Nominative case and AGREE(ment)

Date of Completion

January 2005


Language, Linguistics




This study attempts to understand the nature of Case marking and agreement through the detailed study of Icelandic and Japanese. My investigation focuses especially on nominative Case on certain objects and on non-finite subjects. ^ This dissertation addresses the question of how nominative case can surface on NPs other than finite subjects. In addressing this question, I develop a new theory of Case and agreement based on Chomsky's (2000, 2001, 2004) recent theory of long distance agreement. I propose that AGREE by a single Probe with multiple Goals can take place sequentially (Sequential AGREE). ^ Under the theory of Sequential AGREE, I make two significant assumptions: nominative is uniformly licensed by T and the size of infinitival complement clauses varies (Wurmbrand (2001b)). ^ By investigating Icelandic, I show that not only finite T but also non-finite T licenses nominative. Evidence comes from the person restrictions in dative-nominative constructions, Case-marked PRO, and intervention effects in Icelandic. Adopting Wurmbrand's (2001b) idea, I give a unified account for the distribution of embedded nominative NPs and optional agreement facts between such NPs and the finite verb in bi-clausal dative-nominative constructions in Icelandic. I propose that the optionality comes from whether the infinitival complement is TP or smaller than TP in Icelandic. ^ Based on Japanese facts of scope interactions among the nominative object, potential verb, and negation, I argue that nominative is licensed only by T and not by any other category between T and the nominative argument licenses nominative. Case alternation in potential constructions in Japanese derives from Wurmbrand's (2001 b) idea: when the complement clause is larger than ν*P, accusative is licensed on the object by ν*, when the complement clause is smaller than ν*P, nominative is licensed by the matrix T via Sequential AGREE. ^ Providing relevant data from Icelandic and Japanese, I attempt to argue that nominative NPs in finite clauses are all licensed by the same T and that nominative NPs in non-finite clauses can be licensed exactly in the same way as the ones in finite clauses. Throughout the dissertation, I solidify the claim that regardless of its finiteness, only T licenses nominative Case. ^