Selectional Asymmetries between CP and DP Suggest that the DP Hypothesis is Wrong

要旨

Nearly every researcher in generative grammar now adopts the position that the head of the noun phrase is not N, but is some higher functional projection, entirely parallel to clauses, where higher functional projections dominate the lexical verb. One version of this involves the notion of an extended projection (Grimshaw 1991, van Riemsdijk 1998), where clauses are in some sense extended projections of verbs, and DPs are in some sense extended projections of nouns. However, all facts regarding complement selection and the determination of form indicate that the head of the clause is not the verb; C is (and the head of its complement is not the verb, but some other functional head; and so on down). But the same types of facts indicate that the head of a nominal phrase is not a functional projection like D; it is N. This asymmetry calls into question the correctness of the DP hypothesis. Verbs that select for clausal complements commonly select for C, or other things that are usually re-garded as residing high in the functional domain of the clause. They select for questions versus declaratives (*Sue thinks what time it is); finite versus nonfinite (*Bertrand wants that the world is round); subjunctive versus indicative (Sue asked that the answer be/*is two). In other languages they may select V2 clauses (German) or illocutionary force (exhortative in Korean, imperative in Japanese). No verb ever selects for the verb in the embedded CP, or for modals, auxiliaries, or negation, which can generally appear in any com-

主題