The present study predicted that males arrested for spousal assault participate in one of four paradigms of spousal assault and that males participating in one type of violent episode demonstrate particular person characteristics which distinguish them from males participating in other violent episodes. To examine the hypotheses, thirty-eight violent males, and fifteen nonviolent males in couples therapy, participated in a structured interview discussing episodes of serious spousal conflict and views of their intimate relationships. Subjects were also shown a video depicting a married couple's escalating argument and asked about their reactions to the video using a structured interview and a sixteen-item Likert-type scale tapping subjects' perceptions of the female portrayed in the video. Each case of spousal assault was assigned independently to one of the four paradigm cases by three trained raters using a coding system. The coding system designated what type of interview responses would most likely be associated with males participating in one of the four paradigm cases. Subsequently, subjects' interview responses were also rated and assigned to one of the four paradigm cases using the coding system. Three versions of the Conflict Tactics Scale (Straus, 1979) were also administered to examine (a) the subject's experience with parental physical force directed at him in his family-of-origin, (b) the parental spousal assault that the subject observed in his family-of-origin, and (c) the subject's and his spouse's use of physical force in resolving conflict. The data reveal that thirty-seven of the thirty-eight violent males interviewed have participated in one of the four paradigm cases of spousal assault. Likewise, findings indicate that males participating in a particular violent episode demonstrate particular beliefs and perceptions that distinguish themselves from males participating in the other types of violent episodes. The results suggest that spousal assault is not a unitary phenomenon but occurs in different spousal contexts and for different reasons and by different types of males. The methodological and clinical implications of these findings are discussed.