Learning Effected by the Consequences of Behavior

Abstract

Without any known exception, animals that have evolved a centralized nervous system are able to learn from the consequences produced by their own actions, success acting as a “reward” or “reinforcement,” failure acting as a “punishment” tending to “extinguish” the animal’s readiness to repeat the action just performed. This learning process cannot be explained even hypothetically on the basis of an association between two pre-existent neural systems. An “open program” with a much more complicated structure must be postulated. As discussed in Two/I/1, the system which Heinroth called the arteigene Triebhandlung, species-characteristic drive action, consists of several physiologically distinguishable components; a) appetitive behavior, b) achieving a stimulus situation to which an IRM responds and, finally, c) a consummatory act, in which a phylogenetically programmed action is performed, by which a teleonomic function is achieved and the “drive” or motivation of the action is stilled. This three-link chain of processes is the base on which the ability to learn by success or failure has evolved phylogenetically; also, it still remains an indispensable part of the system achieving this kind of learning. Wallace Craig had fully grasped this when, in his classic paper, “Appetites and Aversions as Constituents of an Instinct” (1914), he described the way in which the performance of the consummatory act induces the animal to seek a very special and complicated stimulus situation in which the act affords a maximum of satisfaction.