An evolutionary model of reinforcer value

Within the field of evolutionary biology, natural selection is often thought to favor traits that lead to individuals behaving as if they were maximizing their evolutionary fitness. The concept of the individual as a maximizer is also popular in behavioral psychology, especially when it comes to theories of operant learning. Here, the individual is taken to adapt its behavior to the local environment, such that the expected amount of reinforcer value is maximized.
Whereas there is a considerable consensus concerning the formal properties of an evolutionary maximand (‘fitness’), there is no generally accepted conceptualization of a corresponding behavioral maximand (‘reinforcer value’). However, such theoretical clarification is crucial to the development and empirical testing of learning theories, since it is impossible to decide whether the concept of reinforcer maximization is adequate, as long as the maximand is not well defined.
This paper presents a formal model of reinforcer value that is consistent with existing work on the nature of reinforcement and provides an explicit link between behavioral psychology and evolutionary biology. The main result is that the reinforcer value of an additional time unit spent at a behavior equals its expected marginal effects on evolutionary fitness. Applying the model to matching behavior, it is further demonstrated how the established link between reinforcer value and evolutionary fitness can be used to derive new hypotheses.