This article attempts to extract the theoretical implications arising from the fact that anthropology not only studies relations, but that the knowledge it produces in the process is itself a relation. It therefore proposes an image of anthropology as an activity founded on the premise that the procedures characteristic of the discipline are conceptually of the same order as those it investigates. Among these implications is the rejection of the contemporary notion that each culture or society embodies a specific solution to a generic problem, filling a universal form (the anthropological concept) with a particular content (the native conceptions). Much the opposite: the image proposed here suggests that the problems themselves are radically heterogenic, and that the anthropologist cannot know beforehand what these will be.
