The following intellectual as opposed to practical reasons for all anthropologists doing fieldwork are examined: fieldwork: (1) records dying societies, (2) corrects ethnocentric bias, (3) helps put customs in their true context, (4) helps get the "feel" of a place, (5) helps to get to understand a society from the inside, (6) enables appreciation of what translating one culture into terms of another involves, (7) makes one a changed man, (8) provides the observational, factual basis for generalizations. None of these (...) is found sufficient to make fieldwork imperative for all anthropologists, although they are quite sufficient to allow that it is imperative for anthropology as a whole that fieldwork in some form by some people continue. In place of the view of fieldwork as an essential preparation for doing anthropology, an alternative role for it is explored: namely as a testing procedure. The implications of this--that the study of problems and the articulation of theories can usefully proceed prior to or even independently of fieldwork--are drawn out, and a new institution of selective fieldwork is proposed. (shrink)