| In a message dated 9/10/2001 6:25:29 PM Central Daylight Time,
lojbab@lojban.org writes: A point I have been trying to stress whenever I say "let usage <Membership in a selma'o may imply NOTHING about its semantics. Membership in NU means only that it serves as the head of a phrase that includes a bridi and is terminated by a possibly elidable kei.> Which gives some prima facie clues about its semantics, unless the included bridi is just window dressing (not a possibility I considered until the{me mi mo'e moi} fiasco). <How about one where it stands by itself, developing its own meaning? A si'o abstraction "means" NOTHING other than that which fills the x1 of sidbo, a ka abstraction is an x2 of ckaji and a few other predicates, a ni abstraction is an x1 of klani, a sedu'u is an x2 of cusku. Deciding what place structures applied to these abstractions did some constraining ofthe meanings, but relatively little. Arguing whether a si'o is a kind ofka is a philosophical argument, of the sort that can never be settled, because it hinges on whether one is willing to consider an idea to be a characteristic of something (i.e., is the x1 of sidbo an x2 of ckaji?) We are NOT going to agree, and the argument is therefore fruitless.> Well, it is not obvious that we will never agree, since it is not obvious that the argument is philosophic rather than linguistic, in which practical features, for example, might play a role. More importantly, it israrely the case that the conclusion is the only fruit of discussion -- a better understanding of what is involved in a particular abstraction or in abstractions generally may arise, as well as some infrmation about all those places that these abstractions fill and the selbri where they fill them. Seems at least as valuable as another dozen banal sentences in safe constructions. <Saying more or less what I said. Equating the places of two different predicates in some absolute manner serves as a metaphysical restrictionon how we look at the universe. Lojban tries to avoid such metaphysical constraints. Semantic conventions are thus to some extent bad thingsif they are rationalized, because the rationalization will always be in view of some particular metaphysical outlook.> The worst kind of pseudo-whorfian claptrap. The argument it claims to be in accord with was a refutation of the claim that {si'o} and {ka} were alike, in support of the claim that {si'o} and {nu} are alike, so not at all what Lojbab said in the first place and he said nothing like the gist of the argument in the second. That aside, some notion of how the semantics of various phrases works is not at all metaphysically constraining, since -- once we know how things goes -- we can account for them equally well ina variety of ways metaphysically. To be sure, a nice picture helps for a while in formulating a theory, but the picture is not the theory. And the theory is not metaphysical. <Nora and I, with less than full reading of all these intertwining threads of abstractors seem to glorkjunkie that si'o is supposed to be the archetype that you are talking about above. If this is the case thenlo'e is a manifestation of that archetype (and not the archetype itself (which manifestation may or may not actually exist in the real world)). We discuss ideas, and not manifestations of those ideas. More specifically, while lo pavyseljirna may not exist, lo'e pavyseljirna is something that people draw pictures and write books about as if such things did exist, and those things are distinct from the ideas/ideals we have of them: the idea of a unicorn is not going to carry a fair maiden, the manifestation of that idea would do so. Now it might still be the case that we should use something like lo'e tanxe in mi sisku lo'e tanxe, because we probably aren't seeking properties, but manifestations that have the property. This is not quite what Jorge had in mind since it does not equate ka and lo'e, but it might clarify/correct what is intended when we talk about seeking a property (which if people recall was introduced to keep people from searching for noda when they were searching for lo pavyseljirna which does not exist).> Thus avoiding all metaphysics! Ideas, ideals, archetypes, manifestations of archetypes, real things, properties... An ontology to make Plato weep -- and Aristotle too fror the opposite reason. I especially like being able to paint a picture of a manifestation of an archetype -- an presumably not ending with a blank canvas. Now, what does all this mean in terms of linguistic usage? Well, we do have the note that maybe we seek manifestations of archetypes rather than properties, i.e. lo'e broda rather than ka ce'u broda (we can at least keep up with the latest decision of usage). But of course we rarely do; any old unicorn will do, even if it far from a manifestation of the archetype (unless, of course, every unicorn is a manifestation of the archetype, in which case we are probably back at the problem that there are no unicorns, again) <My own preferred but totally unofficial rule for zo'e >is that it is a variable bound by an existential quantifier with >maximally narrow scope, so zo'e are bound within the abstraction, >and hence {ro ka broda cu pa mei}. However, if there is no specific >rule for the binding/reference-fixing of zo'e (and if its reference >can be fixed arbitrarily within the abstraction, i.e so that it can't >be exported to prenex of main bridi), then {na ku ro ka broda cu pa >mei}, because there'd be as many {ka broda} as there are construalsof >the zo'e within it. IMO that would be a Bad Thing, because all >abstractions would become intolerably vague, except to glorkjunkies. Nora opines that apparently then you may be stuck with the glorkjunkie version, because when we use ka anaphorically, we appear to get the result you dislike Thus if we are discussing lo ka ce'u lebna loi titla loi cifnu we might later anaphorically refer to le ka lebna where we clearly may want the zo'es to be carried over indefinitely. on the other hand it isn't always the case that we want the zo'es to carry over.> Anaphora is different from scoping problems: we need only a reminder to recover the whole-- complete with it bound terms. <At 06:04 PM 8/31/01 +0100, And Rosta wrote: >{le si'o ce'u broda kei be mi} = my notion of Broda That might be your notion of "le broda", just as {le si'o broda ce'u kei be mi} might be your notion of "le se broda" But let us turn to some abstractions that people often label as Ideas, like "Freedom" and "Peace". I can't figure out whether the ce'u goes in those or why you would want to use one. Yet I have to claim/concede that ce'uless le si'o zifre kei be mi is not the same as ce'uless le ka zifre because the latter does not have "mi" in the place structure, nor is either of these clearly the same as le du'u zifre though the latter two seem closer than the si'o is to either.> Oh goody! We need another kind of idea, Idea. Actually, I think that this is correct, since the properties we have so far and the ideas, too, tend to be very extensional and we need the intensions as well -- at the risk of being accused of metaphysics rather than Logic at this point. |