| In this dissertation, I account for mismatches between the syntax and the morphology of verb agreement in a variety of languages, using a modified form of Lexical-Functional Grammar as my framework.; Chapter one outlines the problem and past attempts to deal with it, and Chapter two outlines the framework used in the dissertation and the assumptions about the nature of verb agreement which underlie it. Chapter three examines a variety of languages, primarily Abkhaz, and shows that these languages can contain "dummy" verb agreement which corresponds to nothing in the syntax or semantics; to account for this, I suggest that verbs in these languages subcategorize separately for agreement markers and for syntactic arguments, with corresponding markers and arguments connected by linking.; Chapter four examines verb agreement in Georgian, particularly the "Inversion" construction, in which the subject is marked on the verb with indirect object agreement and the direct object is marked on the verb with subject agreement. I show that the morphological subcategorization frame introduced in Chapter three allows us to account for the facts of Inversion in a straightforward way, using two devices: a marked linking rule and a lexical rule. These language-specific devices can overrule the universal default conditions Direct Linking and Symmetry, which govern the relationship between syntactic and morphological subcategorization.; Chapter five applies this system to ergative verb agreement in Abkhaz; I show that Abkhaz syntax is not ergative, and that the agreement can be neatly accounted for by a marked linking rule. In the second half of Chapter five and in Chapter six, I examine the agreement systems of Mohawk and Choctaw, in which, among other things, subjects of some verbs are marked with object agreement. I show that the syntax of these verbs is normal, and that the agreement can be accounted for by lexical rules which allow certain agreement markers in unexpected places in the morphological frame. Chapter seven summarizes the conclusions of the dissertation and compares my system to other similar analyses. |
