From @YaleVM.YCC.YALE.EDU:LOJBAN@CUVMB.BITNET Fri Mar 19 07:11:10 1993 Received: from YALEVM.YCC.YALE.EDU by MINERVA.CIS.YALE.EDU via SMTP; Fri, 19 Mar 1993 20:06:51 -0500 Received: from CUVMB.CC.COLUMBIA.EDU by YaleVM.YCC.Yale.Edu (IBM VM SMTP V2R2) with BSMTP id 2964; Fri, 19 Mar 93 20:05:38 EST Received: from CUVMB.BITNET by CUVMB.CC.COLUMBIA.EDU (Mailer R2.07) with BSMTP id 4716; Fri, 19 Mar 93 20:06:51 EST Date: Fri, 19 Mar 1993 12:11:10 -0500 Reply-To: John Cowan Sender: Lojban list From: John Cowan Subject: Re: TECH: QUERY re cmene X-To: Lojban List To: Erik Rauch In-Reply-To: <9303190353.AA25237@relay1.UU.NET> from "Mr Andrew Rosta" at Mar 18, 93 09:15:14 pm Status: O Message-ID: <-q5NwPGdEoF.A.w6.E30kLB@chain.digitalkingdom.org> la .and. cusku di'e > I get the e/o opposition (sort of) and the _/_i/_'i is individual/mass/set, > (right?) Right. > [what are le'e & lo'e?] "the stereotypical" and "the typical", respectively. > but I don't see an a/e/o opposition. o: veridical, that which really is ... e: descriptive, that which I refer to as ... a: onomastic, that which I call by the name of "..." > Why can't words of the e & o series be descriptors of cmene? And why are the > cmavo structured in a way such that the e/o element is excluded from the > descriptor of cmene? If by "cmene" you mean names of whatever flavor, because the opposition veridical:descriptive has no function with regard to names, which are assigned at the whim of the speaker. All names are inherently non-veridical: the predicate "cmene" has (in addition to the name and the benamed) a third place, the namer, which in the case of Lojban name-arguments is invariably the speaker. "la djan." is that which >I< call "John"; should I wish to refer to the thing(s) which George calls John, I must say: le/lo se cmene be la djan. be la djordj. that benamed "John" by George There is simply no difference between what I describe as having such-and-such a name (according to me), and what really does have such-and-such a name (according to me). Per contra, if you mean "cmene valsi", then the answer is historical. Historically, Loglan names appeared in two context: "la " and ""; the former usage produced an argument, the latter a vocative. Lojban has abandoned the bare-name-as-vocative convention in favor of requiring one of several vocative particles of selma'o COI or DOI. An attempt was made at one point to make cmevla the grammatical equivalent of brivla, as you suggest. The practical inconveniences of this scheme, however, were found to outweigh the theoretical advantages. In particular, such a sentence as: la djan. cadzu was no longer a bridi, but a description: the one named a ("John" type of walker) While this is a clever way of rendering "John Walker" into Lojban, which now must be done as "la me la djan. cadzu", the resulting breakage of existing text was too severe to undertake. > Are there cmavo meaning la'o+la, la'o+lai, la'o+la'i? "la'o" itself is "la"-like. To get masses or sets of foreign-named individuals, the converters "lu'o" and "lu'i", which create masses and sets respectively, are the right thing: "lu'o la'o zyz. And Rosta .zyz." is the mass of those named "And Rosta". > OK, so that makes _.And._ a predicate, and also it makes _gismu_ in > _la gismu_ a predicate, but one with a different meaning from the gismu > _gismu_. I think that on the semantic level you are correct; this is a place where Lojban semantics are not fully exposed by the surface level. (What do you want, egg in your beer? :-) ) > Are all cmene unary predicates? I think so, yes. Of the three places of "cmene", x1 is the name (filled), x2 is the thing named (the only surviving place), and x3 is the namer, which in the context of name-arguments is invariably the speaker. (The speaker may derive his name from some remote source, as is usually the case, but is also free to change it by whim.) > My point is that if _(la) .And._ really does have a sense, & this > sense is "entity-named-_And_", so that the use of _la .And._ to > refer to me involves implicit quantification, then > > mi cu (la) .and. > le (la) .and. cu prenu > > ought to make perfectly good sense. They do, if you remove the "la", which is the sign of an argument: they are simply not good grammar. > > lai .and. cu prenu > > the-mass-of-those-named "And" is-a-person > > And _la'i_ would be the set of those named "And"? Yes. > > Since "la lojban." delivers an argument, it may be coerced into a predicate > > by the general method of prefixing "me", which coerces any argument > > into a predicate. > > Does me take a single argument, or can it have an open-ended quantity? A single argument. The resulting predicate has the place structure: x1 pertains to in aspect x2 -- John Cowan cowan@snark.thyrsus.com ...!uunet!lock60!snark!cowan e'osai ko sarji la lojban.