It has been widely and regularly suggested that aggression is not a unitary concept but rather an omnibus term concealing important and distinct subclasses of behavior. Psychologists have offered various categorical schemes hwlighting motivation and fonn as distinguishing criteria. Many writers have also noted the social relativity of d e f ~t i o n s of c k e s of aggression and called for a consideration of lay conceptions of aggressive behavior. The present paper explores the correspondence between psychologists' and laypersons' categorizations of aggressive acts. Ten short scenarios of aggression were generated to embody systematic permutations of form and motive as suggested by the published work of social psychologists. Subjects rated every pairwise combination in terms of perceived similarity. The data were subjected to three-factor, multidimensional scaling. The resulting structure indicated that subjects' ratings were not random and reflected consideration of motive (hostile, normative, instrumental, status) and of direct versus indirect form. The data were thus in close correspondence with the classifications generated by psychologists. The desirability of refming this technique and extending it to other cultural and subcultural groups is discussed.