The long-standing critique of the 'economic model of man' has gained new impetus not least due to the broadening research in behavioral and experimental economics. Many of the critics have focused on the apparent difficulty of traditional rational choice theory to account for the role of moral or ethical concerns in human conduct, and a number of authors have suggested modifications in the standard model in response to such critique. This paper takes issue with a quite commonly adopted 'revisionist' strategy, namely seeking to account for moral concerns by including them as additional preferences in an agent's utility function. It is argued that this strategy ignores the critical difference between preferences over outcomes and preferences over actions, and that it fails to recognize that 'moral preferences' belong into the second category. Preferences over actions, however, cannot be consistently accounted for within a theoretical framework that focuses on the rationality of single actions. They require a shift of perspective, from a theory of rational choice to a theory of rule-following behavior.