A Constraint on Remnant Movement

The aim of this paper is to begin to assess the empirical fit of a constraint on remnant movement that is derived as a consequence of the system proposed by Hunter (2010), the main focus of which was a unification of adjunct island effects and freezing effects. As well as ruling out all extraction from adjuncts and from moved constituents, the constraint proposed in this previous work has the effect of imposing a relatively strict limitation on remnant movement configurations, because of their similarity to freezing configurations. The compatibility of this limitation with the various ways in which remnant movement has been used in the literature has not yet been explored in detail. The rest of the paper is organised as follows. In Section 2 I briefly introduce remnant movement and discuss the derived constraint on it, the “Just Outside Constraint” (JOC). In Section 3 I discuss the empirical fit of the JOC with one of the original, and canonical, uses of remnant movement, namely in analyses of German “incomplete category fronting” phenomena; these analyses turn out to be by and large consistent with the constraint, but there are exceptions, which I present in Section 4. I then turn to some less canonical uses of remnant movement: a collection of proposals which, very broadly speaking, use remnant movement to produce certain word orders that traditionally were thought to require syntactic machinery that (arguably) is best avoided in minimalist grounds (eg. head movement, covert movement, underlying head-final word order). The idea in these works is essentially that since remnant movement is just a combination of “good old-fashioned” overt phrasal movement steps, we should prefer a theory that renders this more elaborate machinery redundant by using remnant movement to emulate their effects. In particular, these proposals address covert movement (Section 5), head movement and SVO/SOV/VSO word orders (Section 6), and verb-second word order (Section 7). I summarise and draw some tentative conclusions in Section 8. The analyses that I will be reviewing form a relatively disparate group: they are unified only by their use of remnant movement, and do not necessarily share all the same framing assumptions. The reader should therefore be aware that certain basic assumptions underlying the analyses being discussed will necessarily change from one section of this paper to the next. For example, some of the works discussed assume that OV is the underlying pre-movement word-order in languages such as Japanese and German, but certain proposals discussed in Section 5 and Section 6 assume universal VO underlying order (following Kayne (1994)); and in Section 3 I review some analyses of