The Multilevel Networks Approach to Kinship -l'approache Multi-réticulaire De La Parenté

This overview of the development of the multilevel network approach to kinship has at its core the definition of a multilevel network as one in which a node at one level may contain a network of relations at the next level (Harary and Batell 1981). The definition is potentially recursive so as to define a particular kind of system, a network with potentially many levels. André Weil (1949), for example, began the mathematical study of kinship networks with the idea of studying kinship as networks of relations among families, embedding at lower levels the distinct relations among individuals as the members of these families (families of orientation being one type of lower level networks and families of procreation another). These definitions open up questions of how different levels of kinship and other phenomena are embedded in or map onto other levels. The multilevel network (multi-reticular) framework opens kinship to visualization, new mathematical theorems and empirical studies, and the testing, in new ways, of theoretical propositions with empirical data. A more recent generalization of the multi-level approach has been done from networks of families and individuals connected by genealogical relations to the much-reduced cognitive networks that encode the relations among kinship terms at a distinctly and self-contained linguistic level, while the terms that define nodes at this level also map onto the relations among individuals or families. Read (1984) was the first to show that the relative-product (in English, for example, brother of father = uncle) relations among terms for kinship define kinship logics at a self-contained level distinct from but potentially applicable (again, in multiple ways) to the relations among persons or families. Aspects of kinship thus can be identified as systems in themselves (such as kin terms, role relations, etc.) but also in relation to other types of relations (Leach 1961, Tambiah 2002), including systems of beliefs (Leaf 2007) and organizations (Leaf 2009). We show how the current reintegration of anthropological theory and findings in the broad areas touched upon by the network approach to kinship were heavily stimulated by the early influence of Lévi-Strauss on Leach (1964, 1961) and how Leach‟s influence and insight were key to the modern formulation of the multi-reticular approach to ethnography and anthropological theory in general. This review also covers the development of anthropological software for kinship and ethnographic network analysis (parental Pgraph software), the use of general purpose network software (e.g., Pajek) for study of kinship and ethnographic network data, and the development of software with alternative frameworks for specialized kinship analyses focused both on networks of families and networks of individuals (Puck software); and explore a range of new results, theorems, and methods that have generated new theoretical and empirical perspectives in anthropology since 1949 and provided integration with earlier paradigms for the study of kinship.