From LOJBAN%CUVMB.BITNET@uga.cc.uga.edu Sat May 13 14:34:35 1995 Date: Sat, 13 May 1995 14:37:23 EDT From: jorge@PHYAST.PITT.EDU Subject: Re: A modest proposal re: tenses To: Bob LeChevalier Message-ID: > It seems to me that a tense like {zu'avi} expresses a relationship between > three things: an area x1 to the left of the reference point x2, a distance > x3 (which should be small), and maybe a direction x4 (if you want to be > more precise about "to the left"). I'm not sure I see what you mean by x4. You mean where to face to determine what is "left"? > There ought to be a way to express > this relationship succintly as a selbri; The gismu {zunle} already exists, and most if not all FAhAs have a corresponding gismu. It doesn't have a place for the distance, though. > something like what {moi} and its > relatives do for numbers. Why not add another cmavo (with its own > selma'o) to serve this purpose? That sounds like a good idea, but it seems too radical. Given that there are already gismu that do that. I like it, though. It goes well with the idea of letting all words function as any part of speech. > (In this message, I'll arbitrarily use > {xoi}, as the only unassigned oi cmavo left.) All xVV and xV'V cmavo were left unassigned precisely for this purpose, to try out new ideas. But if the cmavo is adopted, it would be as a real one, which would have to be of CV'V form, since only of those are there any left. > This would serve at least two purposes: > > a) It would allow more complicated tenses, expressing a specific distance > and/or reference point (using di'o: {di'o lo be'avixoi be le ginka bei > cimitre} Nice. I wonder what's the difference between {bu'u} and {di'o}. Now you can say {di'o lo berti be le ginka}, but expressing the distance would be harder. > b) It would allow you to succintly say something like "the area around > the oven", as I was wondering about: > > litru lo ru'uxoi be le toknu You can say {litru lo sruri be le toknu}. > (I still think the use of {litru} is appropriate here; I don't think > {litru} should carry the connotations of "long distance" that English > "travel" has. I agree, that's why I think the keywords are sometimes a bad idea. I have to make an effort to not let "travel" interfere. > You could also precisely specify a location in reference to your current > position (the default x2 should be the current place, {lo cabu'u > zvacab}): zvacabna > le ri'uxoi be fi remitre > > or "two meters to my left". I think I like it, it would save us form having to learn two words for every direction: ri'u, pritu, zu'a, zunle, be'a, berti, etc. On the other hand, it would be a considerable change, so I don't know. > NOTES: > 1) Of course, the same construction works for time or mixed tenses as > well. > > 2) But motion tenses don't seem to fit in quite as nicely, though perhaps > they still have meaning. Also it's not so clear what it would mean with ZEhAs, ZAhEs, etc. Only for PUs and FAhAs it is clear. > 3) Some of the same holes could be plugged with another cmavo of selma'o > NU, a "position abstract"; then you could say, for instance, "the area > around the oven" with {sruri}, and perhaps using {mitre} in other cases. > {zvati} doesn't have enough places to be useful here. (Naturally, we'd > need to resolve the {li'i} question first.) I'm not sure I understand this point. {le sruri be le toknu} works as is for "the area around the oven". {le sruri} doesn't need to be a material object. > 4) I still see problems saying something like "two feet north and > five feet west". You could say: di'o lo be'axoi be fi lei re gutci be'o je vu'axoi be fi lei mu gutci but I will stick with: be'aku va lei re gutci vu'aku va lei mu gutci > co'o mi'e. dilyn. co'o mi'e xorxes >From lojbab Subject: belated ressponse to Jorge: bu'uvi Responding to Jorge (delayed): >la lojbab cusku di'e >> Actually, having checked, "vi" is normally an ellipsis for "bu'uvi", and >> "zi" an ellipsis for either "bazi" or "puzi". > >What does {bu'uvi} mean? To me, adding a VA to {bu'u} makes as little >sense as adding a ZI to {ca}. They are allowed, because adding >magnitudes makes sense with the other PUs and FAhAs, but semantically >they don't make much sense. What could be the meaning of {bu'uvi}, >{bu'uva} and {bu'uvu}? Or of {cazi}, {caza} and {cazu}? In English, when he say "here" or "at", or when we say "now", we often to not mean exactly what we are saying. Sit "at" the table means to sit close to, but not necessarily so close to even be touching, the table. When I tell my kid to do something "now", I do not mean that they necessarily have to be doing it as I speak, but rather that there is a certain probably small range of time in the near future that I expect them to get started during - the action may continue some time into the future. To people outside the US, I can say that I live "in Washington DC", though I live in a suburb 15 miles outside the city. I thus live "vi" the city to them. To a local, though, I live "va" the city - I am not that close; indeed anyone even a few miles further out has no direct subway or bus connection into the city, and to city residents will be considered "vu" the city. VA and ZI are thus highly dependent on circumstance and point of view. >{vi} is not used as {bu'uvi}, at least I don't understand what {bu'uvi} >means. {vi} is used as the simple {bu'u}. Unqualified "bu'u la Washington DC" means inside the city limits to me. >> > mi ba xruti zi lei mu mentu >> > I'll be back in five minutes. >> >> and I would just use "ba'o lei mu mintu" or more likely "ba'o le mintu >> mumei". > >Yes, but they mean different things. In my case, I'm saying that after >the five minutes I will be back, but maybe I arrived before that. You >are saying that at some point after the five minutes you will be back, >maybe one hour later. Also, you have to assume that the five minutes >are the ones that come right after now. I grant you that that is >probably quite grokable from context, but in other cases one might want >the possiblity of being more precise. I will grant that ba'o claims that you will not arrive before the 5 minutes. I doubt that "zi" would help, since xruti is the returning and not the state of being returned. ba'o does NOT allow an hour to pass by in any normal context. Perfective tenses are strongly tied to their events. ba'o means the immediate aftermath of the 5 minutes, while that 5 minute period is still clearly relevant to ones framework. Very unlikely an hour from now. >Also, consider a slight variation: > > mi ba xruti ba le nu ko'a xruti kei zi lei mu mentu > I'll be back five minutes after he's back. > >What do you do when you need to specify both the origin and the magnitude? For exactitude: mi xruti ba'o le mintu mumei pe ba'o lenu ko'a co'u xruti I return in the aftermath of the 5-minute period after he finishes returning. I would probably actually just say "pe ba lenu ko'a xruti". >> But if "zi" was to be used as you suggest then >> by correspondence "vi" as a tcita would have to label a distance. > >Yes, it would have to, but since {vi} is exceptional even now, my >proposal doesn't really affect its status. {vi le ckafybarja} should >mean by the current interpretation "very close to the cafe", but it is >understood to mean "at/in the cafe". So, if we keep using it like that, >it will be "wrong" whether my proposal is accepted or not. When I say that Nora is "at the store", I do not know for sure that she is inside the store premises. She might be in the parking lot, etc. We commonly use "at" to mean "at or relatively near", and that is my interpretation of "vi". >> I don't understand lambda calculus in the least, but if the need in >> lambda calculus is for more than one variable, THAT sounds like a good >> use for a KOhA with subscripts. > >I certainly prefer a new KOhA to a new PA to do the job, but I doubt >that it is worth it. Even in the examples I gave, {ke'a} could be >ellipsized quite easily. When there is no {ke'a}, and one is needed >to make things clear, the obvious assumption is that the {ke'a} goes >in the first empty slot, so: > > la djan zmadu la maris le ka dunda > John exceeds Mary as a giver. > > la djan zmadu la maris le ka te dunda > John exceeds Mary in being given somnething. > >I don't really think it is a problem to use {ke'a} for this, and I hate >to introduce a new cmavo that is probably practically never going to >get used. However, if there is going to be one, then I'd much rather >have it in {KOhA} space, and not as a modifier of {da}. I'll second this. I tend to want to avoid using "da" unless I wish to explicitly invoke quantificational logic. Chris: >Lojbab: >>and I would just use "ba'o lei mu mintu" or more likely "ba'o le mintu >>mumei". > >But this depends heavily on the use of the intensional; we only know >from context that it's the five mintues beginning in the present which >precedes the act of returning. Probably that's good enough most of the >time, but it's funny that there's no reasonably concise way of talking >logically and unambiguiously about time intervals. How could we say >this without depending on intensionality? Difficult, because 1) there's >no word for "now", and 2) there's no convenient way to talk about the >span between two events. Almost all use of tense is "intensional", as I am sure pc can confirm. If you wish to be logically precise, make your claims about time as separate logical claims. "now" is "cabna ti", which I think even works in relativistic contexts, since "ti" is associated with the speaker's point of view. The time between two events can be specified precisely using temci and the interval non-logical connective. I covered this in my impromptu response the other day. lojbab >From lojbab Subject: Re: Questions >From: jorge@PHYAST.PITT.EDU >Subject: Re: Questions > >> >- the box is bottle-shaped, making ta1 both a bottle and a box; >> >- the box contains one or more bottles; >> >- the box is made of one or more bottles; >> >- the box is the contents of one or more bottles; >> >- the box is the material used for making one or more bottles. >> >> The 4th and 5th examples here perplex me; they violate the only >> meaning I could give to the earlier quote. Anybody? > >The x1 of {botpi tanxu} is a box. I suppose that the 4th example means >that it's a box inside a bottle, a "bottled box", although {se botpi >tanxe} might be more clear. I don't understand the 5th example either. For the 4th, I presume the difficulty is the "one or more". Imagine a nested series of bottles, with a box in the smallest. Actually not a bad description for a matroushka doll, now that I picture it, but I was deliberately going for weird interpretations. For the 5th, I can imagine either cutting up a box and reforming it into a bottle, or simply bottling up something in a container composed of (perhaps glued together) boxes. >> I was asking if it's permissible to use {botpi tanxu} for the fourth >> meaning, or if {se botpi tanxu} is necessary (by the principle that >> "the first place of {botpi} in some way restricts the relation". > >The meanings of tanru are very loose. The only clear rule is that the >places are those of the last component. I would think that the {se} or >{te} are necessary for clarity, but you probably can't totally exclude >those meanings without them. Current thinking tends to require the se/te markers, but the original principle was that such words can be left out in lujvo or a tanru. I think that when no places are being specified for a modifier portion of a tanru, it should not necessarily be considered "bad". I believe that Nick's paper on lujvo place structures uses the presence or absence of the se/te components to make distinctions in place structure. I think that at the moment, Lojban stylistics are leaning to extreme literality. Literality is thhe direction I would prefer things to lean (as compared, say, to Michael Helsem's "Purple Lojban" poetry which is extremely figurative), but perhaps not to the current extent. Everyday language use will probably tend to loosen the extremes to which we seem to be prone right now - it certainly seemed to be the case in our weekly conversation sessions. >> >From the "Diagrammed Summary of Lojban Grammar", line 1619: >> >This construct may be combined with the modal construct discussed just >> >previously to identify a sumti: >> > >> > la djan. ne pu la mark. [ge'u] [cu] melbi tavla [vau] >> > -------- < >. -------- | =========== >> > John, who was (incidentally) before Mark, is a beautiful-talker. >> >> Doesn't this show exactly the confusion about {pu} mentioned earlier? >> {la mark.} is not an event. > >I agree with you. It might mean, I suppose, that John lived before >Mark was born. Why is "la mark." not an event? First of all, it is a named thing, and it is possible that the speaker is simply labeling some event "Mark" (which could be a lifetime, or it could be an act of speaking). If you grant that you can label an event with a name of course, then the default assumption is indeed likely that the event named "Mark" happens to be the lifetime of someone named Mark. Now I agree that "tu'a la mark." might be more logically explicit, but I am not sure that it conveys any additional information - you've simply explicitly said that Mark is a place in some event, and the time comparison is with the event. But it says nothing more about what kind of event (an act of speaking, or a lifetime), so why not just keep things simple. >> What's the proper >> tense indicator for "present, and continuing", without implying any >> termination? > >I would say {caca'o}, to me {ca'o} doesn't imply termination, but others >will disagree (my views on ZAhOs are a bit unorthodox). It is correct that ca'o doesn't logically imply termination, if for example we are dealing with some sort of perpetual state. But all of the ZAhO tenses reflect the perfective paradigm, which looks at events as wholes. To a considerable extent, perfectives are almost certainly going to be pragmatically interpreted the way they are in natural languages, barring contravening context, which includes an implicature of completion (or at least of entirety to ca'o as a tense). I would simply say "ca", which means present, and makes no implication at all about the future. Implications to stress an expected continuation/duration could include an interval marker - "caze'u" meaning that it is continuing over a long "present moment". lojbab