The logic of complex predicates

In the literature, there are two major proposals for resolving the syntax-semantics mismatch characterizing complex predicates. The ‘verb-raising’ approach resolves the mismatch via syntactic movement (or its analog), whereas the ‘argument-sharing’ approach does so by positing merged argument structures for complex predicates at the syntax-semantics interface. Focusing on two types of complex predicates in Japanese—syntactic compound verbs and the so-called -te form complex predicate—I discuss some novel empirical data posing challenges to both approaches in addition to the set of well-known observations from the literature illustrating the tension between the two strategies.The paper then argues for a synthesis of these two approaches within a variant of categorial grammar, taking advantage of the logical perspective on the syntax-semantics interface characteristic of certain recent variants of categorial grammar. The proposed analysis integrates the analytic insights of the two previous approaches seamlessly, and has both theoretical and empirical advantages over the two: theoretically, it clarifies the deeper connection between the theory-neutral analytic intuitions guiding the two approaches; empirically, it provides straightforward solutions for both the old and new empirical problems, enabling a previously unattained unified treatment of complex predicates that has a wider empirical coverage than its competitors.