earticle

영어

The purpose of this paper is to claim that there is little reason to think that control is a special phenomenon of movement. The approaches advocated by Hornstein(1999, 2001) and Chomsky(2000, 2001) have a certain appeal, but they are not without serious problems. Hornstein’s movement analysis of control constructions overgenerates ungrammatical sentences and has theory-internal problems. His “control as movement” approach presents no obvious way to block the incorrect outputs of control. Chomsky’s analysis of control structure as a consequence of the syntactic operation “Agree” also has some conceptual problems based on the movement analysis of PRO, which assumes that after PRO agrees with control T, PRO moves to the Spec of TP to meet the Extended Projection Principle (EPP) and then PRO is assigned with the null Case. The argument against the PRO-movement analyses is based on the empirical evidence from the behavior of floating quantifier all (Baltin(1995)), and conceptual consideration that the EPP is a phonological requirement (Lasnik(1999)). They are discussed to show the evidence that PRO does not move to the subject position of control T which does not allow the EPP-feature unlike other finite clauses and that PRO agrees in the base-generated position in order for its derivation to be converged.