The Pacific Coast Athabaskan (PCA) languages are part of the Athabaskan language family, one of the most geographically widespread in North America. Over a millennium ago Athabaskan-speaking groups migrated into northwestern California and southwestern Oregon from a northern point of origin several hundred miles away, but even after several centuries separated from other languages in the family and in contact with neighboring non-Athabaskan populations their languages changed only incrementally and maintained an essentially Athabaskan character. Beginning in the mid-19th century, disruptions associated with colonization brought closely-related PCA varieties into intimate contact both with each other and with English, leading eventually to their current state of critical endangerment. This dissertation explores these diachronic developments and seeks to understand (a) how the PCA languages are related to each other and to the rest of the Athabaskan language family, (b) the social and structural dynamics of dialect contact between Athabaskan varieties from the mid-19th to the mid-20th century, and (c) the linguistic consequences of bilingualism as Athabaskan communities shifted from their heritage languages to English.
Chapter 1 situates the research in a theoretical and methodological context that informs subsequent chapters. Special attention is paid to outlining competing modes of explanation for diachronic outcomes observed in dialect contact situations, comparing demographic approaches that make reference to population size or density versus indexical approaches that invoke people's strategic deployment of linguistic resources to achieve social aims. This chapter also provides an overview of the PCA languages, including their classification, the extent of their aboriginal territory, and the main sources of documentation consulted in this study.
Chapter 2 addresses questions related to linguistic phylogenesis that took place over a relatively long timespan, focusing specifically on the question of whether or not the PCA languages comprise a well-defined subgroup within the Athabaskan family. This chapter uses recent computational methods that have been adapted from the study of biological evolution to address analogous issues in historical linguistics. Using a combination of lexical, phonological, and morphological characters, the chapter considers how results obtained with this computational approach compare with findings from traditional historical-comparative methods, where it has been suggested that tree-like branching structures are inappropriate models of Athabaskan linguistic relationships. The main empirical finding of this chapter is that the PCA languages emerge as a well-supported subgroup of Athabaskan under a variety of conditions: with lexical or non-lexical characters or a combination of both, and with different model assumptions about rates of evolution across lineages. Other findings of interest include topological differences obtained using multi-state versus binary coding schemes and the extent to which lexical, phonological, and morphological characters give convergent results in phylogenetic analysis.
Chapters 3 and 4 address the linguistic effects of contact between Athabaskan populations that resulted when indigenous people of California and Oregon were dispossessed and consolidated on a small number of reservations in the mid-19th century. Some current approaches to dialect contact predict that koineization leading to leveling of structural differences between input dialects is virtually inevitable, and that the direction of leveling will favor majority variants. Other theories emphasize speakers' agency in selecting the linguistic variables in their environment and allow for a wider variety of outcomes. A close examination of parameters of variation in Hoopa Valley, California (chapter 3), and Siletz, Oregon (chapter 4) shows overall that structural leveling of dialect differences did occur in this period. However, leveling applied unevenly, with some parameters of regional variation persisting into the last generation of speakers to acquire native fluency in the PCA languages, in some cases even for dialects that were a tiny minority. These results suggest that a theory allowing for the maintenance of dialect differences via social-indexical considerations does a better job of explaining the full range of observed outcomes than one relying on deterministic majority-rule demographic principles alone.
Chapter 5 considers the linguistic consequences of contact between English and Hupa, the best-documented PCA language. The chapter focuses on lexical innovations in Hupa introduced to refer to the plethora of new technologies that became part of everyday life in Hoopa Valley in the 19th century. The main finding is that Hupa speakers borrowed very few lexical items from English, preferring instead to adapt existing Hupa words or to coin new ones using productive morphological resources. In non-lexical domains, an examination of main clause constituent order reveals that younger generations of Hupa speakers tended to use a postverbal nominal position less frequently than earlier generations had. However, it is argued that this is due not to direct bilingual interference from English, but rather to pragmatic contraction disfavoring the use of marked word orders. Overall, despite a high incidence of Hupa-English bilingualism already by the turn of the 20th century and the fact that wholesale language shift was underway in this period, Hupa appears to have undergone very few lexical or structural changes that were borrowed or otherwise directly modeled on English. This is interpreted as a continuation of conservative stances towards exogenous linguistic innovations that were a part of pre-contact linguistic practice both within the Athabaskan language family and regionally across California.
Chapter 6 concludes, summarizing the main findings of the dissertation and outlining directions for future research.