T HE sociologist has only to attempt to answer the more penetrating questions of his brighter sophomores to realize that his professional fraternity lacks consensus on the conceptualization of role. If he has assigned two or more readings on the concept, his students may well have found two or more definitions of it. And if he should make so bold as to teach a course which emphasizes research, he is really courting trouble when the floor is thrown open for a discussion of the utility of such a concept in a science of interaction. Inner-directed and other-directed have become almost popular jargon, but they seem to be more closely related to personality and value taxonomy than to interaction analysis.' Parsons says that he conceives the specific-diffuse dichotomy as a system goal concept rather than a classification of status-pair We have, then, the division of statuses and roles into ascribed and achieved, a usage considerably more honored in time and in textbook than in research application. While these are meaningful and useful terms for the description of some institutional behaviors and for the comparison of diverse cultures, they offer us little in the analysis of many social systems: industry, politics, and voluntary associations, for example. Yet, to many of us, is one of the basic analytic constructs, if not the core concept, in sociology and social psychology. Some taxonomy of roles, however simple, would seem desirable. We have been driven by our research to construct a classification based upon the degree of flexibility in the role expectations. The terms were derived for use in a project dealing with occupational roles and will be presented here as they apply to this research. While this dichotomy is subject to the usual difficulties which attend ideal type constructs, it is to be hoped that, as empirical work refines it, the classification may be found more generally useful in the analysis of interaction in other spheres. Sociologists need a theory of occupations. We have a theory of industrial organization by Weber out of Roethlisberger and Dickson, and most recently refined by Gouldner.3 We have a start toward a theory of industrial administration, thanks largely to Barnard, Whyte, and the applied anthropologists.4 Our knowledge of occupations, however, is almost entirely limited to descriptive materials presented without any attempt at theoretic systematization. The bulk of our data on occupations consists of facts which we possess because they were gathered peripherally in research with some theoretical orientation other than that of a sociology of occupations. Studies of the general social structure of industry, for example, through their descriptions of formal and informal organization, have led to an analysis of the stratification of occupations in work plants.5 Community stratification studies