On Mon, Oct 22, 2012 at 3:16 PM, John E. Clifford
<kali9putra@hidden.email> wrote:
Well, that does clarify things a bit: dz is a predicate that extends the scope of a previous binder into the next sentence. Note that it is strictly unnecessary for l (and practically unnecessary in any case).
Here's the way I view it: Predicates like dz and mb=mslf and other anaphoric/deictic predicates are predicates whose semantic value vary from sentence to sentence in the course of discourse. For example, the semantic value of "la'a mba'a" translated as E "I" or "me" depends on who is uttering the sentence. Likewise, the meaning of the predicate "dz" varies depending on what was said in previous sentences. That does not mean that "dz" changes the meaning of previous sentences; it'd be more accurate to say that sentences change the meaning of "dz" as discourse moves along.
Li rlci dkseki is sorta ridiculous; it assumes an ordered class with our group off at one end. It would be more natural, it seems to me, to make the modification in the property rather than the class. If you are worried, as you seem to be, about using up letters, you would be better finding a way to talk about properties -- in both senses -- than this bit of frill.
I have not yet seen an good example in X in which argument places demanding L "ka" are demanding a distinction with an actual difference. Think of L "ka" as being built into the argument place if that seems more agreeable.