CPs Move Rightward, Not Leftward

Abstract

Several recent proposals hold that CP complements appear rightmost in many languages by a two-step leftward movement process: first the CP moves leftward, and then a remnant phrase carries all other material to the left of that moved position. I show here that this analysis faces insurmountable problems. In contrast, a simple rightward movement analysis explains all the facts. Importantly, if binding is computed on the basis of precede-and-command rather than c-command (Bruening 2014), the rightward movement analysis accounts for all binding facts. The remnant movement analysis fails to account for binding, since apparent rightward movement does not pattern with clear cases of remnant movement like partial VP fronting. Furthermore, there is a leftward-rightward asymmetry in syntactic category when CPs undergo movement. I show that it is difficult to capture this in a theory that only has leftward movement. I also address preposition stranding in English, which Moulton (2015) presents as an argument for the leftward movement analysis. The results of a large-scale survey reveal that there is actually significant speaker variability in whether prepositions can be stranded when CPs move to the right. I spell out a non-grammatical account of this variability within the rightward movement analysis. The overall conclusion is that rightward movement is not effected by a two-step leftward movement process, and we need a theory of grammar that allows movement upward and to the right.