A Generative Approach To It-Extraposition Constructions

It-extraposition construction has ever been a widely discussed and highly controversial issue in syntactic literature for recent decades. With regard to the constructions of it-extraposition, the clausal argument (finite clause or nonfinite clause) acting as subject or object is displaced (extraposed) from its normal, canonical position to the right (non-canonical) position, and the original subject or object position is occupied by an expletive it.As far as it-extraposition construction is concemed, many studies have been made in recent literature. In general, there are two representative approaches to it, that is, the rightward movement account and the leftward movement account. They differ from each other in whether the movement involved is rightward or leftward. It seems that the clause supposed to be in the subject or object position actually appears to the right of its structural position. Hence, it is universally held that it-extraposition construction is derived by the process of right-moving subject or object clause and inserting an expletive it as a structural place-holder to the evacuated original subject or object position of the CP argument (Rosenbaum 1967; Stowell 1981). However, the rightwardmovement account runs into a series of serious theoretical problems. First, rightward movement violates the general constraint on syntactic movement that all movements should be upwards and leftwards. Second, rightward movement account runs against Kayne's (1994) Linear Correspondence Axiom, which prohibits rightward movement.On the other hand, in analyzing the formation of it-extraposition constructions, the leftward movement account holds that it undergoes the process of left-moving the expletive it and stranding the CP in situ. This account is based on Kayne's (1994) anti-symmetrical LCA and Chomsky's (1995a, 1998, 1999, 2000)Minimalism, therefore, it avoids such theoretical problems as the rightward movement account encounters and sheds some light on the derivation of it-extraposition constructions. However, this account is not without problems of its own. As we can see, Stroik (1996) does not make the same proposal for subject and object it-extrapositions while analyzing their asymmetry in allowing wh-clausal arguments, which violates uniformity of syntactic researches. This problem is solved in Yoon's (2001) analysis. But he proposes that CPs, like DPs, carry Case features and must be checked in syntactic operation. This thesis provides rich evidence to prove that this assumption is untenable. The major difference between CPs and DPs lies in that the former do not bear Case features and thus are allowed to occur in Caseless positions, which is not true with DPs. Besides, neither Stroik nor Yoon covers all the language facts observed in object it-extrapositions. On the basis of the above analyses and in the light of Kayne (1994) and Chomsky's (1995a, 1999, 1998, 2000) theories and syntactic merge principles, the present paper claims that the expletive it merges in the Spec of the CP and then the merged CP enters into the derivation of the matrix clause. In the process of derivation, PF contains a series of uninterpretable features. In order for the derivation to converge, the expletive it moves overtly to the Spec of AgrSP and AgrOP respectively in subject and object it-extrapositions to check off the NP features of each. The movement of object expletive it occurs at PF and that of subject expletive it takes places in syntax. This results in their asymmetry in allowing wh-phrases. The licensing of the expletive it in object position has not received enough attention and there is no satisfactory account for it. The present thesis makes an attempt to explore optional-obligatory status of the expletive it in object it-extraposition. According to Chomsky's (1981) Projection Principle, we hold that it is the subcategorization properties of the verb that determine whether the expletive it can be licensed or not in object it-extraposition. Transitive verbs may fall into two main types, one with optional NP features such as mention and the other with obligatory NP features as like. The verb imposes different selection restrictions on its complement owing to its bearing features. When bearing NP features, which are strong uninterpretable features at PF and thus must be checked off before spell-out, the verb selects an element with the matching NP features to get the features valued. And the expletive it is generated for feature checking's needs.In short, the proposals developed in this thesis are more economical and can provide a unified account for both subject and object it-extrapositions.