What kind of preference maximization does the weak axiom of revealed preference characterize?

A (multivalued) choice is justified if no two chosen alternatives are preferred to each other, and if all chosen alternatives are preferred to all rejected alternatives. This concept permits a connection between the behavioral property expressed by WARP and a weak form of preference maximization. I am grateful to Michele Lombardi, Michael Mandler, Paola Manzini and an anonymous referee for useful discussions and comments.