Can crosslinguistically variant grammars be formally identical? Third factor underspecification and the possible elimination of parameters of UG

Certain aspects of human knowledge of language (=UG) may well follow from more general laws of, e.g. (biological) computation, or physical law (Chomsky, 1965, 2005). Minimalist derivation allows the possibility of deeper UG-independent explanation via third factor principles concerning “computationally efficient” satisfaction of the interfaces. This line of inquiry could in principle capture crosslinguistic variation, without appeal to stipulated axiomatic parameters of UG describing the variants. Instead possible crosslinguistic derivational variation could in principle be deduced from points of underspecification in the definition of computational efficiency—which allow more than one way to optimally satisfy the interfaces (“a tie for first place”)—just as Richards (2008), Boeckx (2011) suggest. (For previous such analyses see e.g. Chomsky (1991, 2008) and Epstein et al. (1998).) We seek to show that computational efficiency allows different relative rule orderings of: feature inheritance, Agree and wh Internal Merge. The different orders we examine are equally optimal—three rule applications are necessary and sufficient for convergence, but their ordering is irrelevant to the interfaces. Our main focus concerns variation in Complementizer agreement phenomena in Haitian Creole vs Cape Verdean Creole. Agreement variation between English and the Bantu language Kilega is also discussed.