Normative notions in descriptive dialogues

Developments in the theory of individual decision-making have been partly shaped by two criteria: a desire for models consistent with experimental evidence; and a pre-commitment to models built on normatively appealing axioms. This paper explores the compatibility of these two selection criteria. The paper reconstructs and scrutinises an argument due to Friedman and Savage asserting that the normative appeal of axioms provides a source of 'indirect' evidence. I judge their argument questionable and, at best, incomplete. As such it does not provide a convincing rationale for normative appeal to be used as a criterion for selection among descriptive theories.