A philosophy of science perspective on the quantitative analysis of behavior

• B.F. Skinner distinguished between two types of theory in the study of learning: first, a type he suggested could (and should) be avoided; and second, a type he endorsed but felt the field was not ready for as of 1950.
• The quantitative analysis of behavior has given rise to matching theory, which this article suggests is a theory of the second type.
• The attempt to provide a dynamic account of the static regularities of matching theory has produced at least one theory of the first type.
• Within its limited domain, this theory provides a rigorous alternative to cognitive accounts, while now offering predictions that are distinct from those of matching theory.

B.F. Skinner argued that the science of behavior would progress more rapidly without appealing to theories of learning. He also suggested that theories in a quite different sense were possible, but that the science of behavior as of 1950 was not ready for them. The following analysis distinguishes between Skinner’s two concepts of theory. It argues that theory in the second sense has arisen in the quantitative analysis of behavior. The attempt to give a dynamic account of the static regularities of this theory, however, has produced a theory in the first sense. Within its limited domain, this theory offers a rigorous alternative to cognitive accounts of behavior. Rather than distracting attention from actual behavior, it has now led to novel predictions about it.This article is part of a Special Issue entitled ‘SQAB 2014’.