Norms and descriptions

This article addresses the problem of identifying conditions according to which it is possible to distinguish between a descriptive theory and a normative theory. What makes a descriptive theory descriptive and a normative theory normative? The focus is on subjective expected utility theories where it seems open to debate whether the appropriate use is normative or descriptive. My discussion, which takes arguments by Isaac Levi and Hugh Mellor as points of departure, is mostly negative, showing that there are no obvious distinguishing features of theories qua theories. Rather, the theories can be used normatively or descriptively without making them theories of one kind or the other. I also point to situations in prescriptive decision analysis where one should be observant of in what way theories are used.