2003 Volume 68 Issue 2 Pages 226-241
The goal of this paper is to examine the use of "comparison" in contemporary anthropology by analyzing the main features of comparative studies by mid-twentieth-century American anthropologists, which have been considered as old-fashioned since the recent decline of the classic anthropological paradigm that culture does exist and should be grasped in its totality. No anthropologists would deny that comparison is the most basic analytical tool of the discpline. Since the literary turn of anthropology, however, the plan to explicate the particularity and/or universality of human culture by comparative method has tended to be abandoned. In fact, few anthropologists dare to publish world-wide comparative studies while the ethnographic study of a particular culture is in fashion. However, the point that particularity and universality of culture are two sides of the same coin is often forgotten. Now is the time to re-evaluate the use of the comparative method in anthropology. When culture is divided into various items and their functional relations are to be sought by comparative method, as most twentieth-century comparativists did, the project inevitably faces the difficulty that there are too many variables (i.e. items of comparison). In order to avoid this difficulty, some American anthropologists contrived new comparative methods. When a comparativist takes a single variable as independent and the other variables as dependent on it, as White did in his neo-evolutionalist works, the project leads to mere reductionism. If such a reductionist approach is not to be adopted, comparison should be made by controlling the number of variables. The control of variables is realized in two ways: by focusing on a specific cultural institution, as George P. Murdock did in his cross-cultural survey on kinship; or by making a comparison within a limited geographical area, as Margaret Mead did in her studies of three Sepik societies in New Guinea, and Fred Eggan proposed in his presidential paper to the American Anthropological Association. Eggan called the latter method a "controlled comparison", and proposed that American cultural anthropology should borrow the methods and techniques of British social anthropology in order to conduct adequate comparative studies. Close examination of the recent theoretical disscussions on 1) how the unit of comparison should be identified, and 2) how an item of culture can be established cross-culturally, makes it clear that all ethnographers answer these two questions before processing field data. There has been no agreement among anthropologists on the definition of the unit of comparison (i.e. the ethnic unit). The comparability of the ethnic units and their historical independence have been questioned since Francis Gallon criticized Edward B. Tylor's comparative studies on evolution more than a hundred years ago. The final and fatal attack on the problem of identification of ethnic units came from post-modern anthropologists who regard culture as a mere construct of anthropologists. The comparability of cultural items has also been questioned by two influential anthropologists, Rodney Needham and David M. Schneider, both of whom made a fundamental criticism on kinship studies. They argued that there is no such thing as kinship. In order for anthropologists to continue to make comparative studies of "kinship", Needham proposed a more formal analysis, while Schneider suggested that the students of kinship should treat its constructs, which originated from European culture, as a working hypothesis, not as a universal fact. The bottomline is that the comparability of cultural items must be under continuous scrutinization. These valid criticisms on the unit and item of comparison oblige all ethnographers to stop at the entrance into comparative studies. A controlled comparison must be made on the spot before processing
(View PDF for the rest of the abstract.)
