Reversing the perspective on Quantifier Raising

The controversial rule of Quantifier Raising cannot be fully subsumed under the standard definition of the Move operation. We argue that the differences between overt movement and QR can be accounted for if the syntactic computation proceed s top-down (and left-right), rather than bottom-up. In a bottom-up derivation, a phrase must be first merged in argument position, and the remerge position has different properties in the two cases: in particular, the target position of QR is covert, it is not featurally licensed, and it can not freely host a successive cyclic step (whence the clause-boundedness of QR). If we assume instead a top-down + left-right derivation, a radical asymmetry emerges. Overtly moved phrases are first merged in a featurally licensed clausal position, and covertly remerged in argument position. In the case of QR, instead, the computation of the QP in argument position precedes the computation of the remerge position where the QP attaches to its nuclear scope, after the latter has been completely built: thus, QR is necessarily «rightward». The covertness of QR follows from the fact that the system invariably spells out the first merge position; the clause boundedness of QR is shown to be a right roof effect (due to a final phase boundary). The left-right orientation also immediately derives the Leftness Condition on quantifier binding. This rather conservative approach is briefly compared with the continuation-based analysis of quantification outlined i n Barker (2002).