The limits of syntax in inversion

The study of this phenomenon has a long history in generative grammar, where it has generally been taken to be a quintessentially syntactic phenomenon (see, for example, Kayne and Pollock (2001), Pesetsky and Torrego (2001), Rizzi (1996), and Zubizarreta (2001) for some recent analyses). Here I will argue that in at least some types of Romance inversion, the explanation for why inversion occurs seems to lie not in syntax alone, but in the interplay between universal properties of sentence processing and language-particular syntactic properties, with the result that the syntactic mechanisms needed to account for inversion are greatly reduced over what has previously been assumed.