Under the pressure of the three problems mentioned above — memory, motivation, and cognition-most creators of learning theories supplemented Skinner's experimental analysis of environmental and behavioral variables with intermediate variables. Intermediate variables are theoretical constructs whose value is determined through their relationships with various environmental variables, whose General effects they are intended to sum up.
The expectations theory of Tolman
Thorndike, influenced by Darwin's premise of the continuity of the evolution of species, began a transition to a less mentalistic psychology. John B. Watson ended it with a complete rejection of mentalist concepts. Acting in line with the new thinking, Tolman replaced the old speculative mentalist concepts with logically defined intermediate variables.
As for the subject of our discussion (reinforcement), Tolman did not follow Thorndyke's example. Thorndike considered the effects of response to be of extreme importance for enhancing the Association between stimulus and response. He called it the law of effect, which was the forerunner of modern reinforcement theory. Tolman believed that the consequences of a reaction do not affect learning as such, but only the external expression of the underlying processes of learning. The need to distinguish between learning and execution arose in the course of attempts to interpret the results of experiments on latent learning. As the theory developed, the name of the intermediate variable introduced by Tolman, reflecting learning, was repeatedly changed, but the most appropriate name could probably be expectation. Waiting depended solely on the temporal sequence — or contiguity — of events in the environment, not on the consequences of the response.
Pavlov's physiological theory
For Pavlov, as for Tolman, the contiguity of events was a necessary and sufficient condition for learning. These events are physiologically represented by processes occurring in those areas of the cerebral cortex that are activated by indifferent and unconditional stimuli. The evolutionary consequences of the learned reaction were recognized by Pavlov, but not tested in experimental conditions, so their role in learning remained unclear.
Gasri's molecular theory
Like Tolman and Pavlov, and in contrast to Thorndike, Edwin R. Gazi believed contiguity sufficient for learning. However, time-coinciding events were not determined by such broad (i.e., molar) events in the medium as Tolman claimed. Each molar environmental event, according to Gazi, is composed of many molecular elements of the stimulus, which he called the signals. Each molar behavior, which Gazri called "action", in turn consists of many molecular reactions, or "movements". If a signal is combined in time with a movement, that movement becomes completely conditioned by that signal. Learning behavioral action is slow only because most actions require learning many of their constituent movements in the presence of many specific signals.
Hull's drive reduction theory
The use of intermediate variables in the theory of learning has reached its widest development in the works of Clark L. hull. Hull attempted to develop a common interpretation of behavioral changes resulting from both classical and operant procedures. Both stimulus-response coupling and drive reduction are essential components of the hull concept of reinforcement.
The fulfillment of learning conditions affects the formation of an intermediate variable — habits. Habit was defined by hull as a theoretical construct that summarizes the overall effect of a number of situational variables on a number of behavioral variables. The relationships between situational variables and the intermediate variable (habit), and then between habit and behavior, were expressed in the form of algebraic equations. Despite the use of physiological terms in the formulation of some of their intermediate variables, experimental research and hull's theory were exclusively related to the behavioral level of analysis. Kenneth W. Spence, an employee of hull who made a significant contribution to the development of his theory, was particularly careful in defining intermediate variables in purely logical terms.