According to social learning theory, most human behavior is learned through modeling. Under this theory, ideas of how new behaviors are performed are acquired from observing others, and these coded observations become guides for further action. This review of social learning theory as an explanation of spousal aggression focuses on violence in the family of origin, stress, aggression as a personality style, alcohol use, and relationship (marital) dissatisfaction. The author concludes that although the social learning approach to spousal aggression has proven useful in attempting to explain events with grounding in empirical observations, it is not an especially parsimonious approach to spousal aggression. The author believes that aggression, spousal aggression in particular, is a multidetermined phenomenon that requires complex models for prediction and understanding. Such models will require many data sources and many empirical relations. Some integration of a model of spousal aggression and child abuse could occur, because the history of family violence, aggressive personality style, stress, alcohol use, and marital discord relate both to spouse abuse and child abuse. 4 figures, 71 references.