Definite markers, phi-features, and agreement: A morphosyntactic investigation of the Amharic DP

This dissertation has two inter-related goals: (i) to describe and provide novel analyses of three types of important and difficult phenomena within Amharic DPs and (ii) to explore the properties of the syntax-morphology interface. The core phenomena explicated are the unusual distribution of the definite marker, the gender system (which relies heavily on natural/biological gender), and the plural system (which involves the relations between several distinct types of plurals). The dissertation shows how these phenomena have important ramifications for morphosyntactic cyclicity, the morphosyntactic treatment of phi-features, and the relationship between the syntax and morphology of agreement.;I show that the definite marker is best accounted for in two sub-cases. When obligatory, it is a second position clitic, and when optional, it is the reflex of a definiteness agreement process. The operation that places the definite marker in second position is shown to be unable to access previous syntactic cycles, i.e., there is phase impenetrability at PF.;The accounts of both gender and number explore the division of labor between little n (the nominalizing head) and the root. I argue that both natural gender and grammatical gender must be present in the syntax, and propose that grammatical gender is a feature on roots, whereas natural gender is a feature on little n. In the number domain, the differences between regular and irregular plurals (and the behavior of double plurals in Amharic) are best analyzed with two different slots for plural features: they can appear on Number, for regular plurals, and on little n, for irregular plurals. The system of gendered plurals is derived from the interaction of the gender and number features on little n.;Finally, I explore the syntax and morphology of feature sharing, arguing that the plural feature is in fact shared between Number and little n, and developing an account of the morphological realization of shared features. This account (correctly) predicts that all the types of plurals in Amharic are synonymous yet morphologically distinct. I further generalize the feature sharing analysis of plurals to account for all cases of DP-internal agreement.