When learning takes place, be it classical or operant conditioning, it involves the occurrence of certain processes. These include reinforcement, extinction or non-occurrence of learned response, generalisation of learning to other stimuli under some specifiable conditions, discrimination between reinforcing and non-reinforcing stimuli, and spontaneous recovery.
Reinforcement is the operation of administering a reinforcer by the experimenter. Reinforcers are stimuli that increase the rate or probability of the responses that precede. We have noted that reinforced responses increase in rate, while non-reinforced responses decrease in rate. A positive reinforcer increases the rate of response that precedes its presentation. Negative reinforcers increase the rate of the response that precedes their removal or termination. The reinforcers may be primary or secondary. A primary reinforcer is biologically important since it determines the organism’s survival (e.g., food for a hungry organism). A secondary reinforcer is one which has acquired characteristics of the reinforcer because of the organism’s experience with the environment. We frequently use money, praise, and grades as reinforcers. They are called secondary reinforcers. Systematic use of reinforcers can lead to the desired response. Such a response is shaped by reinforcing successive approximations to the desired response.
Extinction means disappearance of a learned response due to removal of reinforcement from the situation in which the response used to occur. If the occurrence of CS-CR is not followed by the US in classical conditioning, or lever pressing is no more followed by food pellets in the Skinner box, the learned behaviour will gradually be weakened and ultimately disappear. Learning shows resistance to extinction. It means that even though the learned response is now not reinforced, it would continue to occur for sometime. However, with increasing number of trials without reinforcement, the response strength gradually diminishes and ultimately it stops occurring. How long a learned response shows resistance to extinction depends on a number of factors. It has been found that with increasing number of reinforced trials resistance to extinction increases and learned response reaches its highest level.
At this level performance gets stabilised. After that the number of trials do not make a difference in the response strength. Resistance to extinction increases with increasing number of reinforcements during acquisition trials, beyond that any increase in number of reinforcement reduces the resistance to extinction. Studies have also indicated that as the amount of reinforcement (number of food pellets) increases during the acquisition trials, resistance to extinction decreases. If the reinforcement is delayed during acquisition trials, the resistance to extinction increases. Reinforcement in every acquisition trial makes the learned response to be less resistant to extinction. In contrast, intermittent or partial reinforcement during acquisition trials makes a learned response more resistant to extinction.
Generalisation and Discrimination
The processes of generalisation and discrimination occur in all kinds of learning. However, they have been extensively investigated in the context of conditioning. Suppose an organism is conditioned to elicit a CR (saliva secretion or any other reflexive response) on presentation of a CS (light or sound of bell). After conditioning is established, and another stimulus similar to the CS (e.g., ringing of telephone) is presented, the organism makes the conditioned response to it. This phenomenon of responding similarly to similar stimuli is known as generalisation. Again, suppose a child has learned the location of a jar of a certain size and shape in which sweets are kept. Even when the child’s mother is not around, the child finds the jar and obtains the sweets. This is a learned operant. Now the sweets are kept in another jar of a different size and shape and at a different location in the kitchen. In the absence of the mother the child locates the jar and obtains the sweets. This is also an example of generalisation. When a learned response occurs or is elicited by a new stimulus, it is called generalisation.
Another process, which is complimentary to generalisation, is called discrimination. Generalisation is due to similarity while discrimination is a response due to difference. For example, suppose a child is conditioned to be afraid of a person with a long moustache and wearing black clothes. In subsequent situation, when s/he meets another person dressed in black clothes with a beard, the child shows signs of fear. The child’s fear is generalised. S/he meets another stranger who is wearing grey clothes and is clean-shaven. The child shows no fear. This is an example of discrimination. Occurrence of generalisation means failure of discrimination. Discriminative response depends on the discrimination capacity or discrimination learning of the organism.
Spontaneous recovery occurs after a learned response is extinguished. Suppose an organism has learned to make a response for getting reinforcement, then the response is extinguished and some time lapses. A question now may be asked, whether the response is completely extinguished, and will not occur if the CS is presented. It has been demonstrated that after lapse of considerable time, the learned or CR recovers and occurs to the CS. The amount of spontaneous recovery depends on the duration of the time lapsed after the extinction session. The longer the duration of time lapsed, the greater is the recovery of learned response. Such a recovery occurs spontaneously.