A Salience Theory of Learning and Behavior: With Perspectives on Neurobiology and Cognition

Abstract

Traditional behaviorists have described behaviors fundamentally as responses to stimuli or, perhaps more liberally, as behaviors under the control of discriminative stimuli or contexts. They have held responses or behaviors to be established, strengthened, sustained, and inhibited or extinguished by contingent events: notably reinforcers, punishers, or the absence of either. In addition, they believed reinforcement acts on the response, the behavior, not on the organism. Here, and in support of Hebb’s view, we advance a contrarian view. A key principle of our framework is that species’ brains are uniquely designed to perceive and to relate stimulus events that are contiguous, salient, and relevant to adaptation. In accordance with what we here view as the constructive biases of species’ brains, stimuli are differentially organized into amalgams that reflect an exchange of salience and response-eliciting properties of component units, which are then integrated to form a basis of knowledge about the organism and its ecological niche. One can then base adaptation on overarching principles and rules, not just on simple associations. Species may create emergent behaviors with no history of specific training, and even new capacities, to service adaptation to both familiar and novel challenges.