Theta-sensitive Dependent Case Assignment in Korean Multiple Case Constructions

Abstract

Korean Multiple Case Constructions (MCCs) have presented a long-standing issue for theories of morphological case. These constructions, in which multiple instances of the same case appear within a clause, directly challenge theories under which a given functional head assigns case to a single nominal. Two types of Multiple Case Constructions in Korean have garnered much attention in the syntactic literature -- Multiple Nominative Case constructions (MNCs) and Multiple Accusative Case constructions (MACs). Up to this point, multiple attempts to account for these case patterns have been attempted under the Agree model, but so far, none have proven to be fully satisfactory. This thesis examines the ability of a different case model, namely, the Dependent Case Model, to account for MCC data. The Dependent Case Model is fundamentally distinct from models under which case is assigned by a functional head. Under this model, morphological case is assigned configurationally, that is, according to the relation between the nominals themselves. Dependent Case, in particular, is assigned when two caseless nominals appear in an asymmetric c-command relation, the c-commanding nominal assigning dependent case to the c-commanded nominal. When attempting to account for MNC data, this model runs into the issue that, according to its case calculus, the higher DP1 ought to assign accusative to the lower DP2. Thus, the Dependent Case Model on its own is not able to account for MNC data. In order to resolve this issue, I utilize the insights of Heycock (1993) and Yoon (2004, 2007, 2015), among others, that the higher DP1 in an MNC is the Major Subject (MS) of a sentential predicate, formed from the lexical predicate and the Grammatical Subject (GS). This predicate assigns its own non-external θ-role to the MS. Attempting to formally encode the lack of assignment of an external θ-role to the lack of dependent case, I propose the Theta-sensitive Dependent Case-assignment Hypothesis. Under this hypothesis, the Dependent Case assignment calculus is amended such that a c-commanding nominal assigns dependent case to a c-commanded nominal if and only if the commanding nominal has been assigned an external θ-role. This accounts for the MNC data by preventing dependent case-assignment from being triggered between the Major Subject and the Grammatical Subject, as the Major Subject is assigned a non-external θ-role by the sentential predicate. This lack of external θ-role is also utilized to explain Nominative Object Constructions (NOCs) in Korean. In addition to MNCs, this new theta-sensitive Dependent Case Model is applied to MACs, and it is found that it is able to account for this case pattern as well. In MACs, the subject is assigned an external θ-role by the vdo head, which triggers dependent case assignment, and results in accusative case on the higher of the two accusative-marked nominals. At this point, dependent case assignment is still triggered, as the external θ-role marked nominal and the lower of the two accusative marked DPs will still be caseless

as a result, the lower DP is able to receive its accusative case. Finally, ECM constructions are analyzed, and it is shown that, while the traditional Dependent Case Model is unable to account for this case pattern as well, the Theta-sensitive Dependent Case-assignment Hypothesis allows the model to account for the data naturally, including ACC/NOM case variation. The Theta-dependent Case-assignment Hypothesis is shown to provide a novel account of the data in all of the analyzed case constructions. Further, it provides a direct link between theta and case, something that has been lacking in the Dependent Case Model up to this point. Application of this theta-sensitive Dependent Case Model to MNC and MAC data allows for a theoretically and empirically satisfying account of the data, and adds to significant insights to the body of literature, both on MCCs and on case in general.