Models of Learning
The scientific investigation of the learning process was begun at the end of the 19th cent. by Ivan Pavlov in Russia and Edward Thorndike in the United States. Three models are currently widely used to explain changes in learned behavior; two emphasize the establishment of relations between stimuli and responses, and the third emphasizes the establishment of cognitive structures. Albert Bandura maintained (1977) that learning occurs through observation of others, or models; it has been suggested that this type of learning occurs when children are exposed to violence in the media.Classical Conditioning
The first model, classical conditioning, was initially identified by Pavlov in the salivation reflex of dogs. Salivation is an innate reflex, or unconditioned response, to the presentation of food, an unconditioned stimulus. Pavlov showed that dogs could be conditioned to salivate merely to the sound of a buzzer (a conditioned stimulus), after it was sounded a number of times in conjunction with the presentation of food. Learning is said to occur because salivation has been conditioned to a new stimulus that did not elicit it initially. The pairing of food with the buzzer acts to reinforce the buzzer as the prominent stimulus.
A second type of learning, known as operant conditioning, was developed around the same time as Pavlov's theory by Thorndike, and later expanded upon by B. F. Skinner. Here, learning takes place as the individual acts upon the environment. Whereas classical conditioning involves innate reflexes, operant conditioning requires voluntary behavior. Thorndike showed that an intermittent reward is essential to reinforce learning, while discontinuing the use of reinforcement tends to extinguish the learned behavior. The famous Skinner box demonstrated operant conditioning by placing a rat in a box in which the pressing of a small bar produces food. Skinner showed that the rat eventually learns to press the bar regularly to obtain food. Besides reinforcement, punishment produces avoidance behavior, which appears to weaken learning but not curtail it. In both types of conditioning, stimulus generalization occurs; i.e., the conditioned response may be elicited by stimuli similar to the original conditioned stimulus but not used in the original training. Stimulus generalization has enormous practical importance, because it allows for the application of learned behaviors across different contexts. Behavior modification is a type of treatment resulting from these stimulus/response models of learning. It operates under the assumption that if behavior can be learned, it can also be unlearned (see behavior therapy).
A third approach to learning is known as cognitive learning. Wolfgang Köhler showed that a protracted process of trial-and-error may be replaced by a sudden understanding that grasps the interrelationships of a problem. This process, called insight, is more akin to piecing together a puzzle than responding to a stimulus. Edward Tolman (1930) found that unrewarded rats learned the layout of a maze, yet this was not apparent until they were later rewarded with food. Tolman called this latent learning, and it has been suggested that the rats developed cognitive maps of the maze that they were able to apply immediately when a reward was offered.