The goal of this paper is to examine whether the recent movement approach to control (Hornstein 1999, Boeckx and Hornstein 2004) gains further cross-linguistic support from object control constructions in Korean. I conclude from conceptual and empirical grounds that the movement analysis may not be an adequate framework for Korean. I contend that this is mainly because object control constructions in Korean basically belong to the arena of non-obligatory control. It is shown that certain facts in control constructions such as binding, NPI licensing, and passivization can only be accounted for by treating a controller and its controllee as two distinct arguments. Reverting to the old pro-based semantic analysis, I claim that some objections to it voiced by proponents of the movement analysis can be countered with a version of the Meaning Postulate (Cormack and Smith 2004) and an underlying structure in which a complement CP c-commands/precedes an accusative object. I show that structural and interpretive properties fall into place under the given analysis, suggesting that obligatory control-like interpretations in some cases be considered as preferred readings thereof facilitated by the scrambling of an object NP.
1. Introduction
2. Object Control Constructions in Lorean and the Movement Theory of Control
3. Some Drawbacks to Movement Theory
4. Proposal: A ‘pro’-based approach an underlying structure
5. Concluding remarks
References
(0)
(0)