Learning Theory

HISTORICAL OVERVIEW
Diane F. Halpern
Beth Donaghey

CONSTRUCTIVIST APPROACH
Mary Lamon

SCHEMA THEORY
William F. Brewer

HISTORICAL OVERVIEW

Learning theories are so central to the discipline of psychology that it is impossible to separate the history of learning theories from the history of psychology. Learning is a basic psychological process, and investigations of the principles and mechanisms of learning have been the subject of research and debate since the establishment of the first psychological laboratory by Wilhelm Wundt in Leipzeig, Germany, in 1879. Learning is defined as a lasting change in behaviors or beliefs that results from experience. The ability to learn provides every living organism with the ability to adapt to a changing environment. Learning is an inevitable consequence of living–if we could not learn, we would die.

The evolution of learning theories may be thought of as a progression from broad theories developed to explain the many ways that learning occurs to more specific theories that are limited in the types of learning they are designed to explain. Learning theories are broadly separated into two perspectives. The first perspective argues that learning can be studied by the observation and manipulation of stimulus-response associations. This is known as the behaviorist perspective because of its strict adherence to the study of observable behaviors. This perspective was first articulated in 1913 by John Watson, who argued that psychology should be the study of observable phenomena, not the study of consciousness or the mind. Watson believed that objective measurement of observable phenomena was the only way to advance the science of psychology.

The second type of learning theory argues that intervening variables are appropriate and necessary components for understanding the processes of learning. This perspective falls under the broad rubric of cognitive learning theory, and it was first articulated by Wilhem Wundt, the acknowledged "father of psychology," who used introspection as a means of studying thought processes. Although proponents of these two perspectives differ in their view of how learning can be studied, both schools of thought agree that there are three major assumptions of learning theory: (1) behavior is influenced by experience, (2) learning is adaptive for the individual and for the species, and (3) learning is a process governed by natural laws that can be tested and studied.

Behavior Theory

The behaviorist perspective dominated the study of learning throughout the first half of the twentieth century. Behaviorist theories identified processes of learning that could be understood in terms of the relationships between the stimuli that impinge on organisms and the way organisms respond, a view that came to be referred to as S-R theories. A central process in S-R theories is equipotentiality. Equipotential learning means that learning processes are the same for all animals, both human and nonhuman. By studying learning in nonhuman animals, the early behaviorists believed they were identifying the basic processes that are important in human learning. They also believed that learning could only be studied by observing events in the environment and measuring the responses to those events. According to the behaviorists, internal mental states are impossible topics for scientific inquiry, and thus are not necessary in the study of learning. For behaviorists, a change in behavior is the only appropriate indicator that learning has occurred. According to this view, all organisms come into the world with a blank mind, or, more formally, a tabula rasa (blank slate), on which the environment writes the history of learning for that organism. Learning, from the behaviorist perspective, is what happens to an organism as a result of its experiences.

Types of behavioral learning. There are two main types of learning in the behaviorist tradition. The first is classical conditioning, which is associated with the work of Ivan Pavlov (1849–1936), a Russian physiologist who studied the digestive processes of dogs. Pavlov noticed that dogs salivated in the absence of food if a particular stimulus was present that had previously been paired with the presentation of food. Pavlov investigated the way in which an association between a neutral stimulus (e.g., a lab technician who fed the dogs), an unconditioned stimulus (food), and an unconditioned reflex (salivation) was made. Pavlov's classic experiment involved the conditioning of salivation to the ringing of a bell and other stimuli that were not likely to make a dog salivate without a previously learned association with food.

In the initial stages of the classical conditioning paradigm, an unconditioned response (UCR; in this case, salivation) is elicited by the presentation of an unconditioned stimulus (UCS; in this case, food). If a neutral stimulus (one that does not elicit the UCR, such as a bell) is paired with the presentation of the UCS over a series of trials, it will come to elicit a conditioned response (CR; also salivation in this example), even when the UCS (food) is absent. In the paradigm of classical conditioning, the previously neutral stimulus (bell) becomes a conditioned stimulus (CS), which produces the conditioned response (CR) of salivation. In other words, the animal in the experiment learns to associate the bell with the opportunity to eat and begins to salivate to the bell in the absence of food. It is as though the animal came to think of the bell as "mouthwatering," although behaviorists never would have used terms like think of, because thinking is not a directly observable behavior.

Even though the original work on classical conditioning was performed using nonhuman animals, this type of learning applies to humans as well. Learned taste aversions and the development of specific phobias are examples of classical conditioning in humans. For example, the first time a person hears a drill at a dentist's office, it probably will not cause the palms to sweat and the heart rate to quicken. However, through the pairing of the sound with the unpleasant sensation of having a cavity drilled, the sound itself may come to elicit symptoms of fear and anxiety, even if one is not in the dentist's chair. Feelings of fear and anxiety may generalize so that the same fear response is elicited by the sight of the dentist's lab coat or the dental chair.

The second type of learning that is categorized in the behaviorist tradition is instrumental or operant, conditioning. The main difference between instrumental conditioning and classical conditioning is that the emphasis is on behavior that is voluntary (emitted), not reflexive (elicited). The target behavior (e.g., a peck at a lever if one is studying birds) comes before the conditioning stimulus (e.g., food), as opposed to the classical model, which presents the conditioning stimulus (e.g., bell) prior to the target behavior (e.g., salivation).

In the instrumental paradigm, behaviors are learned as a result of their consequences. Edward Thorndike (1874–1949) was a pioneer in instrumental conditioning, although he resisted the label of behaviorist. In his view, the consequences of behaving in a particular way controlled learning. Behavior was instrumental in obtaining a goal, and the consequences of the behavior were responsible for the tendency to exhibit (and repeat) a behavior. Thorndike named this principle of instrumental conditioning the law of effect. He argued that if a behavior had a positive consequence or led to a satisfying state of being, the response (behavior) would be strengthened. If, on the other hand, a behavior had a negative consequence, the response would be weakened. Thorndike developed the principles of instrumental conditioning using a puzzle box that required that an animal exhibit a certain behavior (push a latch) to obtain a goal (open a door for access to food). The animal was given the opportunity, through trial and error, to discover the required behavior, and the behavior was reinforced through the opening of the door and access to food. With practice, the animal decreased the time that it needed to open the door. In the instrumental paradigm, the animal learned an association between a given situation and the response required to obtain a goal.

Operant conditioning and reinforcement. B. F. Skinner (1904–1990) is credited with the development of the operant-conditioning paradigm. Similar to instrumental conditioning, operant conditioning requires that an organism operate on the environment to achieve a goal. A behavior is learned as a function of the consequences of the behavior, according to a schedule of reinforcement or punishment. Unlike Thorndike, who used the concept of reward and satisfying states, Skinner emphasized the influence of reinforcers. Reinforcers are events that follow a response and increase the likelihood that the response will be repeated, but they do not suggest the operation of a cognitive component such as reward (or pleasure). Learning is influenced according to the schedules of reinforcement in the operant paradigm. Skinner tested the operant theory by carefully controlling the environment to study behavior and the effects of reinforcement.

According to Skinner, operant conditioning has two laws. The first is the law of conditioning, which states that reinforcement strengthens the behavior that precedes it, which makes it more likely that the behavior will be repeated. The second is the law of extinction, which states that lack of reinforcement for a behavior will make that behavior less likely to reoccur. Reinforcement consists of two types of events, those that are positive, which means that when they are presented (e.g., present tasty food) the probability of a behavior occurring is increased (e.g., press a lever to get the tasty food), and those that are negative, which means that when they are removed (e.g., stop a loud sound or painful shock) the probability of a behavior occurring is increased (e.g., press a lever to stop a loud sound or painful shock). Punishment is defined as an event that weakens the tendency to make a response. Punishment could involve presenting an aversive stimulus (e.g., presenting a loud sound or painful shock), or it could involve removing access to a positive stimulus (e.g., removing a tasty food when a lever is pressed).

Skinner also experimented with different reinforcement schedules, and he found that different schedules produced different patterns of responding. Continuous schedules of reinforcement deliver a reinforcer every time the target behavior is exhibited. These schedules are effective in establishing the target behavior, but the behavior disappears quickly if the contingency is not met. Intermittent schedules of reinforcement deliver the reinforcer on a ratio schedule. For example, an experimenter may decide to reinforce every fourth response that an animal makes, or a reinforcer may be presented after a fixed or random time interval. The two types of intermittent schedules that maintain a high rate of responding and are very resistant to extinction are variable ratio and variable interval schedules.

Strict adherence to the behaviorist tradition excluded analysis of mental or internal events. However, Skinner acknowledged the role of thought. He maintained that thought was caused by events in the environment, and therefore a theory of learning that was concerned with the influence of the environment was appropriate. Like Pavlov and Thorndike, Skinner's work was primarily conducted with nonhuman animals, but the principles of operant conditioning can be applied to humans as well, and they are widely used in behavior therapy and education.

Cognitive Theories

Although behaviorism was a prolific and dominant theory in learning through the early decades of the twentieth century, certain concerns and observations led to a resurgence of interest in cognitive theories of learning. One area of concern was the distinction between performance and learning–that is, does behaviorism describe the factors that influence performance of learned behavior, rather than the act of learning itself? Within the behaviorist literature, evidence of cognitive elements like expectation and categorization exist. Under an intermittent reinforcement schedule, for example, animals increase their rate of response immediately before a reinforcer is delivered, thus acting as though they expect it. Similarly, animals can be trained to distinguish between types of stimuli that belong to different classes. Learning this type of distinction seems to involve classification, which is a cognitive process. Most importantly, scientists who studied learning recognized that the behaviorist theories could not account for all types of learning. Humans and animals can learn something without exhibiting what they have learned, meaning that performance does not always reflect what has been learned.

Cognitive theories grew from the concern that behavior involves more than an environmental stimulus and a response, whether it be voluntary or reflexive. These theories are concerned with the influence of thinking about and remembering experiences or behavior. The assumptions about learning under cognitive theories are not the same as those for behaviorist theories, because thinking and remembering are internal events. Inferences about the internal events such as thinking and remembering can be made as long as they are paired with careful observation of behavior. Cognitive theorists assume that some types of learning, such as language learning, are unique to humans, which is another difference between these two perspectives. Cognitive theories also focus on the organism as an active processor of information that modifies new experiences, relates them to past experiences, and organizes this information for storage and retrieval. Cognitive psychologists also recognize that learning can take place in the absence of overt behavior.

Edward Tolman (1886–1959) was among the first psychologists to investigate the organization of behavior and learning. He conducted research in the behaviorist tradition (objective research on nonhuman species), but he introduced cognitive elements to his explanation of learning. In Tolman's theory, however, the cognitive elements were based on observed behavior, not on introspection. He believed that learning involved more than stimulus and response events; it involved the development of an organized body of knowledge or expectations about a given situation. Tolman conducted many of his learning experiments using rats whose learning task was to run through a maze. By varying the conditions in the maze, he came to the conclusion that learning involved an understanding about events and their consequences, and this led to purposive, goal-directed behavior. Tolman emphasized the role of expectation and its reinforcing influence on the repetition of behavior. He popularized the concept of cognitive maps, which represent an organism's understanding of the relationship between parts of the environment, as well as the organism's relationship to the environment.

In a clear break with behaviorists, Tolman noted that reinforcement was not a necessary component of learning, and that organisms could demonstrate latent learning. Latent learning is displayed only when an organism is motivated to show it. Tolman was also concerned with differences in behavior that might be attributed to internal states of the organism, a consideration that had been largely rejected by earlier theorists. In identical learning paradigms, two organisms can show different behaviors based on their different moods, physiology, or mental states.

Social learning theory. Social learning theory focuses on the sort of learning that occurs in a social context where modeling, or observational learning, constitutes a large part of the way that organisms learn. Social learning theorists are concerned with how expectations, memory, and awareness influence the learning process. Both humans and nonhumans can learn through observation and modeling. Consider, for example, the acquisition of sign language by the offspring of language-trained apes who learn to sign by watching their trained parents. Children learn many behaviors through modeling. A classic experiment by Albert Bandura (1961) allowed one group of children to observe an adult who aggressively pounded on a bobo doll (an inflatable doll used for punching), while another group watched a nonaggressive model and a third group had no model at all. The children who saw the aggressive adult often modeled (imitated) this behavior when given an opportunity to play with the same doll. The children who saw the nonaggressive model showed the least amount of aggressive play when compared to the other two groups. Social learning theorists retain the behaviorist principles of reinforcement and response contingencies, but they also extend the area of inquiry for learning to include components of cognitive processing such as attention, remembering, the processing of information about the environment, and the consequences of behavior.

Appreciation of the cognitive components of learning focused attention on the need to remember an experience over various time intervals. Information-processing theories developed from the cognitive perspective and involve the processes of coding, storing, and retrieving information about the environment. Information processing is used to study the processes of memory, a central cognitive component in modern learning theories. Theories of information processing are a by-product of the computer revolution, and they use the language of computers (e.g., sequential processing stages, input, output) to describe the processes of learning and memory. According to a human information-processing perspective, learning occurs in sequential stages, beginning with encoding information from the environment. Encoding of information involves the process by which information from the environment is translated into usable information. The next stage is storage, which involves keeping the information that has been encoded. Stored information builds the "database" of past learning. The final stage in the information-processing approach is retrieval, which involves accessing the stored information so that it can be used to perform a task. Organisms are seen as active participants in the information-processing model. They do not experience the environment passively or simply absorb information, but instead they seek out certain information, and then manipulate, modify, and store it for later use.

Learning theories have often been used to provide a guide for education. Earlier applications were concerned with the use of appropriate rewards and punishment, concerns that mirrored the major tenets of behaviorist theories. More recently, cognitive perspectives have shaped the field of education, and there has been more concern with learning methods that enhance long-term retention and the transfer of information and skills that are learned in schools to novel problems in out-of-school settings. For example, variability in encoding (learning material in different ways, e.g. video and text) produces more durable long-term retention, even though it is a more effortful (and generally less enjoyable) way to learn. In addition, students can become better thinkers when they receive specific instruction in thinking skills–and when the instruction is designed to enhance transfer. Teaching strategies that enhance transfer include spaced practice (viewing material over time versus cramming), using a variety of examples so learners can recognize where a concept is applicable, and practice at retrieval (repeatedly remembering material over time) with informative feedback.

Learning theories are facing new challenges as people grapple with increases in the amount of available information that needs to be learned, rapidly changing technologies that require new types of responses to new problems, and the need to continue learning throughout one's life, even into old age. Contemporary learning theories supported by empirical research offer the promise of enhanced learning and improved thinking–both of which are critical in a rapidly changing and complex world.