Re: tough-constructions

About this list Date view Thread view Subject view Author view

Robert Levine (levine@ling.ohio-state.edu)
Mon, 19 Dec 94 14:27:47 EST


The idea that _tough_ adjectives do confer a sementic relation on their subjects is not novel; Lasnik and Fiengo in their 1974 LI paper `Complement object deletion' provide excellent examples contrasting _tough_ constructions with true raising-to- subject constructions in this respect, and the contrast between the truth values of It's obnoxious to try to talk to Robin [because, e. g., s/he keeps a pack of starving pit bulls in her/his office] vs. Robin is obnoxious to try to talk to is robust: Robin may be personally lovely and the first could still be true, but it's hard for me to imagine a world in which Robin was personally lovely but the second was true as well. Even more to the point, maybe, is that the syntax of _too/enough_ APs is highly parallel to _tough_ in most relevant respects, but no one I think would want to claim that the subject in Robin is too stubborn to talk to. lacked a semantic relation to the head of the AP. I noted in a paper in 1984 that every single syntactic argument that Chomsky 1981 gives for treating _easy_ constructions as complex predicates (produced mid-derivation by reanalysis) is based on data which is completely paralleled by _too/enough_ examples; yet the reanalysis solution Chomsky gives, which is predicated on the lack of a subject theta-role for tough-subjects, is unavailable for _too/enough_ constructions just because the latter uncontroversially accord thematic roles to their subjects. A parallel treatment of _too/enough_ constructions, for example, would yield a violation of the theta criterion, with the subject picking up theta roles both from the _too/enough_ adjective and its trace following the complex A. Any treatment for _easy_ must extend to _too/enough_ over a large class of data (`violin/sonata' paradigms and so on) and so anything which hinges crucially on the lack of theta roles for _tough_ subjects looks as though it isn't going to work. Bob


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:31:40 PST