Re: ciska bai tu'a zo bai

[Nick NICHOLAS <nsn@vis.mu.oz.au>]

Hu'tegh! nuq ja' Jorge Llambias jay'?

=>For some concepts, we have an abstract sumti place and a concrete sumti place,
=> where the concrete is an argument of the abstract's predication.

=Ok, for example the x3 of djuno is an argument of the predication
=that goes in x2.

=> For some of
=> these, using a different argument of the abstract in the concrete place makes
=> for a different meaning; in that case, we don't have raising.

=This sounds interesting, but I can't think of any examples. What do you have
=in mind?

Well, consider kecti.

.i mi kecti la xorxes lenu la xorxes .ei tavla la nitcion
.i mi kecti la xorxes lenu la xorxes .ei tavla la nitcion

The x2 is an argument of the x3. (Typically; I'd say it's safe to say it always
will be at some level.) If this was raising, then the two sentences would
mean the same. They don't. Contrast with the old definition of rinka:

.i do rinka la xorxes lenu la xorxes tavla la nitcion
.i do rinka la nitcion lenu la xorxes tavla la nitcion

These days, we accept that these two mean the same thing; therefore, this is
raising; therefore, we prefer to say tu'a do rinka lenu la nitcion tavla la
xorxes.

You could argue that they don't mean the same --- that the first implies that
Jorge was the target of some influence. We would probably counterargue that
in that case, the predicate you're after is not rinka, which is only a
cause-effect relation between predications/ states of the world, but xlura.

=Although this issue is also interesting, I was
=thinking of something else. When are object and event allowed _in the same
=slot_, rather than when is there a slot for each. {nelci do} is allowed,
=while {do pluka} is not. Why?

Well, the argument is still based on raising. In pluka, we would argue that
it's what the pleaser *does* which pleases --- that there is no meaning
difference between {do pluka mi ri'a lenu do vitke mi} and {lenu do vitke
mi cu pluka mi}. In that view, the latter phrasing is somehow the 'real'
phrasing, and the former is just a syntactic derivative.

But with nelci, you can like entities qua entities, indepedent of what they do.
.i ka'e go'e .i mi nelci do .enai lenu do roroi darlu

And the distinction doesn't look all that convincing any more, now that I
stated it like that: is it that you can like someone without necessarily
liking something they do? Anyone want to come to the rescue? We had extensive
reasoning of this sort going on last year.

=> All the
=> syntax textbooks I see treat "He seems to be cold" as a raising from "It
=> seems that he is cold", and if they buy a semantic deep structure, it will
=> be SEEMS(COLD(he)). If we acknowledged raising here, we'd say {lenu mi
=> lenku cu simlu} --- since there is no obvious difference between {xy. simlu
=> lenu catra .y'y} and {.y'y simlu lenu se catra xy.} Well, um. There is a
=> difference, isn't there?
=Only one of focus, which could be marked in some other way.

I don't think it's just focus. Assume that .y'y ca'aza'a se catra. It makes
sense to say xy. simlu lenu catra .y'y, but not .y'y simlu lenu se catra xy.
In other words, X simlu P(X,Y) seems to generate the presupposition "it is
not necessarily true that EZ:P(X,Z)" (where E is the existential quantifier),
but not the presupposition "it is not necessarily true that EZ:P(Z,Y)". (This
may be a *conversational implicature* --- but as such, it can be defeated;
that is, cancelled by context. Presuppositions are entailments, and cannot
be so cancelled.)

God I'm confused now.

--
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>.<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
Nick Nicholas. Linguistics, University of Melbourne.   nsn@krang.vis.mu.oz.au
        nsn@mundil.cs.mu.oz.au      nick_nicholas@muwayf.unimelb.edu.au
            AND MOVING SOON TO: nnich@speech.language.unimelb.edu.au