Morphological case systems are frequently described in terms of distinctions related to transitivity. To a first approximation, the case system of Nez Perce nicely fits this bill: one case (ergative) marks transitive subjects, a distinct case (objective) marks transitive objects, and intransitive subjects remain in an unmarked (nominative) form.

However, Nez Perce also shows us a series of circumstances in which the correlation between transitivity and case-marking breaks down. All transitive verbs allow both the case pattern in (1) (ERG subject, OBJ object) and the “caseless” version in (3) (NOM subject, NOM object). Both versions appear to be semantically and syntactically transitive; yet in the caseless version (3), the characteristic subject and object cases of the transitive pattern have disappeared.

What controls the choice of case in transitive clauses? I argue that the deciding factor lies in the grammar of object agreement: all and only clauses with successful object agreement show the ergative and objective case. This finding calls for a theory of morphological case which accords a crucial role not to transitivity itself but to the syntax and morphology of agreement. I propose a version of this view according to which case-markers are morphological realizations of agreement features. If this sort of view can be maintained, case and agreement systems can be handled by grammatical theories which remain relatively featurally sparse. Language does not include related features for case and agreement; these are the same features appearing in distinct morphological environments.

Amy Rose Deal (Harvard) will give this job talk as a candidate for the Linguistics department’s Syntax faculty position.