In an attempt to combat the positivist view that the only legitimate way to conduct social science is in the manner of a natural science, Knight distinguished between positivist and non-positivist categories or levels of interpretation of human-social subject matter. Since each of the categories contained ‘a large element of truth’, Knight argued that any serious analysis would need to embrace a pluralist approach. In this paper I draw on four separate accounts he gave (in 1934, 1940, 1941, and 1942) (...) of these categories in order to arrive at a more complete statement. This fuller description will hopefully promote a deeper appreciation of Knight's emphasis on the need for adopting a pluralist approach to the subject matter of economics. I then apply the perspective of the categories to shed light on Knight's definition of economics and his criticism of Robbins' definition. (shrink)