UC Santa Barbara

This dissertation presents a comprehensive account of the symmetrical voice system in Besemah, an under-documented Malayic language spoken in the highlands of southwest Sumatra, Indonesia. Utilizing a corpus primarily consisting of conversational data, but also including narrative data, this study treats both the syntactic structure and discourse properties of symmetrical voice constructions in Besemah. Previous research on voice in the languages of western Indonesia has sought to understand these languages in terms of well-established systems of voice and grammatical relations, whether that be active-passive or ergative-absolutive. Since the notion of symmetrical voice—a voice system with multiple transitive voice constructions, none of which is clearly the ‘basic’ voice form—was introduced by Foley (1998) just under two decades ago, it has provided valuable insights into the voice systems of the languages of western Indonesia (Riesberg 2014). Drawing on these insights, this dissertation presents a thorough treatment of symmetrical voice in Besemah, which has not been the subject of any in-depth grammatical analysis.

The dissertation has two primary objectives. The first objective is to describe the syntactic nature of symmetrical voice by identifying grammatical relations in Besemah. Based on several ‘diagnostic’ constructions that have been used to provide evidence for grammatical relations in related languages of western Indonesia, this study identifies two grammatical relations in Besemah, primary argument and secondary argument, by utilizing data from the corpus of Besemah. While two of these ‘diagnostic’ constructions, word order and quantifiability, provide evidence for grammatical relations in Besemah, this study demonstrates that many of these ‘diagnostic’ constructions cannot be used for determining grammatical relations in Besemah.

The second objective seeks to answer the following question concerning voice selection in Besemah: at any given point in a conversation, what factors lead a speaker to choose one symmetrical voice construction over the other? In order to answer this question, this study uses advanced statistics to investigate the role of information flow (Chafe 1994), syntactic priming (Gries 2005), and collostructional analysis (Stefanowitsch & Gries 2003) in voice selection. The findings reveal that each of these factors play an important role in voice selection in Besemah conversation.