The Neo-Behaviorists: Guthrie, Tolman, and Hull :

Edwin R. Guthrie (1886–1959), the “most starkly empiricist of all the neo-behaviorists” (J. A. Mills, 1998), defined mind as “a mode of behavior, namely, that behavior which changes with use or practice-behavior, in other words, that exhibits learning” (E. R. Guthrie, 1935/1960). The ability to learn, as C. Lloyd Morgan had suggested, characterized the possession of mind in living creatures. Guthrie’s theory of learning was deceptively simple: Learning occurs through the development of associations between stimuli and responses. These associations are formed by contiguity: “A combination of stimuli which has accompanied a movement will on its recurrence tend to be followed by that movement”. He rejected Thorndike’s laws of effect and of exercise, claiming instead that the apparently gradual nature of learning was a result of a series of one-trial situations in which movements, small muscle responses, rather than acts were learned in response to stimuli. The role of the consequences of responding, whether satisfying or annoying, was to change the stimulus situation, not to strengthen some unobservable bond between
stimulus and response. In contrast to E. R. Guthrie’s molecular approaches to learning, Edward Chace Tolman (1886–1959) offered a molar theory of the psychology of learning. For E. R. Guthrie and for J. B. Watson, descriptions of learned behavior were confined to descriptions of stimulus events and responses. Tolman, in contrast, proposed a theory that interpreted behavior in terms of “motive, purpose and determining tendency” (Tolman, 1922). For Tolman, cognitive events intervened between the antecedent stimuli and their behavioral consequences. Learning and performance were not synonymous (Innis, 1999; Kimble, 1985; Tolman & Honzik, 1930); performance was the observable behavior, while learning was the hypothesized state that accounted for the change in behavior. Tolman described the action of intervening variables on the functional relationship between the independent and dependent variables; that is, between the environmental stimuli and physiological state of the organism on the one side and the overt behavior on the other (Tolman, 1932; see also Innis, 1999; Kimble, 1985). The most important intervening variables were cognitions, defined as expectations about the relationship between signs, stimuli, and significates, rewards or goal objects (J. A. Mills, 1998; Smith, 1986). Tolman hypothesized the formation of “cognitive maps” or cognitive representations of the environment in rats learning a maze. These cognitive maps could be empirically demonstrated in maze experiments in which, for example, blocking a previously used route to a goal resulted in rats choosing the next shortest path to the goal (Tolman, Ritchie, & Kalish, 1946).
Clark Hull (1884–1952) proposed a formal logicodeductive theory of behavior: “In science an observed event is said to be explained when the proposition expressing it has been logically derived from a set of definitions and postulates coupled with certain observed conditions antecedent to the event” (Hull, 1943). Hull’s theoretical treatment of psychology consisted of a set of postulates and corollaries and their mathematical statements to enable quantitative predictions about behavior. Hull’s goal was to develop psychology as a natural science by demonstrating that behavioral phenomena obey universal, quantitative laws that can be stated by equations comparable to physical laws, “of the type governing the law of falling bodies” (Hull, 1950). Even centuries after Kant, Hull was striving to demonstrate that psychology could indeed become a science that met the same standards as the physical sciences. For example, Hull (1934a, 1934b) proposed that the serial position effect in learning a list of words (the phenomena that errors occur more frequently in learning and in the recall of words from the middle of a serial list) exemplifies the same general law that describes the pattern of errors made by rats learning a complex maze (more errors occur in the center of the maze than at the start and the finish). Hull’s research program was directed toward the discovery of such laws and the formulation of the equations that described them. His theory of behavior formulated theoretical
variables in operational terms, defined them by equations, and predicted experimental results. Experiments by Hull, Tolman, and their students were designed to provide crucial tests of predictions from their respective theories. For example, Hull’s theory hypothesized that learning occurred through reinforcement, defined in terms of the extent to which reinforcement reduced a motivational drive; Tolman, on the other hand, argued that reinforcement in this sense was unnecessary for learning (Tolman & Honzik, 1930). Resolution of such theoretical issues was difficult; moreover, the
precise predictions from Hull’s formal theory were frequently not confirmed, and criticism of the theory began to mount from a variety of sources, including Hull’s own students (J. A. Mills, 1998). Differences between the theories of Hull and Tolman came to seem less substantive and more a preference for particular terminology and the reification of intervening variables (Kendler, 1952).