A Parametric Syntax of Aspectually Conditioned Split-Ergativity

Abstract

1. OUTLINE In some languages (e.g., Georgian, Hindi, Samoan, Nepali, etc.), where the ergative system coexists with the accusative one, the alternation between ergativity and accusativity is determined by aspectuality (cf. DeLancey 1981, Dixon 1994, and Ura 2000). In this paper, I will explore the syntactic mechanism of the aspectually conditioned split-ergativity and try to explicate the parameter settings which distinguish the languages with the aspectually conditioned split-ergativity from the ergative languages with no such split and from the accusative languages in general. For the purpose of giving a coherent explanation to some syntactic phenomena found in the languages mentioned above, I will make a Minimalist approach under Chomsky’s (1995, 2000) theory of feature-checking, by refining and updating Ura’s (2000) hypothesis about the parametric syntax of the ergative system, according to which ergative languages have the parameter that enables the logical subject, basegenerated at the Spec of v, to participate in a checking relation directly with v (cf. Massam 2001 for a very similar idea (for some other variations of this idea, see the papers by Carnie, Legate, Massam, Spreng, and Tsedryk in this volume)), which gives rise to a situation where the logical, underlying subject (SUBJ) and the logical, underlying object (OBJ) have their Cases associated with v and Infl, respectively (that is, SUBJ in the ergative system is marked as accusative (= ergative) just like OBJ in the accusative system, and OBJ in the ergative system is marked as nominative (= absolutive) just like SUBJ in the accusative system). Under this hypothesis, SUBJ in the ergative system can move to the Spec of Infl (due to the EPP) after being marked as ergative as a result of its checking of v’s nominal feature, and OBJ in situ can associate its Case with Infl by Agree thanks to SUBJ’s displacement from the Spec of v to the Spec of Infl. This gives rise to a situation where SUBJ and OBJ in the ergative system occupy the same surface positions as their corresponding positions in the accusative system in spite of the difference in their Case forms. Because both SUBJ and OBJ in the languages with the aspectually conditioned split-ergativity show the same syntactic behaviours with respect to their grammatical functions regardless of whether the clause involving them is in the ergative system or in the accusative system, the above hypothesis enables us to give a consistent