Many models of colonial interaction are build from cases of European colonialism among Native American and African peoples, and, as a result, they are often ill-suited to account for state expansion and decline in non-Western contexts. This dissertation investigates social organization and intraregional interaction in a non-western colonial context to broaden understanding of colonial interaction in diverse sociocultural settings. Drawing on social identity theory, population genetics, and social network analysis, patterns of social organization at the margins of the expansive pre-Hispanic Tiwanaku state (ca. AD 500-1100) are examined.

According to the dual diaspora model of Tiwanaku colonial organization in the Moquegua Valley of southern Peru, Chen Chen-style and Omo-style ethnic communities who colonized the valley maintained distinct ethnic identities in part through endogamous marriage practices. Biodistance analysis of cranial shape data is used to evaluate regional gene flow among Tiwanaku-affiliated communities in Moquegua. Overall, results of biodistance analysis are consistent with the dual diaspora model. Omo- and Chen Chen-style communities are distinct in mean cranial shape, and it appears that ethnic identity structured gene flow between ethnic groups. However, there are notable exceptions to the overall pattern, and it appears that marriage practices were structured by multiple factors, including ethnic affiliation, geographic proximity, and smaller scales of social organization, such as corporate kin groups.

Social network analysis of cranial shape data is used to implement a multi- and mesoscalar approach to social organization to assess family-based organization at a regional level. Results indicate the study sample constituted a social network comprised of a dense main component and a number of isolated actors. Formal approaches for identifying potential family groups (i.e., subgroup analysis) proved more effective than informal approaches. While there is no clear partition of the network into distinct subgroups that could represent extended kin networks or biological lineages, there is a cluster of closely related individuals at the core of the network who integrate a web of less-closely related actors. Subgroup analysis yielded similar results as agglomerative hierarchical cluster analysis, which suggests there is potential for social network analysis to contribute to bioarchaeological studies of social organization and bioarchaeological research in general.