A constraint on remnant movement

The aim of this paper is to assess the empirical fit of a constraint on remnant movement that is derived as a consequence of a certain reformulation of the mechanisms underlying syntactic movement. The constraint arises from the similarity between remnant movement configurations, on the one hand, and configurations of the sort that produce “freezing effects” on the other; the reformulated system disallows the latter, and imposes a condition on the former that appears quite restrictive. I show here, however, that a large portion of the uses of remnant movement in the literature work within the bounds predicted by this unusual constraint, including analyses of data concerning VP-fronting, German “incomplete category fronting”, verb-second, and SOV and VSO word orders.