Predication with _beautiful of face_

[Emily M. Bender]

Howdy Fred!

This isn't exactly and LFG answer, but it seems to me
that you could get somewhere by saying that the adjectives
do some semantic funny business --- they introduce two
predications, the one you'd expect (`weak') and also
a possession predication. Then, rather than making their
SUBJ a semantic feature of the weak predication, they
relate it to the other one, while relating the PP
complement to both.  I'm not quite sure what multiple
predications introduced by the same head would look
like in LFG, but I guess it would just be a question
of positing another f-structure somewhere.  This idea
would be pretty straightforward to implement in HPSG
though.  The whole issue raises interesting questions about
how semantic the GF features are in LFG, or, alternatively,
to what extent s-structure has to parallel f-structure.

(Given the general approach, I'd imagine s-structure
wouldn't have to be that tied down, I'm just out of
practice with LFG :-)

Emily

On Tue, Nov 18, 2003 at 11:32:51PM -0600, Frederick Hoyt wrote:
> Hi,
>
> I'm wondering if anyone in the LFG community has considered or could
> reflect upon constructions like the following (these are all examples I
> pulled off the web):
>
> 	"Dietrich was slender of body and of medium height "
> 	"She is slender of build and graceful in movement
> 	"She remembered when she was round of belly, ripe tropical as a
> setting moon"
> 	"The fact is simple: Bush is weak. He is weak of mind. He is
> not very smart. He is weak of body "
>
> and so on.
>
> The issue is how the predication between the subject in such an example
> and the predicate would be represented. Intuitively, saying that "Bush
> is weak of mind" is equivalent (at least truth-conditionally) to saying
> that "Bush's mind is weak" or "Bush has a weak mind." In other words,
> "weak" seems like it should be predicated of "mind," rather than Bush.
> In LFG terms, this should mean that _weak_ subcategorizes for a (SUBJ)
> function, and that _mind_ should fill it, or that it is an adjunct
> modifying _mind_.
>
> However, if I assume that a predicational clause is represented as a
> functional control structure, then I would expect - at least in
> syntactic terms - that _weak_ should be predicated of Bush. Say that
> _be_ is a raising verb, and _weak of mind_ is its XCOMP. Then  _Bush_
> should fill the SUBJ-function in the PRED-function of  _weak_.  But if
> that is the case, then it's not clear to me what role _of mind_ is
> playing.
>
> I would say that in English, examples like this have a slightly bookish
> feel, so maybe they're a marginal phenomenon, but in other languages
> (such as Standard Arabic) they are very productive, and considered good
> style, so even if an analysis might be of marginal concern for English,
> it would be very useful elsewhere.
>
> Thanks for any thoughts,
>
> Fred Hoyt
>
>
> Frederick M. Hoyt
> Linguistics Department
> University of Texas at Austin
> fmhoyt at mail.utexas.edu
>
>