twomodalitiesofcaseassignmentinsakha.docVIP

Two Modalities of Case Assignment in Sakha Mark Baker and Nadya Vinokurova Rutgers University February, 2008 Abstract: There are two competing ideas about how morphological case is assigned in the recent generative literature. The standard view, developed by Chomsky, is that case is assigned by designated functional heads to the closest NP under an Agree relation. The alternative, proposed by Marantz 1991 and others, is that case is assigned to one NP if there is a second NP in the same local domain. We present evidence from the Turkic language Sakha that these two approaches to case theory are complementary. Specifically, we argue that accusative case and dative case in Sakha are assigned by Marantz-style configurational rules that do not refer to functional categories. Crucial evidence for this comes from passives, agentive nominalizations, subject raising, possessor raising, and case assignment in PPs. In contrast, nominative and genitive case are assigned by functional heads (T and D) in the Chomskian way; this is needed to explain the relationship between case marking and agreement in Sakha. We conclude that two distinct systems of case assignment can coexist peacefully, not only in Universal Grammar, but even in the grammar of particular languages. 1. Introduction In the generative syntax literature, there coexist two major ideas about how morphological case markers come to be associated with individual noun phrases in ways that reflect aspects of the syntactic structures those noun phrases appear in. The more common idea is that structural case features are assigned to NPs by nearby functional heads. For example, nominative case might be assigned by (finite) T to the nearest NP that T c-commands. In a similar manner, accusative case might be assigned by (active, transitive) v to the nearest NP it c-commands, genitive case might be assigned by (possessive) D to the nearest NP, dative case might be assigned by (certain) P(s), and so on. This is