Taking Sides: From Coherent Practice in Cohesive Neighborhoods to Macro Organization

We show how simple rules shared by actors acting somewhat independently and with local ratherthan complete global information can nonetheless generate coherent global structures. In the caseof dual organization, from analysis of actual marriage networks and genealogical linkages, wefind many ethnographic instances where two-sided networks and marriage choices go unnoticedby ethnographers because global labels and descent rules for sides are absent. To understandglobal structures and institutions that may be at play, unnoticed, in social systems, it is simplynot sufficient to look for shared labels attached to the parts of global structure: their structuremay reside in patterns of relationships, in their instantiation. What patterns residing inrelationships instantiate, however, is not necessarily a set of local decision rules that are sharedand identically labeled, but rather sets of local outcomes of behavior that contribute – in possiblyheterogeneous even if structurally equivalent ways – to a global configuration. It proves critical, then, to look at the structure or logic of instantiation via network or other formsof relational analysis, if we want to discover patterns and principles that do not already come touse nicely packaged and labeled. One of the keys to discovery of patterns and principles that arenot already obvious is to look at theorems or processes whereby different levels of analysis arenecessarily linked. Micro-macro linkages and principles of emergence may prove central tounderstanding and explaining social processes, the potential for historical transformations, orarchtypical patterns underlying even highly dynamical systems. Further, if we develop models atdifferent levels – cultural, demographic, decision-making, linguistic and social – and look fortheir linkages, it is crucial to study instantiation at each level not only in terms of overall patterns(logics of instantiation) but also to study patterns exceptions (logics of exceptions), since these may be key to how patterns at different levels interlock. The general arguments of this paper are instantiated by a detailed study of dual sided networks inthe genealogical and marriage network of Pul Eliya, Sri Lanka, documented by Edmond Leach, bya reevaluation of the theory of restricted exchange of Lévi-Strauss, and by a critical reexaminationof the literature on Dravidian kinship systems in South Asia.