“Self‐domestication” has been invoked to understand important aspects of human evolution, integrating physiological, behavioral, and morphological information in a novel way. It proposes that selection for reduced aggression on animals undergoing domestication provides a model for selection favoring prosocial behaviors in humans and for a set of seemingly independent features, which arose as a result of developmental correlation. We review the history of the idea and examine patterns of domestication. A lack of empirical studies on evolutionary rates and variation thwarts meaningful comparison with domestication. The neural crest hypothesis for domestication has great explanatory power but it is difficult to test. We suggest a scenario in which the morphological byproducts of domestication can act as an honest signal of reduced xenophobia. Future studies should test if alternative explanations for the features deemed to result from self‐domestication are mutually exclusive and generate data to test predictions of these hypotheses.