A series of investigations that examine the development of representation skills among mentally retarded persons is proposed. Specifically, the focus of inquiry is on learning processes responsible for the development of object-picture representation, spatial relations representation, and motor response representation and for recombinative generalization (the application of rules to promote generative learning). Single-subject research designs will be employed to determine how the environment may be structured to provide representation training that contributes to efficient and generalized learning by mentally retarded individuals. First, receptive picture reading skills of mentally retarded adolescents and adults will be examined to provide a foundation on which to base a series of experimental examinations of representation. Second, a series of experimental studies will be conducted to delineate variables that account for picture representation skills. Tests for transfer of acquired knowledge to untrained combinations of object-picture representations, spatial relations and motor responses will be included. Third, the conditions sufficient to produce a generative representation repertoire with mentally retarded individuals will be examined. The proposed research has implications for developing a functional analysis model of symbolic development and most directly for developing more efficient vocational training procedures that can be used with mentally retarded adults.