Case and NP licensing

This paper considers the theory of Case proposed in Pesetsky (1982) and adopted in Chomsky (1986a) and compares it to that proposed in Burzio (1986). Pesetsky's proposal is that a lexical feature of the verb determines whether or not it assigns accusative Case, and that possessing the feature [+Case] is what allows a head to c-select NP complements. I show that this proposal is empirically inadequate since (i) some heads that can assign Case are predicted by Pesetsky to be marked [-Case] and (ii) [+N] heads which do not assign Case differ as to whether they c-select NP complements. Furthermore, Pesetsky's proposal is stipulatory since there is no general way of predicting which V is marked [+Case]. Burzio's proposal that Case is a structural property of all verbs with specific theta properties is more adequate, both empirically and explanatorily. An additional conclusion is that Grimshaw's (1979, 1981) claim that heads subcategorize as well as select semantically is vindicated.