UC Berkeley

This dissertation provides a description and analysis of the syntax of Kawahíva, a critically endangered Tupí-Guaraní language spoken by approximately 560 people in the Brazilian Amazon basin. The first part of this dissertation includes an overview of Kawahíva documentation and revitalization efforts, its phonology, and a comprehensive grammatical sketch of the language.

In the second half of this dissertation, I analyze two key topics of Kawahíva syntax: clause structure and relativization. In particular, I offer an analysis of VSO word order of Kawahíva clause structure. I argue that verb-initial word order in Kawahíva results from long head movement, a type of syntactic head movement (Harizanov and Gribanova 2019), similar to phrasal movement. This analysis is supported by the hallmarks of syntactic movement in Kawahíva verb movement, including interpretive semantic effects and nonlocality. I also show that two theoretical accounts of the V1 order – Remnant VP Movement and Head Movement – are insufficient to derive the V1 order in Kawahíva. This study is the first to demonstrate that a language can use long head movement as a general principle to create the verb-initial order.

The final chapter of the dissertation investigates relativization in Kawahíva. Several languages are documented with a relativization strategy where a clausal nominalization is used as an adnominal modifier. It follows from this that relativization in these languages is achieved by nominalization morphology itself, without any additional syntactic processes taking place (Comrie and Thompson 1985; Keenan 1985; Andrews 2007; Shibatani 2009). However, I argue that characterizing these structures as nominalizations in Kawahíva is insufficient and that there is additional clear evidence that the nominalized clause also involves a distinct operation, namely relativization. Evidence for this claim will come from showing that adnominal nominalizations exhibit the hallmark properties of relative clauses (and other extraction-based constructions) cross-linguistically: i) sensitivity to island effects and ii) the formation of a long-distance dependency between the gap in the nominalization and the filler (i.e., the modified noun).