Return-Path: <@SEGATE.SUNET.SE:LOJBAN@CUVMB.BITNET> Received: from SEGATE.SUNET.SE by xiron.pc.helsinki.fi with smtp (Linux Smail3.1.28.1 #1) id m0sh1nI-0000ZHC; Sat, 12 Aug 95 00:39 EET DST Message-Id: Received: from segate.sunet.se by SEGATE.SUNET.SE (LSMTP for OpenVMS v0.1a) with SMTP id 058137A1 ; Fri, 11 Aug 1995 23:24:08 +0200 Date: Fri, 11 Aug 1995 13:29:31 -0700 Reply-To: "John E. Clifford" Sender: Lojban list From: "John E. Clifford" Subject: lu'a X-To: lojban list To: Veijo Vilva Content-Length: 4131 Lines: 80 xorxes: There is no problem when the anaphora have a single referent, those are the best sumti there are, they never give any trouble, long live the singular reference! pc: Well, they ain't none in Lojban, but they are a lot easier when they is. Well, hardly any. For the most part, we do have to deal with quantifiers and scope and all that entails xorxes: Anaphorized references are a mess on their own right, and I hesitate to enter into that before having a firm ground to step on. pc: Well, strike "reference," but it is easy to get less than desirable results with some anaphora. Still, remember that "anaphora" just means repetition, so the basic plot is simple, however complicated working it out may be. xorxes: ko'a goi le re gerku ..... ........ ........ i ko'a batci lo nanmu What does the last sentence say? Does it say that each of the two dogs bites its own man, or do they both bite necessarily the same one? In English, a use of "they" there would suggest the same man, but that can only be the case in Lojban if {ko'a} is the mass of the two dogs. I think I would prefer that {goi} have this massifying effect, because then all assigned variables would always have single referents, and it would be much much simpler to deal with them. (In this case, the single referent of {ko'a} would be the pair of dogs, one entity.) pc: If _goi_ is simply anaphorizing then, by the agreed upon interpretation of le re gerku cu batci lo nanmu on the quantifier thread, the last sentence allows each dog to pick its own man, i.e. two men might be bitten. And I can think of no good prior reason to think that _goi_ massifies or otherwise does more than introduce the anaphora marker of choice. Indeed, if it does massify then we get strange results with even the relatively pure cases of individual: ti becomes the mass of this, for example, even if this is not something easily massified (as individuals often are not). So, it looks like it will be hard to get both dogs to hit on a single (well, at least a single) man. Simply going to prenex will not obviously help here, since that prenex is already buried in the scope of a universal (I would say -- as Lojban does) and so cannot be unsubordinated equivalently. Maybe the leaper works, but I do not see what consequences will follow if we say it does and this is an area of strange kinds of dangers. Further, we do not want the man to be totally independent of this pair of dogs, as the leaper might all too easily make him; even the one man is dependent upon which dogs they are. There is not (we hope) one man who gets bitten by whatever pair of dogs takes to biting men. Oh for referential modes or Skolem functions! BTW the referent of ko'a under the massifying effect of goi would not be the pair of dogs but their mass, still one entity. Notice then that in the massifying interpretation ko'a batci lo nanmu would be true if only one dog bit only one man, masses bein g the sort of things they are (note, NOT "the sort of things pc has chosen to define them as"). xorxes: If {goi} does massify (which I now think would be the best thing to do) then {lu'o ko'a} would get a mass of the mass, which is as far as I can tell indistinguishable from the simple mass, so that would even agree with pc's theory. pc: The mass of a mass is, like the massification of other individuals, not something of which we have a very clear sense. It is different from the original mass itself, since it has different components (it has only the one, the mass, while the original mass has all the whatevers that were brought together in it, typically more than one and surely not THIS mass). (BTW, xorxes' interpretation of the lu'a series is starting to have a number of useless forms, by his own admission, just as he claims the more-or-less official version has. Is there a different system here that gets what xorxes wants without cluttering up the usefulness of lu'a, etc. and which does not keep generating odd cases?) xorxes: Multiple reference will breed trouble. pc: So do singular ones, if anaphora is massifying. pc>|83