Syntax of dative-accusative constructions in Japanese
MetadataShow full item record
Both ditransitive verbs and causative transitive verbs in Japanese are linked with the two verbal arguments: the dative phrase and the accusative phrase. Despite this similarity, the syntax of these verbs is in sharp contrast: the ditransitive verb construction involves the mono-clausal structure, whereas the transitive causative construction contains the bi-clausal structure (Kuroda 1965b, Saito 1982; 1985, Hoji 1985, Miyagawa 1989, among others). One crucial instance for such distinction is the behavior of the dative phrase of the two types of construction with respect to the ‘subjecthood’. The subject-oriented anaphor in Japanese (e.g., zibun ‘self’) can take the dative phrase of the transitive causative verb as its antecedent, whereas it cannot the dative phrase with the ditransitive verb as its antecedent (Kuroda 1965b). Notwithstanding the difference, this thesis attempts to proposes a unified account for these two types of dative-accusative constructions in terms of the Phase Theory within the framework of Generative Grammar (Chomsky 2000; 2001). Investigating four subcategories of dative-accusative verbs (i.e., spray/load verbs, give verbs, causative transitive verbs and causative motion verbs), I claim that there are two types of Dative Case Assignment in Japanese: the In-situ Assignment and the Assignment after Movement. The former type of assignment is manifested in VP of give verbs and in that of transitive causative verbs; on the other hand, the latter type of assignment is identified in VP of spray/load verbs and in that of causative motion verbs. In the Phase theory, the probe-goal relation between the functional head and its ccommanding goal(s) (i.e., (Multiple) Agree) governs Case-licensing mechanism. In standard assumptions, the Accusative Case domain is the c-command domain of the light verb v (Chomsky 2004). Following these assumptions, I claim that the two types of Dative Case Assignment can be ultimately attributed to the two distinctive Case features on the functional head v of the four types of dative-accusative constructions: vacc[+multiple] and vdat. If vacc[+multiple] is selected by Merge, the Dative Assignment after Movement is implemented, whereas if vdat is selected, the In-situ Dative Assignment is induced. Hence, the difference in Dative Case Assignment is predicted at which Select picks up vocabularies from the Lexicon in order to set up a reference set for a derivation of the dative-accusative construction. The in-situ dative assignment for the ditransitive construction has been proposed in the literature (e.g., Miyagawa 1996); however, no proposal of a movement-based dative assignment for the ditransitive construction has been made. This is one of the important outcomes of my thesis. However, the most important consequence of my two types of Dative Assignment is the link between two hitherto unrelated phenomena: Dative Case Assignment and the condition on argument alternation. Argument alternation has attracted much attention in the literature of lexical semantics, being independently analyzed from most of the syntactic properties of these ditransitive verbs that I examine in the thesis (Kageyama 1980; 1996, Levin 1993, Kishimoto 2001c, Iwata 2008). However, I show that the condition on argument alternation can be written solely by the syntactic terms without any stipulation of constructional meaning; namely, when vdat is selected in a numeration of a ditransitive verb, the derived verb is never licensed to participate in argument alternation, whereas when vacc [+multiple] is selected, the complex verb is licensed to participate in the alternation. A further contribution of my thesis is to accommodate a new pair within the causative-ditransitive paradigm in Japanese in addition to its already-existing membership between transitive causative verbs and give verbs (Kuno 1973, Miyagawa 1996): a pair of causative motion verbs and spray/load verbs. This new pairing further strengthens the existence of the causative-ditransitive paradigm as a natural class in Japanese. The pairing is solely motivated by the Dative Case Assignment that I propose.