The Syntax Of The Raising Predicate Seem

The present research attempts to provide a comprehensive description and analysis of the syntactic properties of seem as a raising predicate within the general theoretical framework of Chomsky's Universal Grammar. Seem-constructions in English involve a complex interplay between syntax, semantics, and the lexicon, and thus provide a rich source of data with which theories of grammar can be tested.On an observational level, the verb seem can be followed by a DP, AP, or PP. Besides, it can also appear with a TP or CP. It can be observed that the subject of seem-clauses bears no thematic relation of any kind to the predicate seem; rather it is always thematically related with some other predicate in the sentence. It is a traditional view in generative syntax that the formation of seem-structures involves the raising movement of the subject of a lower clause into the subject position of a higher clause. For this reason, the predicate seem is taken be a typical raising predicate. Seem differs substantially from seemingly related control predicates in that the former does not theta-mark its subject, whereas the latter do.In generative linguistics, seem-structures have long been analyzed on an analogy with passive structures. Under the standard view of passives, the surface subject of a passive is base-generated in the postverbal object position at D-structure. The passive morpheme on the verb absorbs the subject theta-role, thereby removing the ability of the passive verb to theta-mark its subject, and at the same time it absorbs the ability of the transitive to be assigned objective Case, thereby detransitivizing the verb. In order to have its Case features checked, the passive object has to move from its Caseless theta position to the surface subject position, a theta-less Case position. Along this line of reasoning, one might conclude that the predicate seem patterns like passive predicates in that both presumably fail to assign an external theta-role and are therefore incapable of assigning structural Case to the complement in the postverbal position. Driven by Case, the complement undergoes overt syntactic movement. In both cases, the specific type of movement involved is called A-movement, which targets an argument and raises it to an empty subject position, another argument position.The traditional account of seem-structures on an analogy with passives works quite well in cases where seem is followed an AP, DP, PP or infinitival clause, but it runs into serious problems when coming to cases where seem is followed by a finite CP. In both passive structures and seem-structures, the subject position is underlying empty. The only difference lies in the fact that the subject position of passive structures is derived empty due to the passive morpheme, whereas the subject position of seem-structures is base-generated empty. In seem-structures, the DP, AP, or PP can be analyzed as the predicate of a small clause, with its logical subject moved into the surface subject position of the seem-clause, where nominative Case is assigned. Much the same can be said about the cases where seem is followed by an infinitival clause: the embedded subject also obligatorily raises into the subject position of the seem-clause. However, problems arise when seem is followed by a finite CP. Specifically, the raising movement of the CP is possible with passives, but not with seem.I argue that these problems could be solved in a unitary way once we abandon one single assumption which lies at the very heart of the theory of passivization: that in those passive constructions which involve finite clauses as complement, the expletive it is simply inserted in the subject position. In the case of seem-structures with a finite CP, I propose that it is the lexical property of seem to select as its complement a combination of expletive it and a finite that-clause, rather than a CP only, contrary to standard assumptions. Since neither the complement position of passives nor that of seem-structures is a Case position, only clauses can remain there, because lexical DPs would need Case whereas clauses do not. By raising the expletive it into the canonical subject position while leaving the finite CP in situ, we derive the usual word order that we see on the surface.The alternative analysis proposed in this thesis has both empirical and theoretical merits. Empirically, it provides a ready answer to the long-standing question of why the embedded clause can be raised to the subject position (in complementary distribution with expletive it) with passive predicates but not with seem. Theoretically, it is consistent with the Principle of Full Interpretation and the Principle of Economy, according to which the expletive it must be derived by movement, contrary to the traditional view of expletive-insertion.