From - Tue Mar 26 16:38:02 1996 Date: Tue, 26 Mar 1996 15:56:17 -0500 X-UIDL: 827875268.006 From: Logical Language Group Message-Id: <199603262056.PAA05171@access1.digex.net> To: pcliffje@crl.com Subject: Re: opacity Cc: cowan@ccil.org Status: U X-Mozilla-Status: 0011 Content-Length: 9279 >Sorry, I just don't get it. The various abstractors create a variety of >different kinds of objects (numbers, functions and situations on the >least discriminating story), so how can they all be treated as the same>or at leat have their differences ignored, with su'u? Suffice it to say that I had a different idea of the effect of abstractors at teh time (whether it is valid or not). Just as it is permissiblke tolook at a predicate as a noun, a verb, an adverb, or an adjective, and it was just a different way of looking at the same thing, just as the Aksionsart are just another way of looking at "nu", I presumed that teh abstractors were collectively just another set of ways of looking at the predication. i.e the value/degree to which something is running is "how it runs", so are the properties of the running, so is the fact that it runs, so is the experience of it running (per Frazee). The house of cards that is su'u was built solidly from the fact that "see how they run" attracted so many different and irreconcilible understandings about what was being "seen", that all we could clearly agree upon was that we were examining some abstract nature of the "running". Hence su'u. su'u was not intended to have anything to do with opaque contexts, which were not recognized as an issue at the time. tu'a was not intended to have anything to do with opaque contexts, which were not recognized as an issue at the time. tu'a was designed for our old friend sumti-raising - to mark a hidden abstractor (which again is an abstractor of an undefined nature). tu'a is not limited to nu, or ka, (or su'u) - we have seen it used in all manner of places. The most obvious is djuno x1 knows x2 about x3 ... In most cases x2 is an abstraction, of which x3 is a sumti. But when you try to apply it to normal English sentences, it gets to be hard to figure out what is x3 and what is x3: I know the score x1 knows fi le [score] or x1 knows tu'a le [score] ? In other words, to know the score does not seem to be to know "about" the score - that is not the natural value for x3, which is probably the game that the score is "about". On the other hand, "I know about the cat" means either mi djuno fi le mlatu or mi djuno tu'a le mlatu where the x2 expressed as an abstraction would be mi djuno ledu'u le mlatu cu co'e where co'e is a compound preduicate that somehow combines all of the different relationships, whose facts constitute our knowledge about the cat. Here tu'a is clearly being used for a ledu'u, because by definition the x2 of djuno is a du'u abstraction. Or rather it is a comglomoration of such abstractions since we will likely never be able to accurately form the co'e that is the implied inner predicate (the facts we know about something is an open-ended and ever changing set - by making a sentence about the cat, we have added a fact that we know about the cat - that it was just talked about by us in this sentence). In any event, du'u got coopted into the non-MEX grammar in order to fill the x2 of djuno. But then we found that wwe were using du'u in other places as well, and somethimes they were NOT The abstract "fact" in the same sense. Sometimes we were using du'u to stand for the predication, and sometimes for the words that expressed the predication. (is the x1 of bridi a text, or the predication that the text communicates? On the other hand, when we express that bridi, we are expressing the text of the bridi and not the relationshiop itself (usually %^). WE thus gave du'u the x1 and x2 places to link these two ways that people seem to in natlang look at expression - both as the text itself, and as the meaning of the text. And we now clearly use ledu'u in some places, and le sedu;u in others. Still no opaqueness issues have been raised. Opaqueness never reared its ugly head until all this stuff was olidly in the language for other reasons, all of which I am sure we discussed with you over the phone, since they were added back then when we were talking regularly, and indeed when you were still coming to LogFests. Opaqueness first entered when Cowan tried to deal with "John seeks a bicycle or a fish", which involves both intensionality, opaqueness, and speaker vs object point of view. But we passed off the opaqueness issue, i think, by using lo or loi. Opaqueness never came up again until Iain Alexander challenged specifically the place struture of sisku - and again this was dealt with (by changing the place structure and forgetting the issue. I think at that point, I worked out a meaning for kalte that did NOT allow for opaque contexts, but it was probably invalid, but in any case kalte was not discussed. Opaquensss came up for the last explosive time when you were alredy on the net, and Jorge started looking for a box in Sept 94. That discussion was never resolved, though I had the clear idea that you thought that your comntext leaper would do it - just couldn;t manage to communicate this to the others who kept on bringing up distracting issues that had little to do with opaqueness (like existence of unicorns, typical (lo'e), quantifiers, etc.) >You (all -- I'm not sure who's included, but I volunteer) Your hired. Solution please %^) >need >to decide how to deal with opaque contexts and intensions generally and >stick to it, rather than patching each fractional problem as it comes >along (often without checking up on the last "solution"). I don't think that the number of problems to be solved is all that large - it is just that we are reading the opaqueness problem into a lot of areas where we used to be merely context to call it an "abstraction " >By the way, >du'u, at least as & seems to use it, looks to be another mare's nest. And refuses to use du'u because he sees a rat's nest. No one has yet been able to explain to me what rats nest he sees, other than the fact that he doesn't like turning a predicate into an abstraction in order to satisfy the places where du'u is used. His kuau or whatever seems to be identical to ledu'u or lodu'u or da poi du'u. he does not acceptthat the sedu'u has anything to do with kuau (I think, but am not sure of this). He is generally opposed to one placed prediucates seems to be his basic justification, and du'u to him is like munje is for others, a one place predicate with by definition singular membership in the x1. I just do my best to ignore this, since it seems a philosophical issue - I could probably find several other bridi where all the places are codependent in such a way that one could say that it was a singular one-place predicate. In any event, you are welcome to make pronouncements, and i suggest like in other situations that you, me and Cowan try to agree before we try to convince other people of the decision (I think that part of the problem in 1994 was that just when we got something settled you made an erroneous choice of cmavo out of rustiness, and in trying to interpret it, people discovered distracting issues.) (Nora of course will be indirectly in the discussion at this end - she is up to chapter 4 of Mccawley and is NOT finding it easy going.) lojbab From - Thu Mar 28 11:19:03 1996 (5.65c/IDA-1.5 for ); Tue, 26 Mar 1996 16:40:08 -0800 Date: Tue, 26 Mar 1996 16:29:22 -0800 (PST) From: "John E. Clifford" To: Logical Language Group Cc: cowan@ccil.org Subject: Re: opacity In-Reply-To: <199603262056.PAA05171@access1.digex.net> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0011 Content-Length: 1440 I have borowed an ftp engine that presented me (with a lot of kicking and screaming, but successfully, unlike either of my offical critters) with the refgrammar files from digex/pub -- hopefully the latest set. I hereby apologize about the abstractors, which were the subject of the first text I happened to read. They actually look to be in pretty good shape as reported in the text; the way they get used is less comforting, though, so maybe the text needs to stress good usage more. The only places I was really uncomfortable were with indirect questions and su'u. I think that Bob's note, to which this is offically a reply takes cae of the latter problem and the question problem is just (I think) lingering fear of leka malglico. Of course, I think su'u in most of its uses is lazy, but then that is what most uses in real language are, so so much the better for it. Refgram does need to deal with opacity and maybe it does elsewhere, in which case it should be cross-referenced at abstraction, where most of the opacities lie. I still have ahalf a cause paper in the box and a start on an "other worlds" paper, which included opacity and all that stuff that JCB is busy screwing up right now (by the way, "under condition", if it is a BAI or so, will do a lot the counterfactual work). I think I am now back to more or less full participation, for whatever that is worth. pc>|83