Feature attraction and category movement

This dissertation presents a new theory of movement in generative grammar within the framework of the Minimalist Program (Chomsky 1993, 1994, 1995). The most important aspect of the new theory is that it departs from the standard view of Move {dollar}alpha{dollar} as a unitary operation.; The dissertation has three main goals. The first goal is to simplify the theory of feature checking in such a way that a single structural configuration (namely, the head-adjunction structure) holds for feature checking in both overt and covert syntax. The second goal is to present evidence for the Split Movement Hypothesis, according to which a set of formal features and its associated category move to separate structural positions in syntax. The claim is that UG makes both Attract and Move available in the form of Attract F and category movement, respectively. The former satisfies the formal requirement of feature checking, and the latter yields the phonological "displacement" effect of overt movement. The theory thus abandons the treatment of Move {dollar}alpha{dollar} as a unitary operation. The third goal is to explore consequences for the proper characterization of locality in syntax. The Split Movement approach opens up a novel way to account for the traditional Subjacency and Condition on Extraction Domain effects (Huang 1982a) that do not fall under Chomsky's (1995) Attract F theory. It is argued that Attract F and category movement are subject to different types of economy conditions, and that the traditional Subjacency and Condition on Extraction Domain effects should not be given a unified account, contrary to the Barriers approach (Chomsky 1986a). The theory also derives a number of previously mysterious properties related to crosslinguistic variation in extractability out of islands, the relative strength of island effects, and the nature of successive-cyclic movement.; This dissertation presents a new approach to movement in syntax which overcomes the conceptual and empirical shortcomings of the traditional approach, while gaining new insights into previously mysterious phenomena and properties of natural languages.