Pavlovian extinction, phobias, and the limits of the cognitive paradigm.

The slow or total lack of decrease in some autonomic responses during extinction in aversive conditioning and concomitant verbalization of fear have remained a problem for learning theories and psychophysiology. Removal of the aversive stimulus should result in a rapid decrement in responding, as it does in cognitive and somatic systems. In laboratory analogues of phobia and clinical neurosis, however, such decreases do not occur in some autonomic responses and reported fear. In this article three areas of research are presented in which dissociations occur between cognitive and autonomic responses: 1) relational learning, 2) phobia, and 3) incubation. The data indicate that there are some important distinctions to be made concerning the properties of different psychological and physiological systems. These distinctions pertain to the differences between cognitive and noncognitive systems, between the two branches of the ANS, and between acquisition and extinction processes. These distinctions lead to a number of hypotheses concerning dissociations between response systems and have important implications for the understanding and treatment of neurosis.