Briggs Buchanan, Marcus J. Hamilton, James C. Hartley and Steven L. Kuhn

We examine the spatial scale of prehistoric social networks represented by point types documented in western North America through comparison with ethnohistorically documented Native American interactive networks at different levels of inclusion. The ethnohistorical data come from Joseph Jorgensen's (1980) Western Indians, which maps tribal boundaries at European contact and the associated language lineage for each tribe. We assume that frequency of interaction follows language relationships. Proximity aside, people will share ideas more often if they possess a language, or part of a language, in common. We use tribal regions and different levels of language affiliation (families, large language groupings, and phyla) that represent increasingly broad spatial scales of social interaction. We compare these measures with the areas calculated for point types in the same general region to determine which level of social interaction recorded ethnohistorically best fits with the point type data. Our analyses show that point type areas most closely resemble the spatial extents of large language groupings and language phyla. The areas of point types are greater than individual tribal regions recorded in western North America at the time of European contact and language families. Based on these results, we suggest that the conflation of point types with prehistoric cultures commonly implied in archeology is not justified. Building on the fundamental ideas of the culture historians, we suggest that point type distributions are a consequence of extensive social interaction networks where combinations of functional and neutral point traits are shared and inherited over a large area.