This paper is about the problem of explanation in anthropology. There are, broadly speaking, three theories of explanation, namely, the scientific theory, the historical theory, and finally what I have decided to call the phenomenological theory, after M. Natanson. The author argues that none of the three theories is adequate by itself to encompass the complex nature of anthropological science. The three theories correspond roughly to at least three different types of questions raised by anthropologists, and this being the case (...) the principle of methodological tolerance seems a natural and sensible principle to adopt. The paper also deals with the problem of reduction, i.e. the problem whether the three theories are different from and logically independent of one another. (shrink)