Date: Sun, 16 Nov 1997 10:06:06 -0500 (EST) Message-Id: <199711161506.KAA09539@locke.ccil.org> Reply-To: bob@rattlesnake.com Sender: Lojban list From: bob@MEGALITH.RATTLESNAKE.COM Subject: Re: `at least one ' vrs `one or more' X-To: lojban@cuvmb.cc.columbia.edu, bob@rattlesnake.com To: John Cowan In-Reply-To: <971115214328_-524820230@mrin52.mail.aol.com> (Pycyn@aol.com) Status: OR X-Mozilla-Status: 0011 Content-Length: 4349 X-From-Space-Date: Sun Nov 16 10:06:09 1997 X-From-Space-Address: LOJBAN@CUVMB.CC.COLUMBIA.EDU Sorry, but this is the LOGICAL language and that means (among other things) that the sentences DO come out as examples of textbook logic. Yes, this is a logical language, but you have not answered my question, which is whether {lo} always expands to {da poi}? As far as I can see, {lo} is a peculiar kind of operator, and does not *always* expands to {da poi}. In particular, {lo} as defined does not always refer to all of its referents, whereas {da poi} does. Yes, it may be the case that bob's arguments seem merely to be bogged down in the ambiguities of English ... but neither you nor _The Complete Lojban Language_ have yet made a clear case. Indeed, I am basing my thesis on _The Complete Lojban Language_. (Incidentally, I agree that if {lo} always expands to {da poi}, then you are right, as are Jorge and everyone else who has commented on this. My question has to do with whether it is proper to claim that {lo} does always so expand.) (Also, by the way, I agree that {to'e nelci} is best for dislike compared to the other forms; and that {loi} would be the term of choice when seeing or liking some cats.) First the argument that {lo} does expand to {da poi}: * {lo} is a `veridicality operator'; that is to say, when you hear {lo} in an utterance, you know that you and your interlocutor have agreed, explicitly or implicitly, on a procedure for determining that the referents actually meet the description of the sumti, and that there is at least one of them. Since {lo} promises there is at least one entity that meets the description (in the universe of discourse agreed upon), and since the description constrains the referent, {lo} does expand to {da poi}: {da} implies existance, and {poi} attaches subordinate bridi with identifying information to {da}. On the face of it, this looks fine and it is certainly how nearly everyone has interpreted this. I agree it is fairly convincing. But now the counter argument: * {lo} is a `veridicality operator', as stated before. {lo} does not tell you the number or try to specify the number of the referents; all it guarantees is that there is at least one referent that meets the description of the sumti. Also, and this is critically important, {lo} do *not* guarantee to refer to *all* the referents. This is where {lo} differs from {da poi}. {da poi} *does* guarantee to refer to all the referents. The reason I interpret {lo} as not necessarily referring to all its referents is the that common English glosses for {lo} are `at least one' and `more or more'; and the default value in Lojban is {su'o lo ro} `at least one of all of those which really are', which is quite different from `all of those which really are'. (See Chapter 6.7) (As Cowan says, this default of {su'o lo ro} is not hard and fast; nor am I saying that all determination procedures do not refer to all; only that there is at least one with that characteristic.) It is clear and agreed that if .i mi nelci da poi mlatu then the negation is `I don't like any cats'. No argument here. As PC says, "for every cat x, it is not the case that I like x." But the question revolves around the negation of mi nelci lo mlatu where {lo mlatu} means `some number of real cats, but not necessarily all'. The negation is translated as "for some number of cats x, not necessarily everyone, it is not the case that I like x." This does not mean `I don't like any cats'! Note I am not saying you cannot ever expand {lo} to {da poi}; indeed, perhaps I should expect people to do that most of the time. What I am proposing here is that {lo} is a `veridicality operator' with certain characteristics that are different from {da poi} which is also a kind of `veridicality operator'; and {lo} is not simply an abbreviation for {da poi}. Moreover, what I am proposing appears to me to be consistent with _The Complete Lojban Language_, and the other interpretation is not. I am sure that if John Cowan had meant `all' rather than `at least one of' he would have said `all', as he did with {le}. -- Robert J. Chassell bob@rattlesnake.com 25 Rattlesnake Mountain Road bob@ai.mit.edu Stockbridge, MA 01262-0693 USA (413) 298-4725