Negative reinforcement

A learning process in which a behaviour increases because performing it removes something unpleasant. The “negative” refers to subtraction, not to the behaviour being bad or harmful.

A horse moving forward when leg pressure is applied and released as soon as they respond is negative reinforcement: the unpleasant stimulus (leg pressure) is removed when the desired behaviour (moving forward) is offered. The same mechanism works across species. A dog coming inside when called because doing so ends the discomfort of standing in the rain is negative reinforcement. A cat jumping off a hot surface, learning to avoid that surface in future, is negative reinforcement.

Negative reinforcement is one of the four quadrants of operant conditioning and is the dominant mechanism used in most ridden horse work and in much of in-hand horse handling. It is also widely used in dog training (lead pressure, body-blocking) and in some applications across other species.

The term is often confused with punishment, but the two are opposites: negative reinforcement increases a behaviour by removing something unpleasant, whereas punishment decreases a behaviour by either adding something unpleasant or removing something pleasant. This confusion is one of the most common in lay discussions of training, and matters because the welfare implications of the two are very different.

The effectiveness and welfare-status of negative reinforcement depend critically on timing. The unpleasant stimulus must end at the precise moment the desired behaviour is offered, or the learner is reinforced for the wrong thing. Inconsistent application of negative reinforcement is one of the most common causes of training problems across species.