Animacy and Inverse in Movima: A Corpus Study

International audienceCoding of transitive clauses in Movima (lowland Bolivia) including a speech act participant reflects a person hierarchy, 1 > 2 > 3; when a lower-ranking person acts on a higher-ranking one, the verb is marked as inverse. When a transitive clause contains only third person arguments, direct vs. inverse marking of the verb depends on relative discourse prominence and on an animacy hierarchy; in actual discourse, the inverse never occurs in clauses where a higher-ranking entity acts on a lower-ranking one, while the inverse is frequent when a lower-ranking entity acts on a higher-ranking one. Further research is needed to clarify the exact discourse factors involved. 1. Introduction. This article investigates the role of animacy in the direct-inverse system of Movima, an unclassified, endangered language of southwestern Amazonia (Bolivia). In Movima, there is neither case nor agreement marking; instead, the linear order of constituents in transitive clauses is fixed, and participant roles are assigned by verb morphology. The ordering of the arguments depends on their referential properties, which include their relative status in a hierarchy of person and animacy, and on their discourse prominence. The participant roles of the arguments are indicated by direct or inverse marking on the verb. The present study examines the conditions of argument coding in transitive clauses containing only third person referents. While in scenarios involving a speech-act participant, the expression of the arguments and the corresponding direct or inverse marking on the verb can be predicted from the position of the referents in Movima's person hierarchy (in which first person is highest: 1 > 2 > 3), this is not the case when two third persons are involved. Here, the question is which principles determine their occurrence in either the one or the other syntactic slot. Basically, these principles are most adequately captured by DeLancey's (1981) notions of " viewpoint, " which comprises animacy and discourse prominence of the referents, and " attention flow, " which accounts for a tendency for forms coding the actor to precede forms coding the undergoer. Analysis of a firsthand oral text corpus shows that, while speakers are, in principle , free to choose the viewpoint from which to present an event, they usually prefer that of the actor–which is then expressed in the first position after the direct-marked predicate–but that animacy also plays an important role in this choice. The remainder of this article is structured as follows. Section 2 presents the structure of pragmatically unmarked transitive clauses in Movima and show