From LOJBAN@CUVMB.CC.COLUMBIA.EDU Sat Mar 6 22:50:50 2010 Return-Path: Delivered-To: veion@XIRON.PC.HELSINKI.FI Received: (qmail 967 invoked from network); 7 Mar 1997 03:22:44 -0000 Received: from segate.sunet.se (192.36.125.6) by xiron.pc.helsinki.fi with SMTP; 7 Mar 1997 03:22:44 -0000 Received: from segate.sunet.se by SEGATE.SUNET.SE (LSMTP for OpenVMS v1.1a) with SMTP id <11.AAC81AC3@SEGATE.SUNET.SE>; Fri, 7 Mar 1997 4:22:44 +0100 Date: Fri, 7 Mar 1997 00:18:19 -0300 Reply-To: "Jorge J. Llambias" Sender: Lojban list From: "Jorge J. Llambias" Subject: Re: respectives X-To: lojban@cuvmb.cc.columbia.edu To: Veijo Vilva Content-Length: 4687 Lines: 118 Message-ID: la lojbab cusku di'e > >> 1a The people each read **a different** book. > > le prenu cu tcidu lo frica cukta > >No problem here, because the scope of "lo" is within that of "le". You > >cannot however rephrase as: > > lo frica cukta cu se tcidu le prenu > >which does not mean: A different book is read by each of the people. > > I resist this kind of scoping difference, though I have not analyzed it. How could I hope to argue against that? :) > Not trying to override the refgram, or even what "needs to be", I > presume in general that all sumti are equal scope regardless of order > (possibly excepting this when there are explicit external quantifiers). What do you mean here by "equal scope"? Are you saying that in le prenu cu tcidu lo cukta Each person read at least one book. there is no subordination of "lo cukta" to "le prenu"? So they all must have read at least one book in common? Actually, it seems to me that what you mean is that there is no way of telling how each of the people relate to each of the books from that phrase, i.e. you understand it the way I understand this other one: lei prenu cu tcidu loi cukta The people read some books. Here I don't specify any relationship between each individual person and any book. All I say is that the group of people I have in mind is in relationship "tcidu" with some books. > If I want to specify scope, I put it in the prenex. This may make me > "wrong", but I'll probably need to be corrected a LOT of times before I > accept it, because it does not fit my internalization of the language. I don't really know what is your internalization of the language, so to argue against or to agree with it I will have to wait to see examples of use. If "le prenu cu tcidu lo cukta" and "lo cukta cu se tcidu le prenu" mean the same to you (when talking of more than one prenu) then we are not speaking the same Lojban. To me the first means that the people read at least one book each, the second that there is at least one book read commonly by all. If we disagree on this really very basic sentence, I don't think there is much point in discussing more subtle cases, is there? > Furthermore, many BAIs and even more so discursives were added to the > language specifically to allow a short cut avoidance of what logical > precision might require. Especially this tricky scope stuff. Could you give a list of the many BAIs that can be used to avoid logical precision, please? The only one I was aware of that was proposed to do so is pa'a. As an aside, for example in the case of ki'u with negation, the refgrammar goes to great lengths to explain how logic is preserved in spite of the direct English translation suggesting otherwise. "mi na klama le zarci ki'u le nu mi xagji" does not mean "I don't go to the market because I'm hungry", but rather "it is not the case that (I go to the market because I'm hungry)". So what is the list of the many BAIs that do break the usual rules? > >> Mark Vines responds: > >> > >> 2a le vrici mamta be la xorxes. je la .and. > >> > >I don't think that works. In 2a you're describing each of the things as > >being vrici mamta of both Jorge and And. > > The "je" might > cause problems, though, since it implies expansibility into two > sentences. So you agree with me that it should be grammatical, though? ui!!! > >> > > 2a the mothers of Xorxe and And > >> > >> le mamta be la xorxes .e la .and > > > >I think that has to be a mother of both. It does not expand to le mamta > >be la xorxes ku e le mamta be la and. > > Why not? Because {le mamta be la xorxes e la and} has to be some {da} such that {da mamta la xorxes e la and}, i.e {da mamta la xorxes ije da mamta la and}, the same da mamtaing both of them. This is not equivalent to {le mamta be la xorxes be'o e le mamta be la and}, which allows for two different mamta. > The plausible alternative is that it expands to "le mamta be la xorxes > je mamta be la .and.", It doesn't exactly expand to that because that's a tanru, but that is approximately the meaning, I agree. > since the "le" is attached to a unitary > description. What unitary description? It could be many of them, as long as each is mamta of both. > But I suspect that we never have ruled between the two, > and indeed could not find an example of be with connection in a quick > search of the refgram. The refgram is extremely careful with these things. You can hardly find any example where "le" refers to more than one thing, or where "lo" is used in sentences with more than one sumti. Most examples in the refgram are trivial with regard to scope issues. co'o mi'e xorxes