earticle

영어

Fong and Ginsburg (2013, 2014) propose a new theoretical framework for Agree relations, namely the stack mechanism. Its core idea is the stack. It is a kind of buffer which stores Syntactic Objects with an unvalued feature temporarily. The stack mechanism is argued to be more economical than the probe-goal search of Chomsky (2001) because, with the stack, there is no need for a probe to check all of the labels within its search domain. Fong and Ginsburg (2013) try to make their stack mechanism empirically motivated by demonstrating that it can successfully account for the notorious violin-sonata paradox. Their analysis, however, has two drawbacks: the indetermination of stack order in the case of doubling constituents and the wrong assumption that there is no A'-movement in tough constructions. To solve these problems, this paper suggests the A-over-A Constraint on Stack Order, and shows that Hicks’ (2009) doubling structure is more appropriate for doubling constituents. With these minor modifications, the stack mechanism is still a viable solution for the notorious violin-sonata paradox.