Remnant movement and complexity

Linguists in the transformational tradition have recently been exploring the extent to which previous empirical coverage can be maintained with fewer grammatical devices. For example, Kayne (1994) develops some analyses in a framework that requires constituents to begin in specifier-head-complement order, with only leftward movement, and Chomsky (1995) explores how the structure building operations in such a system can be bottom-up and driven by lexical features. In more recent work, Kayne (1998) begins the exploration of the possibility that there is no covert movement, and Koopman and Szabolcsi (1998) show how head movement is not needed in the analysis of verbal complexes. With the restricted grammatical options these theoriesmake available, “remnantmovement” has come to have increasing prominence, where a “remnant” is a constituent from which material has been extracted. Moving a constituent from which material has already been extracted means that traces of earlier movements may be carried to positions where they are no longer c-commanded by their antecedents, something which was banned in earlier theories. Interest in remnant movement has a longer history. For example, den Besten and Webelhuth (1990) proposed analyses like the following: