|Rethinking Reinforcement: Moving Beyond Response Strength
|Saturday, May 28, 2022
|4:00 PM–5:50 PM
|Meeting Level 1; Room 156C
|Area: PCH; Domain: Theory
|Chair: Carsta Simon (University of Agder, Norway)
|Discussant: Timothy A. Shahan (Utah State University)
Behavior interacts with its environment both during an organism’s lifetime, through operant selection and, across generations, through natural selection. The radical behaviorist’s attempt to explain operant selection usually both depends upon and denies the implicit, hypothetical process of reinforcement by response strengthening. In this symposium, we discuss what is problematic about strengthening by reinforcement and suggest how the interaction between behavior and the environment is better described without this unfalsifiable concept. In our talks, we put forward an assembly of alternatives, all replacing response strengthening by approaches questioning the traditional distinction between consequences and antecedents. These alternative explanatory approaches, namely, induction, signaling, and selection, not only explain food-maintained behavior, but also shock-maintained behavior. They explicitly connect behavior analysis and evolutionary theory and do not rely on hypothetical constructs. We will show data supporting our conceptual arguments and indicating that our alternatives to strengthening by reinforcement might not only improve conceptual cohesiveness but might also benefit both our basic science and our technology of behavior.
|Instruction Level: Intermediate
|Keyword(s): avoidance, induction, response strength, selection
|Avoidance, Induction, and the Illusion of Reinforcement
|WILLIAM M. BAUM (University of California, Davis)
|Abstract: Environmental events that impact reproductive success may be called phylogenetically important events (PIEs). Some promote reproductive success, like mates and food; others threaten reproductive success, like predators and injury. Beneficial PIEs induce activities that enhance them, and detrimental PIEs induce activities that mitigate or avoid them. Free-operant avoidance relies on electric shock as a proxy for injury, a PIE. A popular theory takes avoidance behavior to be reinforced by its reducing shock rate. An alternative explanation is that avoidance is induced by the PIEs it usually prevents. Four parametric data sets were analyzed to show that avoidance is maintained, not by shock-rate reduction, but by received shock rate. Avoidance is not reinforced at all; avoidance is induced by its failures. Induction explains not only avoidance itself, but also phenomena unique to avoidance, like warmup and unavoidable shock. The concept of induction explains behavior more generally than reinforcement, because induction explains not only food-maintained behavior, but also shock-maintained behavior. Received shocks induce operant avoidance, and received food induces operant activity producing food. Not only negative reinforcement, but reinforcement in general fails to explain behavior when reinforcement is defined as due to consequences. Induction erases the distinction between consequences and antecedents.
Control by Past, Present, and Potential Depends on Generalization Across Multiple Dimensions
|SARAH COWIE (University of Auckland, New Zealand)
Research suggests that behavior depends on likely future environmental conditions, as generalized from past experience and present environment. We assessed how generalization across the various dimensions of important events – reinforcers – affects the degree to which behavior tracks changes in likely future reinforcer availability. Pigeons worked on a concurrent-choice task in which one response was more likely to produce a reinforcer before a stimulus change, and the other was more likely after the stimulus change. We manipulated the nature of the stimulus (time or number) and the nature of the response (e.g., to a location, color, or flash duration). Regardless of the nature of the stimulus and response, choice was well described by a model that assumes that control by the structure of the environment is weakened by generalization across the various dimensions of a reinforcer (or other important events). We suggest that generalization contributes substantially to the degree of control by likely future events.
Reinforcement in Action: A Disequilibrium Theory of Reinforcement With Practical Aims
|KENNETH W. JACOBS (Salem State University)
There are no “reinforcers” or “punishers” as we colloquially know them in practice – there are only reinforcement and punishment effects. Response Disequilibrium Theory (RDT) is an approach to reinforcement that makes no assumptions about response strengthening or the value of stimuli to reinforce. RDT is the culmination of Timberlake and Allison’s (1974) response deprivation hypothesis and should be differentiated from motivating operations. The disequilibrium approach to reinforcement is a general theory of behavior that asks the following question: What do individuals usually do during their free time—without any constraints on their behavior—and what would happen if we disrupted that behavior with a contingency? Unlike other theories of reinforcement, RDT predicts the direction and magnitude of behavior change. RDT a priori specifies the necessary conditions for reinforcement, which makes it testable and practicable. Response Disequilibrium Treatment (RDTx) is an emerging application that can compete with the expediency of calling a stimulus a reinforcer. RDTx is defined by (1) the assessment of unconstrained activities in context and (2) the scheduled disruption of activities to induce socially significant behavior change. The current presentation aims to describe RDT’s terms, methods, and applications.
|How the Conceptualization of Behavior-Environment Interaction as a Selectionist System Relates to Material Reality
|JACK J MCDOWELL (Emory University)
|Abstract: Conceptualizations of behavior-environment interaction, and the theories that are developed from them, may be more or less readily related to material reality. To take examples from another discipline, the relation between Newtonian mechanics and the material world is readily apparent for the most part, whereas the relation between quantum mechanics and the material world is problematic. In this symposium presentation, behavior-environment interaction will be conceptualized as a selectionist system that consists of three processes: selection, reproduction, and mutation. A computational theory based on this conceptualization, and the evidence supporting it, are presented in detail in a SQAB tutorial at this conference. In the present symposium, this computational theory will be described briefly, but the focus will be on the theory’s material interpretation. It will be argued that the algorithmic operation of the theory’s three elements supervenes on, but is not necessarily identical to, material operations of the brain. In other words, if the theory agrees fully with experiment, then whatever the material operations of the brain may be, they must be functionally equivalent to selection, reproduction, and mutation. Other examples of this type of supervenient realism will be presented to support the plausibility of the argument.