Two individuals are involved in a conflict situation in which preferences are ex ante uncertain. While they eventually learn their own preferences, they have to pay a small cost if they want to learn their opponent's preferences. We show that, for sufficiently small positive costs of information...