Control as Anaphor Movement

Abstract

Whether the control module independently exists or not has been a hot theoretical issue in recent syntactic debates in generative grammar. In the standard theory (Chomsky 1981, Landau 2003), raising and control have been treated as distinct processes: Whereas raising involves movement, control involves a nile of construal relating PRO to its antecedent. Hornstein (1999) and Boeckx (2000), among others, have recently claimed that control is just an instance of raising. Despite some theoretical merits, the latter claim, called control-as-movement, can only thrive at the cost of the standard theta theory in that this claim is forced to allow the movement into a theta-position. This paper attempts to combine two approaches which are deficient on their own into a new proposal in which problems of these two can be reconciled. The major point of this middle path is that the obligatorily controlled PRO is a reflexive clitic similar to se in the Romance languages, and that it moves toward its antecedent, called control-as-anaphor-movement. As a consequence, this approach maintains the standard theta theory as it is and still captures the movement properties of control while eliminating the control module.