Case vs. positions in the locality of A-movement

This paper contrasts two families of approaches to the ban on A-movement out of a finite clause (hyperraising). One line of approach attributes the constraint to the positions involved in the movement chain (e.g., improper movement), the other derives the constraint from the case properties of the moving element (e.g., Activity Condition). Based on novel evidence from Hindi-Urdu, this paper develops an argument in favor of position-based approaches. In a nutshell, I argue that (i) A- and Ā-movement in Hindi-Urdu exhibit the same locality contrast as in English, that (ii) both apply to already case-marked DPs, and that (iii) they differ in the position they target in the same way as in English. As a result, the locality difference between the two movement types in Hindi-Urdu can be attributed to the positions that these movement types target, but not to their case properties. These results indicate that when case and positions diverge, locality tracks positions, not case. This conclusion supports the view that a comprehensive account of the locality of A-movement requires reference to syntactic positions in some form. The results also raise the possibility that case can be eliminated as a constraint on the locality of A-movement.