon moving, not old Sol. Because our daily experience is so uniform, it's hard to remember that at the North Pole the Sun rises in the South, and does so only once per year... and that if most humans happened to live there, "the Sun rises in the South, once a year" would be the common wisdom. If it seems ridiculous to you that any human community could possibly be so parochial as to suppose that the Sun behaves over the entire planet as it does at the North Pole, I invite you to consider the concept "simultaneous". In the universe we live in, in a general setting, there is no such thing as "simultaneous", and Newtonian mechanics generally is absurd. These ideas are useful in practice because humans currently live in a very peculiar arrangement: * They think glacially slowly; * They all live in essentially the same place; * They are all going essentially the same direction at the same speed. If you relax any of these conditions, "simultaneous" ceases to mean much. To draw an analogy, if Terran civilization were based on sentient trees who all lived near the equator and faced East, we would not need any concept of "North" and "South" as distinct from "left" and "right", and it would be unambiguous to say "Tacoma is 35 miles right of Seattle". As it is, with people facing all directions at random, "left" is in the eye of the beholder, and such a statement is fairly useless. Thus, we introduce the observer-independent notion of "south", and say that "Tacoma is 35 miles *south* of Seattle". In similar fashion, "simultaneous" is, for a general group of observers, purely in the eye of the beholder. Two events A and B may be simultaneous to one observer, may happen in the sequence A,B for a second observer, and may happen in the sequence B,A for a third observer. For useful communication to take place, an observer independent vocabulary is needed. Time, fundamentally, is about causality. The most important implications of "A happened before B" are that A may have caused B, and that B could not possibly have caused A. In the Newtonian approximation to reality, the speed of light is infinite, interactions can happen instantly over arbitrarily large distances, and a neat trichotomy holds: given any two events A and B, either: 1 A happened before B, and potentially influenced B. 2 A happened after B, and potentially was influenced by B. 3 A was simultaneous with B. Newtonian physics allows no other possibility. Out in the real world, light needs a thousand picoseconds just to crawl a foot, and things are more complex. Given two events A and B, there are fundamentally four observer-independent possibilities: 1 Light from A could have reached, and potentially influenced, B. 2 Light from B could have reached, and potentially influenced, A. 3 A and B could have been co-incident -- same time and place. 4 A and B could have been so distant in space and close in time that neither could have influenced the other. In the Newtonian approximation, the latter two cases are indistinguishable, but in our universe the distinction is vital. Perhaps a couple of examples? Suppose we are debugging the Terrific Teraflops 2001 supercomputer. The tracelog shows that a glitch at time/location B reversed the clock polarity and fried 10E12 molecular switches, but why? Another event A, a nanosecond earlier and ten feet away, looks suspicious -- should we investigate? Newtonian intuition says "yes", since A was before B, hence might have caused it. Physics says "no", since A "4" B, and could not possibly have influenced B. To perform an automatic search for candidate causes for glitch B, we would ask for all suspicious events X such that X "1" B, not X "before" B. Again, suppose you are wondering who murdered Ishtar. If you are told that one month prior to the murder, Thshpck broadcast an offer of three megacredits to anyone who'd kill Ishtar, you might decide to check out the local guns for hire. But if you were also told that the broadcast took place a light-year away from the murder, you could immediately dismiss the broadcast as causally irrelevant to the murder. The bottom line is, any time you are interested in reasoning about the universe we actually live in, the event-relationship categories 1-4 above are the ones useful for deriving valid deductions. I don't really grok what follows, and anyone *really* interested in understanding spacetime should pick up a good book on the subject, but for those who want just a *bit* more and aren't allergic to a *gasp* formula: What's really going on is that the separate concepts of "distance" and "elapsed time" need to be merged into the single concept of "interval". In the Newtonian formulation, an event A happens at point Ax,Ay,Az, and independently at time At. Similarly with event B at Bx,By,Bz, Bt. The distance between the two events is the Pythagorian 2 2 2 0.5 ( (Ax-Bx) + (Ay-By) + (Az-Bz) ) and the elapsed time is independently 2 0.5 ( (At-Bt) ) In Newtonian physics, it is assumed that the distance and elapsed time can be computed independently, and that the results will be the same for all observers. No such luck! If we insist on applying Newtonian metrics to a relativistic world, we encounter the notorious "relativistic contraction" of moving objects, &tc, and the descriptions of events vary wildly from observer to observer. To leave Newton for the company of Einstein, we must swap our Euclidian space for a Minkowski space, and the above two independent metrics for the single interval metric 2 2 2 2 0.5 ( (Ax-Bx) + (Ay-By) + (Az-Bz) - (At-Bt) ) Even though independent observers will compute different distances and elapsed times for a given pair of events, it turns out that (in flat space and for freely falling observers, at least), the above interval metric is observer independent, so multiple observers can again sensibly compare notes. Note that, while time now enters into our "space" measurements, it does so with the sign reversed -- time is still a distinguished dimension. For two events which are distant in space and close in time, the above square root will be of a positive quantity, the result will be real, the interval will be "space-like", and causal interaction between the two events is impossible. For two events which are close in space and distant in time, the above square root will be of a negative quantity, the result will be imaginary, the interval will be "time-like", and causal interaction between the two events is possible. (So now you know what an imaginary distance is :-) Note that two events which happen at the same time and place are separated by a zero interval, but the converse is not true: Two events can be seperated considerably in time and space and still be separated by a zero interval. In the latter case, it is exactly possible to send a photon from A to B, with no time left over. What the two cases have in common is that there is no time for thought between the two events: any observer present at both events will perceive them as simultaneous and co-incident... and in fact, the distinction between the two cases is in the eye of the beholder, and won't appear in a physically sensible tense system. A quantitatively oriented language might relate the names for its four tenses to the above arithmetic properties of the interval measure: past: -imaginary future: +imaginary here-and-now: zero distant: real Simple tenses, in fact, might be simply permissible abbreviations of a more complete quantitative interval measure. (Loglan, of course, far from being a quantitatively oriented language, follows in the anti-quantitative natural language tradition of pretending that only black and white exist :-) :-) -- all functions defined in the language must return TRUE/FALSE. Were natural languages designed by bigots to make others into bigots? Only time travel will tell :-) :-) [2] by Bob LeChevalier (response to [1]) My first reading of Jeff P.'s tense commentary gives this (I may say more upon further reflection): 1. Newtonian mechanics is a simplification of relativistic mechanics, and in most natural linguistic situations works fine. Therefore, by Lojban's maximim optionality assumption, relativistic effects should be ignorable if not important. 2. As Jeff notes, not everyone 'groks' relativistic tense, and it is irrelevant to most people. It cannot therefore be a core concept that MUST be learned in order to learn Lojban, unless either we come up with a schema that makes it naturally and intuitively obvious. Lojban's claim to be a simple language depends upon this, and we can't realistically limit Lojban's usefulness to those who understand relativity theory. Nor can we be expected to teach relativity to anyone learning the language. Logic is going to be bad enough. 3. Lojban tense structures are non-mandatory, and in fact are subordinate claims that could be stated in a separate predicate. Thus ko'a ba broda ko'e x will razzlefratz y is equivalent to ko'a broda ko'e .i la'edi'u balvi ti x razzlefratz's y and this(the referent of the last sentence) is in the future of this (i.e. here and now) Tense structures must be simplistic, and can take only 1 sumti as incidental information without horrendously complex grammar, or expanding into multiple sentences as in the example. You can call this a grammar limitation; it is usually acceptable. 4. If you expand the tense into a separate predicate, then relativity is implicit in Lojban, and in fact is NOT limited to tense. EVERY predicate has the capability of adding, via sumti tcita, extra places, which can include expression of causality and observer. The latter two can also be expressed separately in a 3rd and 4th predicate, if the sumti tcita structure is too limited. thus ko'a broda ko'e .i la'edi'u balvi ti ga'a ko'i ri'anaiku x razzlefratzes y, and this is in the future of here-and-now to observer z, non-causally. 5. Note that Lojban separates causality from tense entirely. THis is partly because Lojban has several kinds of causality, and you'd have to complicate Jeff's description by an indication of what kind did or did not apply. The example I gave used physical causality. [3] by Arthur Protin (response to [2]) I would like to add my comments to this discussion, even though I am not up to speed on learning lojban. [4] by Jeff Prothero (response to [3]) Well, I'm not, either, of course... and have no ambition to be... [5] by Arthur Protin (continuation of [3]) Many of the arguments on the relation of relativity and causality are important and valid BUT probably should be dealt with by only providing for clear and non-awful hooks. [6] by Jeff Prothero (response to [5]) The specific proposal was just that Loglan should provide for all four tenses, and adopt the physically significant definitions for "before" and "after"... which coincide with the Newtonian definitions in the Newtonian limit. Should be at least as clear and non-awful as the approach Loglan actually has, does, and will continue to take :-). (As I stated in my original tense post, this is a personal pet peeve which I bring up every few years.) [7] by Arthur Protin (continuation of [5]) My understanding of our language (which dates back to the 1970's) is that it is a language of appearances. Thus the normal should be to assume a speaker frame of reference, with provisions for adding things to shift that frame to: 1) any arbitrary listener, 2) a specified listener, 3) an inertial reference, 4) a time travelers reference, 5) God's reference, and 6) any other reference frame that can be described. (that's not too much to ask, is it? :-) I would expect than that it also makes sense to tie any notion of causality to someone infering that relation, usually the speaker. Thus, the simplest (least modified) lojban expression conveying that "A happened before B" should mean that "as I witness/experienced it A happened before B". Some simple modification of that expression should yield "As best as I can infer (from my knowledge of physics and the events), A happened before B". It should be possible with half a dozen syllables to say "according to NAMEd paradigm, A happened before B". I expect that each of these types of qualification should be available for all statements, or at least all observations. [8] by Jeff Prothero (response to [7]) A most peculiar and unanticipated response, if I follow correctly -- you are arguing that Loglan should actually *favor* observer-dependent descriptions?! Thus, to take a specific case, you would *prefer* that Loglanists say "Tacoma is to the right of Seattle" rather than "Tacoma is south of Seattle" (with hooks so that it is easy to specify *which* observer the observation is relative to), since this emphasizes "appearances", and makes the normal a "a speaker frame of reference". Similarly, presumably, you would prefer that in Loglan altitudes be given relative to the speaker rather than sea level, nearby sizes should be given in multiples of the height of the speaker rather than meters, mass should be expressed relative to the strength of the speaker rather than in kilograms, distant sizes should be given in terms of the percent of the speaker's visual field they cover, time should be the local solar time (or time since breakfast?) rather than speaker-independent timezone time, and so forth? There's certainly nothing intellectually wrong with your proposal (although, obviously, it runs diametrically opposite to the Western-rationalist-scientific thrust of the last few centuries to uncover and propagate objective observations which can be interpreted and verified without reference to the observer!), but I'm curious as to how you arrived at the conclusion that this is a part of the spirit of Loglan... both the spirit of formal logic and JCB's aspirations to "carve reality at the joints" seem to run directly counter to this ...? (For what it's worth -- nothing, in my opinion -- JCB was very enthusiastic about an earlier "relativistic tenses" piece I did for The Loglanist, and it was only with difficulty that I got him to pull it when Keith Wright (?) correctly pointed out that the equi-interval curves I had drawn as circles should in fact have been hyperbolas...) [9] by Bob LeChevalier (response to [8]) The default observer reference frame in a lojban sentence is unspecified, and hence vague. sumti tcita can be used to add places to any predicate to specify an observer (ga'a) or a frame of reference (mai) or both. I think this will suffice for all of Prothero's posed problems and Arthur's specific examples. [10] by Arthur Protin (response to [8]) (Your message makes you sound like such an extremist. ;-) [11] by Jeff Prothero (response to [10]) Yes, this is perfectly true, I'm about three sigmas off the norm. E.g., I still like a priori languages, I think a human language based on arithmetic would be interesting, I think Loglan has no chance of beating Esperanto at its own game, and I think "human" and "person" are orthogonal concepts... :-) [12] by Arthur Protin (continuation of [10]) I must clarify something: "Appearances in an observer reference frame" IS NOT identical with "observer dependent". [13] by Jeff Prothero (response to [12]) Maybe you must, but you don't :-) ! If an observation is relative to an observer reference frame, then it cannot be interpreted without knowledge of the observer... which is extra work and a hassle, hence the universal tendency inside and outside the sciences to communication real-world information in observer-independent terms. [14] by Arthur Protin (response to [13]) Sure, but there are a great many ways that a report can be observer dependent without reflecting what was apparent to the observer. [15] by Arthur Protin (continuation of [12]) When I was in Seattle, I did not observe Tacoma to be to the right (are they really that much more conservative there.) I did find it to be "before" Seattle. But that had only to do with how I approached it. The only observer reference frame descriptor that accurately describes to the observer that orientation necessary to observe Tacoma while in Seattle is "South". [16] by Jeff Prothero (response to [15]) Which statement is only useful given more information about the observer, as you note. "South" is a completely observer-independent concept; the statement "Tacoma is south of Seattle" remains true no matter where the observer is or what direction s/he/it is facing. The relativistic "interval" measure has the same property: it remains the same for observers with different positions and velocities. "Time" and "distance" measurements don't have this property. Thus, observers in a relativistic context will prefer to report intervals to times and distances for exactly the same reasons that you prefer to report "South" to "right" in the case of Seattle/Tacoma. [17] by Arthur Protin (continuation of [15]) As far as altitudes are concerned, I am currently at a loss to define the "normal" because I am totally oriented toward the instrumentation that measure altitude. I never considered it before I was taught to think of it relative to sea level. Sizes should never be given in multiples of the height of the speaker because, if my experiences are representative, the speaker can not truly OBSERVE his/her own height. [18] by Jeff Prothero (response to [17]) You don't have to "observe your own height" to notice that the rock is twice as big as you, do you? (It's only the poor auditor who must scurry around doing extra work because of your laziness...) [19] by Arthur Protin (response to [18]) You most definitely do. You can only crudely approximate the comparison of anything else to your own height, but you can do much better at comparing the height of something with the height of something else placed near by. My use of rulers (in either feet+inches or meters) is not dependent on my desire to communicate those measurements to someone else. I also use measurement devices where their sole benefits are precision, accuracy, and repeatability. [20] by Arthur Protin (continuation of [17]) Trying to answer each instance of your misinterpretation, reminds me that we are taught how to observe and how to express those observations. None the less, the language design, as it was presented to me all those years ago, was that it was not to embed the moral, religious, or philosohical models but to be based on the appearances of the common experience. [21] by Jeff Prothero (response to [20]) Which is to say that it should be observer independent -- interpretable without knowledge of the moral, religious or philosophical properties of the observer. [22] by Arthur Protin (continuation of [20]) As much as my beliefs are scientific, and I want to bias my childrens thoughts toward scientific models, I am willing to give up biasing the language to my beliefs in order to keep it from being biased by anyone elses beliefs. Thus a chair is a chair not because it has the soul of a chair but because it appears that it can function as a chair. Recognizing it as a chair is the act of the observer. I have had the experience of not recognizing a chair as a chair because it accomplished its function in a manner that I had not seen prior. It became a chair (something I could sit on) only when the observer (myself) could visualize how it could function as a chair. My friend's chair is not necessarily a chair to me. [23] by Jeff Prothero (response to [22]) I think you interpret this opposite to JCB! I think JCB was trying to come up with an observer-independent description of chairness dependent only on the publicly-available properties of the chair itself, whereas you seem to be headed for a private notion of the predicate which is uninterpretable execept with respect, not just to a particular speaker frame of reference, but in fact to a particular speaker. Now Loglan is descending into a Babel of private languages! [24] by Arthur Protin (response to [23]) This line of reasoning sure seems like a mess now. A "description ... dependent only on the publicly-available ...." looks like a good thing but it has the key to the problem. Where is this definitive repository of "publicly-available" information? And is the public really going to avail itself of that availability? [25] by Jeff Prothero (response to [24]) Physical phenomena, as distinct from cultural convention. E.g., "sphere" and "cube" can be distinguished in culture free fashion, but "freedom fighter" vs "terrorist" (or "authoritarian" vs "totalitarian") can be distinguished only by phoning the Ministry of Truth. (To pick a topical example, Saddam Hussein moved from being a genial authoritarian leader to being a dastardly totalitarian dictator when he switched from invading Iran to invading Kuwait. Alien observers equipped only with satellite photography might be hard-pressed to distinguish the two actions...) I think JCB wanted to keep the "sphere"/"cube" distinctions in Loglan, but stamp out the purely connotative distinctions. This would make much American political discourse virtually untranslatable into Loglan, of course, since "Joe believes the Evening Star is the Morning Star" just isn't the same as "Joe believes Venus is Venus" :-) [26] by Arthur Protin (continuation of [22]) Rather than try to perfect the model of the universe before defining how to talk about it, I back the approach of defining the language to be as free of the model as possible and to make the role of the model explicit such that evolving and conflicting models can coexist for a sufficient length of time for the better one to prevail. [27] by Jeff Prothero (response to [26]) I think this makes your point of view clearer... language seems to me inherently unsuited to this role, cinema verite' seems much closer to presenting a raw, uninterpreted sense-stream for the audience to interpret as they wish... communication of data with no transfer of understanding. (Personally, I think transfer of understanding is fascinating and laudable :-) [28] by Arthur Protin (response to [27]) I like trying to transfer understanding as well, but there appears not to be any way of assuring that that transfer occurs. Communication seems to be a probabilistic endeavor which is very dependent on the speaker and listener sharing a great many common experiences, enough to have developed shared common abstractions and models. I don't have a problem with providing for the additional tenses needed for relativity, but will the expansion exhaust the pool of available tenses. How will we provide for the many strange ones that are required for discussing time travel? [29] by Jeff Prothero (response to [28]) (Only one additional tense, in my formulation.) I seriously doubt that time travel will become a real-world problem in a physical sense, but if Loglan were to replace Latin as the standard funny-looking SF language, it might conceivably become a real-world linguistic problem :-). Comments: * Since they are pure fantasy, there are as nearly as many models of time travel as there are writers about it. Hard to cover them all. * I once proposed, and RLC agreed (a unique occurrence ? :-) that Loglan should reserve a "do not define" bank of LittleWords for experimentation. These can be used for local-to-a-story tenses. E.g., you might have separate tenses for separate parallel worlds, or for self-created causal loops ("All You Zombies..."). (To the folks who will immediately note that English time-travel SF gets by fine without extra tenses: Maybe the stories that really need them don't get written.) * The difference between Newtonian space and Newtonian time is precisely that you can travel freely through one but not the other. In the limit of Newtonian timetravel, time simply becomes a fourth spacial dimension. This suggests that in the limit, you should simply drop the tenses and use the spatial-position LWs. [30] by Bob LeChevalier (response to [29]) We already allow for 4-dimensions in the spatial tenses. Does this mean that we've covered your relativistic problems? [31] by Jeff Prothero (response to [30]) 4 dimensions? How did you justify LWs for that? Not really, the limit is a long way away. In the limit, a circle becomes either a line or a point, but it's still nice to have a word for circle :-) But it does count as a point against time travel. [32] by Bob LeChevalier (continuation of [30]) Can anyone come up with a REAL example of a sentence that would use this hypothetical 4th tense, including sufficient context as who is saying it to whom, and why none of the other tenses are correct, provided that the tags 'to observer ...' and 'in reference frame ...' (ga'a and ma'i) are specified. Can you argue that the examples of such usage are numerous enough to justify use of a cmavo or in a very limited list of available words left? Can this '4th tense' be predicated instead of abbreviated with a cmavo (all tenses are considered to be abbreviations of some subordinate predication, as jimc is oft wont to point out)? [33] by Jeff Prothero (response to [32]) I'd hoped my computer example provided such a context... some further comments: It's always *possible* to do it with epicycles, the Copernican model only wins on elegance. The argument for 4th tense as LW comes more from symmetry and elegance than Zipf. If you are heavily using a set (simple tenses) with exactly four members, it is inelegant (and potentially misleading) to have names for only three of them. Fortunately or otherwise, arguments from elegance are more likely to move mathematicians than businessmen. It is definitely unfortunate -- IMHO, a design blunder -- that Loglan is designed around small, fixed-size lexeme slots that easily overflow. It would not have been difficult to make all lexeme slots infinite, so that the number of defined members of each lexeme could be governed more by logical and physical necessity, and less by the structure of the human vocal tract. A good exercise for the Handbook of Human Language Engineering. [34] by Bob LeChevalier (response to [33]) The fixed size is based on a limited word space for the cmavo. The predicate word and borrowing word spaces are infinite. It is arguable that far more than half the cmavo we have now, including ALL of the tenses, are Zipfean abbreviations for predicates. A limit on abbreviations seems much less offensive than Jeff's accusation, now, doesn;t it? Yet we are fully prepared to add to certain lexemes, and some lexemes have equivalents that may be concocted ad hoc from the predicat words (including tenses. So if you can express the relativistic and/or time travel concepts as predicates, we can always use this method to cover them for tenses). Thus even some of our cmavo lexemes are effectively open-ended. This sounds like an argument for making the language some kind of perfectionist ideal TOTALLY DIVESTED FROM THE BASIC PURPOSE OF LANGUAGE, which is human communication. If human vocal tracts (and hearing and multi-track processing and other limits caused by our being human) are not allowed for in the language design, it is NOT a human language. Or do I misunderstand your argument, Jeff? [35] by Jeff Prothero (response to [34]) (I *don't* propose changing Loglan's design in this respect. I *do* propose that follow-on designs try to do better.) You say the glass is half full: I say it is half empty :-): something approaching half the cmavo aren't such abbreviations! The point is just that, if you're starting from scratch, you can make all word classes variable-length without much effort. I think you do! My point is that (e.g.) the number of logical connectives in your language should not be dictated by (say) the number of vowels humans can clearly enunciate. It is perfectly possible to design a language with such a constraint (and 1960 Loglan had a flavor something like this, no?), but it is also perfectly possible to design a language *without* such a constraint, and have it be just as human-usable. Riding the point to one of my favorite outre' ideas: It is perfectly possible to design a pronounciation system for your alphabet which makes all sequences of letters equally pronouncable. This approach lets word design take a much more logical course, without irritating special cases due to the fact that your regular word-construction algorithm produced an unpronouncable result. My own experimentation goes further and adopts an arithmetic basis... but we're wandering away from Loglan. My basic point was just that Loglan's word morphology means that we are forever in danger of painting ourselves into a corner somewhere in LittleWord space. It is always possible to get out of such a corner, at the cost of a few footprints, but a better initial plan could have eliminated the problem once and for all, no? Language prototypers take note. (This belongs on a Language-Design mailing list, but there's not enough traffic to justify one yet, and lojban-list seems the closest current approximation.) [36] by Bob LeChevalier (continuation of [32]) In short, I haven't seen why this is needed, and the discussion doesn't seem to provide enough information on how it would be used in order for any teacher to teach it. [37] by Jeff Prothero (response to [36]) Well, I think Jim considers this topic just short of too trivial to discuss, Art thinks I'm probably but not yet provably daft, and you think this is a waste of valuable time. Maybe if I can convince Art a fourth tense is worthwhile, we can hammer out something specific enough to satisfy you... :-) [38] by Bob LeChevalier (continuation of [36]) There is a different standard held up for assignment of cmavo (LWs) than for brivla (Predicate words) in Lojban. We can easily devise words as brivla, even on the fly. People who don;t undersatnd can ask the speaker what the word means. cmavo, though, have to have a grammar assigned to them (lexeme or grameme), possibly have a rafsi (affix) assigned, and take up some of a very limited word space. In addition, they appear in a dictionary, as well as in LogFlash and must have a definition. (The same is true, of the gismu - the root 'primtives', by the way.) These lists will be baselined, and we are planning in even the earliest books to have examples of the use of each word, so that people have enough sense of what they mean to use them, since we expect them to memorize them. They will prefereably BE useful, and have a definable use. Theoretical usages that cannot be explained are garbage. (This is not to say that all possible Loglan expressions have to have a pre-defined semantics. It means that we have to present enough information in our prescription of the language that someone who wants to figure out the semantics for themselves, can do so.) [39] by Jeff Prothero (continuation of [27]) (I think) there's an inherent quantitative problem in your attempt to squeeze the model out of language: human perception supplies hundreds of millions of bits of data per second, but human language conveys only about ten bits of data per second, so any human reporter is necessarily providing only a very abstract (== model-dependent) transcription. By breaking a continuous world up into discrete objects, and then further breaking this continuous spectrum up objects into a few discrete classes of objects (predicates), a human language is (IMHO) necessarily imposing a very strong model on the universe -- a strong base for the Whorf Hypothesis. In the case of tenses, which are particularly deeply embedded in the language, I thought it might be nice to have the model be a physically sensible one instead of the traditional fiction :-) By "the better model" you mean whichever one in fact prevails, right? Or do you have a more fundamental measure?) By this logic, it seems you should favor a fourth tense, since it lets the Newtonian and Special Relativistic tenses compete on an even footing. Or would this require seven basic tenses so that the auditor always explicitly knows which model the speaker is using? [40] by Jim Carter (response to [27]) Physically you have no choice but to use observer-dependent descriptions, since the act of observation randomizes what is observed. Of course in everyday life there is so much correlation, in time and space, between the randomized part of a system and the remaining parts that you can get away with a statement "The creature is moving toward you" as opposed to "I saw the creature's left upper tentacle move toward you just now". [41] by Jeff Prothero (response to [40]) (Um, we're wandering from relativity to quantum...) I'm not sure this has anything to do with observer-dependence. The fashion and degree to which the observed system are randomized depend on the system and the probe, not on any characteristic of the observer who launched the probe. It means you have are guaranteed a certain minimum amount of problem reconciling successive observations of a system, but this isn't observer-dependent: you have exactly the same problems whether the successive observations are collected by different folks or the same observer. [42] by Jim Carter (continuation of [40]) In talking about emotion, we're all aware that people perceive the same event differently and so many (all?) of the Lojban root words about emotions include an explicit case for the experiencer; you don't even have to use a modal phrase ("sumti tcita", yucky terminology). As for relativity, the reference frame is in fact irrelevant for a causal connective. Starting at a "causative event" X, the universe is divided in three parts by the null (lightwave) cone centered on X into X's future (events reachable at less than or equal to lightspeed, which it could conceivably cause), X's past (which could reach it at less than or equal to lightspeed), and X's "present", which can neither cause nor be caused by X since the required communication would be faster than light. The null cone is invariant under arbitrary coordinate transformations anywhere (because the metric is a tensor) and hence choice of a reference frame cannot move an event between these three categories. It's true that any "present" event could have a time coordinate before or after X depending on the frame, but that sign is irrelevant causally; it's still a "present" event. [43] by Jeff Prothero (response to [42]) Your interpretation is defensible, but not the only one and (to my eye) not the most natural. In English, at least, when one says that a person (or thingie) is "present", we mean precisely that it is practical to interact with that person (or thingie). Your definition reverses this, and defines the "present" to be that portion of spacetime with which we *cannot* interact! Perhaps my viewpoint becomes clearer if we reflect that the "causative event" X is in practice usually not a spacetime point, but rather a volume. One can justify this by noting that most real-world events, such as a bullet shattering a window, or the fall of the Roman Empire, have a nonzero duration in time and extent in space. (Could one also make this point by an appeal to the Uncertainty Principle?) This makes the null cone less of a negligible boundary (as you are treating it) and more of an interesting volume in its own right -- the "present", in my four-part treatment. In this treatment, the "past" and "future" are as you described, but the "present" is the null cone, and (for non-pointlike events X) includes that set of spacetime points which the event (potentially) *interacts* with: both (potentially) influences, and (potentially) is influenced by. (Your "present" becomes my "fourth tense".) To modify your definition: Starting at a "causative event" X, the universe is divided in four parts by the null (lightwave) cone(s) centered on X: X's "future": Those events reachable at less than lightspeed, which it could conceivably cause; X's "past": Those events from which X could be reached at less than lightspeed; X's "present": Those events reachable from X, or from which X can be reached, following a geodesic at lightspeed, -- the null cone itself. For a non-pointlike event X, this fuzzes from a (hyper) surface into a (hyper) volume containing events which X both potentially influences, and also by which X is potentially influenced. (One can also look at this as the set of points composing X, or the set of points at which X is present, or the set of spacetime points co-incident with X for some possible observer.) X's "????": Those events which which can neither cause nor be caused by X since the required communication would be faster than light. (Should we call "????" X's "heaven"? X's "ignorance"? X's "protopast"? X's "exfuture"? ... :-) [44] by Jim Carter (continuation of [42]) It would be satisfactory to me to define Lojban "present" to mean "events separated from the reference event by a spacelike interval". In daily life the thickness of the present is too small to observe, so Lojban would not be spoiled for daily life, yet a precise definition is there for applications, like ultra-high-speed computers and space travel, where speed-of-light delay becomes significant. [45] by Bob LeChevalier (response to [44]) Per Prothero's comment on English meaning of 'present', I think that the word has split into 2 (or more) meanings, and he seems to be jumping between them. When, in English, "someone is present", this means indeed that you can interact with them, but this is due to colocation in SPACE (and only incidentally in time). The synonym is 'here'; i.e. "someone is here". [46] by Jeff Prothero (response to [45]) Well, as I see it, at least, the question is how to extend the English "present" tense from the Newtonian limit into a more general setting. Since humans, to date at least, haven't had to deal with relativistic effects a lot in daily life, this is necessarily as much of a creative act as a logical one... The meaning of a word can be defined denotationally as a set of situations (as Montague mostly does), but is most interestingly described as the set of relevant reasoning rules: knowing that a certain word describes X, what are we entitled to conclude? We (I?) would like to generalize the current definition of the present tense so as to preserve the sorts of reasoning now possible with it. (Since relativity views the world in terms of events and lightspeed communication, it would be nice to frame our definition in that language.) What is the most essential difference between "exists", "existed" and "will exist"? I was arguing that it is the potential to interact: If someone tells me that the Mississippi *is* a great river, I feel entitled to suppose that I can, at least potentially, go visit it. If someone tells me that Gibraltar Falls *was* one of the greatest scenic wonders on the planet, I'm rather inclined to conclude that it's too late to see them. (Last I heard, it's currently believed that there never was a Gibraltar Falls. Sigh.) If someone tells me that WW III *will be* impressive to watch from Luna, I'm inclined to believe that nobody has seen it yet, and that I can't count on just hopping a shuttle to enjoy the show. On this reading, the English present tense includes events out to infinity in all directions precisely because, in the standard Newtonian model, you can potentially interact with anything "currently" existing, if only you get there fast enough. (This intuition seems so strong that people appear to somehow feel robbed by the lightspeed barrier, that it must be a cruel trick, that somehow getting out to everything out there "now" must just require a little engineering. Conservation of energy doesn't seem to arouse quite the same indignation. Is this because English tenses don't suggest you *should* be able to get something for nothing? :-) Following this reading, the essential semantic property of the present tense which should be preserved in its new incarnation is that it selects those events which we can potentially interact with. Which is more or less what my definitions have attempted, and which JimC's latest paraphrase may have done most neatly. The "problem" with this definition is that it leaves a (vaguely donut-shaped) hole in spacetime, consisting of those events which are neither past, present, nor future, for which I need a fourth tense... [47] by Bob LeChevalier (response to [45]) Now I can recognize a whole bunch of arguable connotations about time that can certainly support "present" in the sense of here matching "present" in the sense of "present tense", but I for one don't think of them when I use the term in the 'here' context. In Lojban, we have the two gismu 'zvati' x is present/'here' at and 'cabna' x is simultaneous with for the two meanings. (I recognize that I may have totally misunderstood both Prothero and Carter in your usages of 'present', but I never made it through relativistic physics, which is why my degree in astrophysics has made me 'only' a computer-type.) In any event, we have to teach the language, and relativistic effects are a bit too tough for any textbook I'll ever write. [48] by Arthur Protin (response to [45]) An example that might illuminate this noise (warning: pun in progress) is the phenomena of lightning and thunder. Two separate manifestations of one event, which due to their propogation delays reach us at different times (actually due to the assignment of terms before understanding the event, lightning refers both to the flash of light and the discharge that causes the flash with a bang). Is thunder the sound wave that starts simultaniously with the flash, or is it the experience of hearing that sound many seconds after the seeing the flash. I consider it to be the latter (the hearing) extrapolated to the unobserved formation of the sound (the observation is of the wave radially propagated many seconds (and thankfully dissipated by that radial propagation)). Do the flash and the bang occur at the same time? No, and we would not describe them are simultanious, but we all understand that they originate together. [49] by Jeff Prothero (response to [45]) I think you're panicking before the attack is sounded, Bob! Loglan can have a word for "twistor" without requiring you to teach quantum mechanics in the textbook, it can have a word for "gravitational lens" without requiring you to teach General Relativity in the textbook, and it can have a fourth tense without requiring you to teach Special Relativity theory in the textbook. Heavens, we can even have a word for "pulsar" without covering stellar evolution! As long as you keep your textbook examples in the realm of Newtonian mechanics, the current three primitive tenses remain unchanged. As I see it, at least, the subjects of the tense discussion are (1) Should Loglan define the tenses *away* from the Newtonian limit? and (2) Does the general Special-Relativistic case require additional tenses? I think at least two or three other folks believe (1); I'm not sure if I've yet convinced anyone else of (2). Since I've been through this discussion a couple of times before without having the slightest influence on Loglan, I'm not really holding my breath in expectation of a fourth tense. Might pop up in Guaspi, though, since Jim understands this stuff considerably better than I do... [50] by Jim Carter (response to [49]) The four-part division is interesting for extended events, An Earth-bound person would be tempted to define "present" as those points in the future relative to parts of the event and in the past relative to other parts. [51] by Jeff Prothero (response to [50]) Sounds like a concise, precise definition. (Maybe that's why you're in math and I'm in anatomy :-) Drek. Maybe we need to go to *five* tenses. In my previous post(s) I was very careful to say only that my fuzzed-out null-cone "present" *included* those events which could interact with (both influence and be influenced by) our extended event X, since this "present" clearly doesn't *consist* of such events -- it includes many more. I'm beginning to think we need to distinguish these two "present"s: Coincident-present ------------------ Given an extended spacetime event X, the "coincident-present of X" consists of all (non-extended) spacetime events Y such that (some part of) X is co-incident with Y for (some possible) observer. This is the union of the null-cones emanating from X. Interactive-present ------------------- Given an extended spacetime event X, the "interactive-present of X" consists of all (non-extended) spacetime events Y such that Y is both in the future of (some part of) X and in the past of (some part of) X This is the set of events which can *interact with* (both influence and be influenced by) X. This is a proper subset of the coincident-present, and corresponds the most closely with the normal English/Newtonian "present" (IMHO). [52] by Jim Carter (continuation of [50]) But when you start charging around at a substantial fraction of the speed of light, the volume which you have to consider which is neither past nor future becomes the majority. Is he coming in range and I don't know it? [53] by Jeff Prothero (response to [52]) I'm afraid I'm not following you. Presumably we can treat ourselves as being at rest, so your "when you start charging around at the speed of light" can be treated equivalently as "when you start interacting with objects travelling near lightspeed". (Our "Earthbound person" is encountering photons, of course, but not commonly interacting with them as enduring and individual objects.) But I don't see how the volume which is neither past nor future can "become the majority" (of spacetime, by "volume", I presume) unless the extended event itself occupies the majority of space or time. (Taking myself as an example: An extended event (should we make "thing" a technical term? :-) a few feet (light-nanoseconds) by a few decades (hundreds of megaseconds) across will have an "interactive-present" a few light/decades in radius no matter how you cut it... in a universe a light/giga-decade or so in radius. Since any chord through this "present" is an observer-invariant interval, I (naively?) presume the "volume" of this "present" is also observer-invariant, and always much smaller than spacetime. The "coincident-present" of the above extended event will be much larger (one can visualize it as a decades-thick sphere expanding outward through spacetime into the past and future, no?) but still fills only an infinitesimal fraction of spacetime. Perhaps you are thinking of the "present" as including all of any extended objects which pass through it? I'm not -- my "present" is *defined* by an extended object, but *consists* of classical point-events. The elder George Burns is in my "present", but not the young one. [54] by Jim Carter (continuation of [52]) Similarly in real life, a conversation would consist of a timelike, not spacelike chain of sub-events (speeches), and it's neither useful nor polite for both parties to transmit at the same time. Relative to the last speech, the union of your two present zones are points where the listener should be listening, not talking. [55] by Jeff Prothero (response to [54]) (Well, there was a time when they sang four part harmony with different words, even in different languages, for each part! :-) My apologies: I'm afraid your point completely escapes me. What do the pragmatics of conversation have to do with tenses, relativistic or otherwise? Where did *two* present zones come from, here? [56] by Arthur Protin (response to [43]) Now that I am really starting to think about tenses as verses thinking about relativity, I find I am even more confused by your proposal. Please give me examples of the tenses that you propose opposite the tenses that exist right now complete with the transliterate English. These examples should be complete phrases if not sentences. [57] by Jeff Prothero (response to [56]) Well, my bits on the computer glitch and space murder were supposed to have provided such examples. I'd have coded up some specific Loglan phrases, but I don't even know what "pa na fa" have become these days... Let's return to the computer example. We are talking to each other at time T, trying to deduce why the prototype BioFlow 2000 computer just fried it's crossover, by searching the various RNA logs. The immediate cause of the failure is obvious -- a glitch G reversed the clock phases, resulting in a dead short in all active crossover switches. We're both aware that such a glitch is "utterly impossible", of course, and are trying to figure out what might have led to it... You discover a fluke F on a related RNA log, and we discuss it relative to glitch G: [1] "Look, this module had just fluked!" [2] "No, it was just *going* to fluke -- check the logtime!" [3] "You're forgetting, this log had been running ten nanoseconds slow for nearly a millisecond -- the module had been fluked for two nanoseconds at this point." [4] "Oh, right. But look -- the module is ten feet [light-nanoseconds] from the crossover. The module *tense-four* fluked. Any other flukes?" (These are all compound tenses in English since in speech we're required to do a bank shot from time-of-speech off time-of-glitch to get to time-of-fluke. In at least some loglans, these could be simple tenses because we would just adopt G as our time of reference.) Statement [1] notes that F was before G, using the past tense. The unstated implication is that F might have caused G, hence deserves further investigation. Statement [2] protests that F was after G, using the future tense. The unstated implication is that F couldn't possibly have caused G, and might in fact have been a consequence of the general collapse of the machine after G, hence isn't worth wasting any time on. Statement [4] points out that, even though F was before G in a wall-clock / Newtonian sense, F was further away in feet than it was previous in nanoseconds (light travels almost exactly one foot per nanosecond, as Grace Hopper observes), hence F lay in the "oblivious zone" of G, and could neither have influenced, nor have been influenced by, G. Note that neither past, present, nor future tense could make this point properly! [58] by Arthur Protin (response to [57]) Thank you for the example because it starts to provide evidence that the "special" tense that you ask for may be unneeded. I finally saw what you were asking for a distinction between present tense interactable and present tense non-interactable. Thus, the all-important fourth line of your example is not correctly represented in English. It should be a compound of the past tense with the present non-interacting, more like: "The module was independently fluking". [59] by Jeff Prothero (response to [58]) Right, I should have said "The module had *tense-four* fluked". [60] by Guy Steele (response to [59]) No, I think it should be "The module did *tense-four* fluke." [61] by Jeff Prothero (response to [60]) Maybe! I examined: "had fluked" ( at past reference point) "had been going to fluke" (future relative to past reference point) "had been fluking" (continuing at past reference point) "had been fluked" (past relative to past reference point) [62] by Guy Steele (response to [61]) No, this last one is a case of passive voice. It is confusing that the participle and the past tense have the same form. Your table should read: "has fluked" ( at past reference point) or "fluked", or "did fluke" "had been going to fluke" (future relative to past reference point) "had been fluking" (continuing at past reference point) "had fluked" (past relative to past reference point) Right? [63] by Jeff Prothero (response to [62]) I grab my American Heritage Dict to look up "participle", "gerund", "nominative" ... then carefully re-read:] Right! I stand corrected. [64] by Jeff Prothero (continuation of [61]) The pattern appears to be: "add a 'had' to reach past reference point", with various operators added to refer to the past, present or future of that reference point. So I extended the pattern with: "had *tense four* fluked" (oblivious zone relative to past refpoint) with the trailing "-d" probably inappropriate, but an aid to intuition. [65] by Guy Steele (response to [64]) The problem is that English uses both suffixes and pre-positional auxiliaries. I chose the "did" form so that everything could be consistently cast as auxiliary operators in the same left-to-right order that Lojban uses. [66] by Jeff Prothero (continuation of [64]) It's not clear to me that one can do the same sort of temporal bank-shots quite as naturally with "did": "did fluke" (present-of-past) "did already fluke" (past-of-past) "did ready to fluke" ("did about to fluke"??) (future-of-past) "did *tense four* fluke" (obliv.-zone-of-past) (Obviously *all* of these examples misuse the noun "fluke" as a verb. This is not an unrealistic extrapolation, given the steady Haigspeakification of the language... and it simplified the example dialogue.) [67] by Guy Steele (continuation of [60]) Hm. My next thought was that if one can compound a past tense with tense-four, then such auxiliary verbs as "had" also ought to come in four flavors (that would fall out naturally from the orthogonal structure of Lojban, I hope, but it bears pointing out explicitly for the English-pidgin version). So I ought to be able to say, e.g., "The module *tense-four* *tense-four* fluke"; in other words, the module, in an event tense-four-related to the speaker, has a fluke event tense-four-related to the event under discussion. [68] by Jeff Prothero (response to [67]) Yes. In Old Loglan, we had pa (past zone) na (present zone) fa (future zone) to which I can add for purposes of discussion qa (oblivious zone) yielding the Old Loglan psuedo-examples (I just invented the predicates) "le modlu na fluke" The module is fluked/fluking. "le modlu nana fluke" The module is now fluked/fluking. (?) "le modlu pa fluke" The module was fluked/fluking. "le modlu pana fluke" The module was then fluked/fluking. (?) "le modlu pafa fluke" The module was about to fluke. "le modlu papa fluke" The module had already fluked. "le modlu papafa fluke" The module had already been about to fluke. "le modlu qa fluke" The module *tense-four* fluked/fluking. "le modlu paqa fluke" The module had *tense-four* fluked. "le modlu paqaqa fluke" The module had *tense-four* *tense-four* fluked. Old Loglan syntax allows the simple tenses to be chained as deep as the speaker desires. I don't know the details of the current Lojban treatment, but I'm confident it is similar in spirit. (Bob?) [69] by Guy Steele (continuation of [67]) But my third thought was that such sentences, though conceptually useful, ought to occur only in fantasy and science fiction, because the speaker is making a claim about an event with which he cannot have interacted and therefore of which he can have no knowledge. [70] by Jeff Prothero (response to [69]) Certainly, you cannot *know* what is currently happening in your oblivious zone, that being a defining property of the oblivious zone. [71] by Guy Steele (response to [70]) Actually, that was the very core of my observation: that you cannot influence it at all. [72] by Jeff Prothero (response to [71]) (Ok, so maybe you want to banish the past tense to SF? :-) :-) I think we can agree that of my four spacetime zones (past, present, future and oblivious), the oblivious zone is of the *least* immediate importance -- we get information from the past, interact with the present, and try to influence the future, but interact with the oblivious zone only one step removed, by guessing influences our past actions may turn out to have exerted on it, and anticipating information we may eventually receive from it. [73] by Jeff Prothero (continuation of [71]) But you can extrapolate such activities, much as when you say "Aunt Sue should be landing in L.A. just about now" (or as in my example of a sister taking finals in a different solar system), and you can you can talk in hindsight about what *was* going on in someone's/something's oblivious zone, as in the computer-debugging example under discussion. If you think about it, I think you will find that your third-thought objection applies with equal force to the future-zone tense and the oblivious-zone tense, given that we are equally unable to have any direct knowledge of events in the future or in the oblivious zone. (The difference between the two, that we can hope to influence events in the future, but not in the oblivious zone, doesn't appear to enter into your objection.) Do you seriously propose that the future tense, although conceptually useful, ought to occur only in fantasy and science fiction? [74] by Guy Steele (response to [73]) I was overstating my case. Maybe I mean that such a tense in English would almost always necessarily appear in the subjunctive mood. [75] by Jeff Prothero (response to [74]) A fixpoint: I agree -- the fourth tense will likely prove gregarious, found rarely alone, usually with a subjunctive or in a compound tense. [76] by Jeff Prothero (continuation of [73]) (This seems unlikely to me, leading me to suspect that I have yet another person on my hands who is really arguing against the unfamiliar. While I'm making irrelevant comments: I will consider it a real milestone when the fantasy crowd starts using relativistic tenses! Sort of like seeing the first personal-computer retail shops appear, and hearing the first personal-computer radio commercials -- a sign that the revolution has reached the mainstream.) In general, I can easily imagine that keeping track of projected events in your oblivious zone might be essential any time that you *have* an oblivious zone... whether you are clearing buffer space (in an amplifying, soliton-mode optical-fiber loop, perhaps? -- the spirit of the mercury delay line may yet live!) for projected incoming traffic from the optical net, or anticipating action from opposing orbital laser battlestations (WW III may be essentially decided before first news of the start *reaches* Earth :-( ). [77] by Arthur Protin (continuation of [58]) As of right now I would not plead ignorance and abstain but would vote a definite NO. I have several reasons: This change caters to the personalities prone to the logic flaw "Post Hoc Ergo Poster Hoc" (It follows this, therefore it is caused by this). [78] by Jeff Prothero (response to [77]) I think I understand your point, although I don't agree with it. At the risk of putting words in your mouth, you see me as unnaturally slicing up the "present" to cater to some obscure technical needs which, while perhaps valid, are uncommon, and which screw up everyday reasoning in the process: The game isn't pointless, but it isn't worth the candle. In my opinion, of course, you are a Whorfian-mindlock victim who prefers a familiar fiction to an unfamiliar reality. :-) If we peer at the relativistic world through conventional Newtonian glasses, my proposal does more than slice the conventional present into a near (interactable) and far (non-interactable) present: these "presents" (which I call the present and the oblivious zone respectively) widen out with distance, eating into the naive Newtonian past and future. (In the limit of infinite lightspeed (and zero reference-event duration), this widening is negligible and the present collapses to an infinitesimally thick cut through spacetime.) The cost is a paradigm shift -- some non-monotonic learning -- and the benefit is that statements made using the tenses carry information more likely to be useful in everyday reasoning: IN A RELATIVISTIC CONTEXT, the relativistic tenses allow the auditor to immediately make deductions about potential causation -- deductions not possible if the naive Newtonian tenses are used. (In a Newtonian context, three relativistic tenses become identical with the standard three, and the fourth is unused... barring metaphorical use, which I consider likely among folks who have made the mental paradigm shift...) [79] by Arthur Protin (continuation of [77]) Interactions can be through a wide varienty of agents each with their own separate properties, not the least of which is propagation rate. [80] by Jeff Prothero (response to [79]) Yes, and while I haven't commented on this for fear of muddying waters which most folks already find sufficiently opaque, it would be possible to use relativistic tenses (and related predicates, of course) in a loose metaphorical way when other communication rates dominate the discussion. In a world in which the fastest communication is by mail carried by sailing ship, Africa is in London's oblivious zone for purposes of day-to-day business. As remarked above, I suspect a community of people acculturated to the new tense paradigm would indeed find such usages useful. The difference, of course, is that such barriers are technological, but the lightspeed barrier is not. (The speed of sailing ships does not enter into the ratio of mass to energy in a particle, for example, but the speed of light does...) [81] by Arthur Protin (continuation of [79]) While I accept relativity as the best expanation currently available to us and its models are sufficient for describing those of our experiences recorded (and reproducable) to date, I do not consider it to be the last word on the subject. (I have already heard theories that may be candidates to succeed relativity). [82] by Jeff Prothero (response to [81]) Are you still waiting to find out if the Earth is approximately spherical? :-) :-) Modern science has an *immense* amount of experience with the realm described by Special Relativity. General Relativity is fairly well tested, but this is very hard. By contrast, every particle accellerator, cosmic ray detector &tc built *must* take Special Relativistic effects into account in order to work at all. If all of these gadgets turn out *not* to have been working for the last century, I will be as surprised as if New York city turns out to have been a media hoax. (I am *not* asserting that either is absolutely impossible. But I've been to New York and I've spent years studying particle interactions in bubble chambers, and you'll have to come up with an awfully good story to convince me I've been had...) If these gadgets *have* been working as described, then Special Relativity correctly describes some very important properties of the universe we live in. We may someday find better ways of describing these properties, but they won't go away, and we might as well give them names. One of them is the oblivious zone. [83] by Arthur Protin (continuation of [81]) I will rule out neither faster than light travel nor time travel; and I will not require them to be linked. [84] by Jeff Prothero (response to [83]) Sure, and clocks may start running backwards and fairy godmothers may fill the air. Nobody is claiming otherwise, we're just trying to figure out how to deal with the things we *know* about. Rod Serling can fend for himself. Unlearning "truth"s which fail can certainly be an uncomfortable experience. :-) [85] by Arthur Protin (continuation of [83]) Thus, I am quite willing to leave as easily undistinguishable that which is concurrent and immediately interactable and that which is concurrent and remote. How I infer concurrence should not be part of the language. [86] by Jeff Prothero (response to [85]) Are you willing to leave as easily undistinguishable that which is future and that which is past? Can you present any argument other than blind habit as to why the past/future distinction is more deserving of a LittleWord than the present/oblivious-zone distinction? (Some linguist shoot me down, please: is it Japanese which lacks an explicit future tense?) [87] by Jeff Prothero (continuation of [57]) Note that a world-wide optical network connected to sub-nanosecond computers will shortly smash the traditional Netwonian assumption that you can communicate so much faster than you compute that communication can be treated as instantaneous: the local planet is 10^8 light-nanoseconds around. The human cycle time being aproximately a tenth of a second, this is analogous to holding a conversation in which it takes sound (and light :-) a year to make it all the way across the table. (To complete the analogy, we may assume that conversations last a million years and humans live many billions of years... a year is still a long time to wait for the punchline! :-) Dealing with the oblivious zone may become an everyday experience. (Letters already cross in the mail, but in principle you can get around that by phoning...) The above example is still pretty Newtonian in most respects: all the events are in the same inertial frame of reference, so our RNA logs can still be synchronized sensibly: we can maintain the illusion of "simultaneous" to some extent. With a more complex scenario involving observers on independently moving relativistic spacecraft, (or a computer designed using optically coupled relativistic electrons confined in suitable magnetic structures) we could remove this prop as well. Taking a different tack, if your sister radioported to Altair IV to study compact-matter engineering and hasn't yet returned, you can't really say that she *was*, *is*, or *will be* close to finals since future and past tense clearly don't apply and present tense really only makes sense (on human scales) for events local to your solar system -- you really need to say that she *tense four* finals. The inappropriateness of the present tense would be intuitively clearer if local stars, or many observers, had relativistic velocities... perhaps it suffices to note that she is so far out in your oblivious zone that for all you know she will turn out to have washed out of engineering a year before finals and do-*tense four*ing a thesis on the stylistic influence of DNA art on cetacean doppler music. Recapping with considerably less precision than JimC would tolerate: Newtonian tenses consist of: past present future We can give physically meaningful definitions of these (which are equivalent to the standard defintions of them whenever we can treat the speed of light as being effectively infinite) like so: An event is in our past if we could have received some sort of signal from it (canonically, a photon) but not vice versa. An event is in our future if it could receive some sort of signal from us, but not vice versa. An event is in our present if we can *exchange* signals with it. Relativistically, this mostly makes sents if "we" are an event with a non-zero duration in time (and, probably, space), so that the "present" includes more than a single point of spacetime. In the Newtonian limit, this "present" includes all spacetime points "simultaneous" with us, since light can get from us to any point and back again, instantaneously. Unfortunately, if we apply the above trichotomy to the universe we actually live in, we wind up with a big left-over chunk of spacetime: Those parts of the universe which could neither have sent us a signal, nor have recieved one from us -- our "oblivious zone". If we are living on a timescale marked in picoseconds, then our oblivious zone comes within microns of us. If we are living on a timescale marked in seconds, then our oblivious zone comes nearly as close as the Moon. If we are living on a timescale marked in decades, then our oblivious zone creeps within a few stars of us. This "oblivious zone" chunk of spacetime is just as significant as the other three, and just as deserving of a name. Just as we need a future tense to refer to future events, a past tense to refer to past events, and a present tense to refer to cotemporaneous events, so we need a *tense four* to refer to events in our oblivious zone. Is this at all worthwhile, or are we wearing a circular track in the rug? [88] by Bob LeChevalier (response to [57]) I see that this is an example where the choice of observer frame of reference may lead to a different truth value for the claim 'A is before B', in short 'A is before B to observer C (in reference frame D?)'. That the observer and reference frame are necessary to the truth value does not negate that the statement is either true or false given the observer and reference frame. Or at least I don;t understand th example if it is indeterminite given such knowledge. [89] by Jim Carter (response to [88]) Around a specified reference event the null (lightwave) cone divides the universe into past, future, and a large zone between them with can only be reached by faster-than-light travel and which hence can neither influence nor be influenced by the reference event (if we lack FTL). It is very reasonable to me to call this the "present". The null cone is independent of coordinates so the "past" and "future" need no reference frame specified. However, by proper choice of coordinates the time of any "present" event can be forced before or after the reference, but that time difference has no physical significance. In such a situation it would be better to say that "A is relativistically disconnected from B"; of course the gismu now assigned to "present" or "simultaneous" would be used. In daily life the thickness (duration) of the present is negligible, but not in astronomy, computing, satellite communication, etc. For example, the maximum diameter of an Ethernet system (1500 meters) is controlled by how much delay can be tolerated for the thickness of the present, so transmitters which start up "simultaneously" (+- thickness) become aware that the other transmitter has spoiled their signal before assuming that the packet has gone out. Jeff has some interesting ideas concerning events that extend over a region of time and space. However, given that Lojban tenses are somewhat extendible, I think we can wait until star travel arrives to deal with that problem :-) [90] by Jeff Prothero (response to [88]) One of the themes I've been playing (more of a bagpipe drone :-) is that, with a relativistically sensible tense set, "A is before B" is observer-independent -- it will be true for one observer only if it is true for all observers. It is only the Newtonian tenses which have the irritating property of being partly in the eye of the beholder. (Irrelevant comment: I'm certainly not slamming Newton's accomplishments! Anyone at the top of Gauss's list of great minds...) [91] by Bob LeChevalier (continuation of [88]) A predicate language has to be fairly careful in its place structure policy, and I think this is handled by ours. For another, more down to earth example the gismu 'bajra' (run) was once x1 runs from x2 to x3 via x4, modelled after klama (go). But then someone asked about running in circles and/or running in place, where some of the places have NO true value, so we changed the place structure (it is now x1 runs on limbs x2, and if you want to go somewhere running, you make the lujvo bajrykla (runningly-go) which analytic ally gives the other place structure. But why only 4 tenses? You just argued that there might be 5 needed. And given some of the strage theories like superstring theory with its 8 dimensions, why not 8 tenses (assuming that tense is relevant to the others). [92] by Jeff Prothero (response to [91]) I thought 10 was the most popular number ... ? Anyhow, the "extra" dimensions are rolled up too small to matter. How wide is the universe? Smaller than a proton -- if you travel through any of the last six dimensions, har har! [93] by Bob LeChevalier (continuation of [88]) And what if relativity is found to be just another successive approximation to reality, as seems likely to me, at least? What happens to tenses when you have tachyons? [94] by Jeff Prothero (response to [93]) Absolutely no question of this -- (classical) quantum mechanics doesn't handle gravity, and relativity says zip about quantum effects. Contemporary science has virtually no idea what happens when they interact... albeit lots of interesting leads. (It bothers me that none of the quantum approaches seem to have a good explanation for the Equivalence between inertial and gravitational mass, which Newton went to great pains to verify, and which General Relativity explains beautifully. Either I misunderstand, or a beautiful insight is being ignored...?! Superstrings and twistors may escape this criticism, as they have a strong geometric component...) But, as Asimov points out, new science doesn't invalidate old insights, it just adds more decimal places. Any child quickly discovers that the ground isn't flat on the meter scale. It has been known for millenia that, moutains aside, the ground is flat to a percent or two on a scale of tens of kilometers. At least a few folks have known for the last couple of millenia that the ground is spherical to within a few percent on a scale of thousands of kilometers. None of these insights invalidates the previous insight. Astronomy progresses, but the Sun keeps rising in the East (for us nonpolar types :-). Special Relativistic effects are just a real and well established as gravity -- the oblivious zone is no more likely to vanish than is gravity. I find it hard to take tachyons seriously. You can look through your legs and see the world upside down. You can look at electrons as positrons going backwards in time. Yawn. In real-world terms, the speed of light *is* infinite, in many useful senses. For example, in the Newtonian model, as you get closer and closer to infinite speed, you can travel further and further in a second, without limit, and exactly at an infinite speed, you can get anywhere instantly. In relativistic terms, this remains true, with "lightspeed" substituted for "infinite speed": as you get closer and closer to lightspeed, you can travel further and further in a second (as measured by your wristwatch), with no limit whatsoever. In the limit of actually reaching lightspeed, you can travel anywhere in the universe instantly, as measured by your wristwatch. What could it possibly mean to go faster than that? You really expect to see your wristwatch run backwards?? The intuitional problem comes when you travel "instantly" to another galaxy, buy a candy bar, then hop "instantly" back to find the Sun has burnt out. Well, I *told* you "simultaneous" doesn't make sense and that the present tense doesn't go that far. (Hotel Solaria: "You can leave the present any time you like, but you can *never* check back in!") If you tune your language and tense system to emphasize that "now" only includes suitably nearby things, this double hop won't surprise you at all. As long as you think in a Newtonian tense system with a "now" that goes to infinity, you're going to keep finding that the universe doesn't behave as your language does, and feel indignantly that the universe is cheating! *I* think that this is (arguably) a wonderful example of a Whorfian construct at the heart of the language getting such a strong mindlock on millions of people that they can't imagine an alternative and honestly blame the universe for not working like their language! (For example, we say "The Megallanic Clouds *are* our closest neighbors," implying that they are part of the present universe just like New York and Iraq. But they aren't, in as fundamental, physical a sense as the past and future aren't part of the present -- as the above double-hop shows.) [95] by Bob LeChevalier (continuation of [93]) Lojban is already VERY avant garde among human languages in making time and space tenses almost completely symmetric gramamtically. We haven't left in the time travel tenses, but there is an obvious spot in the grammar if they are ever needed. Otherwise, the structures for time and space are identical. But separate in the default because MOST people don;t think in terms of time and space in the same breath. Lojban at least allows you to. An essence of Lojban's design philosophy, carrying out one of JCB's ajor ideas, is that Loglan attempts to "remove constraints on thought". [96] by Jeff Prothero (response to [95]) Aye, and there's the rub: To think and speak relativistically takes four tenses, to think and speak in the Newtonian model takes only three. Loglan adamantly restricts you to three. Repeat after me: You *will* think Newtonianly! You *will* think Newtonianly! You *will* think Newtonianly! ... [97] by Bob LeChevalier (continuation of [95]) Wherever practical, we;ve designed choices into the language that seem to remove a constraint - to make something obligatory in most langauges optional in Lojban. Our tense system is arguyably the most powerful such in human langauge, though I occassionally find places where given infinite lexeme size we might do better. I challenge Jeff and anyone else to actually look at the tense grammar and the lexeme members we have. You'll blow your mind. (In Lojban you can orbit a point in time, whatever that might mean.) If Jeff and others can come up with a way that breaks it and is needed for human communication, though, then we should consider further change. But the window of time to do this is small, so start studying!!! I reserve judgement about whether you are just daft, Jeff, or maybe ornery :-) I also have no idea how trivial the question is. We've spent far longer discussing more trivial points in the Loglan design before. If a change to the language design is really needed, it is not too trivial. Because I don't know, it is NOT a waste of time. The baseline grammar is up he flaws in the language before the textbook is done and the dictionary, and the language design is frozen, probably indefinitely for points like these. I would urge people to study the grammar, and look for more significant questions (they are certainly there - because we've made undocumented decisions that no one has discussed publicly). [98] by Jack Waugh (response to [91]) The four tenses sound good to me as a reflection of our understanding today of time, communication, etc. The fifth tense is not so fundamental; it's just the start of a countably infinite series: The two things are close enough in space and long enough in duration so communication at light speed can get from A to B and back to A, the same thing and back to B, the same thing and back to A, and so on. The response someone gave about what about if our understanding of physics changes again, ties in with my ideas of another letter today, on having two engineered grammars. If tense is only expressed with predication in the deep grammar, it will be poised to grow with physics. You can put in ideas like the fourth tense without worrying about Zipf, because for the deep grammar you don't care. The shallow grammar would abbreviate expressions that need frequent expression, such as the tenses that are used in everyday life. [99] by Robert Chassell (response to [91]) I'm not yet convinced lojban needs more tenses/spatials/relatials than it has, but will observe (`will' as in its original sense, of `want/wish to'.... 'observe' used metonymically ...) that an acquaintance is working on a computer with nanosecond switching times. His circuit boards are 18 inches across and speed of light delays are a design/debugging issue. Protin's imaginary debugging session in the example in his recent message is not fantastical. [100] by Bob LeChevalier (response to [1]) I just spotted something in Jeff P's earlier writing on relativistic tenses that raised red flags (Wed 2/13 16:55) >Time, fundamentally, is about causality. The most important >implications of "A happened before B" are that A may have caused B, >and that B could not possibly have caused A. > >In the Newtonian approximation to reality, the speed of light is >infinite, interactions can happen instantly over arbitrarily large >distances, and a neat trichotomy holds: given any two events A >and B, either: > 1 A happened before B, and potentially influenced B. > 2 A happened after B, and potentially was influenced by B. > 3 A was simultaneous with B. >Newtonian physics allows no other possibility. I assert that the first sentence is an assumption that Jeff is making that is not necessarily true, and that the conclusions that result from it are therefore invalid. [101] by Jeff Prothero (response to [100]) You're touching on some fundamental issues, Bob, but I think you're reading them backwards... I'm certainly arguing from a set of propositions which I presume we hold in common. (There is no alternative except violence. :-) Among the propositions which I'm assuming we agree on are: * I'm assuming that Loglan is intended to be useful for humans who want to discuss events in the physical world. If your stance is that Loglan is purely for discussing logical propositions, then we indeed are arguing at cross-purposes. In that latter case, the inclusion of tenses in Loglan is very hard to understand. *Nothing* in pure logic requires tenses. * I'm assuming that we agree that there is a pragmatically significant difference between time and space. I have done a crtain amount of pondering on this, and it seems clear to me that the sorts of properties and symmetries we observe along the spatial axes are indeed quite different from those we observe along the time axis. I summarize these differences in a nutshell by observing that causality operates along the time axis, but not along the spatial axes. I'm certainly willing to argue in support of either of the above propositions! [102] by Bob LeChevalier (continuation of [100]) Specifically, the assertion of relativity seems to differ - that "A happened before B" does NOT imply that A caused B if B is not in the light cone of A. The alternative is to assume that the definition of "happened before" is something in particular that justifies the implication - in short, assuming the conclusion. [103] by Jeff Prothero (response to [102]) When *using* a language, logic or theory, we use given statements to determine conclusions about the world. When *designing* a language, logic or theory, we work from known properties of the world to determine what mechanisms we should include. It is perfectly true that the direction of synthesis is opposite in the two cases. This does not imply that either process is invalid. In designing our tenses, we are partitioning spacetime up into a few (three or four, in this discussion) mutually exclusive sets. There are an infinite number of ways of doing this. (Not just a coutably infinite, or even Aleph-1 infinite number, either!) We do *not* select a partition at random! We select that partition which we think will be most useful to our audience. How do we do that? Obviously, we cannot run a program to try all possibilities and return the one that maximizes utility. So we set up some set of criteria, and construct a partitioning scheme according to them. The criteria I've focussed on are: (1) It would be nice to reduce to the traditional tenses in the Newtonian limit. (2) It would be nice if the auditor could make deductions from statements using the tenses without having to know anything about the person who made the statement. (This is certainly entirely in the spirit of pure logic, and comes close to capturing what we mean when we say "formal".) (3) It would be nice if the auditor could make deductions about potential causality on the basis of statements made using our tense set. Given the above three design criteria, I believe (and, to a great extent, science has demonstrated that) my proposed tense system is a best solution, and nearly a unique solution. (If jimc were to provide his design criteria for a relativistic tense system, I think they would be substantially the same, and he and I treat the "present" in nearly diametrially opposite fashions, so there's at least some room for disagreement within the framework.) Is there a better set of design criteria, or does my design fail to meet them? Obviously, a concrete exposition of a better set of criteria is more constructive than a simple rejection of the above set. So: Yes, my tense definitions are *deliberately* designed to allow the auditor to conclude from "A before B" that "A potentially influenced B". Being able to draw new deductions from a given set of propositions is entirely characteristic of logic, and enabling this is an entirely proper design goal for a logical language. The real question is, why should we give up this capability by adopting a naive-Newtonian tense system? What do we gain in return for our loss? The answer seems to be that we are giving up logical deductive powers to avoid using one cmavo slot in LittleWord space. A poor bargain, in my opinion -- like giving up logical negation to free up a cmavo. [104] by Bob LeChevalier (continuation of [102]) It is possible that in English, time and causality have some deep structural relationship that justifies Jeff's assumption. But Lojban, as a logical language, need not, and indeed I believe does not. [105] by Jeff Prothero (response to [104]) The link between time and causality is in the physics of the world we live in, not in the language we speak. English has developed among a population of people who care, on a continuing and everyday basis, about things like catching trains, planting crops, and not getting killed... concerns which require accurate communication of information about time and space. I've been presuming that Loglan will be used by people with similar interests, but there is of course nothing in pure logic which justifies such a presumption. (If you *really* believe that there is no logical connection between time and causality in Loglan, then you must believe that there is no way for a Loglanist, told a that "A happened after B", to conclude that "A did not cause B". Are you willing to commit to this stance? If so, Loglan detectives are going to suffer! If not, Loglan's reasoning rule's *do* link time and causality, since the first statement concerns only time, and the second only causality.) From the point of view of pure logic, we can drop the tense system altogether, or use the "future tense" to denote events happening on odd-numbered days, the "past tense" to denote events happening on even-numbered days, and the "present tense" to mark events which never happened at all. Logic would be perfectly happy with this arrangement. Would *you* be perfectly happy with this arrangement? I think you would *not* (whether you will explicitly admit it or not :-) because such a tense system would not be tuned to conveying the information you are most likely to want to convey. I think that if you have the insight and honesty to explicitly dissect out *why* the classical Newtonian tense system you favor is more useful than the even-odd tense system above, you will find that it is because the classical Newtonian tenses have associated with them powerful reasoning rules which allow your listener to form additional valid and useful conclusions from your statements, and that these conclusions are centrally concerned with potential causality. For example, scholars give primary credit to the *first* publication of a fact or theory. Logic would be as happy with the *last* or *best*, but from the standpoint of potential causality, the *first* publication is uniquely positioned. I don't think this consideration is *central* to priority disputes and most other discussions driven by tense-carried information. Since neither you nor anyone else has tried to formulate a formal semantics for Loglan, I can't examine the reasoning rules *you* have postulated for the tenses, and am necessarily reduced to offering my own and guessing at yours... letting you arbitrarily dismiss mine and object that my guesses are mere straw men. My formulation of tenses is designed to allow the listener to make valid deductions about potential causality. In the relativistic regime, the naive Newtonian formulation does not allow this, and the four-tense relativistic formulation does. What does the naive Newtonian formulation offer in return? Or is conveying useful information to the listener completely irrelevant to Loglan's mission? [106] by Guy Steele (response to [104]) If you ask me to take an absolute, black-or-white position on time and causality, then I would have to conclude that, yes, every event has a causal influence on every event in its future light cone and is causally influenced by every event in its past light cone. But most of those causal influences are negligible. They are of order 10^-40 or less relative to other causal influences. It would be wrong of me to claim that astrology is bogus because the stars cannot influence us; the very fact that we are aware of their existence is evidence that they can influence us. But I might claim that their influence is negligible with respect to the phenomena with which astrology attempts to deal. So we must ask ourselves whether Lojban is to make itself pragmatically useless by insisting on 100% pedantic accuracy, or whether we ignore negligible relationships most of the time. (Consider, by comparison, the English idion "the lion's share". If you have read Aesop, you know that the lion's share is, in fact, the whole thing, 100%. But most people use it to mean 80% to 99%; in other# words, the lion's share (sense #2) of references to "the lion's share"# use sense #2, even though it is not true that the lion's share (sense# #1) of references to "the lion's share" use sense #2. This is because a colorful idiom for "99%" is much more useful than one for "100%". [107] by Bob LeChevalier (response to [105]) My essential disagreement with you is that I do not admit that Lojban's tense system is inherently Newtonian. What it infers is subjective time which is relative to a single observer. Tense (unless explicitly noted) should not be referring to God's omnipresent and objective (presumably) point of view. I'll append a posting to this I was going to post separately that elaborates. Re the association between time and causality: you and I may assume it, but is it a mandatory epistemological/metaphysical assumption that must be embedded in the language. Noting that Pynchon's Gravity's Rainbow is based on the possibility that causality is NOT time sequenced (or rather that it is possible that an event in the future could cause one in the past), it is at least conceovable that the opposite holds. John Cowan, Athelstan, Nora and I (and eventually pc by phone) got together this weekend. Among the topics was the lively (40 printed pages so far) discussion of relativistic tenses in Lojban. Some highlights: Tense is essentially subjective time; it is not a claim about absolute time from God's (omnipresent) point of view, because the speaker may not know what God's point of view is. This is essentially the same as what Arthur Protin was saying about speaker orientation. To a peasant with no astronomical training, the stars are 'just above the sky' and most certainly are there in the present tense, regardless of what physics says. In Lojban, all truth and knowledge has a place for epistemology. The language allows the creation of predicates that specifically address measured or objective time, and this is the preferred means for relativistic references. It was agreed that tense makes no claim about causality, as I've stated. Indeed, I might generalize this to say that in Lojban, as a new language with a bunch of only nebulously defined words so far, you can make NO >assumptions< that predicate A necessarily implies predicate B merely on the face of the definition, unless the definition explicitly includes that implication. There was consensus that if, in the future, a case was made to justify a relativistic tense of the sort proposed, it is merely the addition of a cmavo to lexeme PU. The interaction of this tense with the rest of the tense structure is of course unknown as to semantics. pc, whose specialty is tense logic, listened to our summary of the discussion, and indicated that he believed it to be a non-problem. He gave a reference that will send jimc and Jeff P. back to the research books. Apparently, a researcher investigated the problem of relativistic tense thoroughly in the 1960's. His PhD thesis is considered the landmark study in the area, and has not been significantly challenged. The conclusion was that no special modifications needed to be made because of relativity. pc did not have the exact reference, but the following should be enough for jimc (who is at UCLA) to find, and probably Jeff as well. Nino B. Cochiarella is the author, date approx. 1967 UCLA thesis title something like "A System of Relativistic Tenses in Tense Logic" As to a formal semantics, our consensus was that there is no consensus on a semantic theory and that to adopt one would be a strong metaphysical bias. pc pointed out that it would hardly solve the problems that Jack Waugh is concerned about, anyway. We would need a formal theory of pragmatics for that, and pragmatics theory is much less advanced than semantics. (Pragmatics includes such issues as jimc's problems of anaphora resolution.) [108] by Robert Chassell (response to [105]) Consider: `The flower grew and then I watered it.' ``Obviously'' (in some world views), the flower grew _because_ it expected me to water it. Newtonian and Einsteinian views of physical causality are culturally recent innovations. (Interesting and valuable innovations, in my opinion, but that is beside the point.) Loglan has words for different kinds of causality: physical, entailment from rules of game, moral, and so on. The tense system should not presume one or other is primary. If so, Loglan detectives are going to suffer! Many detective stories use a `means, motive, opportunity' format for figuring `who dun' it---Aristotle's "final cause" is the type of cause considered, as well the "material cause" and "efficient cause". Loglan should be able to handle a world view that presumes that all entities in the universe, including rocks, are sentient in some manner and that all occurrences may be explained in a telelogical manner, in which the causal events occur _after_ the caused event. Of course, like the presumption that events described in a paragraph are occuring at and after the time suggested by the last tense marker, speakers will by default presume one or other causality system. Once you stretch beyond the contemporary tense system of the culturally recent past, tenses in English begin to look weird: The use of a form of desire to indicate the future, as in "I _will_ go to the store." The use of spatial movement to indicate the future, as in "He is going to build a house." Perhaps equally useful weirdnesses will arise in lojban. The idea that a littleword should be reserved for some kind of relational other than the conventional contemporary tenses is interesting. I am not sure whether the relational should be a tense or a spatial or something else. On the other hand, the number of littlewords is very small. As a practical matter, I think that a block of littlewords should be reserved for future use, without specifying what they are for. [109] by Bob LeChevalier (continuation of [104]) Lojban's system of expressing causality is independent of its system of expressing time. On the other hand, the system of expressing time and space are highly interactive. I justify my claim by looking at another area where causality overlaps a distinct feature in English - that of implication in logic. [110] by Doug Landauer (response to [109]) I don't think Jeff is suggesting either that A before B implies A caused/influenced B or that there should be any dependence between causality and tense. It seems to me that all he was saying is that it would make sense to make the set of tenses more complete; and that it would be a bonus, in a practical sense, that completing the set of tenses from the three newtonian ones to the four relativistic ones would allow us to say A "tense-four" B implies A cannot have caused B and A before B implies A *might* have caused B. As you can probably guess, I agree with him that this addition makes sense. In English, when we say "If A then B" we often mean that "A causes B". The example used in Loglan discussions historically is "If you water it, it will grow." Those same historical discussions (most recently in one of the Volume 5 issues of The Loglanist - I believe Richard Kennaway was one of the authors, and is on this list - care to comment, Richard?), clearly indicate that in Lojban, the conditional is a purely predicate logic one, and that causality is not part of the definition. Elaborate efforts were then made to allow parallel and independent assertion of causality in a similar and parallel structure to the logical connective form to specifically delineate the semantic opposition between the two concepts. We retained this distinction in Lojban. Thus, for our example: ganai do dunda loi djacu ko'a gi ko'a banro the conditional Either it is not the case that you water it or it grows. If you water it, then it grows. ri'agi do dunda loi djacu ko'a gi ko'a banro the causal Because you water it, it grows. which both use lexer_G_935 of the lexer grammar to form the first compound. A SIMILAR OPPOSITION EXISTS BETWEEN TIME TENSE, SPACE TENSE, AND CAUSALITY. Indeed it can even appear in the same sentence form: bagi do dunda loi djacu ko'a gi ko'a banro the tensal (?) After you water it, it grows. or pugi do dunda loi djacu ko'a gi ko'a banro After you water it, it grows. or cagi do dunda loi djacu ko'a gi ko'a banro Simultaneous with your watering it, it grows. or even for space tense vigi do dunda loi djacu ko'a gi ko'a banro Where you water it, it grows. It turns out that causals are permitted anywhere else that tenses are permitted, and indeed all sumti tcita of lexeme BAI and their conversions and negations can occur anywhere that tenses are permitted. The only constraint, which is dictated by the LALR1 grammar, is that you cannot use a nonce sumti tcita using FIhO inside lexer portions of the grammar, since a FIhO 'modal' (to use jimc's word) can carry a full predicate with all of its sumti and other grammatical garbage into the modal, requiring in effect the full grammar to be recursively embedded in one lexeme. (It is these same sumti tcita that allow specification of observer and reference frame. Enjoy: sega'agi do dunda loi djacu ko'a gi ko'a banro Observing your watering it, it grows. ma'igi do dunda loi djacu ko'a gi ko'a banro In the reference frame of your watering it, it grows. Whatever these mean, they are not causals, or tenses, or implications.) In Lojban, any predicate claim has an optional potentiality component. The unmarked form is formally vague on whether there is a claim of actuality or potentiality; pragmatic context determines the implied tense in such a statement. But there is at least a claim of potentiality in the ellipsized tense. I would maintain that discussion of potentiality, which is carried in the grammar separately from the tenses (although part of what might be called the overall tense structure) is also independent of both causality and time/space tense in Lojban. If tense relationships in Lojban do not imply causality or potentiality, is there still a need for these extra tenses? [111] by Jeff Prothero (response to [109]) Your points, while valid, appear irrelevant to the immediate issue: 1 Why do we define the standard three Newtonain tenses the way we do? Put another way: sorts of deductions does this formulation support? 2 Given the above, how can we extend those definitions into the relativistic regime while doing minimum violence to them -- while preserving their logical properties as much as possible, so that the sorts of deductions we are habitually accustomed to making will continue to be valid? I think we're failing to communicate because you are taking the naive-Newtonian tenses as God-given (that is, your interpretation of "A before B" as inviolable, even in the relativistic regime), and reasoning from there, whereas I am taking the logical properties of the tenses as a given and deriving the logical spacetime partition as a consequence. "A before B" really means that "A is the in *past* part of spacetime as partitioned at B". Whether this is true in a particular case depends on just what rules we use to slice up spacetime. You are arguing for a naive-Newtonian partition (apparently) because it saves a LW, I am arguing for an alternate partition because it comes as close as possible to letting people continue to use their familiar Newtonian-regime tense-related reasoning patterns even in the relativistic regime -- it preserves the logical properties of the tenses, to the extent that it is possible to do so. As long as we refuse to explicitly acknowledge that there *are* formal properties of tenses that people rely on, we can treat the loss of those properties as a negligible issue. When and if we try to specify the formal semantics of Loglan, keeping the formal semantics small and regular will loom just as important as keeping the formal grammar small and regular, and adding one extra LW will seem a small price to pay for perhaps halving the formal semantic description of the tenses. [112] by Arthur Protin (response to [111]) Since long before Newton, the present tense was used to describe things that are going on now, that is, things you can not possibly see the beginning of but may if you are "fast" enough see the end of. Yes, I am sure that you could take long enough to get there that the Mississippi would no longer be flowing, all its water stolen by the Achefallia (I forget how it is spelled). Back when fast was horse back, "now" still meant that you might have to be infinitely fast to observe it. The notion of causality and interaction were IMHO far less significant in the understanding of "now" than was the concept of linear time. "Now" will still mean what it did even for events outside of the zone of interactability. Let me offer the following SF situtation as an example. I set up my banks and banks of lasers years ago and fired them such that now even as we speak their moon is exploding. (and if you were omnipresent you would be able to both observe my statement and its truth.) The event light years away is going on now, I caused it years ago and I will observe the feedback of the event in years ahead. Any good change to "now" has to deal with a non absolute time. It is not important to the notion of "now" that the senses may witness the effects of two events in the wrong order, just that there is AN order. If time is not monotonic, then the notion of "now" will have to be revised. As long as time is monatonic, then there is one true ordering of the events and while we have only a limited vantage point to observe the reality, we are omnipresent in our model of that reality. "Now" means concurrent with the speacker in that absolute time. I can't find any better way of dealing with this (as of now). [113] by Arthur Protin (response to an unavailable message by Jeff Prothero) You appear to misunderstand me. I was in no way suggesting that those devices did not exist and work as described. I am saying that as Newtonian physics described some of the available experiences but proved to be a special case of relativity, so relativity may be only a special case of a more complete and more correct model. One that may add to or modify the limitations demanded by relativity. The speed of light may not be the true limit. [114] by Jeff Prothero (response to [113]) Certainly Newton, who initially appeared to have virtually exhausted physics, in fact left many things to be said, and later physicists have gone on to say some of them. But Newton's fundamental laws and insights stand as firm today as the day he wrote them. (On the physics front, at least. His alchemical and theological positions don't seem to have fared as well.) It is not fair to claim that "The speed of light may not be the true limit", implying that this would be no more surprising than the insights Special Relativity added to those of Newton's Principia. The role of the speed of light is as central to Special Relativity as universal gravitation is to Principia. A *fair* comparison comparison would be: Just as it may turn out that gravititational attraction is *not* proportional to the masses and inversely proportional to the square of the distance, so it may turn out that the speed of light is not a fundamental limit. And, in fact, General Relativity (and we're now in *much* deeper waters than Special Relativity) *did* replace Newton's Principia as the reigning description of gravity. Gravitational attraction no longer propagates instantly, but is limited by the speed of light (funny how it keeps popping up, no?), and the attraction doesn't have to be *quite* inverse-square, because space turns out not to be *quite* flat. (From General Relativity, we now know that the angles of any physical terrestrial triangle do *not* add to exactly 180 degrees, and that any physical parallel lines near Earth do *not* maintain a constant distance between them. But Euclid's geometry still serves us as well in practice as it did him, two millenia ago -- it is true to a very high degree of accuracy, as the Greeks discovered and nobody since has disputed. I haven't heard any proposals for striking the word "parallel" from English *or* Loglan, whatever its status in Nature's geometry...) BUT: General Relativity or no, you can still plot the orbit of a spacecraft just fine using Newton's laws, just as Newton plotted the orbits of comets -- and in fact, NASA does so. Newton's laws are still taught to students and used by engineers on a daily basis, all over the planet. Nothing in General Relativity would make Newton hang his head in shame, feeling that he had thoroughly screwed up. (For example, he explictly never claimed to know if his gravitational force was real, or how it might be propagated, and would undoubtedly be delighted to hear that progress had been made on this front. Newton, the alchemical heretic, was *amazingly* prescient. His later writings give light a wavelike periodicity to explain diffraction, and a spatial assymetry to explain polarization. In retrospect, one is almost surprised that he didn't formulate a good wave theory of light, saving the rest of us a couple of centuries... just imagine if his alchemical obsession had led him to play with batteries, electricity and magnets!) Equally with Special Relativity: The fundamental role of the speed of light in our universe is beyond doubt, but one can certainly pick nits, and in fact can continue to learn new physics from examining the lightspeed limit in more detail: Light travels more slowly in dense media than in vacuum, so we can and do observe particles travelling faster than light in liquids and solids, emitting characteristic Cherenkov radiation -- a photonic "sonic boom". Recent theoretical calculations suggest light travels something like 1 + 10E-19 (?) times faster when within a micron (?) of a charged flat plate and travelling perpendicular to it, due to the altered nature of the vacuum. Photons may yet turn out to have nonzero rest mass, hence travel slightly slower than the relativistic "speed of light". (In which case we will probably start calling it the "speed of neutrinos", or if *they* have rest mass, "the speed of gravity". We still have almost no idea why particles have mass, much less why they have the particular masses they do...) You can send light the long way around the planet and take a short cut, arriving "faster than light". Clever, no? It is conceivable that the universe has a more complex topology than normally supposed, providing short-cuts ("wormholes") for some paths -- again, if you send light the long way and take the shortcut yourself, you may be able to arrive "faster than light". (You still couldn't get through the wormhole faster than light could, and theoretical attempts to construct wormholes result in conditions so extreme -- tidal effects that disrupt atoms, &tc -- that nothing *but* a photon seems likely to make it through. The wormholes also tend to collapse after a microsecond or so. Details.) None of the above has physicists doubting the essential validity of Special Relativity, or the reality of the effects it describes. [115] by Arthur Protin (response to [114]) I am beginning to believe that we have become distracted by many issues of relativity. Even if relativity is replaced by a more complete theory that includes a non-null explanation, the practical effects predicted by his work will still be observed. Einstein never said (anything like) "consider this deep structure for the universe and see how that prevents us from measuring the speed of light relative to its medium (the ether)" rather he said (something like) "let us accept the observation that we can not measure the speed of light relative to its medium and see what the mathematical implication of that failure is") [116] by Jeff Prothero (continuation of [114]) The utility of the proposed "fourth tense", like the utility of our particle accelerators, doesn't depend on the eternal acceptance of Special Relativity as the best available description of phenomena like the oblivious zone, just on the *existence* of those phenomena: * To destroy the practical utility of Newton's laws, you have to eliminate something like gravity; * To destroy the practical utility of Einstein's formulation of Special Relativity, you have to eliminate something like the fundamental role of the lightspeed limit -- maybe E != mc^2, and isn't even close? * To destroy the practical utility of the fourth tense, you have to eliminate the oblivious zone to which it refers. Honestly, Art, I think you're primarily exhibiting humanity's age-old xenophobic reaction to anything new. [117] by Arthur Protin (response to [116]) While that may be close, it is probably the effects of having grown up around a saleman, and I now can not be sold anything (I buy things only because I decide to). Even my physics instructors in relativity class, and quantum mechanics could not get me to accept things just because they said them nor even just because they were in print in the book. If this idea of yours is a good idea, then only the cold logic of its justification will move me. [118] by Jeff Prothero (continuation of [116]) Try a null test on yourself: All your arguments about the possible collapse of relativistic physics apply with equal force to Newtonian physics. Thus, if your arguments require the exclusion of the relativistic tenses from Loglan, they also require the exclusion of the current Newtonian tenses from Loglan. If you'll provide a list of possible ways in which relativity could collapse to render the relativistic-tense partition of spacetime nonsensical, I will undertake to match you one-for-one with a list of possible ways in which Newtonian physics could collapse to render the Newtonian-tense partition of spacetime nonsensical. In my opinion, there will be time enough to reform the Newtonian component of the Loglan tense system when and if Newtonian physics *does* collapse (as opposed to collecting marginalia), and there will be time enough to reform the relativistc component of the Loglan tense system (if we get one) when and if relativistic physics collapses. Either eventuality would be an unprecedented disaster (to what would you compare it -- the failure of astrology??), and trying to guard against such a historically unprecedented disaster strikes me as being much like worrying about the possible sudden breakup of the planet while building your house. Even defining what one *means* by the phrase "faster than light" is a vexing problem, if one makes any attempt at all to make it consistent with what we know about the universe. For example: travelling faster means covering a given distance in less elapsed time, by ordinary intuition. Travelling at lightspeed means covering any given distance in *zero* subjective elapsed time. Thus, travelling faster than light presumably means covering a given distance in less than zero elapsed time. This is a difficult concept for me... Again, any massive observer boosted to lightspeed has infinite mass, as measured by a naive observer. A massive observer boosted to more than lightspeed would thus presumably have more than infinite mass, which is again a difficult concept. Anyone travelling FTL will appear to be travelling backwards in time to some observers. If you really let timetravel in, it is going to be very difficult to keep *any* conservation laws in effect, no? As I commented earlier, if time travel is possible, it is quite difficult to explain why the extreme conditions of the Big Bang didn't propel particles forward in time, to rain down throughout our contemporary universe... everywhere one looks, things seem to be coming unstrung. It's like adding the axiom "2+2=5" to arithmetic and then frantically trying to patch all the problems that follow. [119] by Arthur Protin (continuation of [113]) I have not had time to review the texts on relativity, and since you claim to work with it, can you post any examples of situations that require time to not be monotonic, and/or prohibit an absolute ordering of events (in an inhabitable reference frame)? [120] by Jeff Prothero (response to [119]) Let's correct a possible mis-impression: I'm not a physicist, I get my physics from Isaac Asimov & Co just like most folks, and any *real* physicists reading the list are probably snickering from time to time. The undergraduate Computer Science program at the UW opened up after my freshman year, and I never took another physics course. I *did* hold a no-brain student helper job at the UW, scanning photographs of neutrino events in the Fermilab two-meter bubble chamber. The accelerator wouldn't work without relativistic corrections, and some of the particles I saw decaying have lifetimes so short that they wouldn't have made it far enough to be observable in the absence of relativistic effects. Excluding from discussion things like the quantum foam of spacetime as observed near the Heisenberg limit, where all bets are off (for example, quantum fluctuations can create masses large enough to change the topology of spacetime! -- this is where quantum effects and General Relativity meet, a theoretical mess), and restricting ourself to the scale and era Special Relativity is designed to describe: Any pointlike observer can always assign an order to all observed events, simply by time of observation. Non-monotonic time is the sort of paradoxical concept that makes no sense in a straightforward interpretation. Time is measured by the successive states of some physical system. In the absence of time travel (or faster-than-light travel), no clock will see another clock running backward, although other clocks may appear to tick too slowly or too quickly. [121] by Arthur Protin (response to [120]) Excellent! Given that the universe is such a physical system, then there should be an "absolute" reference for time (as marked off by the states of the universe). Even if we can not experience it directly, we can assume its existance and attempt to define the correction from our experience of time to that conceptual absolute!!! [122] by Jeff Prothero (continuation of [120]) The problem is that the observed order of events often depends on the observer, rather than the events. [123] by Arthur Protin (response to [122]) Excellent! Given that the universe is such a physical system, then there should be an "absolute" reference for time (as marked off by the states of the universe). Even if we can not experience it directly, we can assume its existance and attempt to define the correction from our experience of time to that conceptual absolute!!! [124] by Guy Steele (response to [123]) No, you have fallen into a trap here: the use of the word "state" begs the question. Implicit in the word "state" is the notion that you are taking a "snapshot" of the entire universe *at the same point in time everywhere*--but that is exactly the notion we are trying to get a grip on! Observers in different inertial frames will differ on what constitutes a state. [125] by Arthur Protin (response to [124]) Here you say that states are subjective, but the various subjective states that you referred to are misdirection. [126] by Guy Steele (response to [125]) Yes; I was sloppy in making a transition and I am sorry I confused you. In effect, I first cited the "intuitive" notion of state (a snapshot) and pointed out that that this definition will not do because it is observer-dependent. I then proceeded, in the paragraph beginning "What is a state?", to appeal to a more basic intuition about state without recourse to the notion of global simultaneity that is implicit in the word "snapshot"; then I used this more basic intuition to *redefine* the word state so as to have an observer-independent meaning. Note that such states do not have unique successors; they are only partially ordered, not totally ordered. State X is earlier than state Y if and only if there do not exist two events x (in X) and y (in Y) such that x lies within the future-light-cone of y. [127] by Arthur Protin (continuation of [125]) There is either determinism (and we can keep our physics) or there isn't. I will assume for at least this lifetime that there is determinism! [128] by Guy Steele (continuation of [124]) What is a state? It is the minimal amount of information needed to imply future states (assuming determinism--here I ignore quantum mechanics). I argue, then, that to include in a single state information about two causally related distinct space-time events (locations) would be redundant, because some information about the causee would be implied and could be omitted. (This argument is not air-tight, but perhaps could be made so.) It follows that a state ought to consist of information about a set of events that are not causally related. In other words, the set should be "space-like". And this is in fact about the best we can say in a relativistic theory, independent of any particular inertial frame: that any "spacelike" slice through the four-dimensional universe may be validly regarded as a state. But another observer may regard different parts of that slice as belonging to different states as he observes them. So, I am sorry, but the "successive states" model simply does not necessarily define absolute global time. It does define local time, and that is consistent with relativity. [129] by Jeff Prothero (response to [125]) The simple fact is that many things which seem "self-evident" are not, and indeed a good number turn out not to be true. (Part of the delight of science! Was it Eddington who said "The universe is not only stranger than we imagine, it is stranger than we *can* imagine"? But even if we cannot imagine Nature's wonders, we can discover them, if we keep a sharp eye and open mind...) It is "self-evident" that objects at rest tend to stay at rest, and that objects in motion tend to come to rest. Anyone playing with toy boats in a bathtub can observe this. It took the genius of Galileo Galilei to realize that in fact there is no such thing as absolute motion or rest -- motion can be measured only relative to some other object(s) -- and that all inertial (unaccelerated) frames are equivalent... and it took Newton to codify this neatly. It is "self-evident" that different observers can agree when two events are simultaneous. It took genius to realize that in fact the simultaneity of two (spatially separated) events is dependent on the observer's inertial frame of reference, that different observers in different frames will see things differently, and that nature doesn't favor any particular frame -- observations made in one frame are just as "true" and valid as those made in any other. (I'm not sure if Einstein gets full credit for this insight... the Lorentz Transforms were common currency, and there is general agreement that if Einstein had not done Special Relativity, someone else would have done it shortly. Not to knock Einstein -- he has forgotten accomplishments overshadowing the complete careers of most physicists...) [130] by Arthur Protin (response to [129]) So the problem with our tenses is not talking about interactivity (which you have succeeded in convincing me belongs in separate words) and observing relativistic events but in communicating about events with beings in a different inertial frame of reference (BIADIFOR). [131] by Jeff Prothero (response to [130]) Useful communication between observers in different inertial frames is certainly a central issue here! [132] by Arthur Protin (continuation of [130]) When communicating with BIADIFOR, what tags will be useful for translating between the DIFOR's? [133] by Jeff Prothero (resopnse to [132]) There is a fundamental problem here: tenses convey *qualitative* information. When we use tenses to describe the relationship between two events A,B to somebody else, we usually start with quantitative information -- we normally have a pretty good notion of the distance in time and space between the two events -- but we throw away most of this information, and tell our auditor only which (of a small fixed set: *before*, *during*, *after*...) conceptual pigeon-hole into which the relationship fell. (In statistics, this is called "binning", and Numerical Recipes recommends delaying it as long as possible, because of the information loss.) When we reduce our quantitative information about the relationship between A,B and ourself to a simple tense, we throw away the information needed to transform this description into a different frame of reference. I see only two ways around this problem: (1) Include in our statement enough quantitative information about A,B and ourself for the auditor to transform into any desired inertial frame of reference. In my opinion, this approach violates the spirit of the tense system, will be unacceptably verbose, and will require more arithmetic ability than most Loglanists are willing to acquire. (I *like* the idea of a language which forces a little arithmetic on the user, but I don't think RLC would consider this to be in the spirit of Lojban. :-) (2) As I have been advocating, use a set of pigeon-holes which don't *need* translation from one inertial frame of reference to another, because they use boundaries which are physical invariants, appearing the same to all observers: the speed-of-light null-cones defined by one of the two events A,B. If we define A's "future" to be those events lying within A's forward null-cone, and A's "past" to be those lying within A's backward null-cone, then all observers will agree on which events call for the past or future tense with respect to A, with no need for explicit transformations between different inertial frames of reference. [134] by Arthur Protin (continuation of [132]) When communicating with BIADIFOR's becomes a common experience (at least common enough that every body has met or spoken with some one who has communicated with a BIADIFOR ("complete second hand exposure")) it will be important that we have the tools to make those translations easy. That gives you some time to make a few false starts (as obviously you have and will again) before you find a good provision for it. (If you still think that relativity and interactivity affect our notion of "now", realize that the size (in four dimensions) of the zone ("cone") of commonly agreed "now"ness is a function of the difference between the velocity vectors of the two observers IFOR's!) [135] by Jeff Prothero (response to [134]) Um, my formulation uses tenses to describe the relationship between a pair of events, using *one* as the reference: "now" doesn't depend on pairs of observers. I think my relativistic "now" is as well defined as the classical Newtonian "now", but won't claim to have found a clearly ideal formulation. ("Now" is admittedly a heuristic concept, in the Newtonian *or* the relativistic formulation: "now" may cover an interval of a billions years or a femtosecond, depending on the subject at hand, with the auditor expected to guess correctly.) [136] by Arthur Protin (continuation of [134]) My position is: 1) The tenses are a crude reflection of the sequencing of events. [137] by Jeff Prothero (response to [136]) They are a qualitative classification of possible relationships between pairs of events. Any qualitative presentation of quantitative relationships may be legitimately termed "crude" :-) [138] by Arthur Protin (continuation of [136]) 2) Since that sequencing is dependent on the inertial frame of reference (IFOR) of the observer with no known physical bias toward any one IFOR and no known way of removing the IFOR from the description, our language will remain in the speaker IFOR and may someday have hooks for ease of translation. [139] by Jeff Prothero (response to [138]) The naive-Newtonian tense classification is dependent on the observer's inertial frame of reference; the relativistic tense classification is not -- thus, there *is* a "known way of removing the IFOR from the description". (Finding descriptions which are independent of any particular coordinate system is a central theme of General relativity in particular -- hence the focus on tensors -- and of physics generally. At the risk of entering my zone of ignorance, we may define as "geometric" precisely those properties which are invariant under change of coordinate system.) [140] by Arthur Protin (continuation of [138]) 3) Attempts to encode a speed of light (SOL) distance into the tenses is bogus because: a) The second point (the origin) for the measurement is not always obvious, [141] by Jeff Prothero (response to [140]) Tense is used to describe pairs of events; One of the two serves as the reference point (origin). In compound tenses, the reader is asked to guess at (or accept as arbitrary) one of the reference points; this is true in both Newtonian and relativistic formulations. [142] by Arthur Protin (continuation of [140]) b) It is insufficient to aid the translation into other IFOR's, [143] by Jeff Prothero (response to [142]) In a sense, but since such translation is not needed, the point is moot. [144] by Arthur Protin (continuation of [142]) c) It biases our thinking toward only those interactions that are direct and at SOL, [145] by Jeff Prothero (response to [144]) It makes physically important distinctions, as opposed perhaps to politically or theologically important distinctions: I think this charge could be leveled against any tense system. It tells us more about physically possible influences and interactions than it tells about the other kinds of influences and interactions... I suppose you could consider this a bias. :-) Again, I think this charge could be leveled against any tense system. [146] by Arthur Protin (continuation of [125]) Only one state description will correctly imply the future. That state description is one of a sequence that is the absolute time reference. That we can not measure, observe, or even correctly infer those state descriptions will not deter me from the belief in their existence. (A much more compelling proof on their non-existence will be required to move me.) [147] by Guy Steele (response to [146]) This is the flaw in your reasoning. In a relativistic theory, observer-independent states are not totally ordered, only partially ordered. [148] by Arthur Protin (continuation of [146]) Now what proof have you that we can not ever infer the "true and absolute" event sequence for the events we do observe? [149] by Jeff Prothero (response to [146]) None! I don't even have a clear notion of what a "true and absolute" sequence would be, given what we know about the universe, much less how it might be established. (A verifiable angel stating that "Allah is the only God, Mohammed is his Prophet, and Mecca defines his Inertial Frame," perhaps? :-) Can you suggest an experiment that would establish such a frame? I also don't have any proof that we will not one day discover that the Earth is indeed at "true and absolute" rest, with the rest of the universe rotating around it, just as Ptolemy supposed. In both cases, it is clear that any such proof would be of very little practical importance -- it is *abundantly* clear that in practice, the Earth is most conveniently treated as being just one more planet, and that in practice the local inertial reference frame is most conveniently treated as being just one more typical inertial reference frame. For example, even if the Earth is "proved" to be at absolute rest, and epicycles are "proved" to constitute the "true and absolute" description of planetary motion (again, I haven't the foggiest notion what such a proof would look like), NASA will go right on plotting planetary spacecraft courses using the Copernican Sun-at-rest model and Newton's laws, just because they are so much simpler and more convenient -- they capture important truths, if not the only truths. We're moving from science (descriptions of what we actually observe) to religion (unshakable a priori beliefs) here. No problem -- religious beliefs are an accepted part of contemporary human culture. But is it really fair to ask Loglanists to use a tense system based on phenomena which cannot be measured, observed, or correctly inferred? Would it not be more conservative to use a relativistic tense system which depends only on what a normal person can reasonably observe and deduce? Should ESP be a pre-requisite to learning Loglan? [150] by Jeff Prothero (continuation of [122]) Any person or system interested in surviving goes to great lengths to ensure that, as far as possible, descriptions of events depend on the events themselves and not the observer. For example, the human visual system goes to great lengths -- *cheats*, to put it bluntly -- to ensure that the percieved color of an object depends on that object, and not on the color of the illumination. Similarly, the visual system goes to great lengths to ensure that the perceived size of an object depends only on the object, not on the distance between the object and the observer... in spite of the fact that the retinal image varies dramatically with distance. Mistaking a lion for a kitten can be fatal! In similar fashion, most people, most of the time, attempt to make their descriptions observer-independent. We don't say "it was a little wee train, shorter than my finger", just because it was distant, nor do we say "it was a giant train, taller than the Space Needle", just because we were next to it. We *care* about the difference between "There's a kitten in the bedroom, dear." "There's a lion in the bedroom, dear." even if they kittens and lions sometimes look alike to some observers, neglecting size and distance. We expect the speaker to make a good-faith effort to distinguish the two cases for us, and provide an observer-independent description, not to provide unedited sense-data. Thus, I have no doubt whatever that *real* people, dealing routinely with *real* relativistic effects, would carefully distinguish between real "A before B", when that relationship is true for *all* observers, and mere apparent "A before B", when that relationship only indicates observational happenstance. In summary: * "Non-monotonic time" means almost nothing to me. I could invent a meaning, but it wouldn't be very physical. * Any *single* observer certainly *can* impose a subjective order on events. * There is *no* absolute ordering of *all* events that all observers can impartially agree on. [151] by Arthur Protin (response to [150]) Right, there will always be people that can not understand the laws that affect their perception of the absolute reality, and as such will never be able to correct their perception for those distortions. Further, we may have to refine our models to more completely correct our observations. [152] by Jeff Prothero (continuation of [150]) * There are *some* events whose ordering all observers will agree on. (These are the only cases where *I* would use the past and future tenses. I don't like to confuse apparent size with real size, and I don't like to confuse apparent ordering with real ordering.) * For the remaining events, whose order lies purely in the eye of the observer, my proposal uses a fourth tense. JimC would (ick! :-) use the present tense for these, and the naive Newtonian tense approach (deceptively, in my opinion) uses the regular past and future tenses, lumping them in with the truly ordered events. ("Kittens, lions... hey, they're all cats, right?") [153] by Arthur Protin (response to [152]) No! I am sorry, but we just showed that there is an absolute time that can be infered and subjective observations can be translated into that absolute reference!!! [154] by Arthur Protin (continuation of [119]) One of those "other reactions": I don't like cluttering up a simple notion of time with (what I believe is) an orthogonal notion of interactablity. [155] by Jeff Prothero (response to [154]) Riposte I: If you truly believe that "time" and "interactability" are orthogonal, let us construct a table: A was long before B | A was during B | A was long after B --------------------+----------------+------------------- A potentially | | influenced B | | --------------------+----------------+------------------- B potentially | | influenced A | | --------------------+----------------+------------------- A potentially | | interacted | | with B | | --------------------+----------------+------------------- A could not | | possibly have | | interacted with B | | --------------------+----------------+------------------- Since the two axes carry orthogonal labellings, it should be possible to find a pair of events A,B to go in each box. Would you care to do so? Riposte II: It seems to *me* that the naive-Newtonian tenses, lumping apparent event orders indiscriminately together with real event orders, in fact have a more complex semantics than the straightforward relativistic tenses. In particular, the relativistic semantics doesn't have to deal with the position and velocity of the speaker and auditor, which the naive-Newtonian tenses do -- they having no other way to distinguish apparent event orders from real event orders. Riposte III: I've explained what *I* think the primary use of tense information is -- to provide the auditor with the information needed to make qualitative deductions about potential causation and interaction. For example, *I* regard the differences between "There was a lion in the bedroom, dear." "There is a lion in the bedroom, dear." "There will be a lion in the bedroom, dear." as crucially involving the *interactions* I can expect if I open the bedroom door, and the results they might *cause*. I place great value upon being able to deduce such potential interactions, and would be quite upset if the tense system were changed so as to reduce or eliminate my ability to make such deductions. (The linguistic mechanisms we now use to convey tense obviously *could* be used to convey much different information, such as the relative and/or absolute social status of speaker and listener. I suspect that there are human languages which in essentially do this. Presumably Loglan has good pragmatic reasons for making the choice it did...) If you really *don't* see deduction of possible interactions as a significant use of tense information, can you explain what information you *do* get from the tenses in the above examples? Can you specify the design criteria for an ideal tense system? Can you explain why you care about tense at all? What difference would it make in your behavior and understanding if people completely stopped using tenses in their communication with you? [156] by Arthur Protin (response to [155]) I can easily fill out your table of time versus interactability (this is becoming tedious), but you would complain that the interaction was skewed. For example, the window shatters and an hour later the boy brings home his new BB gun (obviously the BB traveled much faster and more directly into the liviing room). We need to be able to talk about the observation and the origin independently. When we say "There will be a lion in the bedroom, dear." we are not limited to thniking about the interaction between the lion and the speaker, the listener, the bedroom, but may be interested in any or all the things that may be in the bedroom at any point of the lions stay as well as anything that may be left or altered by the lion during his stay that can interact with anything that enters the bedroom after he has left. Your energetic discussion of this topic has convinced me completely that the requested "fourth" tense is bogus. I want to be able to think about time and the sequences of observations even when I can not understand the causes and effects. I do not want my language corrupted by the forced linkage of time with arbitrary scopes of interactability. [157] by Carl Burke (response to [156]) As I understand it, this debate which has been clogging my in-basket is raging over whether or not to incorporate a fourth "time" cmavo, for a set of "past, present, future, and *not applicable*". The fourth tense refers to (in a relativistic sense) events outside the perceptible area of space-time; *fourth tense* events cannot possibly interact with the speaker, at least in the location/interval in question. Is this an adequate non-technical summation? [158] by Jeff Prothero (response to [157]) An excellent summation, except I prefer "*other*" to "*not applicable*". "*Not applicable*" suggests that the statement cannot sensibly be evaluated in context -- in computer terms, that it deserves a "compile error" (or evaluates to "bottom".) But it is not senseless to discuss nonpresent, nonfuture, nonpast events in a neighboring galaxy -- it just requires a fourth tense. Certainly, we could concievably abuse a "*not applicable*" flag by using it instead of a fourth tense, but this would be a semantic mess. If someone asks what tense correctly describes the relationship between pride and the color green, *then* I would invoke *not applicable*, since pride and green are not spacetime events describable by tenses. [159] by Carl Burke (continuation of [157]) This seems to be a useful concept, but it would find applicability outside the strictly physics-related relativity frame. For example, a criminal's alibi is an attempt to assert a *fourth tense* relation with the crime, under the prevailing conditions. Locked room mysteries, by limiting the communications means, isolate the "world-line" of the crime; the detective must identify the means by which communication occurred to reach the goal of "solving the case". In the limit, barring FTL communications a la Bell's Theorem, you have the cases argued (ad nauseum) so far. [160] by Jeff Prothero (response to [159]) (Bell's Theorem / EPR demonstrates "nonlocal effects", but does not give us FTL communication. Let's not thrash *that* out here! But I'll nominate it for best example of the universe being stranger than we could have imagined...) Certainly, a literal tense-four relationship to a crime would be an excellent alibi -- for example, showing that you were born in the oblivious zone of the crime would be as good as showing you were born after the crime. But a locked room doesn't mean much to a physicist: neutrinos, gravity waves etc freely carry information into and out of human rooms. (A black hole, now, ...) Such a use of the fourth tense would be very loose and analogical. In English, at least, I have the impression that such loose use of the tenses themselves is rare to the point of non-existence, although the corresponding predicates are fair game. [161] by Carl Burke (continuation of [159]) This would appear to be every bit as useful as the addition of a third value to traditional Boolean logic: True, False, *unask the question*. [162] by Jeff Prothero (response to [161]) I think there are actually a number of candidates for a "third" logical value: "unknown" springs to mind, plus "true but unprovable" and "orthogonal -- can be assumed true *or* false without danger of contradiction or inconsistency". I suspect pc could provide a dozen more. Want to make up a list and run them by RLC? They deserve at least compound predicates, I should think. [163] by Carl Burke (continuation of [161]) The third value holds where neither a true or false value is applicable; the traditional question is "When did you stop beating your wife?" I seem to recall that there is another way to assert this third truth value in Lojban, but I do not recall the method. If this is available, then there is no need (other than shorthand convenience) for the fourth tense; you merely state that *neither true nor false* *actor* *relation* *arguments* ***at all times/places, if you must specify*** [164] by Jeff Prothero (response to [163]) The entire Lojban tense/locator system is (as RLC likes to remind us) a shorthand convenience for things we could say via (appropriate) predicates, and one can certainly handle the fourth tense this way, with or without a third logic value. But you lose symmetry, compactness and compound tenses involving the fourth tense -- you add enough verbosity to practically guarantee that nobody will actually use the fourth tense: "Omit needless words! Omit needless words! Omit needless words!" (Will Strunk.) [165] by Jeff Prothero (general summary) RLC/lojbab wanted a simple rule for using the fourth tense on an everyday basis. I think I understand it well enough now to give supply a twenty-five words or less rule (-: Whenever you find yourself about to use the present tense to refer to the fixed stars, use the fourth tense instead. (21 words!) Rationale: In everyday use, the "present" is usually less than a year long, which puts the fixed stars clearly out in the oblivious zone. With the exception of the fixed stars, contemporary daily life rarely if ever has occasion to refer to events in the oblivious zone.