From nobody@digitalkingdom.org Sat May 13 09:14:42 2006 Received: with ECARTIS (v1.0.0; list lojban-list); Sat, 13 May 2006 09:14:43 -0700 (PDT) Received: from nobody by chain.digitalkingdom.org with local (Exim 4.61) (envelope-from ) id 1Fewl5-0005Ri-JL for lojban-list-real@lojban.org; Sat, 13 May 2006 09:14:23 -0700 Received: from web81306.mail.mud.yahoo.com ([68.142.199.122]) by chain.digitalkingdom.org with smtp (Exim 4.61) (envelope-from ) id 1Fewl3-0005RY-PI for lojban-list@lojban.org; Sat, 13 May 2006 09:14:23 -0700 Received: (qmail 83417 invoked by uid 60001); 13 May 2006 16:14:20 -0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=sbcglobal.net; h=Message-ID:Received:Date:From:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=Rg7cZHxOgeTvQ4JCBNGd+3IDQEVqbkVY+Rb+lgTk2LOo/ycquh9Ke9zfi2BR6DYZoNx/73SQQwqxCTRUaExXNh1z9x6H7+Agm5JiQt2w8M1veE650xhMrp1aDg29UYHp1CsOfz1L68eXtARq7qYuWaoGfxfwccBL/eB5wDWhF7s= ; Message-ID: <20060513161420.83415.qmail@web81306.mail.mud.yahoo.com> Received: from [70.237.223.173] by web81306.mail.mud.yahoo.com via HTTP; Sat, 13 May 2006 09:14:20 PDT Date: Sat, 13 May 2006 09:14:20 -0700 (PDT) From: John E Clifford Subject: [lojban] Re: Uses of {le} and {lo} To: lojban-list@lojban.org MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Spam-Score: -0.7 (/) X-archive-position: 11535 X-ecartis-version: Ecartis v1.0.0 Sender: lojban-list-bounce@lojban.org Errors-to: lojban-list-bounce@lojban.org X-original-sender: clifford-j@sbcglobal.net Precedence: bulk Reply-to: lojban-list@lojban.org X-list: lojban-list Maxim Katcharv: < wrote: > --- Maxim Katcharov > wrote: > > > On 5/12/06, John E Clifford > > wrote: > > > --- Maxim Katcharov > > > > > wrote: > > > > I hold this position, and you're arguing > > against it and for only > > Position 2. I'm arguing for including both this > > position /and/ > > Position 2. > > Goodness, why? I took you to be arguing that > this position was incompatible with something > every one does hold, but that we held to this one > too. Now it turns out that you are the one > holding incompatible positions -- by your own > argument. This is all very strange. They aren't incompatible. They can exist mutually, just not at the same time, within the same word. I had thought that you implemented both with {ro}, it turns out that you implement position 2 with {ro}, and position 1 either by saying something nonsensical within the context (which doesn't guarantee success), or by being needlessly verbose ("poi ... and (such that are in our context and such that are not in our context)").>> Well, since I have been mistakenly attributing things to you, I suppose it is only fair for you to mistakenly attribute things to me. But still I don't recognize anything I have said in this brief characterization; perhaps it is there in the long form somewhere. I'll leave it to you to accurately characterize how you want to change ("improve"? "correct"?) Lojban usage. I'll try to do the same for what I take to be current usage -- keeping in mind that some of this is open to disagreement. So, we are talking about bears, maybe very generally, maybe in a fairly restricted way (those in a zoo or an area, those of a single species, et. We may even be talking about specific bears (in which case we can -- but need not -- use {le}). In any case, there will be some bears that count -- that can be cited as evidence for or against claims, for example -- and others that don't (even though, in some sense they are bears). Once this group of bears that count in the discussion is established, we can refer to them (strictly, to groups of them of unspecified size from one to all) by {lo cribe} (and anaphoric forms to get the same group a second time). All of them would be referred to by {lo ro cribe}. But sometimes it is desirable to change the group of bears that count in various ways, to focus down on a subset or to expand to include some previously not included bears. This can be done in a variety of ways, depending upon what the change is and how aware the discussants are of the parameters of the established group. But all of these ways come down to describing the new group, perhaps in terms of the old group, perhaps absolutely. In any case, {lo cribe poi...} or {lo broda cribe} or some combination. this can be done with greater or lesser care: for going to a subclass we may say either just {lo cribe poi ...} or "the bears we have been discussing that are ...", and similarly to expand we can say simply {lo cribe poi...} or "the bears -- whether ot not we have been discussing them -- that ..." We almost always take the first option, reserving the second move for the cases where the first fails to move the right way (for the restricting case we also have {lo cribe ku poi...} as well and for the expanding case we can drop back to quantifiers + variables, but both of these are seen as unnecessarily complex -- even when the first shot at this fails). This doesn't look much like either position 1 or position 2 nor does it look like your proposals nor does it eem in need of "correction" or "improvement." On the last issue, it is true that attempts to change domains sometimes fail, i.e., the hearer ends up taking the domain to be different from what the speaker intended. In any given case of this, there is probably another description the speaker could have used that would have succeeded and, indeed, the speaker -- once he becomes aware of the problem will produce other descriptions until he achieves the change desired. But using the description that finally works at the beginning is not really an option. For one thing, we do not generally know beforehand what it will be or even that the description we do use will fail (else we would not have used it). The description that works will generally be more detailed than the one that failed (though not necessarily -- choice of detail can be as important as number) and to use a lot of details when fewer would work is a violation of convention again. On the other hand, no level of detail is guaranteed generally to succeed, so we might give an involved description and still fail. That is, the move of changing the domain is inherently fallible and no rule about the level of detail involved in cases of change will make it certain (nor even improve its success to effort ratio significantly). <<> Since I am now very worried about what you mean, > I am leery of agreeing with you about anything. > But, yes, following the flow of discourse is a > large part of the way a speaker understands the > less than explicit parts of the conversation. > > > I'm proposing that we give the speaker the > > option to be definitely > > precise, meaning that the listener doesn't have > > to rely on context, > > and can take the words of the speaker exactly > > as they are. > > That is, we force the speaker to go against the > linguistic habits of the last at least ten > thousand years and not leave out anything that > could possibly be needed to fix those previously This is like saying that by introducing predicate relationships with up to 5 (!) arguments, we're going against the grain of the last 10000 years of linguistic habit.>> No, we deal with predicates like that all the time and always have -- indeed with more arguments than 5 (though the frequency declines rapidly above 3 -- as it does in Lojban too). And,even if it were the case that we never had such complex predicate relations, I don't see how this -- a curiosity of vocabulary -- relates to a standing convention of successful conversation. <> And what prevents you? The only objection is to always spelling everything out when there is reason to think that somehting less will work as well (the reasons may be wrong, of course, but that is not enough to justify beginning with the whole load). <<> implicit parts of the conversation. I am not > sure that is possible in general, but it can be > done in some particular (and I think rather > peculiar) cases. And, if it could be done, it <> I think we are talking at cross-purposes here. I am taking you to be saying that we always have to specify in detail and that there are general rules about this, but it turns out that you are only saying that in each case we can find a description that will work -- which no one denies. I do think you are also saying that we should start with that description and, in that case, my point is just that we don't know beforehand which description that is and that going into too much detail is as bad (though in a different way) from giving too little (or, it turns out, the wrong ones). <<> > > > Sure it does. If the listener properly > > understands {L_ cribe} to mean > > an unspecified bunch of bears, maybe one, > > maybe[...], then he has > > correctly picked out all the relevant bears. > > No, not at all. When that is the way things go, > what {lo cribe} is about (it does not pick out > anything -- it is not {le} even then) is some > relevant bears, not necessarily all of them -- So {lo cribe} is {lo su'o cribe} ("some bears") then? I doubt that this is what you mean.>> Well, as xorxes points out, it is a bad principle to insist on implicit quantifiers. But, as far as I am concerned, {lo cribe} in primary occurrences does imply that there are bears being talked about (xorxes differs on this, though just what he means by his contrary cases is a mystery he has kept to himself for over a decade, I think. In any case, they are rather unusual circumstances which seem generally to be accountable for by other, natural means). <<> and no specified ones at that. I don't think that > going off to talk about a few of the relevant > bears makes the rest irrelevant, they are just not > the ones were are talking about at the moment. They're clearly irrelevant to the current talk.>> Not necessarily; it depends upon what is said. And they may come back into the conversation at any point without all the domain shifting we talked about earlier. < > "Nonsensical" is a little strong. It does > violate a convention to restate the obvious, as > this would do if I meant to refer to the > currently relevant bears. Since I am assuming > that no conventions are being violated (always > the proper assumption until the evidence > overwhelms you) it must not be stating the > obvious and so saying something new. To call > this nonsensical is to miss the role that it > plays constantly in conversation; this is a > standard way of changing domains (and context). I understand the role it plays in conversation. It's a very important role, but it's still implemented by having the speaker say something that is nonsensical.>> Well, "rude" maybe and being rude is against the conventions as well. But it is not any othese; it is just the standard move in shifting domains -- specifying the new one (what the old one was does not really enter in). <<> > < moving into some new domain. This is a hack.>> > > For what? It is a standard move and has been for > countless ages past. What is the real move we > keep not using and how did everyone miss it > until you came along? I assume that you're just ribbing me, but I feel that I should mention that it's a fallacy to assume that an argument is wrong because the arguer has no known clout within the field.>> Yes, it is the flip side of ad verecundiam. But, like ad verecudniam, it contains a grain of truth: we expect more in the way of proof from someone who does not have his credentials in order. You have not given us as much evidence as we would expect even from a recognized expert. You have just said that something that we use all the time is wrong but have neither demonstrated this nor given us an alternative that is demonstrated to be free(r) of the flaws of established practice. <> As usual, I have to ask "Who is it that holds that {ro} in these contexts refers to "relevant all" as opposed to "all absolutely"? Who, of course, on the other side really means "absolutely all"?" <<{__ ro cribe poi nenri [this zoo]} {__ ro cribe poi nenri [this forest]} {__ ro cribe} {__ ro cribe poi mi'o ponse} {__ ro cribe poi ...}>> And what do these illustrate? They are now so out of context as to be a mere list of forms, all grammatically correct (up to the point where they cease to be Lojban) and as such meaningful, but what are they meant to mean other than what they do mean (which is a bit hard to specify out of sentential context, let alone larger frames). All but the middle one would be naturals for shifting domains, though of course they have other uses as well. I take it you want to use them only for shifting domains? <<> take care of himself.) What did you do if not > spell out exactly (as much as need be) where you > wanted to go. Maybe your case did not involve > repeating anything obvious, but the move is still > essntially the same. and thus, I suppose, It's not the same. Saying something that is nonsensical within the context sets the context to some undefined sensible context, this method just says "the referent is this. Forget about context, you need not rely on it to know what I'm talking about".>> Yes, that is what I was trying to say by objecting to your "nonsensical." Sorry, if I wasn't as clear as you needed to get the point. To be sure, the incongruity has a role, but it is not what decides the issue. <<> another hack. On the other hand, if you did not These could be considered a hack if they did something not-quite-proper and still got the job done in some way. But they don't.>> Ah, but they do; as witness the fact that we use them successfully all the time. Of course, they fail sometimes, but so do all approaches to this matter. <<> spell out where to go, how does your move get you > there? > > <<> > > > > Yep. These factors are used to help the > > > listener pick out which things > > > are being spoken of. Another (or at least > > > another use of) context is > > > using it to place things relative to it: if > we > > > didn't have this > > > context, we wouldn't know what "now", or > > > "before", or "tomorrow", or > > > "here" meant, because they're all relative to > > > the current context. The > > > latter is necessary just about always, but > the > > > former is only > > > necessary when you aren't being precise > enough > > > (when you aren't > > > restricting enough). > > > > Actually, in the theory, most of these examples > > are relative not to the whole context but only > to > > the occasion of utterance, a relatively > > controllable component of the context. > > > > Sure. My point is that they're part of the > setting-context, and not > the domain-context.>> > > What exactly (Hell, approximately) is the > distinction here. I just haven't seen this > terms before, so far as I can remember. The setting context lets you put things relative to it. Right now, when I say {mi}, you know that I'm talking about myself, or if I say {nau}, you know that I'm talking about the time/place of this writing. That's the setting-context. The domain-context is (roughly), if we've been talking about 20 bears, those 20 bears.>> There are probably better terms for this, but these will do. Typically the setting context is called the occasion (who, where, when, what is uttered). I take it the domain-context is just the domain of discourse, the relevant things. as you notre, this is not too important usually in domain shifting (for obvious reasons). <<> > If your normal means fail, you lack the facility > for this sort of > precision. I try to show this in my caged bears > example. > > But of course you failed to show it, precisely Er, yes, but it's also possible that you failed to understand it or consider it properly. But as far as I'm concerned, since it seems that we're both not crazy, and are rational, each of us has simply (equally) not yet done a good enough job of explaining and/or of understanding. Pointing this out regarding any particular person usually incites something that isn't at all an argument.>> Well, I may have failed to understand your point, but, insofar as I do understand it -- and what you build from it is consistent with my understanding -- you have failed to make your caes in just the way I describe here. What did I miss? <<> because you could set up the example. All that > is needed now is the (possibly extraordinary -- > but I don't really think so ) step of copying out > what you did in setting up the example to > describe the case. Pop! I don't understand what you're saying. I gave you an example where both of these positions could be used, and are used, in what I would consider to be common conversation. If they're both used, then both should be addressed in some sensible way by Lojban, if possible (and it is).>> Now I am afraid I have lost the cases again. both of what positions? using the current domain to partially explain the new domain (how the current one is restrictd or expanded) and defining the new domain without reference to the current domain? But why can't Lojban do it -- as I said -- simply by copying your explanation of what these cases are? And, if it can do that, then there is no problem of doing these two things in Lojban. You owe us some evidence that Lojban can't just do this and I don't see it. At most I see cases where Lojban has not done it one way or the other, but that doesn't show that it can't do it. To unsubscribe from this list, send mail to lojban-list-request@lojban.org with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if you're really stuck, send mail to secretary@lojban.org for help.