These comments are intended not as criticism of Steven Lukes' conclusions, with which I mainly agree, but to raise some questions about their further implications and suggest some answers to these questions (1).
The questions arise when we stand back and look at the entire set of positions on the matter of rationality represented by Lukes' five paradigms and his own argument. They fall easily into two groups, composed of the first four and the last two, counting Lukes' position as the sixth. The first group consists of the views of practising anthropologists primarily concerned with understanding and explaining certain quite specific sorts of peculiar beliefs. In approaching their task they make various assumptions about rationality, about the relation of modern scientific thought to primitive magico-religious thought, and about the type of attitude towards these alien forms of mentality which is appropriate to the scientific endeavour. The second group contains the views of philosophers reflecting on the methods and theories of anthropologists, emphasising the philosophical character of the problems they take up out of this reading, and which they treat with the techniques of linguistic or logical analysis. They are less interested in the specific peculiarities of certain primitive beliefs than with the formal criteria by which any different sets of beliefs can be logically classified and compared. And they are less concerned with understanding and explanation than with the assumptions made about the philosophical status of, and typical relations between, different sets and sorts of beliefs when understanding and explanation are seen as problematic. Thus the second group concentrates more exclusively on a higher-order inquiry than that with which the first group was centrally concerned (despite the fact that Lukes has successfully isolated this level of inquiry out of the less focussed methodological reflections of the anthropologists). The question can therefore be raised, how far the positions on rationality are detachable from the various explanatory methods and theories which are associated with them in the first group, and what implications for understanding and explanation, if any, follow from the positions on rationality adopted by the second group.