ABSTRACT
As Wagner (this volume) pointed out, current theories of human memory distinguish between automatic and controlled processing. An automatic process is thought to involve an associatively connected set of routines in long-term memory, the execution of which proceeds rather inflexibly once triggered and is largely beyond voluntary control (Shiffrin & Schneider, 1977). In contrast, according to Atkinson and Shriffrin (1968) controlled processes: “are not permanent features of memory, but are instead transient phenomena under the control of the subject; their appearance depends on such factors as instructional set, the experimental task, and the past history of the subject [p. 106].” The distinction between controlled and automatic processes actually has a long history in writings about animal behavior. It is therefore quite familiar even though most students of animal learning would employ different terms. Hobhouse (1915/1973), for instance, was careful to distinguish between habitual modes of processing and those that appeared to involve more “intelligence.” Later researchers were divided with respect to how much importance to attach to the two kinds of processing in animals; it is well known that workers in the Thomdikian tradition relied on passive processes in their accounts of conditioning (Spence, 1937), whereas Tolman (1948) was inclined toward more active processes in his views of animal behavior. For himself, Thorndike chose to emphasize the direct and automatic effects of rewards on association formation and to discount the importance of “rehearsal or reconsideration” of events in conditioning situations (Thorndike, 1931).
