2000 Ergative Agreement Systems

Can ergative agreement occur without ergative Case? Most previous approaches to ergativity have treated ergative agreement as a unitary phenomenon, answering this question with either an unqualified ‘yes ’ or an unqualified ‘no’, but new typological information indicates that there are two distinct types of ergative agreement systems, which differ radically in their relationship to Case: one is parasitic on Case (e.g. Hindi) while the other is independent of Case (e.g. Jacaltec, Selayarese, Abkhaz). This supports the view of Marantz 1984b and Johns 1996 that ergativity is not a unitary phenomenon, at least with respect to agreement. The correct theory of agreement must be sufficiently restrictive so as not to be able to generate the first type in the absence of ergative (or dative) Case, but nonetheless able to generate the second type in the absence of ergative Case. This can be accomplished within the structural approach to agreement of Chomsky 1995, with some help from Optimality Theory (Prince and Smolensky 1993). In the Hindi-type of ergative agreement, agreement simply cross-references nominative arguments; the pattern is ergative only when there is a lexically Cased subject and a nominative object. In the Jacaltec-type, the ergative agreement pattern has a completely different source. It is the result of a preference for using a clitic rather than agreement to cross-referenc