Tom Hukari (hukari@uvic.ca)
Mon, 22 Jun 1998 17:39:17 -0800
I actually wrote a paperoid on the construction a few years back. (I think it appears in my department's working papers.) I concluded that the antecedent is some species of theme. Speakers' judgments vary, but some people find that even a PP can control the gap: I presented Sandy with Jojo to play with on the plane. This aside, isn't there a danger of introducing an excessive degree of nonlocality into complement selection if ARG-ST is a head feature? In very, very clear cases of subcategorization which no one would dispute is selection by a head, do we need to look at the arguments of a complement? Tom
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:38:20 PST