Dick:
> Chet,
> Like And, I'm impressed at your ability to supply Inuktitut examples from
> memory! (Though I do notice that the eskimo for 'large' turns out to be
> "large"!) Very interesting they are too, aren't they?
>
> >Very interesting, and I'm grateful that you took the time to give a full
> >explanation of the problem.  I think you proposed account (or something
> >else) would also be required to handle so-called "noun-incorporation".
> >E.g. in Inuktitut (Eskimo), the verb for "to hunt" is siuq- and with
> >it one can hunt various things:
> >(1) tuktu-siuq-puq  "I'm caribou hunting"
> >    caribou-hunt-1SgIndic
> >etc.  Here there is no problem as tuktu is not required to have
> a referent
> >and need not be involved with any other part of the sentence.  However,
> >an adjective with proper case-marking can appear, and there there are
> >problems similar to the clitic example, or at least it would
> seem so to me:
> >(2) large-mik tuktu-siuq-punga  "I'm hunting for a large caribou"
> >	 -Acc
> >(with apologies for not supplying the number -- I'm at home and all of my
> >Inuktitut dictionaries are at work).  Here is a similar example with a
> >relative clause (headless):
> >(3) sini-ju-mik nanuu(q)-siuq-punga  "I'm hunting for a sleeping bear"
> >    sleep-Dep-Acc
>
> ## In principle, I don't see why these shouldn't be handled directly in
> dependency analysis, with 'large' (or whatever) depending directly on
> 'caribou' even though this takes its place within the verb. Presumably the
> problem for surface word order is that some other dependent of
> the verb can separate 'large' from 'caribou', which produces tangling; but
> that's ok if we maintain the principle that morphology takes priority over
> syntax (which is Sadock's principle). I.e. tangling in surface structure
is
> ok provided it's caused by morphology - or, putting it another way, all
parts
> of a single word count as being in the same place as far as surface
> structure is concerned:
>
> 	large often caribou-hunt-I = 'I often hunt large caribou'
> 	large <-- caribou, but in surface structure 'caribou' = 'hunt'
> 	often <-- hunt, but in surface structure 'caribou' = 'hunt'
> 	caribou <-- hunt

I've never seen data like this. Can anyone confirm that examples exist in
any language? I can't remember whether Sadock allows it. He does say that
morphology overrules syntax, but this is subject to some (IIRC, tenative)
constraints the nature of which I can't remember at all.

Certainly my assumptions of the nature of the syntax--phonology
interface are incompatible with "large" depending on "caribou", because
syntax cannot tangle under any circumstances, and all the syn-phon
interface can do is cause adjacent words to share the same phonological
expression.

> This example actually makes a nice contrast with the French one,
> because it
> shows that what you might call a 'morphologically demoted' word can have
> dependents, but if it does they have to behave as normal. So I don't see
> how we can say that "fromage" is forced to be null just because "le" is
> morphologically demoted.

I'm not sure whose remarks this is addressed to. It doesn't seem to
invalidate my sort of explanation for the nullity of "fromage", e.g.
a rule that says if \le\ coenounces with a (not necessarily immediately)
following \X\ then the resulting phonological form is composed of /le/ + the
form of X. You thus get a conflict between trying to form /le fromage/ and
/le mange/, which can be reconciled only by invoking a different rule which
allows \le\ \fromage\ to have the shape /le/, thus making /le mange/
possible.

--And.




%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%