This thesis has a number of aims. The first is to give an account of the social organization of a particular Amazonian society that of the Yaminahua, an indian group of South Eastern Peru. The descriptive part of the thesis gives an account of such things as: the organization of work and production, the structures and ,developmental cycles of households, communities and local groups, the organization of kinship and marriage. All these things present features common to many Amazonian societies: an economic system based on swidden agr iculture, hunting and fishing; small communities which regularly move to exploit new hunting grounds and garden sites; extended family residencial units; a fluid political system based upon the .consensus mobilized by headmen; a simple classificatory kin s hip terminology associated with bilate~al cross-cousin marriage; in short? a simple society with every appearance of extreme fluidity and minimal "structur-e". From a theoretical perspective, the thesis aims to reconcile this fluidity of social practice with the fact that there also exists, within Yaminahua ideology, a highly elaborate and systematic set of ideas about their own social stru~fure. These centre around a system of personal-name categories and moieties, which are in turn linked to concepts of a dual order inherent in the constitution of the world as a whole. The ·central question of the thesis, is; "How are we to understand this apparant paradox?", a question which it tries to answer by considering, firstly , the historical tran sformations of Yaminahua society over the last one hundred years, and secondly, at a more theoretical level, the intrinsic nature of the "fit" between this type of ideology (highly symbolic, highly logical - · - what used to be called a totemic system) with social practice an~ the exigencies of social life. Here the thesi s addresses certain issues arising from the debates about Levi-Straussian str ucturalism and the other types of symbolic analysis which have dominated this field in recent years. One of the assumptions of this work is that these have misrepresented Amazonian ideologies by flattening them out into a static web of analogies and symbols. It argues that these ideologies can only be under stood when their symbolic content is reintegrated with their conceptual content, and when their highly systematic aspects (ritual and totemic systems) are reintegrated with the much less systematic aspects, closer to the domain of every-day social concerns. This line of argument is demon strated with an extended consideration of shamanism, looking at its conceptual bases in theories of the person, spirit and causality. The discussion shows how these are intimately linked to the ideas of structure and order in the world already mentioned but also how, as a practical activity, it interweaves ideas and symbols quite independently of that structure and can, in no sense, be Understood as merely its expression or a means of reasserting it.