Non-Nominative Subjects in Russian and Lithuanian: Case, Argument Structure, and Anaphor Bindin

In this dissertation, I aim to provide a comprehensive view of subjects in Russian and Lithuanian that have a morphological case other than nominative. Non-nominative subjects (NNSs) provide a window into the notion of subjecthood because they are lacking in some core subject properties and yet are still considered subjects. While the constructions that have dative, genitive, and accusative subjects seem to vary greatly, I show that they fall into two groups: those that have subjects with inherent case and those that have subjects with structural case. I propose a uniform way to account for how case is assigned in each construction and for the apparent ''subject'' movement that they undergo. In a later portion of the dissertation, I report on an investigation into anaphor binding by these subjects that relies on experimental methodology.;In Chapter 2, I focus on inherent non-nominative subjects, which in both Russian and Lithuanian are dative Experiencers. I argue that the two kinds of predicates in these languages with dative Experiencer subjects, psychological verbs (e.g. 'to like', nravit'sja in Russian and patikti in Lithuanian) and non-verbal psych predicates (e.g. 'sorry', zal' in Russian and gaila in Lithuanian), have different argument structures. While both datives are specifiers of an Applicative Phrase, the Theme of a psychological verb is the specifier of a vBEP (following Cuervo's 2003 proposal for Spanish) and the Theme of a non-verbal psych predicate is the complement to the predicate. In this chapter, I also account for how NNSs seem to undergo subject movement because they appear pre-verbally in discourse neutral contexts. I argue that T can inherit its uninterpretable ϕ-features from Rizzi's (1997) Fin head separately from the EPP feature via a modification to Feature Inheritance (Chomsky 2008, Richards 2008b) called Split Feature Inheritance. The NNS moves to Spec FinP to check EPP, and the [uϕ] probe on T is free to undergo Agree with any active DP.;In Chapter 3, I turn to the remaining constructions, which I argue to have subjects with structural case. In Russian, sentences with an infinitive as the predicate (i.e. Fleisher's 2006 ''Main Clause Infinitivals'') have dative subjects. I argue that these are bi-clausal (following Fleisher 2006 and Jung 2008), but that these subjects are assigned dative case by an embedded non-finite Fin head and then raise to the matrix clause. Here, Feature Inheritance has not occurred and Fin still has the [u?] feature bundle which agrees with the subject DP. I argue that this is also the source of the dative case assigned to the subjects of adjunct participial clauses in Lithuanian. The subject in these clauses moves to Spec FinP and is assigned dative via Agree with Fin. Contra Arkadiev (2012), I propose that these clauses are tenseless and therefore lack a TP layer. When these participial clauses are embedded under a matrix verb like sakyti 'say' or matyti 'see', Split Feature Inheritance is triggered and an Aspect head is what inherits [uϕ]. Because this [uϕ] is now in the verbal domain of the clause, the subject is assigned accusative case via Agree. Finally, I propose that the genitive subjects of the Lithuanian Inferential Evidential construction are subjects of a gerundial DP and assigned structural genitive case via Agree with D (cf. Lavine 2000, 2010).;The focus of Chapter 4 is an acceptability judgement experiment investigating the ability of dative NNSs to bind anaphors. Russian and Lithuanian both have subject-oriented reflexives, making anaphor binding a common subjecthood diagnostic in these languages (see Rappaport 1986). In addition, pronouns in these languages are anti-subject-oriented in that they can only be bound by non-subjects. The experiment tests predictions made by Nikolaeva's (2014) theory of the (anti) subject-orientation of anaphors in Russian, Index Raising, wherein anaphors and pronouns are the spell-out of an index that raises to certain positions in a clause. I find that dative subjects of psychological clauses cannot bind reflexives, but they can bind pronouns. Dative subjects of Russian infinitival sentences and Lithuanian participial clauses, on the other hand, bind reflexives. Sentences with these subjects binding pronouns are judged to be neither completely unacceptable nor completely acceptable. I show that we can begin to account for this pattern of binding if we assume the argument structures I propose in Chapters 2 and 3.;Chapter 5 concludes the dissertation with a review of the proposals and findings from Chapters 2 through 3.