ABSTRACT
A simple extension of the ideas of social learning theory would suggest that the general preference level for a particular version of a role is a direct function of the anticipated total profit associated with its performance. In the real world, terms that have become part of the basic vocabulary of the social scientist should have generally accepted meanings. The ultimate dependent variable in social-psychological theory is social behavior, and if roles refer to actual behavior there would be little for roles to explain. Sociological social psychologists have discussed an opportunity for rehearsal that has not been emphasized by Bandura. Though many social learning experiments involve children engaged in play like behavior, the theory does not give much specific attention to play and games as contexts in which rehearsal and innovation take place. Much of the sociological literature on role learning starts out by describing one or more of the contexts in which socialization occurs.
