Pied-piping in A-bar extraction represents a straightforward case of cross-linguistic variation within certain limits. For example, there are some languages, like English, which can ask wh-questions like the one in (1a); many more languages that can ask the wh-question in (1b); and (1c) is apparently unattested:
(1) a. [Who] are you talking with?
b. [With whom] are you talking?
c. *[Talking with whom] are you?
Cable (2007, 2010) introduced the functional projection QP, which dominates wh-phrases and is the phrase that actually undergoes wh-movement. On his account, the question of how much material may pied-pipe is simply a question of the distribution of QP.
I will try to derive Cable’s observations about QP from Contiguity Theory (Richards to appear), which seeks to predict the distribution of various kinds of overt movement by positing conditions on the relation between syntax and phonology. In addition to facts like the ones in (1), I will try to develop an account of differences in the pied-piping possibilities for different kinds of movement (for example, the differences between restrictive and non-restrictive relativization), and some cross-linguistic differences.