good news from generative grammarians

[Robert Levine]

Also, how does a framework whose data structures and licensing
conditions date back to the Old Stone Age of computation theory---when
tree structures and structure-to-structure mapping were the *only* way
to represent syntactic properties and dependencies and information had
to be horsed around trees by actually moving it from place to
place---wind up `presupposing' *anything* about frameworks whose
representational language is based on contraint logics defined on
structured algebraic models? In what sense is a framework that still
operates with the fossilized vestiges of the assumption that all
terminal elements are introduced into structures by rules---that is,
that *there is no lexicon*, leaving morphosyntactic dependencies to be
captures by hauling morphemes or feature complexes around graphs
containing ten times as many nodes as morphemes (as in late GB and
minimalist dogma respectively) ---`presupposing' anything about
frameworks in which information about morphosyntactic properties is
specified by lexical entries linking and interpretation, and regulated
by general conditions on feature sharing in general classes of
configurations? *My* question is, what purpose is served by this kind
of empty bragging?

Bob