Operator Islands, Maraudage, and the Intermediate Step Corollary

Relativized Minimality and the Minimal Link Condition are conceptually related, and there is a certain amount of empirical overlap. These constraints can account for operator island effects (specifically, wh-island and topic island effects). I argue that both constraints are conceptually problematic from a minimalist point of view; furthermore, they can be shown to be empirically problematic because they cannot predict asymmetries between movement types as they show up both with the island-creating item and the long-distance-moved item. In contrast, I suggest exploiting a consequence of the PIC-approach to CED islands developed in Müller (2010). A side effect of this approach is that intermediate movement steps of successive-cyclic movement end up in non-outermost specifiers; they precede the final movement operation triggered by regular structure-building features of a head. I call this consequence the Intermediate Step Corollary. The Intermediate Step Corollary makes a new account of operator island effects possible that does not rely on the concept of intervention: It suffices to postulate that certain types of goal features on categories (like wh-features) automatically trigger checking once they enter a suitable configuration (as they do with intermediate movement steps); this is an instantiation of a more general concept of feature maraudage. As a consequence, operator island effects are derived (as is their variable nature). Acknowledgements: For comments and discussion, I am grateful to Klaus Abels, Marcel den Dikken, Gisber