Abstract
We propose a descriptive theory of lexicographic choice, lexicographic majority, where a decision maker’s ranking of attributes is not necessarily strict. Thus we allow an agent to assign the same rank to a set of attributes. If a decision maker assigns the same rank to a set of attributes, it is assumed that the simple majority heuristic is used, within the set, to discriminate between pairs of vectors. Our model is general enough to account for intransitive preferences. Moreover, it includes as special cases: pure lexicographic preferences, simple majority preferences and lexicographic semiorders. We justify lexicographic majority preferences by providing an axiomatization in terms of behavioral properties. As a further justification we characterize lexicographic majority preferences as a subset of the class of weighted majority preferences.