ARG-ST on phrases (long)

[Carl Pollard]

Hi folks,

In response to Ivan: I certainly didn't mean to downplay the empirical
side of things. But there may be ways to handle locality questions
besides in terms of feature geometry. In fact, as my previous message
was meant to suggest, I don't put much stock in the representation in
terms of feature structures; as far as the algebra of categories is
concerned, all that matters is the algebras up to-isomorphism. It
doesn't really matter what categories are.  This is a point that has
been made in various forms by many people over the years, including
Drew Moshier, Mark Johnson, Gert Smolka, and Paul King.

Carl