Wh-movement and the theory of feature-checking argues that cross-linguistic variation in wh-constructions reduces to the availability of different lexical instantiations of a +wh C0 both across languages and within a single language, and the way in which such lexical elements are syntactically identified, either via movement or base-generation. Evidence from a wide range of patterns including wh-expletive questions leads to the conclusion that wh-feature checking may sometimes be effected non-locally and ‘at a distance’ (long-distance wh-agreement), and that movement in general takes place for two related but discrete reasons: both to identify and activate an underspecified licensing head and in order for an element to occur in the checking domain projected by its relevant licensing head. Developing and generalizing the proposals beyond wh-phenomena, the study also goes on to argue for a Minimalist model of syntax in which feature-dependencies are in fact all licensed in the overt syntax and where there is no need for any further level of LF.