fate of SPEC and mutual selection

[David Baxter]

Hi fellow HPSG-ers,

We are implementing an HPSG-based English grammar and hoping to avoid the
cycles introduced by mutual selection such as that between a determiner
and noun using the SPR and SPEC features in Pollard & Sag 1994. We notice
that in Sag's recent "English Relative Clauses" paper, this selection
appears to be only one-way - the SPEC feature is gone.

Could someone please send us (a pointer to) an argument motivating the
elimination of SPEC and/or an explanation of how the work done by SPEC in
PS94 is done in grammars without SPEC?

We are also interested in any other cases where such mutual selection
might be needed in an HPSG-based English grammar.

Thanks in advance.

Sincerely,

David Baxter


==================================================
David Baxter                          Cycorp, Inc.
baxter at cyc.com                      (512) 342-4044