Agree and EPP in Bantu

Abstract

C agrees in the φ-features of an operator in its Spec but not one in situ in many Bantu languages. Despite this I argue that a closest c-command based account (Chomsky 2000) is superior to a Spec-head agreement analysis, for two reasons. First, feature-valuation under closest c-command permits a unified treatment of Bantu operator agreement and West Germanic complementizer agreement with subjects. Second, it explains a common requirement that subjects be in situin Bantu A′-movement constructions: a subject raised to Spec, TP would intervene between C's φ-features and the operator in outer Spec, vP that must value them. Like wh-movement, subject raising always coincides with agreement in Bantu; I propose that Bantu uninterpretable φ-features have EPP features. These are distinct from the classical EPP of T which, I argue, underlies a verbal agreement requirement giving every Bantu inflectional category a specifier. I show that nominative Case-checking is independent of both agreement and EPP in Bantu. © Springer 2005.