From lojban@cuvmb.bitnet Sun Mar 31 23:57:53 1996 Received: from punt4.demon.co.uk by stryx.demon.co.uk with SMTP id AA11138 ; Sun, 31 Mar 96 23:57:49 BST Received: from punt-4.mail.demon.net by mailstore for ia@stryx.demon.co.uk id 828156746:08417:0; Sat, 30 Mar 96 03:32:26 GMT Received: from cunyvm.cuny.edu ([128.228.1.2]) by punt-4.mail.demon.net id aa07824; 30 Mar 96 3:30 GMT Received: from CUVMB.CC.COLUMBIA.EDU by CUNYVM.CUNY.EDU (IBM VM SMTP V2R3) with BSMTP id 5241; Fri, 29 Mar 96 22:30:04 EST Received: from CUVMB.COLUMBIA.EDU by CUVMB.CC.COLUMBIA.EDU (Mailer R2.07) with BSMTP id 0570; Fri, 29 Mar 96 22:30:41 EDT Date: Fri, 29 Mar 1996 19:15:43 -0800 Reply-To: "John E. Clifford" Sender: Lojban list From: "John E. Clifford" Subject: Re: change 46 X-To: lojban list To: Multiple recipients of list LOJBAN Message-ID: <828156687.7824.0@cunyvm.cuny.edu> Status: R So the answer to my second inquiry (what does the branching quantifier use of the q broda/q lo broda distinction have to do with the existential iport use of the distinction) is "Nothing at all and further the existential import use has not been established" On the last part, I am sorry, but I thought we had come to as near to an agreement as we ever seem to get that one of these two would follow the q da poi interpretation and the other the q da ganai scheme (restricted and unrestricted) and the only problem left was which was which. Sorry I was wrong, that is, and that we still have not gotten agreement about anything. Still, the existential import issue only concerns "all" (and maybe "no" and "not all" -- an undecided issue itself) and the branching issue only strictly effects numerical quantifiers, there are no neceessary conflicts between the two uses, though they are about as related as putting h and ng in the same phoneme in English. And there would be a conflict if we wanted to use the q/q lo distinction as a general afterthought device, with "all," particularly. So we do not seem to have a general afterthought brancher yet, even with John's new device. Nor do we have one even in prenex: xorxes' original use of connectives does not work, on transformational grounds, though I think he did shift from e to joi, which might work (again in an apple cum orange sort of way). John's device will work for prenexes and is not markedly worse than joi -- without the possibility of misinterpretation. And it will work for clumps, so only unclumped sumti remain a problem. Or, of course, we can do without any short form for restricted quantification (or unrestricted, but I don't see that option getting very far) and use the q/q lo distinction just for the the branching problem. (By the way, why does no one ever point out that le q is always on an independent branch?) In this case, John's new device would only be useful in prenexes, since the other device works better in the matrix. The answer to the my first inquiry (how does q/q lo solve the branching problem?) is, as often, John's fiat, without any one else being involved and independently of (even at variance with) other discussions in the community. And now for the theoretical question: how do we deal with quantifiers that branch relative to some other but not to all? In prenex, John's device can handle these to just about any level of complexity that I have found, but none of the devices seems to work regularly with unclumped sumti in afterthought mode. Nor would any of the leaper devices, come to that. pc>|83