The history of 2-process learning theory is described, and the logical and empirical validity of its major postulates is examined. The assumption of 2 acquisition processes requires the demonstration of an empirical interaction between 2 types of reinforcement contingencies and (a) response classes, (b) reinforcing stimulus classes, or (c) characteristics of the learned behavior itself. The mediation postulates of 2-process theory which argue that CRs are intimately involved in the control of instrumental responding are emphasized, and 2 major lines of evidence that stem uniquely from these postulates are examined : (a) the concurrent development and maintenance of instrumental responses and conditioned reflexes, and (b) the interaction between separately conducted Pavlovian conditioning contingencies and instrumental training contingencies in the control of instrumental behavior. The evidence from concurrent measurement studies provides, at the very best, only weak support for the mediational hypotheses of 2-process theory. In contrast, the evidence from interaction studies shows the strong mediating control of instrumental responses by Pavlovian conditioning procedures, and demonstrates the surprising power of Pavlovian concepts in predicting the outcomes of many kinds of interaction experiments.