1) A procedure that is used to investigate recognition of stimuli. For example, in a simple recognition procedure, a pigeon may be presented with three keys. A triangle is projected onto the centre key. The triangle is the sample stimulus. To ensure that the bird attends to the sample, it is required to peck the sample key. When this happens, two sidekeys are illuminated with a triangle on one and a square on the other, called the comparison stimuli. If the bird pecks the comparison stimulus that corresponds to the sample, this behaviour is reinforced and leads to the presentation of a new sample. Pecks to the noncorresponding stimulus result in extinction and the next trial.

2) A procedure in which the choice of a stimulus that matches a sample stimulus is followed by a reinforcer. Typically, in the matching-to-sample procedure, the organism touches a key on which the sample stimulus appears. The performance on the sample stimulus is reinforced by the appearance of two stimuli on two other keys. These stimuli correspond to the sample. The final reinforcement occurs if the organism chooses the key on which the stimulus corresponding to the sample stimulus appears. A time out or the reappearance of the sample stimulus occurs if the organism chooses the key which does not correspond to the sample.

3) A simultaneous conditional discrimination procedure, or the performance maintained by such a procedure. As it is typically arranged for pigeons, a sample stimulus is presented on the n-dddle key of three keys. A peck on it turns on comparison stimuli on two side keys (see observing response, sample-specific behavior). A peck on the matching side key is reinforced (perhaps according to some schedule); a peck on the other side key may produce a timeout or invoke a correction procedure. When the criterion for matching is physical correspondence (as when a pigeon must peck a green comparison given a green sample and a red comparison given a red one), the procedure is sometimes called identity matching, though accurate matching may be based on features other than the identity relation, such as stimulus configurations. When the matching is based on arbitrary sample-comparison relations (e.g., as when a pigeon must peck a circle given a green sample and a triangle given a red one), the procedure is sometimes called arbitrary matching (symbolic matching, an alternative terminology, has the disadvantage of suggesting that the sample and comparison stimuli have additional stimulus functions besides those in the matching-to sample procedure).