Date: Sun, 20 Oct 96 16:16:21 EDT
Message-Id: <9610202016.AA10115@JESPERSEN.CS.NYU.EDU>
From: Ruth Reeves <reevesr@jespersen.cs.nyu.edu>
To: linconf@tamvm1.tamu.edu
Subject: General comment
This is just a general comment stemming from my reading of the papers
and comments so far. This is neither a question nor a criticism --
it's just intended as a clarification. As Broadwell specifically
states and Zlatic's proposal implies, it matters whether or not the
notion "argument" is included in Binding Theory's definition. It seems
to me, though, that some of the discussion about the notions
surrounding this term has been cross-talk. It has seemed as if
"A-bound" vs "A-bar-bound" was a syntactic and perhaps purely
structural notion while "argument" is a semantic notion. But after
all the "A" stands for argument, where this notion does indeed involve
theta-roles -- a position that gets one is an argument position.
Invoking this aspect of theta-theory has zero implications for the
geomtry versus theta-hierarchy debate. Some kinds of anaphoric
elements require A-bar binding (Broadwell); some require A-binding
(Zlatic). If the difference betwenn "A"s and "A-bar"s can be
structurally defined, as suggested in one of Seely's comments, then a
perhaps a geometric account which does not include the notion of
theta-assignment at all can be acheived. Absent such a result, if we
don't accept that there are various marks on syntactic objects which
determine their identity vis-a-vis other syntactic objects, then
neither the distinction between NP vs VP nor between NP(i) vs NP(j)
can be made. Invoking the notion A-bar is also an appeal to this
rudimentary defintion of argument, though as far as I know, a
definition which would divide up the class of syntactic objects once
A-positions have been excluded such that A-bar positions are defined
is yet to come.