Date of Completion


Embargo Period



unaccusativity, Russian, syntax, morphology, animacy

Major Advisor

Jonathan David Bobaljik

Associate Advisor

Andrea Calabrese

Associate Advisor

Susanne Wurmbrand

Associate Advisor

Željko Bošković

Field of Study



Doctor of Philosophy

Open Access

Open Access


This dissertation investigates the effects of animacy in the syntax and morpho-syntax of language. More precisely, I examine cases in which minimal pairs, varying only in the animacy of the subject NP, show different morpho-syntactic behavior. This study suggests that a significant part of the explanation of these effects lies in the syntactic representation of argument structure—that animate and inanimate NPs, under certain conditions, receive different thematic roles, and thus occupy different structural positions at the syntactic level(s) where argument structure is represented. The findings of this project are shown to constitute a part of a broader generalization and find empirical support in a cross-linguistic perspective.One of the primary findings in this dissertation is an empirical generalization concerning animate arguments of unaccusative verbs in Russian which display a previously unnoticed type of ‘variable behavior’ for unaccusative diagnostics. I propose that animate arguments must be interpreted as Experiencers whenever possible, and can only be Themes when an Experiencer interpretation is unavailable. The thematic distinction correlates with a structural difference that explains the variable behavior under unaccusativity diagnostics.Another important contribution of this study is a uniform account of Russian and Italian data involving typical unergative predicates which are problematic for the original formulation of the Unaccusativity Hypothesis. I propose that the parallels between the Russian and Italian facts occur due to an alternation in perspective structure or the framing of the event, which is syntactically represented as a choice between two argument structure frames, determined by such factors as contextual inference, the verb's lexical semantics and general knowledge.