Cognitive Psychology

An Overview for Cognitive Scientists

Lawrence W. Barsalou

• Publisher: Psychology Press
• Year: 2014
• eBook ISBN: 9781317783671
• Topic: Psychology
• Subtopic: Cognitive Psychology & Cognition
• Index: Psychology

This text presents the basic concepts of modern cognitive psychology in a succinct and accessible manner. Empirical results, theoretical developments, and current issues are woven around basic concepts to produce coherent accounts of research areas. Barsalou's primary goal is to equip readers with a conceptual vocabulary that acquaints them with the general approach of cognitive psychology and allows them to follow more technical discussions elsewhere. In meeting this goal, he discusses the traditional work central to modern thinking and reviews current work relevant to cognitive science.

Besides focusing on research and theory in cognitive psychology, Barsalou also addresses its fundamental assumptions. Because the cognitive approach to psychology is somewhat subtle, often misunderstood, and sometimes controversial, it is essential for a text on cognitive psychology to address the assumptions that underlie it. Therefore, three of the eleven chapters address the "meta- assumptions" that govern research and theory in cognitive psychology. These meta-chapters provide a deeper understanding of the content areas and a clearer vision of what cognitive psychologists are trying to accomplish. The remaining eight "content" chapters cover the central topics in cognitive psychology.

This book will be of value to a variety of audiences. Ideal for researchers in computer science, linguistics, philosophy, anthropology, and neuroscience who wish to acquaint themselves with cognitive psychology, it may also be used as a text for courses in cognitive science and cognitive psychology. Lay readers who wish to learn about the cognitive approach to scientific psychology will also find the volume useful.

1

INTRODUCTION

Imagine that an alien spaceship leaves Earth after a short visit and mistakenly abandons several robots as it heads for other exotic points in the universe. Several days later, a hiker discovers the robots, which are rushed to a research university. Over the next few months, scientists and engineers from around the world converge to dismantle one. To their amazement, they discover that the robot’s internal structure far exceeds current scientific understanding. Neither the robot’s components nor their organization resemble anything ever seen on Earth before. Moreover, the complexity of the robot’s intricately related components is overwhelming. After extensive examination, the scientific community concludes that decades, if not centuries, of research will be necessary to explain the robot’s physical structure.

Inability to understand the robots physically does not prevent understanding them in other ways. First, scientists could attempt to understand the robots behaviorally. How do the robots behave when placed in particular situations? When placed in sunlight, for example, the robots might “sunbathe” for several hours before actively pursuing goals. After being placed in darkness for long periods, the robots might seek sources of light. By discovering systematic relations between situations and the robots’ responses to them, scientists can learn much about the robots’ behavior.

Second, scientists could attempt to infer internal mechanisms in the robots that produce these behaviors. For example, sunbathing and light seeking suggest that the robots have a rechargeable solar battery. Although scientists might not be able to identify the physical system in the robots that captures, stores, and utilizes solar energy, their behavioral experiments may consistently show that the robots have such a system. Further behavioral research might identify the system’s functional parameters, such as the amount of light it stores, the wavelengths of light it absorbs, the rate at which it absorbs light, and so forth.

Behavioral research could also provide insights into other aspects of the robots, such as their perceptual abilities. Scientists could present the robots with different types of perceptual information and observe which types produce responses. For example, the robots might respond to visual, tactile, and kinesthetic information but not to auditory, olfactory, and gustatory information, leading scientists to conclude that the robots have perceptual systems for seeing, touching, and moving but not for hearing, smelling, and tasting. Further studies could assess specific aspects of the robots’ perceptual abilities. What wavelengths of light do the robots’ visual systems detect? Do the robots detect the same wavelengths of light as humans? Do the robots perceive different wavelengths as different colors? Do the robots detect other wavelengths not visible to humans, such as infrared light?

Behavioral research could assess whether the robots have a memory. Imagine that scientists place a robot in a dimly lit maze and that the robot eventually discovers a path out of the maze to sunlight, after initially running into several dead ends. Further imagine that scientists later place the robot in the same maze and that it exits the maze faster than it did the first time, avoiding the dead ends and immediately retracing the direct path to sunlight. Faster time to exit the maze the second time would indicate that the robot has memories of the paths that produced failure and success on the previous occasion. If the robot learns to be more efficient, it must store information from previous experiences and use it to improve later performance. Further studies could assess specific aspects of the robot’s memory. Does it remember visual information that marks the path? Or does it remember the movements that led to sunlight? If the robot remembers its movements, it should find its way even when visual information along the path changes. But if the robot uses visual information to find its way, then changes in visual information should worsen performance. By manipulating aspects of the task in this manner and observing the robot’s behavior, scientists can develop increasingly detailed accounts of the robot’s memory.

After much behavioral research, scientists might develop a fairly complete account of the robots’ abilities, including extensive accounts of the internal mechanisms that underlie these abilities, without having any knowledge of their physical bases. Knowledge of these internal mechanisms could nevertheless be extremely useful both scientifically and practically. It could explain the robots’ past behavior. It could predict the robots’ future behavior. It could allow people to use and interact with the robots optimally. It could allow people to keep the robots from getting into dangerous situations. It could guide explorations of the robots’ internal structure, perhaps suggesting that scientists look for physical systems that store solar energy, detect particular kinds of perceptual information, and store particular kinds of memories.

As you have no doubt been thinking, this scenario is similar to the problem psychologists face with respect to humans. Ultimately, psychologists seek to understand the physical structure of the human nervous system and how it operates to produce intelligent behavior. Although many important facts about the brain and its peripheral nervous system have been discovered, they constitute only a very small proportion of what there is to know. For example, scientists have learned much about the local behavior of neurons, but they have learned little about how large collections of neurons generate the more global brain activities that underlie intelligence. Scientists have learned little about how the brain produces memory, learning, language, and thought.

Yet, for about a hundred years, psychologists have studied human behavior scientifically, knowing little about its physical basis in the brain. In the process, they have collected bewildering amounts of data that describe behavior in its myriad aspects. The scientific study of psychology began in several different European movements. In the late 19th century, a group of scientists known as introspectionists attempted to describe the contents and composition of conscious experience systematically (Boring, 1953, 1957; Dellarosa, 1988; James, 1890; Titchener, 1910). At the same time, another group of scientists known as psychophysicists attempted to describe systematic relations between conscious experience and physical information in the environment (Atkinson, Herrnstein, Lindzey, & R.D. Luce, 1988a; S.S. Stevens, 1975). In the early 20th century, gestalt psychologists attempted to describe how people organize fields of information during perception, memory, and thought (Duncker, 1945; Koffka, 1935; Kohler, 1940; J.M. Mandler & G. Mandler, 1964; Wertheimer, 1945/1982). Concurrently, behaviorists attempted to describe how humans and other organisms adapt to their environments (Bower & Hilgard, 1981; Domjan & Burkhard, 1986; Honig & Staddon, 1977; Rachlin, 1970; Skinner, 1953; Watson, 1930). For a good review of the developments that led to cognitive psychology, see R. Lachman, J.L. Lachman, and Butterfield (1979, chaps. 1–3).

Since these early movements, scientific psychology has flourished. Today, most universities and colleges have sizable psychology departments that promote psychological research and train graduate students as future scientists. Extensive literatures on nearly every aspect of human behavior have evolved, as several generations of experimental psychologists have refined their empirical and theoretical methods. Developmental psychologists have described the behavioral development of humans from birth through old age (M. Cole & S.R. Cole, 1989; Heatherington & Parke, 1986; Sroufe & R.G. Cooper, 1988). Social psychologists have described human behavior in groups and various social settings (R. Brown, 1986; Shaver, 1987). Clinical psychologists have described abnormalities in human behavior (Goodwin & Guze, 1984; Kaplan & Sacuzzo, 1984). Perception psychologists have described how humans pick up physical information from the environment (Atkinson et al., 1988a; Boff, Kaufman, & Thomas, 1986; Bruce & P. Green, 1990; E.B. Goldstein, 1989; Sekular & Blake, 1990). And cognitive psychologists have described how humans store, manipulate, and use information (J.R. Anderson, 1990; Atkinson, Herrnstein, Lindzey, & R.D. Luce, 1988b; Ellis & R.R. Hunt, 1989; Glass & Holyoak, 1986; Lindsay & Norman, 1977; Neisser, 1967, 1976; Reed, 1988; Solso, 1988). By no means are these behavioral accounts complete, and a comprehensive theory of psychology remains elusive. Nevertheless, experimental psychologists have learned a tremendous amount about behavior in all of these areas, and their theoretical formulations are becoming increasingly sophisticated.

Most behavior remains poorly understood at the physical level. Scientists typically have little understanding of the underlying neuronal mechanisms that produce specific behaviors. Nevertheless, our understanding of the brain has also proceeded at a rapid rate. During the 20th century, physiological psychologists and other neuroscientists from biology, chemistry, physics, and medicine have made important advances in understanding the human nervous system (Adams & Victor, 1989; F.E. Bloom & Lazerson, 1988; N.R. Carlson, 1986; Heilman & Valenstein, 1985; Kolb & Whishaw, 1980; Lezak, 1983; Pinel, 1990). In certain areas, scientists have established close ties between behavioral phenomena and physical mechanisms. In perception, for example, the behavioral properties of sensory systems have become increasingly understood in neurological terms. Most scientists expect our understanding of the relations between the brain and behavior to grow extensively in coming years. Currently, however, relatively sparse relations exist between these two levels of theory.

1.1 INTERNAL CONSTRUCTS IN PSYCHOLOGY

As we saw in the robot example earlier, behavioral data often suggest internal constructs, independent of physical evidence for them. Similarly, in the absence of physical evidence, psychologists have proposed a wide variety of internal constructs to explain their behavioral observations of humans. Freudian theory provides a well-known example. On the basis of behavioral evidence observed clinically, Freud (1933) postulated the id, ego, and superego as internal mechanisms that produce behavioral normalities and abnormalities. Since then, scientific psychologists in virtually every area except behaviorism have proposed internal mechanisms to explain behavior.

In most cases, psychologists had no neurological evidence on which to base these mechanisms, but proposed them solely on the basis of behavioral evidence.

Proposing internal mechanisms in psychology has been quite controversial. In fact, this issue has dominated the history of modern psychology, dictating research and theory in fundamental ways. In the early years of this century, a group of psychologists known as behaviorists argued that internal constructs were unscientific. Because psychologists can not observe these constructs directly, behaviorists believed that including them in scientific explanations produced only unrigorous and shoddy science. Behaviorists proposed instead that psychological theories should exclusively address the physical stimuli that an organism encounters and its observable behavioral responses to them. Theoretical explanations should abandon internal constructs in favor of functional laws that express mathematical relations between observable stimuli and responses.

Typically, such laws in behaviorism have been the laws of classical and operant conditioning. In classical conditioning, a conditioned stimulus (e.g., a bell), associated with an unconditioned stimulus (e.g., food), comes to produce a conditioned response that is normally produced by the unconditioned stimulus (e.g., salivation). In operant conditioning, a reinforcer (e.g., food), associated with an arbitrary behavior (e.g., press the left response key), comes to control the behavior when a cue signals the reinforcer’s availability (e.g., a light is on). Traditionally, behaviorists characterized the organism as a tabula rasa or blank slate, whose behavioral habits and skills primarily reflected the particular environmental events that occurred in conjunction with behavior. Thus, behaviorists argued that a complete learning history of an organism’s behavioral responses to environmental stimuli should, in principle, be sufficient to explain its behavior (Bower & Hilgard, 1981; Domjan & Burkhard, 1986; Honig & Staddon, 1977; Rachlin, 1970; Skinner, 1953).

Behaviorism dominated scientific psychology in the United States from about 1910 to 1950. During this period, postulating internal constructs was taboo: Research did not address the presence of internal constructs, and theories rarely included them. The behaviorist edict to eschew internal constructs has few parallels in the history of science. Researchers across the sciences have nearly always proposed internal constructs that they could not observe directly. In the early history of genetics, for example, no one had ever observed a gene. Yet, this construct produced a revolution in biology, because it greatly improved biologists’ ability to account for data, to generate new research, and to develop useful applications. The direct observation of genes years later was not surprising, given the tremendous amount of indirect support that the construct had accumulated. Similarly, physicists have frequently proposed theoretical constructs for non-observable entities. Although no one has ever observed quarks and other subatomic particles, these theoretical entities perform important scientific functions. The presence of unobservable theoretical constructs is certainly not unique to psychology. Unobservable constructs have a venerable history of success across the sciences (Hacking, 1983; G. Maxwell, 1964; McMullin, 1978; Toulmin, 1961).

Although the stature of behaviorism has declined considerably over the last 40 years, it nevertheless remains a central and important part of psychology. Classical and operant conditioning have been observed in a wide range of species and capture fundamental aspects of how organisms adapt to their environments. Moreover, behaviorists developed experimental design and control to a high art, and the quality of research in virtually every other area of experimental psychology has benefited from these advances (11.2). The sense of rigor and skepticism that behaviorists brought to experimental psychology has made it a stronger science throughout.

Three developments around the middle of this century greatly reduced behaviorism’s influence on psychology and led to what many have termed “the cognitive revolution.” One important development occurred in linguistics. When the eminent behaviorist, B.F. Skinner (1957), tried to explain language learning from a behaviorist perspective, a young linguist, Noam Chomsky (1959), wrote what was considered a devastating critique. Chomsky’s arguments convinced many that a theory of language learning must include internal constructs. A theory that only considers the observable stimuli and responses in linguistic interaction is not sufficiently powerful to account for the structural properties of human utterances.

Chomsky proposed alternatively that the systematic patterns in human language primarily reflect the presence of an internal grammar. To explain language behavior, researchers must study this internal grammar empirically and include accounts of it in theory. Even though Chomsky had no physical evidence for such a grammar in the human brain, he argued that behavioral data provided a compelling argument for it. This proposal was heresy from the behaviorist perspective and was received as such. Many psychologists, however, had become worried that the behaviorist framework was not yielding progress in understanding the higher cognitive functions of language, memory, and thought. Chomsky’s ideas not only provided a provocative theory of language, they also suggested that addressing internal mechanisms would be necessary to account for the higher cognitive functions. The formal elegance and rigor of Chomsky’s particular theory also convinced many that, in general, theories of psychology could have these qualities (Chomsky, 1957, 1965; Chomsky & G.A Miller, 1963; G.A. Miller & Chomsky, 1963).

Information theory constituted a second important development that contributed to the demise of behaviorism and the rise of cognitivism. During the 1940s, applied mathematicians developed information theory to model electronic information systems, such as sonar and radar (Shannon & Weaver, 1949). An information theory account of a system typically assumes that a transmitter sends a message through a channel to a receiver, with different information systems varying in the characteristics of these basic components. For example, transmitters vary in their rate of transmission, messages vary in their complexity, channels vary in their signal-to-noise ratio, and receivers vary in their rate of decoding. By combining such variables mathematically, information theorists described important aspects of information processing.

During World War II, many experimental psychologists assisted the military in optimizing the ability of human operators to interact with electronic information systems. These psychologists found it natural to view an operator’s ability to transmit and receive information in terms of information theory. They observed, for example, that human operators were limited in their rates of transmission and decoding, as well as in the complexity of the messages they could process. Moreover, an operator’s limitations in transmitting and receiving appeared to reflect innate constraints rather than insufficient learning ex...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• Preface
• 1 Introduction
• 2 Categorization
• 3 Representation
• 4 Control of Information Processing
• 5 Working Memory
• 6 Long-Term Memory
• 7 Knowledge in Memory
• 8 Language Structure
• 9 Language Processes
• 10 Thought
• 11 Approaches to Cognitive Psychology
• References
• Author Index
• Subject Index