Reversal Without Remapping: What We Can (and Cannot) Conclude About Learned Associations From Training-Induced Behavior Changes.

The "cognitive revolution" in psychology is often framed as a departure from associationist principles rooted in animal learning research, yet it is clear that these principles have immediate relevance for contemporary questions in cognitive and social psychology. Intuitions about the consequences of learning procedures can easily be misleading, making these principles particularly important. To illustrate this point, we identified recent examples of studies applying a particular learning paradigm-response-reversal training-to the study of three different psychological problems (e.g., why objects in the right side of space are preferred to those in the left in right-handed people). The strategy of each study was to alter a typically encountered contingency once in the laboratory, in order to reverse a hypothesized learned response. Yet, contrary to intuitions, we demonstrate that behavior changes can be observed without the reversal of a prior association. Further, many different associative changes can underlie response reversals. We focus on these examples of response-reversal training, but our broader aim is to help connect the animal learning literature to problems in cognitive and social psychology in an effort to strengthen the inferences that might be drawn about learned associations in these contexts.