Chapter 7
Chapter 7: Learning
Chapter Review
THE PERSPECTIVE OF LEARNING THEORY
- The empiricist philosophers argued that learning involves the forming of simple associations. More complex learning just involves a great many associations, each layered upon the others. From this perspective, all learning depends on the same mechanisms, and so all learning should be governed by the same principles.
HABITUATION
- The simplest form of learning is habituation, a decline in the response to stimuli that have become familiar through repeated exposure. In dishabituation, the organism learns that a previously predictable stimulus has now changed, causing the organism to renew its attention to the stimulus.
CLASSICAL CONDITIONING
- In classical conditioning, animals learn about the association between one stimulus and another. Before conditioning, an unconditioned stimulus (US, such as food) elicits an unconditioned response (UR, such as salivation). After repeated occasions on which the US follows a conditioned stimulus (CS, such as a buzzer), this CS alone will begin to evoke the conditioned response (CR; here again, salivation).
- When a CS-US relationship is well established, the CS can be preceded by a second, neutral stimulus to produce secondorder conditioning.
- Trials in which the CS is presented without the US lead to extinction. However, the phenomenon of spontaneous recovery shows that the CR is masked, not abolished, by extinction.
- Because of stimulus generalization, the CR can also be elicited by stimuli that are similar to the CS. To train the animal to discriminate among stimuli, one stimulus (CS+) is presented with the US, while another (CS–) is presented without the US.
- Several lines of evidence suggest that the CS serves as a signal for upcoming events. This fits with the fact that learning is less likely if the CS is simultaneous with the US, or (worse) follows it. The signal value of the CS is also evident because learning occurs only if there is some contingency between CS and US; mere contiguity between these stimuli isn’t enough. Animals can also learn about the absence of contingency, and animals clearly prefer environments in which stimuli are predictable.
- Animals seem sensitive to relationships among probabilities, but this is not because they’re tracking the probabilities directly. Instead, animals develop expectations about upcoming events and adjust their expectations whenever events surprise them. In this way, the expectations are gradually adjusted until they are accurately in tune with probabilities in the environment. The role of surprise is directly evident in the blocking effect.
- The CR is not identical to the UR. Instead, the CR seems to be a means of preparing for the US. Sometimes the preparation takes the form of a compensatory response, and this point may be crucial in understanding drug addiction and drug tolerance.
INSTRUMENTAL CONDITIONING
- When training an animal using instrumental (or operant) conditioning, the trainer delivers a reward or reinforcement only after the animal gives the appropriate response. According to Thorndike, learning in this situation is governed by the law of effect, which states that the tendency to perform a response is strengthened if it’s followed by a reward and weakened if it’s not.
- Operants are voluntary responses, strengthened by reinforcement; but acquiring them may call for some initial shaping, through a method of successive approximations.
- Some reinforcers are stimuli whose reinforcing power is unlearned. Other conditioned reinforcers acquire their power from prior presentations with stimuli already having that capacity. The magnitude of a reinforcer depends on several factors, including the magnitude of other reinforcers that might be available. This effect, which is reflected in the phenomenon of behavioral contrast, may be one source of findings sometimes attributed to intrinsic motivation. Many theorists, however, believe that intrinsic motivation involves a separate set of principles, different from those that govern operant conditioning.
- During partial reinforcement, the response is reinforced only some of the time. The rule that determines when a reinforcer is given is called a schedule of reinforcement. In ratio schedules, reinforcement is delivered after a number of responses; the ratio used may be fixed or variable. In interval schedules, reinforcers are delivered for the first response made after a given interval since the last reinforcement; this interval, too, can be fixed or variable.
- Learning involves more than a change in behavior; it also involves the acquisition of new knowledge. This principle is evident in many settings, including latent learning.
- Operant conditioning results when reinforcement is contingent on a response, not just when reinforcement happens to be contiguous with responding. Organisms’ sensitivity to contingency can be demonstrated in the phenomenon of learned helplessness, where animals seem to learn they have no control over the events they’re experiencing.
OBSERVATIONAL LEARNING
- Many animals can learn simply by watching other individuals and learning from their example. This is evident in vicarious conditioning, and it’s also clear in learning through imitation. However, the impact of media violence reminds us that imitation can be a source of undesired behaviors as well as a source of new skills.
VARIETIES OF LEARNING
- According to the early learning theorists, just about any CS can become associated with any US, and just about any response can be strengthened by any reinforcer. This assertion is challenged by the fact that certain conditioned stimuli are more readily associated with some unconditioned stimuli than with others, as shown by studies of taste aversion learning. These studies suggest that animals are biologically prepared to learn certain relations more readily than others. Similar effects occur in instrumental conditioning, where some responses are more readily strengthened by some reinforcers than by others.
- Certain forms of learning are species specific. In humans, specialized forms of learning may include our capacity for learning language as well as our remarkable ability to learn by observing others.
- Animals vary in how (and what) they learn; but in some aspects of learning, diverse species also exhibit striking similarities. These similarities probably arise because all organisms live in the same world, and the nature of the world creates a need in many species for the forms of learning we call classical and operant conditioning.
THE NEURAL BASIS FOR LEARNING
- In recent years, investigators have made considerable progress in understanding the neural bases for learning. These bases involve diverse mechanisms, such as presynaptic facilitation, and postsynaptic changes, such as long-term potentiation (LTP). Still another mechanism involves the creation of new synapses—made possible by the growth of new dendritic spines, which act as “receiving stations” for synapses and open new lines of communication between cellular neighbors.