Whither head movement?

We argue that head movement, as an operation that builds head-adjunction structures in the syntax, has been used to model two empirically distinct classes of phenomena. One class has to do with displacement of heads (fully formed morphological words) to higher syntactic positions, and includes phenomena like verb second and verb initiality. The other class has to do with the construction of complex morphological words and is involved in various types of word formation. Based on the very different clusters of properties associated with these two classes of phenomena, we argue that they each should be accounted for by distinct grammatical operations, applying in distinct modules of the grammar, rather than by the one traditional syntactic head movement operation. We propose that the operation responsible for upward displacement of heads is genuine syntactic movement (Internal Merge) and has the properties of syntactic phrasal movement, including the ability to affect word order, the potential to give rise to interpretive effects, and the locality associated with Internal Merge. On the other hand, word formation is the result of postsyntactic amalgamation, realized as either Lowering (Embick and Noyer 2001) or its upward counterpart, Raising. This operation, we argue, has properties that are not associated with narrow syntax: it is morphologically driven, it results in word formation, it does not exhibit interpretive effects, and it has stricter locality conditions (the Head Movement Constraint). The result is a view of head movement that not only accounts for the empirical differences between the two classes of head movement phenomena, but also lays to rest numerous perennial theoretical problems that have heretofore been associated with the syntactic head adjunction view of head movement. In addition, the framework developed here yields interesting new predictions with respect to the expected typology of head movement patterns.