Neural networks: II. Unified learning theory and behavioral psychotherapy

Abstract

Part I, Tryon (1993), demonstrated that neural networks provide a unified theoretical perspective from which to integrate at least 11 important theoretical/philosophical schisms which have unproductively divided psychology in general and clinical psychology in particular (Staats, 1983, 1991). Part II extends unification to neural network learning theory (NNLT) and discusses implications for behavioral psychotherapy. The cognitive revolution within clinical psychology addressed the need to provide a proximal causal explanatory mechanism mediating stimuli and responses of Skinner's radical behaviorism. Neural networks provide a learning mechanism capable of addressing the full spectrum of psychological and behavioral phenomena. Neural networks are trained rather than programmed, which gives them a developmental history. To fully integrate neural network learning theory with operant conditioning, their development is approached from an evolutionary perspective. The success of the present synthesis is demonstrated by showing that NNLT is a cognitive model which is fully consistent with behavioral values. Integrative success is more strictly demonstrated by providing a proximal causal mechanism capable of explaining the functional properties of variation and natural selection, on which operant conditioning is based. Implications for behavioral psychotherapy for selected DSM-III-R disorders are discussed.