Instrumental Learning in Animals: Parameters of Reinforcement

Abstract

The first 10 volumes in this series have general chapters on learning, the most recent of which, by Kendler, was published in 1959. Since then, there has been no comprehensive review of the literature on animal learning, al­ though parts of it have been touched upon under a variety of more special­ ized headings. Our purpose in this chapter is to summarize the work of the intervening years on the role of reinforcement in the instrumental learning of animals. Research on the process of classical conditioning is treated only insofar as it bears on the interpretation of the instrumental data, while re­ search on larger problems of generalization, discrimination, and attention is reserved for subsequent consideration. Even so, the scope of the literature reviewed makes exhaustive citation impractical, and we must rely on repre­ sentative studies to illustrate the principal findings. An experiment on instrumental learning is one in which the occurrence of some event ( reinforcement) is contingent upon the occurrence of some specified response (the instrumental response), and the effect of the contin­ gency upon the occurrence of the response is measured. It is useful to dis­ tinguish two main categories of instrumental training: in the Thorndikian case, the instrumental response increases the probability of reinforcement; in the avoidance case, the instrumental response decreases the probability of