Contemporary Learning Theories

Volume II: Instrumental Conditioning Theory and the Impact of Biological Constraints on Learning

Stephen B. Klein, Robert R. Mowrer, Robert W. Mowrer, Stephen B. Klein, Robert R. Mowrer, Robert W. Mowrer

• Editorial: Routledge
• Año: 2019
• ISBN:
  9781317728146
• Edición: 1
• Categoría: Psicología
• Categoría: Historia y teoría en psicología

This unique two-volume set provides detailed coverage of contemporary learning theory. Uniting leading experts in modern behavioral theory, these texts give students a complete view of the field. Volume I details the complexities of Pavlovian conditioning and describes the current status of traditional learning theories. Volume II discusses several important facets of instrumental conditioning and presents comprehensive coverage of the role of inheritance on learning. A strong and complete base of knowledge concerning learning theories, these volumes are ideal reference sources for advanced students and professionals in experimental psychology, learning and learning theory, and comparative physiology.

I INTRODUCTION

1 A Contrast Between Traditional and Contemporary Learning Theory

Robert R. Mowrer

Angelo State University

Stephen B. Klein

Fort Hays State University

I. INTRODUCTION

The chapters in this text discuss the process of learning and the modern theories designed to explain this process. In particular, this volume focuses on instrumental learning and the biological constraints on learning. Learning is defined as “a relatively permanent change in the ability to exhibit a behavior; this change occurs as the result of successful or unsuccessful experience” (Klein, 1987, p. 2). Investigations of the learning process have occupied psychologists for over 100 years and remains a central research topic to this day. Many different views of learning have been proposed by a group of psychologists known as behaviorists and this chapter briefly examines the early or traditional conceptualizations of the learning process. These early theories were intended to explain all aspects of the learning process and may be referred to as “global” theories of learning. While this approach provided substantial information about learning, the more contemporary learning theorists investigate and describe more specific aspects of the learning process. After a brief description of traditional learning theories, we look at the reasons why many contemporary learning theories focus on specific learning principles rather than attempting to explain all aspects of learning. The remaining 8 chapters present the ideas and research of a number of noted contemporary learning theorists regarding instrumental conditioning and biological constraints.

II. EARLY CONCEPTS OF LEARNING

At the crux of learning theory is the school of psychology known as Behaviorism, which emphasizes the role of experience in controlling behavior. According to this position, all organisms possess innate motives which drive instinctive behavior, but the more important determinants of behavior are those that are learned. These acquired or learned drives come to motivate behavior and are learned through the organisms interaction with the environment. The primary goal of the behaviorist (and learning theorist) is to determine the rules or laws governing this learning and use these laws to predict, control, and understand behavior. Although a number of ideas have impacted on the behavioral view, the concept of a reflex has been especially important.

Descartes and the role of reflexes. According to the 17th-century philosopher, Rene Descartes (1662/1972), all human and nonhuman behavior is a reflexive response to environmental events. An external stimulus excites a given sense, then sends this excitation to the brain, which in turn conveys this excitation to the muscle, resulting in behavior. This reflexive system is purely mechanical in that the behavior is an automatic response to the environment. The reflex concept was particularly important in the early investigations of classical or Pavlovian conditioning, which is detailed in the companion volume “Contemporary Learning Theories: Pavlovian Conditioning and the Status of Traditional Learning Theory.”

The Role of the British Associationists. Locke (1690) argued that all ideas are developed through experience and that the neonate begins with an absence of any ideas (tabula rosa) and that ideas are imprinted on this blank slate as a result of experience. According to Locke, the slate is imprinted with associations or links between events. He suggested that events that are contiguous become associated and the activation of one event will automatically activate its associate. Likewise, David Hume (1739) posited that causal associations could also be formed as a result of experience. A causal association is formed when it is believed that one event causes the occurrence of another event. The concept of an association is central to the learning process and was especially so for the work of Edward Thorn-dike (1898).

III. A BRIEF REVIEW OF TRADITIONAL LEARNING THEORY

Thorndike’s Role. Thorndike’s studies established that an animal’s behavior could change as a result of experience. In his classic puzzle box experiments with cats, Thorndike noticed that if he placed a hungry cat in the box with food located outside, the cat would engage in a variety of behaviors in its attempt to obtain the food. The cat eventually happened onto a response which resulted in its escape from the box and access to the food (pulling a string or pressing a pedal). With each successive trial, the time to escape decreased.

Thorndike reasoned that the animal formed an association between the stimulus (box) and response (string pull or pedal push). This S–R association developed because the hungry cat was rewarded with food following the response, resulting in a satisfying state of affairs. The association’s strength increased with each successive trial. Thorndike labeled this strengthening of an S–R bond by pleasant events the “law of effect.” Thorndike further proposed that learning was a trial and error process. His cat simply changed behaviors until it discovered an effective one. Reward then functioned to strengthen the correct response. This point, that reward (or reinforcement) was essential for learning, resulted in Thorndike becoming the first “reinforcement” theorist. In contrast, two other influential learning theorists, Edwin Guthrie and John Watson (whose work was very similar to Thorndike’s) appealed only to the contiguity between events such as a stimulus and response to explain the learning process, relegating reinforcement to only a secondary role.

The Influence of Watson and Guthrie. John Watson regarded all learning as classical conditioning. He believed that all organisms are born with certain S–R connections of reflexes. In addition, any new S–R connection was learned through the process of classical conditioning described by Pavlov (1927) or Bekhterev (1913).

Two important learning principles described by Watson were frequency and recency. Frequency refers to the notion that the more frequently a certain response is made to a stimulus, the more likely that response will occur again in the presence of that stimulus. The recency principle states that the more recently a given response was made to a stimulus, the more likely that response will occur again when the stimulus is presented. Why does one response occur over others? Unlike Thorndike, who believed that the response which was followed by reward would be strengthened, Watson felt that the response that resulted in the greatest change in the stimulus situation would gain strength. For example, in the puzzle box problem, pulling the string allowed the cat to escape. This response changed the stimulus situation such that the cat was no longer in the presence of the stimulus (box). It was this repeated change in stimuli following a response that stamps in the S–R bond in accord with the rules of frequency and recency. How does reward function in Watson’s theory? According to Watson, reward has no function and as long as a response occurred in a specific context, an association would develop regardless of whether or not reward occurred. Further, Watson felt that all behavior, even complex behavior, could be explained by stamping in a series of S–R bonds.

Edwin Guthrie’s position on the learning process was very similar to that described above for Watson. Like Watson, Guthrie rejected Thorndike’s law of effect and felt that contiguity between a stimulus and response was sufficient to establish an S–R association. He felt that anytime a response occurs in the presence of some stimulus they automatically become associated. Unlike Watson (or Thorndike), Guthrie believed that the S–R association was at full strength following a single pairing. That is, the formation of an S–R bond was all or none, and there was no gradual strengthening or stamping in of an association through repetition. To explain the apparent gradual strengthening of an S–R bond, Guthrie appealed to three principles. First, while numerous stimuli may be present in a given situation, only a portion of these are being attended to by the organism and thus available for association. Since the particular stimuli attended to will vary from trial to trial, for a particular stimulus to produce a response it must be attended to on the previous and subsequent trials. Thus, the change in behavior from trial to trial reflects an attentional process not a learning process. Second, any number of different stimuli can become associated with a certain response. Over trials, as more and more stimuli become associated with the response, the likelihood of making a response on later trials increases. This increase is not due to a strengthening of the S–R bond but rather to an increase in the number of stimuli available to elicit the response. Finally, a complex behavior consists of many separate responses. For responding to be effective, each behavior must become associated with the stimulus. This process occurs gradually over trials and results in the apparent strengthening of the association.

What then is the role of reinforcement? Guthrie, like Watson, claimed that the response that changes the stimulus situation is the one which gets associated. According to Guthrie, reinforcement changes the stimulus situation (internal or external) and ensures that the response immediately preceding it enters into the S–R association. This change in stimulus context also prevents new responses from being associated with the original stimulus.

While Thorndike is considered the first reinforcement theorist, a second reinforcement theory was developed by Clark Hull. Hull’s reinforcement theory stands in contrast to Watson’s and Guthrie’s contiguity version of learning theory.

Hull-Spence Learning Theory. From the 1930s to the early 1960s the most elegant and dominant theory of learning was that developed by Clark Hull and later modified by Kenneth Spence. This theory is detailed in the companion volume in this set, so we only engage in a cursory examination here.

The critical component of this theory was the concept of DRIVE (Woodworth, 1918). A drive is defined as an intense internal force which motivates behavior. According to Hull (1943), this intense internal arousal, drive or D, automatically motivates behavior and further, could be either unconditioned or acquired.

All organisms are motivated to maintain biological homeostasis. When a given biological system is “out of balance,” the organism is automatically motivated to adjust in order to restore the biological system to normal. For example, in the absence of food, an organism will utilize stored energy to maintain normal functioning. If this absence persists, behavior will be activated to resolve the deficiency (i.e., seek food sources). In addition to maintaining homeostasis, organisms are also motivated to avoid intense environmental events such as shock or loud noise. That is, these events activate the internal drive state and the organism behaves appropriately. All organisms behave in order to reduce this internal drive or tension.

Hull suggested that through classical conditioning, previously neutral environmental events can acquire the ability to produce internal drive states resulting in conditioned drives. What determines the behavior elicited by various drive states? For Hull, while drive is a general state of tension, each instrumental action depends on the environment. When a given response results in a reduction of drive, the stimuli present become associated with that response; thus, the likelihood of that response occurring again increases when in the presence of the stimuli. The strength of this association was termed habit strength and increased each time the response resulted in drive reduction. Unsuccessful behavior causes the drive to persist and behavior may be suppressed temporarily by a process referred to as reactive inhibition (similar to fatigue). Reactive inhibition eventually decreases and an animal again will engage in behavior intended to reduce the drive state. When behavior continues to be unable to reduce the drive, a more permanent suppression of responding may occur due to conditioned inhibition.

Hull initially felt that the intensity of the reinforcer influenced habit strength: the greater the reinforcement (larger reduction in drive), the stronger the habit strength. Studies by Crespi (1942) and others indicated that reinforcer magnitude had a significant effect on an animal’s motivation; these experiments led Hull to add the notion of reward magnitude or K to his theory. The specific role of K or incentive was detailed by Kenneth Spence. Briefly, when an animal traversed a maze or alley and was met with reward in the goal box, environmental stimuli in that goal box became associated with reward and elicited anticipatory goal responses on the next trial. These anticipatory goal responses are accompanied by internal stimulus changes which motivate instrumental behavior. This mechanism essentially moves back through the maze or alley until even stimuli present in the start box will elicit the anticipatory goal response and stimuli, thus motivating behavior. For Spence, the larger the reward, the stronger the anticipatory goal mechanism and the higher the incentive motivation of K. While Hull’s reinforcement theory had wide application, it failed to explain a number of learning phenomena. Many of these phenomena evolved out of the work of Edward Tolman.

Tolman’s Expectancy Theory. Tolman introduced to learning theory the concept of an expectancy. This notion was in wide contrast to the mechanistic views of behavior briefly described above. Indeed, for Tolman, behavior was not considered an automatic response to the environment but reflected direction and purpose. According to Tolman, an organism’s behavior is goal-directed in that they are motivated to approach or avoid a specific situation depending on whether that situation is rewarding or aversive. In addition, organisms are capable of understanding the structure of the environment and behaving appropriately. Importantly, it was not necessary that the organism be aware of the purpose or direction of behavior only that they behave as if they were aware.

For Tolman, behavior is not only goal-oriented, but specific outcomes are expected to follow specific behaviors. How does the organism know which behavior to exhibit? Tolman suggested the certain events in the environment convey information about where goals are located and the organism behaves appropriately only after learning the environmental signs leading to reward or away from punishment. As organisms explore their environment, a mental representation or cognitive map develops. This map guides exploration through the environment to obtain reward or avoid punishment.

Like Guthrie and Watson, Tolman did not feel that reward was necessary for learning to occur. The best example of this idea may be seen in his famous place learning studies, which also provided a great deal of trouble for the Hull-Spence theory of learning. The simplest design for such a study involved two group...