The phonology-syntax interface and polysynthesis: a study of Kiowa and Saulteaux Ojibwe

Abstract

This dissertation examines phonology and its interface with morphology and/or syntax. Current models diverge in terms of what phonology may reference in order to delimit domains for phonological processes: primarily phonology (the Phenomenon-Based Approach), a mix of morphology and syntax (Relational Mapping), or syntax only (Syntax-Driven Mapping and the Syntactic Spell-Out Approach). While comparisons exist between some -- but not all -- approaches, no comparison evaluates the approaches using the same language data (e.g. Selkirk, 2011; Vogel, to appear). The primary contribution of the current work is to fill this gap. To do this, data from two languages are presented as test cases: Kiowa and Saulteaux Ojibwe. Selected because they are traditionally described as polysynthetic, the languages’ extreme levels of morpho-syntactic complexity often blur the lines between words, phrases, and clauses, offering a crucial test for interface approaches. ☐ Through comprehensive phonological analysis of each language, focusing on syllables and larger strings up to full clauses, it is shown that Relational Mapping (c.f. Nespor and Vogel 1986) is the only approach which successfully accounts for the data in both languages. Relational Mapping assumes a strict separation between morphology and syntax, offering new insight to the ongoing debate regarding the necessity or even existence of a morphological module of grammar (see Bruening 2018). It is argued that while Relational Mapping is successful, a syntax-only approach is theoretically preferred. Finally, a new theoretical proposal is put forward which accounts for both languages with the same level of success, but which references only syntax. The approach yields new advantages and disadvantages compared to Relational Mapping, which are discussed in detail. ☐ A secondary contribution of the current work is that of language-specific analysis and documentation. In addition to their theoretical value for this line of questioning, each language is endangered. Unless otherwise cited, the language data presented here comes from fieldwork in each language community. Thus, this body of work serves as a major step forward in the documentation of each language - particularly of phonological processes. For Kiowa, the phonology has not been studied closely since the 1980s, and Saulteaux phonology has never been analyzed to this degree.