Any consequence that follows a behaviour and increases the likelihood of that behaviour happening again.

Reinforcement is one of the two primary categories of consequence within operant conditioning (the other being punishment). It comes in two forms: positive reinforcement, in which something the animal finds pleasant is added, and negative reinforcement, in which something the animal finds unpleasant is removed. Both forms increase behaviour. The “positive” and “negative” refer to addition or subtraction, not to the moral quality of the behaviour or the reinforcer.

The defining feature of reinforcement is the effect on behaviour. If a consequence is applied with the intention of reinforcing a behaviour, but the behaviour does not in fact increase, then the consequence has not been reinforcing for that individual on that occasion. This is one of the most useful diagnostic insights from learning theory: reinforcement is defined by its effect, not by the trainer’s intention. A wither scratch that one horse finds rewarding may be neutral or even mildly unpleasant to another.

The choice of reinforcer matters substantially. What functions as a reinforcer for a given animal depends on the species, the individual, the moment, and the surrounding context. Food typically works as a primary reinforcer across most species because it connects to fundamental physiological needs, but its effectiveness varies with hunger, palatability, and competing reinforcers in the environment. Effective trainers attend carefully to what is actually reinforcing for the individual animal in front of them.