Spence's Theory Of Discrimination Training

In operant conditioning stimuli are associated with a reinforcer or punishment. Generalization occurs when an individual responds to similar stimuli in the same way and discrimination occurs when they respond to some stimuli but not others. Stimulus discrimination training refers to the strategy in which an individual is taught to engage in certain behaviors in the presence of particular stimuli. This type of training can either be conducted using classical or instrumental conditioning procedures. In instrumental conditioning procedures the stimulus that signals the availability of the reinforcement is referred to as the S+, while the stimulus that signals the lack of reinforcement is called the S-. The S+ and S- are called discriminative stimuli once they have successfully gained control of the behavior of an individual. The S+ is the discriminative stimuli for performing an instrumental response while the S- is the discriminative stimuli for not performing such a response. Dr. Kenneth W. Spence produced a theory of discrimination learning, which aimed to gain a better understanding on what is

learned during discrimination training.
According to Spence’s theory of discrimination learning, reinforcement of a response in the presence of the S+ conditions excitatory response tendencies to S+. If a subject responds to a stimulus and the response is reinforced, the tendency to respond to that stimulus again increases. This is referred to as excitation. In contrast, not reinforcing a subject’s response to a stimulus decreases the tendency to respond to the stimulus again, in what is known as inhibition. With stimulus generalization, after a response to a stimulus is reinforced, that response tends to occur in similar stimuli. The greater the similarity to the stimulus, the stronger the response will be. In contrast, after not reinforcing a response to a stimulus, that response tends to be suppressed in the presence of similar stimuli. As a result, the greater the similarity, the stronger the suppression. In the stimulus generalization gradient, the gradient of responding is plotted as a function of how similar the test stimuli are to the original training stimulus. For example, if birds are exposed to two different tones (S+ or S-) and must discriminate between them, responding is less generalized due to the competing tone producing no reward. This results in the shape of the excitatory gradient becoming steeper and narrower near the top.
The theory of discrimination learning suggests that discrimination-training results in excitation to the S+ being conditioned while inhibition is conditioned to the S-.This theory can be supported with summation tests to determine if the S- is capable of inhibiting instrumental responding. For example, a pigeon can be used an experiment in which there would be an S+ (tone) and an S- (light). If a tone is produced and the pigeon pecks, they receive food. If a light comes on and they peck, they receive nothing. To test the effect of the S-, the tone should be paired with the light so that if S- has inhibitory properties, responding will be driven down. When the tone is presented alone there should be high responding and when it is presented with the light responding should be suppressed.
In addition, the theory of discrimination learning also takes into account how S+ and S- can influence the other, particularly when they are somehow related.

In a training procedure called an intradimensional discrimination, the S+ and S- are similar except for one feature. Intradimensional discrimination can produce an event, which accounts for the theory of discrimination learning, known as the peak shift effect. The peak shift effect is a shift in the peak responding that moves from the original S+ in the direction away from the S-. This effect occurs when both the excitatory (S+) and inhibitory (S-) stimuli are conditioned and the inhibition changes the shape of the gradient. As a result, the maximum amount of responding occurs to a stimulus that was not previously trained as the S+. For example if the inhibitory stimulus (S-) is to the right, the peak shifts to the

left.