Head movement: an even uglier duckling
Head movement returns
Head movement is an operation which—in contrast not only to A-bar displacement, the most evident case of displacement with interpretive effects, but also to A-movement, with semantic consequences less sharply visible and more dependent on specific assumptions about the architecture of the grammar and the nature of representations delivered to the C-I component—remains by and large a mechanism whose presence leaves clear traces on the A-P side of the derivation and which remains much more hypothetical as far as the workings of narrow syntax are concerned. Its fate has been as a consequence even more convoluted than that of phrasal chains. Despite attempts to unify all displacements under a single umbrella of (descendants of) the Move-а rule, peculiarities of head movement remained, leading to several theoretical difficulties, summarized succinctly in Chomsky (2001):
Other considerations have to do with the nature of the head-raising rule, which differs from core rules of narrow syntax in several respects. It is an adjunction rule; it is countercyclic in ways that are not overcome along the lines discussed earlier; the raised head does not c-command its trace; (...) identification of head-trace chains raises difficulties similar to those of feature movement, since there is no reasonable notion of occurrence; it observes somewhat different locality conditions. (Chomsky 2001: 38)
A radical suggestion was to expel head movement from narrow syntax entirely, relegating it to the realm of postsyntactic operations taking place on the road to PF, thus following the route indicated already with regard to phenomena related to V2 in Chomsky (1995: 368)—a proposal which led to developments and modifications of the PF-theory (e. g. in Boeckx and Stjepanovic (2001) or Platzack (2013)), attempts to derive effects of the head movement from phrasal displacements (as e. g. in Koopman and Szabolcsi (2000)), ‘mixed’ proposals striving to save partly the syntactic nature of the phenomenon by attributing only some of its effects to morphological processes (as in Matushansky (2006)), and hypotheses which ascribe to the operation properties otherwise absent from standard phrasal displacement (as on accounts allowing the displaced head to become the head of the structure; see e. g. Donati (2006), and the overviews of the whole issue in Roberts (2001, 2011), Barrie and Mathieu (2014)). The matter remains unsettled also in the most current developments of the minimalist framework: although the nature of head movement is discussed at some length in Chomsky (2015b) and the operation itself is incorporated into the theory of syntactic structure building as one of processes decisive for establishment of interpretable structures which may be operated upon by the C-I component in particular, the question about the applicability of syntactic head movement analysis to all phenomena involving head-head relationships lacks a definitive answer so far (see the remarks in Chomsky (2015b: 15)). Yet it is a crucial ingredient in a label- based framework of Chomsky (2015b), an operation which—together with another kind of head-head relationship, viz. feature inheritance—is a key factor in modeling the derivational process as proceeding by phases and determining instructions for the C-I component at each stage of the computational process.