Emotions play a crucial role in moral behavior. The present paper does not contest this point but argues that qualifications of certain feelings such as shame and guilt as moral emotions should not exclusively be based on a proximal analysis of their function. A proximal analysis details how moral emotions produce moral behavior. Emotions are qualified as moral when they are elicited by concerns for others rather than the self and produce prosocial action tendencies. Although researchers have acknowledged that moral emotions may also have an ultimate function that details why it is in the individual interest that these moral effects occur, they have neglected to translate such ideas into testable hypotheses. Using guilt and shame as an example, we show how an analysis of ultimate functions accommodates recent findings, which contest the view that guilt is more moral than shame and provides new insights as to when and why moral emotions will produce moral effects.