In applied behavior analysis, the terms replacement behavior, differential reinforcement, and extinction are often used interchangeably in popular dog training discourse. This imprecision creates confusion in both planning and implementation, leading to inconsistent outcomes and misinterpretation of normal learning processes. Although these concepts are functionally related, they are not synonymous, and understanding the distinctions between them is essential for designing effective, ethical, and durable behavior change programs.
A replacement behavior is best understood as a topographical alternative to a target behavior. It describes what the learner is expected to do instead of the behavior of concern. Importantly, a replacement behavior is not inherently effective simply because it is incompatible or desirable. For a behavior to function as a true replacement, it must be capable of producing reinforcement that is equal to or greater than that historically produced by the target behavior. Without this functional relationship, a replacement behavior exists only as a suggestion, not as a viable behavioral alternative. Teaching a dog to sit instead of jump, for example, does not reduce jumping unless sitting reliably contacts reinforcement that jumping no longer produces.
Differential reinforcement (DR) refers not to a behavior, but to a class of reinforcement contingencies arranged to change the distribution of responding. In DR procedures, reinforcement is delivered for one class of responses while being withheld for another. Variants such as DRA (differential reinforcement of alternative behavior), DRI (incompatible behavior), and DRO (other behavior) specify which responses contact reinforcement, but all share the same defining feature: reinforcement is made contingent on some responses and unavailable for others. Differential reinforcement is therefore the mechanism through which replacement behaviors are strengthened, not the replacement behavior itself.
What is often underemphasized—or actively misunderstood—is that extinction is inherently embedded within differential reinforcement procedures. Extinction occurs whenever a response that previously contacted reinforcement no longer does so. When reinforcement is shifted toward an alternative response, the original behavior is necessarily placed on extinction. This is not a separate intervention layered onto differential reinforcement; it is a byproduct of changed contingencies. Attempting to use differential reinforcement without extinction is conceptually incoherent, as both processes occur simultaneously when reinforcement is redistributed.
The behavioral effects commonly attributed to “failed training” during DR interventions—such as escalation, variability, or regression—are in fact well-documented extinction effects. An extinction burst, characterized by a temporary increase in response frequency, intensity, or duration, reflects the learner’s history with the behavior and its prior effectiveness. Response variability often emerges as the organism samples alternative behaviors in search of reinforcement, and spontaneous recovery may occur following time gaps or contextual changes. These effects are not pathological; they are expected outcomes when reinforcement contingencies shift and should be anticipated in program design.
Misinterpreting extinction effects frequently leads practitioners to introduce positive punishment prematurely, often in response to escalation during extinction bursts. While punishment may suppress behavior temporarily, it also interferes with the learner’s ability to contact and stabilize alternative reinforcement contingencies. From a behavior-analytic perspective, early punishment can obscure functional relations, suppress adaptive variability, and prevent the replacement behavior from achieving sufficient reinforcement density to compete with historical contingencies. The result is often suppression without replacement, rather than durable behavior change.
Function and motivation further complicate the distinction between these concepts. Replacement behaviors must be selected with consideration of the function of the target behavior and the motivational variables maintaining it. Behaviors maintained by access to movement, social engagement, sensory feedback, or high arousal require reinforcement capable of competing at that same functional level. A low-value reinforcer delivered for a replacement response cannot reasonably be expected to outcompete behaviors maintained by strong biological or environmental reinforcement. Thus, failures attributed to “extinction” are often failures of reinforcement design.
In summary, replacement behaviors describe what the learner does, differential reinforcement describes how reinforcement is arranged, and extinction describes what happens to behavior that no longer produces reinforcement. Confusing these concepts leads to inappropriate expectations, mistimed interventions, and unnecessary frustration for both trainers and learners. When properly understood and implemented, these processes work together to produce clear, functional, and ethically sound behavior change. Extinction is not something to fear or avoid; it is an inevitable and informative component of learning under differential reinforcement.
References
Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied Behavior Analysis (3rd ed.). Pearson.
Lerman, D. C., & Iwata, B. A. (1995). Prevalence of the extinction burst and its attenuation during treatment. Journal of Applied Behavior Analysis, 28(1), 93–94.
Vollmer, T. R., & Iwata, B. A. (1992). Differential reinforcement as treatment for behavior disorders: Procedural and functional variations. Research in Developmental Disabilities, 13(4), 393–417.
Sidman, M. (2001). Coercion and Its Fallout. Authors Cooperative.
