This dissertation provides new empirical discoveries with consequences both for how case is assigned and the range of possible types of cases. In this dissertation, I explore the relationship between Voice, case and subjecthood through the lens of Lithuanian, a Baltic language. Evidence from the active existential construction shows that the structural accusative case can be assigned in the absence of a higher c-commanding nominal. Specifically, I demonstrate that Lithuanian exhibits an active existential Voice – a Voice which assigns accusative case to a grammatical object and is realized by active morphology, but whose external argument is not syntactically projected. This finding counterexemplifies Burzio’s(1986) Generalization, its alternative versions (e.g., Kratzer 1994, 1996; Legate 2014) and related theories such as Dependent Case Theory (Marantz 1991; Woolford 2003; McFadden2004; Bobaljik 2008; Preminger 2014). I demonstrate that accusative case assignment is a property of a functional head independent of the projection of a specifier, and propose anew flavor of active Voice, one that assigns accusative case and yet semantically introduces the initiator as existentially bound rather than projecting a specifier. The properties of Voice are also examined by contrasting two constructions: the-ma/-ta impersonal and the canonical passive. I argue that while both constructions overlap morphologically, they are syntactically distinct. Although the Lithuanian impersonal patterns with the Ukrainian cognate -no/-to passive in allowing an auxiliary, it behaves like an active voice with a null projected initiator - a pattern found in the Polish-no/-to impersonal and other impersonals crosslinguistically (Blevins 2003; Maling and Sigurjónsdóttir 2002; Lavine2005, 2013; McCloskey 2007; Legate 2014). I show that the Lithuanian passive lacks a syn-tactically realized initiator and selects for a type of Voice without a specifier (in line with Bruening 2013; Legate 2014; i.a. contra Collins 2005). Empirical work on case has established a distinction between two cases, structural vs.non-structural (Chomsky 1981, 1986; Woolford 2006; Pesetsky and Torrego 2011; i.a). My dissertation challenges this dichotomy by identifying a type of case, namely marked structural, that falls between these categories depending on the syntactic environment it is realized in. Normally, non-structural cases (inherent, inert, lexical) are all assigned along with aθ-role. I demonstrate that marked structural case is like a structural case in not being assigned thematically. Rather, it is assigned by a thematic Voice head (for a similar approach in Icelandic see Schäfer 2008; E.F Sigurðsson 2017). However, this case also behaves like inherent case in that it must be obligatorily assigned and its assignment is insensitive to the featural makeup of the thematic VoiceP e.g., active vs. passive. This dissertation contributes to Case Theory by showing that there exist mixed cases like marked structural case, which constitute an intermediate step between structural case and non-structural case. Lastly, this dissertation provides important insights for subjecthood theories by identifying two types of non-nominative subjects in the language. Non-nominative subjects are normally assigned non-structural case lexically determined by a specific class of predicates (Zaenen et al. 1985; Sigurðsson 2002, 2004; i.a.). I demonstrate that non-nominative subjects can vary in terms of their case assignment and do not constitute a homogeneous class. I establish a number of syntactic tests for subjecthood in the language. Using these tests, Ishow that the genitive subject of the evidential construction behaves like a canonical nominative subject and is assigned a structural case by a functional head. In contrast, the dative subject of lack-class predicates shows only a subset of subjecthood properties and its case is non-structural assigned by a lexical verb. The contrast between the two non-nominative subjects provides independent evidence for the separation of syntactic case from its morphological form (for a syntactic approach to case see Vergnaud 1977/2008; Chomsky 1981,1995; Legate 2008).