Abstract
The problem of moral luck relates to the question whether contingencies ought to influence our moral judgments. This article argues that while our moral assessments of agents ought to be independent from luck, there are strong non-moral reasons for letting contingencies influence the extent to which we expose agents to negative reactions. First, I offer an account of the problem of moral luck. Subsequently I argue that legitimate moral blame is immune from luck. Morally blaming an agent for unintended harm conflicts with the principle of the universalisability of moral judgments, and it rests on a confusion between evaluative and normative moral judgments. Nevertheless it can be rational to expose an agent who causes unintended harm to negative reactions. So there seems to be a tension between moral reasons and rational reasons. The assumption that moral luck exists can be seen as resulting from the attempt to dissolve this tension.