Reentrancy in feature structures

[Antonio Branco]

Carl Pollard wrote:

> Mike Maxwell wrote:
>
> >
> OK, then how are reflexives handled under a system where co-indexing is
> not syntactic (and, I presume, therefore inaccessible to the
> morphology)?

>
> Perhaps some readers who advocate nonsyntactic theories of the kinds
> of facts that have been handled under the rubric of "Principle A"
> and "Principle B" (however formulated) could address these issues?
>
> Carl

Hi,

I am not sure that it is possible to advocate a completely nonsyntactic theory
for binding (though I suppose my approach is definitely much more
semantics-driven than any other current alternative I am aware of),
but in my paper for the HPSG2001 conference I propose an account of
binding that does not resort to indexes or coindexing (at least as these have
been used after the P&S94 book). If you are interested in that paper,
you can get a copy of it at  http://csli-publications.stanford.edu/HPSG/2/Branco-pn.pdf

There is also the article that was recently published in CL, with a detailed
discussion on why handling binding principles via "syntactic" coindexing leads to
problems of efficiency and how these problems can be overcame:
A. Branco, 2002, "Binding Machines", Computational Linguistics 28-1, pp.1-19.

I am currently working further on integrating and expanding these ideas. This
draft is not available yet in my webpage, but I can send a copy of it to those
interested in seeing it.
Best,
António