Behavior analysis initially developed through research with rats and pigeons in the laboratory. Within the curricula of psychology, it is typically identified as "learning theory" despite the fact that it addresses most domains of psychology (both human and non-human) and is concerned at least as much with the maintenance of behavior as with its acquisition. Thus, behavior analysis is often juxtaposed with another, quite different tradition in the study of animal learning, in which inferred associative bonds are taken as the fundamental units underlying behavior. To lump behavior analysis together with animal learning and to view the aggregate as a unitary field is to ignore the conceptual divide between the analysis of nonhuman behavior and associationist learning theory. To be sure, the two approaches share some research techniques, their data sets occasionally resemble each other and appear in the same journals, and they often compete for the same research funds, faculty positions and other resources; nevertheless, their experiments have nearly always addressed fundamentally different questions.

The essays in this part attempt to deal explicitly, if not always constructively, with the resulting tension between the two positions, in a sequence spanning several decades of that strained relationship. Skinner’s review of Hull’s Principles of Behavior,addressing a seminal work in the alternative tradition at the time of its original publication, recognizes the need for systematizing and organizing data but finds fault with Hull’s hybrid of hypothetico-deductive and functional analyses. Although giving a respectful nod to Hull’s experimental studies, Skinner criticizes Hull's speculative appeals to neurological events and the inconsistencies arising from mixed levels of explanation. In arguing that Hull had not “made full use of an outright functional analysis,” Skinner in effect faults Hull for having failed to become a Skinnerian.

Wearden’s retrospective review of the same book attempts a more constructive approach. It treats not only Skinner’s critique, regarding it as milder than suggested by our characterization above, but also some conciliatory observations offered by Hull's student, Kenneth Spence (1966). Spence found the approaches of Hull, Tolman and Skinner similar by virtue of their focus on behavior rather than on mind, consciousness or physiology. Wearden finds special value in Hull's contributions, arguing that they identify issues and research that deserve more emphasis in contemporary behavior analysis.

Another student from the neo-Hullian tradition is Neal Miller, whose work is reviewed by Rescorla, himself a major proponent of research in the associative tradition. Acknowledging that Miller consistently adhered to hypothetico-deductive methods focused upon theoretical constructs, Rescorla recommends Miller's work to behavior analysts for its emphasis on empirical findings. Whereas behavior analysts present behavior-environment relations as primary phenomena even for an understanding of physiology, Miller often tried to elucidate behavioral findings using physiological manipulations: “To Miller, the tie of physiology with behavior is an integral part of the discovery of laws of behavior.” Thus we see a ramification of the difference between mediational and nonmediational theories.

The remaining two reviews bring us to specific issues that have been vehemently argued between behavior analysts and proponents of associationist theory. Weisman’s review of Mackintosh’s Psychology of Animal Learningis a valiant attempt to sift out what is useful to one's viewpoint when the material in question has been presented in an implacably antagonistic way. By distinguishing between "empirical generalizations from the literature, constructive theorizing, and muddled acrimonious critique," Weisman invites behavior analytic readers to consider work they might be tempted to disregard because of its unpalatable packaging. His review also conveys some flavor of the separation between the two traditions: Mackintosh dismisses dynamic interactions between behavior and environmental events as hopelessly complex, even though his own experiments use procedures that arrange for such dynamic interactions. In contrast, the associationist's elaborate arrangements for inferring differential connections between representations of stimuli (or of stimuli and responses), often seem arcane and misguided to the behavior analyst.

Finally, Mackintosh’s book of a decade later, Conditioning and Associative Learning,provides the occasion for Williams to convey the gist of advances within both associationist and behavior analytic studies of nonhuman learning. Williams examines the divergent histories of the two viewpoints, and then details conceptual and empirical bases for constructive contact between them. Differences between within-subject research designs with extensive observations of single subjects and statistical research designs based upon brief observations of many subjects within a group have imposed barriers of both style and substance. In Williams' view, both approaches would be strengthened if they drew more freely from each others' work, and he sees the failures to do so as having occurred on both sides. He shows that, rather than reflecting arbitrary theoretical taste, some contention arises from the problem of reconciling differences between episodic analyses that proceed trial-by-trial and more molar analyses that accept rates as fundamental.