ABSTRACT
About five years ago, we began to develop a mode of social network analysis that would build on the considerable success anthropologists had achieved using kinship-based methods but would be applicable in situations where kinship was not of overriding importance. We consciously intended that our methodology would arise from several features of anthropological theory. In a general sense, this objective is similar to that of earlier anthropological network researchers, although our program differed insofar as we sought to preserve a rather different mix of features of the older kinship-based studies and weighted them rather differently (see Foster, 1979). Like the earlier network researchers, we focus on individuals' behavior within the constraints imposed by their social networks, and we insist that our approach be useful crossculturally. Unlike them, however, we wished to work with models in which individuals and sets of individuals are the units of analysis in a very direct way; Levi-Strauss (1963), who calls such models "mechanical models," has argued that they are a fundamental characteristic of anthropological analysis. Moreover, we wished to formalize the individual behavior aspect of the analysis far more than had
been done in the earlier network research and to do so in such a way that it could be fully integrated with the method of structural analysis. This formalization was to be achieved by incorporating some features of cognitive anthropological decision theory which, in addition, contributed to the crosscultural applicability of the method. Although the point cannot be developed here, we wish to emphasize that each of these features is motivated by a large body of anthropological theory (see Foster, 1979; Foster and Seidman, 1981).
