Cognition: From Memory to Creativity

Chapter 1

Introduction to the Study of Cognition

Before you even arrived at your first class this morning, you had engaged in numerous cognitive acts: recognizing the sound of your alarm clock and the time depicted on its face, saying good morning to your roommate, and categorizing your cereal as a breakfast food. You also had to remember the day of the week so that you knew which classes to attend, you decided which clothes to wear, and you paid attention as you crossed the road to get to your first class. Perhaps you even engaged in some creative thinking as you doodled while waiting for class to start. These are all examples of the cognitive processes—the mental processes—at work. Cognition both allows us to operate in the real world, and makes life richer.

Humans are captivated by how the mind works, and this fascination makes its way into popular culture. Stories about cognitive functioning and the connection between the brain and the mind are in newspapers and on TV all the time. Films about memory—whether the loss of memory (Memento) or implanted memories (Total Recall, Inception)—have become top-grossing hits. Books about consciousness (Dennett, Consciousness Explained, 1991), intelligence (Herrnstein & Murray, The Bell Curve, 1994), language (Pinker, The Language Instinct, 1994), memory (Foer, Moonwalking with Einstein: The Art and Science of Remembering Everything, 2011), and the relation between talent, practice, and success (Gladwell, Outliers: The Story of Success, 2011) were bestsellers. Articles in popular magazines discuss insight in problem solving (Lehrer, The Eureka Hunt, 2012a) and creativity in business (Gladwell, Creation Myth, 2011). The appeal of the mind holds even for scientists: Since 2001, psychological topics related to cognition or neurocognition have made the cover story of Scientific American magazine numerous times. The discipline of cognitive psychology has historically encompassed the study of the cognitive or mental processes, and provides the research upon which so many popular films and bestselling books are based. However, more recently, there has been a broadening of research on cognition to include neuroscience, computer science, linguistics, and philosophy, which has spawned a new discipline: cognitive science.

While much of the research on cognitive processes takes place in laboratories, for the cognitive scientist, life itself is an experiment in cognition: Everywhere one looks, it is possible to see evidence of mental processes at work. Dr. Weisberg's daughter used to be a competitive ice-skater, and every day she would go for practice sessions. The ice would be full of skaters, practicing the jumps, spins, and other moves they would need for their competitive programs. The practice sessions were not purely athletic endeavors; we can dissect what is happening at a cognitive level as each skater practices on a crowded ice rink.

First, memory is involved (Chapters 2–4). The main task facing those skaters is to master their material, so that they remember the correct sequence of jumps, glides, spins, and twists in their programs. Sometimes during a competition a skater begins to move in an erratic way, losing synchronization with the music: The skater has temporarily forgotten the program. The pressure of competition often causes skaters to forget or misremember a sequence of movements that was remembered easily many times during practice.

A second cognitive task facing the ice-skaters involves visual and spatial processing (Chapter 5): Each skater has to know the boundaries of the skating rink and the spatial configuration of their routine within those boundaries. They must also recognize other skaters as people to be avoided and determine their own and others' speed and direction, to determine if any collisions are likely. Sometimes younger skaters run out of space and cannot perform a jump because they are too close to the wall. Such skaters are not able to accurately calculate the space available for the move they hoped to carry out. This occurs much more rarely with experienced skaters, indicating that those visual-processing skills have developed over years of practice. This is one example of the general importance of knowledge in cognitive functioning.

Third, attention is involved in our skaters' practicing (Chapter 6). To a spectator, the scene on the ice has a chaotic quality, as all those youngsters zoom this way and that, each seemingly concentrating only on improving his or her own skills. And yet there are very few collisions; the skaters are typically able to practice their routines while avoiding each other. This requires both selective attention—each skater pays attention to his or her own skating routine while ignoring the practice routines of others—and divided attention (i.e., multitasking). As each skater is attending to his or her own routine, he or she must determine where other skaters are headed, so as not to be in the same place at the same time as anyone else. While watching a group of skaters of mixed levels of expertise, one quickly sees that the inexperienced skaters have problems with the multitasking demands of the practice session; they cannot concentrate on practicing their programs while at the same time attending to and avoiding the other skaters. The more-experienced skaters, in contrast, are able to avoid collisions while at the same time working on a jump or spin. So one of the consequences of the development of skill is an increase in the ability to multitask. Another way to put this is to say that the knowledge of the experienced skaters is useful in dealing with the attentional demands of the practice session.

Additional cognitive skills can also be seen in the skaters' practice sessions. Sometimes, one hears a coach give instruction to a skater: Do you remember how crisply Jane does that tricky footwork at the end of her program? It would be good if you could move like that as you do yours. Presumably, the coach and the skater are able to communicate because both of them can recall Jane's appearance as she skates. They are able to use imagery (Chapter 7) to remember how Jane looked as she did her footwork. The coach can use the memory of how Jane looked as the basis for judging the quality of the skater's footwork, and the skater can use her memory of Jane's performance as the basis for her own attempt to do the footwork.

Other cognitive skills necessary for optimum ice-skating performance are the acquisition and use of concepts (Chapter 8) and language processing (Chapters 9 and 10). A coach may revise a routine by saying, I'd like you to insert a Biellmann spin here—it's a layback where you pull your free leg over your head from behind. This example makes it evident that language is an important vehicle through which we acquire concepts. The skater will recognize a layback and use the coach's elaboration to understand what must be added to produce a Biellmann spin. In so doing, our hypothetical skater has just acquired a new concept. Also, the skaters' coaches constantly monitor the skaters' performance on the ice. One may hear a coach call out, Keep that free leg up while the skater spins, and one sees an immediate change in the posture of the skater. The skater processes the coach's linguistic message and adjusts his or her movements accordingly.

Finally, sometimes a coach and skater will change the routine during the practice session. The coach might decide that something more is needed in the way of jumps, for example, or that the choreography needs refinement. Or the skater might ask for some addition to the program, perhaps to make it more challenging. In these examples, the coach or skater has made a decision under uncertainty concerning the structure of the program (Chapter 11). Neither the coach nor skater is certain that the proposed changes will be helpful, but they have weighed the available information and decided that it would be beneficial to make a change. When changing the program, the coach and skater have identified problems to solve (Chapter 12) and creative thinking plays a role in producing changes in the program (Chapter 13).

These examples are by no means extraordinary. Surely each of us could compile, from any randomly selected day, a long list of phenomena in which cognitive processes are centrally involved: seeing a friend today, and picking up the thread of a conversation begun yesterday; using directions acquired online to drive to a new restaurant; being impressed with the creativity of a new song produced by your favorite group. Cognitive processes are at the core of everything we do.

In the past 30 years there has been an explosion in the study of human mental processes, and the momentum shows no signs of slowing down (Robins, Gosling, & Craik, 1999). New developments in the study of cognition have come from many disciplines, and are now best encompassed under the general term cognitive science. First, many areas which researchers had in the past studied only peripherally, if at all, such as imagery, language processing, and creative thinking, have come under investigation and have begun to yield their secrets. Second, in many areas, interdisciplinary cross-fertilization has occurred. Cognitive psychologists and neuroscientists regularly collaborate in the study of the relationship between the brain and cognitive processes, to determine the specific cognitive skills lost when a patient suffers a stroke or accident, or to discover, for example, which parts of the brain are most active when someone learns or recalls information. Those studies have increased our understanding of both normal and abnormal neurocognitive functioning. Linguists, cognitive psychologists, and computer scientists have made advances in our understanding of language processes. Philosophers of mind contribute to the study of cognition by clarifying the concepts and theoretical issues within cognitive psychology, including issues related to consciousness and the relation of mind and brain. Third, cognitive scientists have developed new ways of analyzing how we learn, organize information, and carry out cognitive tasks, most notably the computer-based information-processing perspective.

Why Do We Need to Study Cognition Scientifically?

A psychologist once remarked that being considered an expert in the field of psychology is difficult because since everyone has psychological states, everyone thinks that they know everything there is to know about psychology. When students are introduced to the scientific study of cognition, including much new terminology and numerous new concepts, they sometimes wonder why it is necessary to study cognition scientifically. Don't we all know how memory functions, since we each use our memory all the time? Don't we know about attention, from our own experiences attending to events in the environment? We all possess what we could call a commonsense cognitive psychology. Why do we need to learn all this jargon to describe and explain phenomena with which we are already familiar?

The scientific study of cognition is of value is because, contrary to what laypeople believe, they do not know very much about their own cognitive processes. Nisbett and Wilson (1977) found that humans often are extremely bad at giving accurate explanations for their own behavior. A recent bestseller, Blink, begins with an example of art-history experts knowing that a supposedly ancient Greek sculpture is a fake, but even the experts could not explain how or why they could detect the fraud (Gladwell, 2005). Thus, even experts in a field cannot discern the processes that underlie their cognitive abilities.

In many places in this book, we discuss research findings that are surprising or counterintuitive. The dangers of texting while driving are well known, and 39 states have banned the practice (Governors Highway Safety Association, n.d.). However, one example of a nearly universal lack of knowledge about cognitive processes is seen in recent legislation in many states banning the use of hand-held cell phones while driving. Such laws seem totally reasonable: Statistics have shown that using a cell phone while driving increases the risk of accidents, and most people assume that the dangerous aspect of cell-phone use is taking one hand off the steering wheel to hold and dial the phone. Legislators then enact laws banning hand-held cell phones. However, experimental studies of people driving in a simulated vehicle while talking on a cell phone have found that hands-free cell phones are just as dangerous as hand-held phones (Strayer, Drews, & Crouch, 2006). Driving while talking on a cell phone—hands-on or hands-free—is as dangerous as driving drunk (Strayer et al., 2006; these findings are discussed further in Chapter 6), and increases the risk of a collision fourfold (Redelmeier and Tibshirani, 1997). The problem with talking while on a cell phone is not that your hands are occupied—it is that your mind is.

Only 10 states in the United States have passed laws prohibiting cell phones while driving for all drivers, but not a single state bans hands-free phones (as of 2012; http://www.ghsa.org/html/stateinfo/laws/cellphone_laws.html). That means that no state has a policy that is consistent with the research findings (several additional states ban all cell phone use by those under 18 only). The legislators' lack of knowledge about and/or understanding of the cognitive issues underlying cell-phone use could have tragic consequences (Redelmeier & Tibshirani, 1997). This real-life example illustrates why we have to study cognition scientifically; although we each possess the cognitive processes and use them all the time, in actuality most of us do not know very much about the finer points of how they work.

Outline of Chapter 1

This chapter has several purposes. We first examine two uses of the term cognitive psychology, to set the stage for discussion of the development of modern cognitive science over the past 150 years, culminating in the recent ascendance of cognitive psychology as a major area within contemporary psychology. Many disciplines contributed to what has been called the Cognitive Revolution in the 1950s and 1960s, in which the study of mental processes supplanted behaviorism, which had been opposed to the study of consciousness and mental events. As part of our discussion of the cognitive revolution, we will consider the question of how cognitive scientists can study mental processes, which cannot be seen, and which may not be accessible to us at a conscious level.

As we have already noted, the modern study of cognition is made up of many different domains of academic inquiry, ranging from traditional research in psychology, to modern techniques for the study of brain and behavior, as well as theories and methods from areas outside psychology, such as linguistics and computer science. The final major portion of the chapter provides a more detailed introduction to how those disciplines have come together in the contemporary study of cognition.

Cognitive Psychology: A Subject Matter and a Point of View

The term cognitive psychology has two uses: It describes a subject matter, and it also describes a point of view or philosophy concerning how one studies that subject matter. The subject matter of cognitive psychology is the mental processes. These include memory; perceptual processes, such as pattern recognition (e.g., recognition of objects, words, sounds, etc.), attention, and imagery; language, including comprehension and production, and related phenomena, such as conceptual knowledge; and the class of activities traditionally called thinking, or the higher mental processes, including problem solving and creativity, and logical reasoning and decision making. Cognitive psychology as a point of view, or a scientific philosophy, refers to a set of beliefs concerning how those topics are to be studied (e.g., Neisser, 1967). According to the cognitive perspective, understanding behavior—such as remembering your mother's birthday, solving a math problem, or reading words on a page—requires that we analyze the mental processes that underlie that behavior. This perspective can be contrasted with behaviorism, which was based on the belief that behavior could be understood by determining the external stimulus conditions that brought it about, and not worrying about internal mental processes.

Studying Hidden Processes

Accepting the cognitive point of view raises a difficult question: How can one study cognitive or mental processes, which occur internally and therefore cannot be examined directly? Students often propose a simple method for studying internal processes: Have the person report on what he or she is thinking. That is, perhaps we can use subjective reports as the basis for studying hidden processes. This is a reasonable suggestion, but there is a basic difficulty with subjective reports. Suppose I tell you that right now I am imagining a dollar bill. How can you tell if my report is accurate? I may be lying about what I am thinking, or perhaps I am mistaken (and am really thinking about the candy bar I'd like to buy with that dollar). The fact that subjective reports cannot be verified—that is, the fact that we cannot tell whether they are accurate—means that they cannot be used as evidence for internal processes; other types of evidence must be found. Instead of subjective reports, we need objective data.

The question of whether and how one can study mental phenomena—which cannot be seen directly—had been a point of disagreement among psychologists for 100 years, until the advent of the cognitive revolution. We discuss this question in detail later in the chapter, after we place it in historical context. In our view, the cognitive scientist's study of hidden mental processes is no different than the activities carried out by scientists in many disciplines (e.g., biology, chemistry, physics) or, indeed, the activities carried out by ordinary folks in our understanding of events in the world. We deal with hidden processes all the time.

Psychology as a Science

Wundt and Introspection

The beginning of psychology as a science is traced to Wilhelm Wundt's establishment of the first psychological laboratory in 1879, in Germany (Boring, 1953). Until that point, the sorts of phenomena now studied by psychologists were investigated by researchers in physics and biology, as well as in philosophy. Students from all over the world came to study with Wundt, and many of those new psychologists returned to their home countries and established their own laboratories. Wundt and his followers could be considered the first cognitive psychologists, because they were interested in several mind-related topics, including consciousness. However, there were a number of important differences between Wundt's psychology and modern cognitive psychology.

First, the specific topics of Wundt's research differ from the topics of contemporary cognitive-psychology experiments. Wundt and his followers were interested in determining the basic elements or structure of conscious experiences, in the same way that chemists of that era were attempting to determine the basic elements of chemical compounds. While many modern cognitive scientists are also interested in the study of consciousness, the subject matter of modern cognitive science encompasses many other phenomena, such as those outlined above. Second, the methods of studying cognitive processes have also changed significantly over the 125-plus years since Wundt, his students, and colleagues began their work. In those days, it was believed that one could study consciousness by training observers to analyze their own experiences into their basic components and to report on them. This method was called introspection, which means looking inside. An example of a task used in introspectionist investigations of consciousness would be to present the names of two animals, say dog and cow, and ask the participant to judge which animal is larger in size, and then to provide an introspective report of what occurred between the presentation of the task and the production of a response.

Introspection required more than a casual report, however. The observer had to be trained to avoid the stimulus error, which was reporting the unanalyzed conscious experience in terms of commonsense, everyday language, rather than analyzing it into more basic components (Mandler & Mandler, 1964). For example, if, after making the judgment that a cow is larger than a dog, the observer reported, I imagined a cow standing next to a dog, and mentally compared their heights and lengths, that would be an example of the stimulus error. If the observer correctly engaged in introspection, he would convey more raw perceptual impressions, and might say something like: An image of a large nonmoving bulk and smaller one.… A feeling of movement.… An image of one end of the small bulk, and then the other.… A verbal image ‘the cow is bigger.’ …Production of the verbal response.

The Imageless Thought Controversy

When introspection was applied to the study of conscious experience, several difficulties arose. First, the results obtained in different laboratories were not consistent. Some investigators, such as Titchener, one of Wundt's earliest students, insisted that virtually all thought relied on imagery, based on the results of his introspection studies, while others reported that their studies showed that thinking could also be carried out without imagery (see discussion in Mandler & Mandler, 1964). Those conflicting findings raised questions about the usefulness of introspection, since seemingly identical investigations had produced opposite results. Whether imageless thought could occur became a major controversy, and resulted in many psychologists becoming dissatisfied both with the focus of psychology being the mind, and with the use of introspection as a scientific technique. One outcome of the imageless thought debate was the rise of a group of psychologists who wanted psychology to be a science of behavior—the behaviorists (Leahey, 1992).

Behaviorism and the Question of Consciousness

The strongest reaction against attempts to use introspection to analyze the structure of conscious experience came from John Watson (1913), the founder of American behaviorism. Watson wrote forcefully against the value of studying conscious experiences, because of the already-noted problems with verification of introspective reports. He proposed that psychology should follow the example of the established sciences, such as physics and chemistry, whose methods were only concerned with phenomena that were observable and directly measurable. When physicists studied the effects of gravity on falling objects, for example, they measured the height of the fall, weight of the objects, and time to fall. In his behaviorist manifesto—Psychology as the Behaviorist Views It—Watson (1913) advocated a similar perspective for psychology: psychology must…never use the terms consciousness, mental states, mind, content, introspectively verifiable, imagery and the like… (pp. 166–167) because the scientist cannot directly observe those things. Psychologists should study only observable events: environmental stimuli and behavior.

Watson promoted the now-familiar stimulus–response (S–R) approach to the analysis of behavior. He believed that there was a law-like relationship between environmental stimuli and behavioral responses, with every behavioral act being brought about by one measurable stimulus, and each stimulus producing only one response. Therefore, it should be possible to analyze behavior to such an exact degree that, for any response that occurred, the psychologist could know exactly what the stimulus had been; and if a given specific stimulus occurred, one could say exactly what the response would be. In Watson's view, the main task of psychology was to be able to predict and control behavior through presentation of environmental stimuli. One should not try to measure hypothesized internal psychological states, which might not even exist. Furthermore, Watson proposed that by strictly controlling the environment in which an organism grew up, he could determine the trajectory of a person's life:

Give me a dozen healthy infants, well-formed, and my own specified world to bring them up in and I'll guarantee to take any one at random and train him to become any type of specialist I might select—doctor, lawyer, artist, merchant-chief and, yes, even beggar-man and thief, regardless of his talents, penchants, tendencies, abilities, vocations, and race of his ancestors. (Watson, 1930, p. 104)

Thus, Watson adopted a radical stance to the study of psychology, claiming that there are no mental processes that play any causal role in a behavioral chain.

The second major advocate of behaviorism was B. F. Skinner, who championed what is known as operant conditioning. Based on the ideas of Thorndike (1898, 1911), Skinner proposed that the consequences of behaviors—reinforcements and punishments—would determine whether those behaviors increased in frequency and intensity, or whether they decreased (Skinner, 1938). If a behavior was reinforced, it would become more likely to happen in the future; if punished, less likely. Like his predecessor Watson, Skinner rejected mentalistic explanations of behavior as unscientific. Skinner's principles of operant conditioning were derived from maze running and key pressing in animals. The book Verbal Behavior (1957) marked his attempt to apply conditioning principles to complex human behaviors, such as the development of language in a child. In Skinner's view, children acquire a language by mimicking what they have heard and by being reinforced for their utterances (e.g., by delighted parents or by more quickly receiving what they want). Thus, language learning is brought about by the same learning mechanisms that are evident in lower-level animals; there is no need in a scientific theory for mentalistic or cognitive explanations of any skill or behavior. As we shall see, the inability of the behavioristic framework to account for complex phenomena, such as language, problem solving, and creativity, would eventually lead to the paradigm's loss of favor (a paradigm is a theoretical framework that helps to guide research within a topic area).

Toward a New Cognitive Psychology

The development of behaviorism resulted in reduced interest in the study of cognition in the first half of the 20th century, particularly in America. However, even at this time there was still interest in cognitive processes among some psychologists and philosophers. As one example, William James, a philosopher with interests in the study of behavior, presented discussions of complex human psychological phenomena. Second, there were a number of centers of psychological research in Europe in which the full-fledged study of human cognitive processes went on. Thus, this work was available in books and journal articles when other psychologists began to become interested once again in cognition. Finally, developments in several areas outside of psychology, including linguistics and computer science, provided psychologists with new ways of analyzing complex psychological phenomena. Those new perspectives greatly stimulated the development of modern cognitive science.

Cognitive Stirrings in America

William James was an American philosopher who wrote a two-volume survey of psychology, Principles of Psychology (1890/1890), in which he addressed many issues that were to become important to modern cognitive psychologists (e.g., Estes, 1990). James provided detailed descriptions of his own phenomenological experiences, that is, his personal experiences of psychological phenomena. For example, he described the experience of selectively attending to some event or object at the expense of paying attention to others. Attention was the taking possession of the mind, in clear and vivid form, of one out of what seems several simultaneously possible objects or trains of thought.… It implies withdrawal from some thing in order to deal effectively with others (1890, pp. 403–404). James also described his phenomenological experiences of remembering, and presented descriptions of experiences, which led him to distinguish between primary and secondary memory (approximately corresponding to the distinction between short- and long-term memory that has been studied extensively by today's cognitive psychologists). Both these areas—attention and the question of the structure of memory—became foci of research in modern cognitive psychology.

It should be noted that James's use of phenomenological analysis is not the same as the introspection carried out by Wundt and his followers. James was not interested in analyzing his conscious experience into its component parts, but, rather, attempted to present a detailed and accurate description of the conscious experiences themselves. This work was important because James discussed complex cognitive phenomena, such as shifts in attention or remembering, not merely simple sensory experiences. James is also considered to be a functionalist, because his explanations often emphasized the purpose or function of psychological and mental phenomena (Leahey, 1992), and how they allow people to adapt to their environment: Man, whatever else he may be, is primarily a practical being, whose mind is given him to aid in adapting him to this world's life (James, 1898).

The Study of Cognition in Europe

A number of European investigators were engaged in research on topics within the realm of cognitive psychology, not only during the late 1800s, but even when behaviorism dominated American psychology during the first half of the 20th century.

Ebbinghaus and the Study of Memory

Hermann Ebbinghaus was a German psychologist who is credited with bringing scientific techniques to the study of memory. He insisted on using material that was not associated with any previously learned information, and thus devised nonsense syllables, meaningless consonant-vowel-consonant strings such as REZ and TOQ, to determine how many repetitions he needed to learn new lists of items, and how long he could retain the information after having learned it. He used the method of rote rehearsal—simply repeating the items again and again. With this method, he could objectively measure the amount of time needed to memorize a list. However, from a modern perspective, Ebbinghaus's analysis was lacking, as he did not make any inferences about the internal processes that accomplished remembering. Ebbinghaus also was the first researcher to systematically study forgetting. He retested his memory for the lists he had learned after 1, or 2, or 30 days. In this way, he was able to measure the amount of memory loss (or forgetting) as a function of time (Ebbinghaus, 1885/1885).

Ebbinghaus's work on memory had a great influence on the study of cognition many years after his death. Through the middle of the 20th century, a number of American psychologists who wanted to study human functioning without having to appeal to mental processes used Ebbinghaus's (1885/1885) research as their model, because of his rigorous scientific methods (see, for example, chapters in Cofer & Musgrave, 1961, 1963). Although Ebbinghaus's approach ignored the study of underlying mental processes, his work did demonstrate that one could study memory in the psychological laboratory. Ebbinghaus's research and that of those who followed him brought the study of memory to the attention of many experimental psychologists.

Donders's Subtractive Method

F. C. Donders, a researcher from Holland who was a contemporary of Wundt, developed techniques for mental chronometry—the measurement of the time to carry out basic operations within an act of cognition (Donders, 1868/1969/1868/1969). For example, Donders might seat a person in front of a light, and tell him to press a button whenever the light came on. Imagine that it takes, on average, 250 milliseconds (equal to ¼ sec.) for a person to detect the light and respond to it. The reaction time (RT) to the light is thus 250 milliseconds (msec). In another condition, the person would be told to press a button on the left if light A came on, but a button on the right if light B came on. This takes (hypothetically), on average, 400 msec. The second task requires detection of the light, discrimination of whether light A or light B has turned on, and the response. Let us say that the experimenter wants to determine the length of the discrimination stage in the second task. The subtractive method allows one to do that. The second task takes 400 msec; the first task takes 250 msec. The only difference between the tasks is discrimination, so that process must require the additional 150 msec. Thus, we have decomposed those tasks into their parts and measured the time needed to carry out one of them. (Note the similarity to Wundt's attempts to decompose conscious cognitive phenomena.)

The subtractive method provided a way of measuring mental processing that was based on objective measurement—that is, on the time needed to carry out various tasks. The subtractive method became important in modern cognitive psychology in the 1960s, when Sternberg (1966) used reaction time to measure how we recall information from memory. The logic of Donders's subtractive method has influenced the design of many cognitive psychology experiments, and RT is now a common measure used by many researchers. The subtraction method also plays an important role in neurocognitive research, as we shall see shortly.

Gestalt Psychology: Perception and Problem Solving

The Gestalt psychologists, who worked in Germany and then in the United States, mainly during the first half of the 20th century, carried out investigations of several areas of human cognition. They were interested in the study of perceptual situations in which the organization or form of the whole situation produced an experience that could not be anticipated from analysis of the elements or parts that made it up. The term Gestalt, German for form, has entered our ordinary vocabularies, as well as being a part of the technical vocabulary of psychology. An example of a situation of interest to the Gestalt psychologists is presented in Figure 1.1a: The perceptual experience of a triangle is accomplished by focusing on the organization, or Gestalt, of the elements (rather than the individual parts themselves). Thus, we impose an organization on the three Pac-Man–type figures by mentally filling in the lines between them.

Figure 1.1 Gestalt demonstration: These figures illustrate how the perceiver is involved in the interpretation of a stimulus. (a) Kanizsa triangle. (b) Reversible figure: Vase/faces.

The Gestalt psychologists also investigated reversible figures, such as the one shown in Figure 1.1b. When one studies such figures, it is common to see a sudden reversal, from a vase to two faces in profile (and back to a vase). Thus, a reversible figure is one stimulus that produces two responses. The existence of such figures disproves the behaviorists' belief, proposed by Watson (1913), that it would be possible to specify precisely a single response to any individual stimulus. Cognitive psychologists believe that the ambiguous faces/vase picture can produce two different responses because the person can cognitively analyze it in two different ways. Reversible figures are very simple illustrations of the necessity to analyze internal processes in order to understand observable behavior (e.g., the person first reports seeing a vase, then the two faces).

The Gestalt psychologists also carried out research on more complex human cognition, most notably problem solving and creative thinking (e.g., Duncker, 1945; Maier, 1930; Wertheimer, 1923, 1959). They believed that complex thought processes could not be broken down into simple elementary processes, and that the performance of lower animals (e.g., pigeons, rats) on simplified versions of problem solving tasks would not shed light on human cognitive abilities. The Gestalt psychologists also emphasized the method of collecting verbal protocols, where participants were instructed to verbalize their thoughts, providing a stream-of-consciousness verbalization as they solved problems. Verbal protocols were different from the reports obtained by Wundt and Titchener during introspection, since the participants were not trying to break their conscious experiences into basic elements. Research has demonstrated that protocols can provide a useful record of thought processes that can be verbalized (Ericsson & Simon, 1980) and used as a supplement to other means of assessing cognitive activity.

Bartlett's Analysis of Memory

Sir Frederick Bartlett (1932, 1958), an English psychologist, carried out a long series of investigations of memory during the first half of the 20th century. Bartlett proposed that remembering information depends on more than the passive stamping in of the information in the person's memory. He suggested, instead, that people are active participants in cognitive processing and that they use their knowledge to interpret and remember information. Bartlett's view thus contrasted with Ebbinghaus's (1885/1964) adherence to rote rehearsal to memorize meaningless nonsense syllables. Bartlett theorized that much of what people remember consists of their interpretations of the material, rather than the material itself, and thus they actively construct their memory. He demonstrated that the person's interpretation of the material that is to be recalled plays a crucial role in remembering. When memorizing a verbal passage, for example, we most likely use a schema—a cognitive structure that helps us organize and make sense of the new material. Please perform the sample experiment in Box 1.1 before reading further.

Box 1.1: Bransford and Johnson (1972) Passage

Have a paper and pencil ready before going further. Please read the following passage once, at normal speed, and then try to recall it on paper.

The procedure is actually quite simple. First you arrange things into different groups depending on their makeup. Of course, one pile may be sufficient depending on how much there is to do. If you have to go somewhere else due to lack of facilities, that is the next step; otherwise you are pretty well set. It is better to do too few things at once than too many. Remember mistakes can be expensive. At first the whole procedure will seem quite complicated. Soon, however, it will become just another fact of life.

Write the passage as well as you can from memory. After you do that, go on reading the text.

Now read the passage again, with the hint that the passage is about washing clothes, and again try to write as much as you can from memory.

The passage in Box 1.1 is so designed that it is almost impossible to understand or to recall fully without being told what it is about. Bransford and Johnson (1972) asked participants to study the washing clothes passage and others like it for later recall. Half the people were given the title Washing Clothes, to make the passage easier to understand. Presentation of the title before the passage made it much more comprehensible and also increased recall greatly. Providing the title after the passage did not facilitate either comprehension or recall. These results indicate that recall of the passage depended on activation of a schema (e.g., your knowledge about washing clothes), which improved comprehension of the passage as it was being read. Providing a framework for comprehension after the fact did not help. Bransford and Johnson's results are strongly supportive of Bartlett's view that one's interpretation of events plays a crucial role in memory for those events. Bartlett's emphasis on active processing of information became very important in psychologists' explanations of many cognitive phenomena, as we will see in Chapter 3.

Toward a New Cognitive Psychology: Summary

We have just seen that there were researchers—James, Donders, the Gestalt psychologists, Bartlett—interested in the study of cognition even when most American psychologists accepted the behaviorist viewpoint. The work of those individuals provided a foundation for the development of the cognitive revolution in psychology around the middle of the 20th century. Critical developments in psychology, linguistics, and computer science helped propel the study of cognitive processes to the forefront, as we will see in the next section.

The Cognitive Revolution

Despite European openness to the study of mental structures and processes, resistance toward mentalistic explanations of psychological phenomena remained high among many psychologists in the United States until the middle of the 20th century. At that time, dissatisfaction with strict behaviorism among psychologists, as well as developments in several disciplines outside psychology—most notably linguistics and computer science (Kendler, 1987)—culminated in a new orientation to the study of psychology, which Simon (1980) referred to as a revolution, now known as the cognitive revolution.

Revolt Against Behaviorism

Many psychologists interested in understanding complex behaviors, such as language, memory, and problem solving, began to view strict behaviorism as inadequate to the task. Even from within the ranks of behaviorists, some suggested that mentalistic concepts and analyses of what was taking place internal to the learner might be critical in explanations of human (and even animal) behavior. For example, E. C. Tolman (1932) studied the behavior of rats in a maze similar to the one depicted in Figure 1.2. He first allowed the rats to explore the maze, then he put them in the Start Box, and reinforced them with food for running down the straight pathway (Path 1) to the Goal Box. Once they had learned that task, he blocked the pathway at Block Point A. The rats typically then avoided Block A by using the triangular Path 2. However, if Block Point B is used, only Path 3 will get the rats to the Goal Box (and to food). When they encountered Block B, the rats would run back and take Path 3 (ignoring Path 2). Behaviorism predicted that the rats' responses should be dependent only on the strength of pathway/reinforcement contingencies, but they were not. The rats chose pathways based on the most efficient way to the goal box. Tolman proposed that the only way to explain those results was to hypothesize that the rats had developed a cognitive map during their exploration of the maze, and that the internal map was being used to guide their behavior.

Figure 1.2 Tolman maze.

Given what we now know about how animals efficiently forage in the wild for food (MacArthur and Pianka, 1966) and find their way home after traveling long distances (Gould & Gould, 2012), the notion of a cognitive map does not seem revolutionary. However, during Tolman's time, it was a significant change in psychological theorizing, because it postulated a mental representation—an internal version of the environment—that played a critical role in an organism's response to a stimulus (such as the maze). Tolman and others working within a behaviorist framework (e.g., Woodworth, 1938) helped turn the tide to what became known as S-O-R psychology (stimulus-organism-response). In this cognitive elaboration of behaviorism, any law-like connections between environmental stimuli and behavioral responses are assumed to be filtered through the knowledge and habits of the organism. The door was opened in the United States for cognitive processing to become part of scientific inquiry.

Using Behavior to Infer Inner States

Behaviorism had what one could call two negative effects on early psychology: (1) a rejection of the study of consciousness and related mental phenomena, which had been the subject matter of interest to many of the founders of psychology; and (2) a shift away from the study of complex human activities, such as thinking, problem solving, and decision making. However, behaviorism also made a positive contribution to psychology through its emphasis on tying all concepts to observable behaviors. Although cognitive psychology considers the study of mental processes to be the key focus of scientific inquiry, it does use the behavior of people and animals as the basis for theorizing about cognitive processes. For example, if we are interested in studying whether a person bilingual in French and English has equal facility with both languages, we could ask her to read aloud identical words in either French or English, and measure her response time. If she was faster reading the French words, we could conclude that she was more facile with that language. Thus, her behavior in the word-reading task would shed light on her underlying cognitive processes and skills. We could also measure brain activity in areas known to be connected to language and word recognition, to determine if there is a difference in neural activity as our person reads French versus English words.

Most of the research detailed in this book uses behavioral responses as the basis for making inferences about underlying cognitive processes, and to develop theoretical models of cognitive abilities. A quick perusal of the graphs depicted within the book also shows that reaction time (RT) is a popular way to analyze the time sequence, and processing constraints, of cognitive processes.

Chomsky and Linguistics

The linguist Noam Chomsky became a major voice in the early years of the cognitive revolution. His work was important in two ways. First, he published a strongly negative review of Skinner's book Verbal Behavior (1957), in which Skinner had attempted to explain language acquisition using classical and operant conditioning principles (Chomsky, 1959). Within that review, Chomsky managed to present a negative critique of the whole enterprise of attempting to explain complex human behavior on the basis of conditioning. Second, he offered a view of the structure of language and of language development that had important implications for psychology. Chomsky argued that, contrary to Skinner, language development is based on innate language-specific principles, not on simple learning mechanisms that apply equally to a rat's learning how to run through a maze and a pigeon's learning how to peck a key to receive food. Furthermore, he considered those innate language abilities to be human-specific, since we are the only known species that possesses language.

In Chomsky's view, the most impressive aspect of human language was its creativity: We almost never say exactly the same sentence twice. Furthermore, we have no trouble producing new utterances as needed, and understanding the new sentences that others produce. As an example of the novelty in language, consider the following sentence: George Washington was the King of England. It is highly unlikely that anyone has ever spoken or written that sentence before, but it is a perfectly grammatical sentence, and you were able to read it and understand it (while recognizing it as false). Chomsky proposed that the ability to produce an unlimited number of grammatical sentences is due to human language being a rule-governed system; that is, we all learned a system of rules when we learned to talk. Furthermore, in discussing how children learn to speak, Chomsky concluded that the linguistic input to the developing child—the language that the child hears around her—is too simple to account for the complexity of the young child's speech. That is, an average 3-year-old's sentence-production abilities are too advanced to be accounted for purely on the basis of what she has heard. This idea has been called the argument of the Poverty of the Input (Pinker, 1994): The linguistic input to the developing child is, by itself, too poor to enable language to develop as richly as it does. If the environmental input is not sufficient to support the normal language development of children, then, Chomsky argued, nature must provide a language-specific set of guidelines or rules that are activated by hearing speech, and these help a child organize incoming speech.

Chomsky's theorizing had several important effects on psychology. If human language was rule-governed, and could not be understood in terms of conditioning, then many psychologists were led to question the basic assumptions concerning behavioristic explanations of other complex behaviors. In addition, since Chomsky's innate language-learning principles are specific to language, he introduced the concept of cognitive modularity—the rules for learning and carrying out one skill (in this case, language) are located in a specific module, or processing unit, separate from the rules for other skills (such as vision or problem solving; see also Fodor, 1983). To illustrate, learning the grammatical rules of a language does not help someone to develop the mathematical competence necessary to balance a checkbook.

The question of whether cognition depends on specific modules versus general processing mechanisms is one that has stimulated debate in modern cognitive science, and we will have occasion to address it numerous times in later chapters. Chomsky's concept of modularity and his view of language in particular (and cognition in general) as a rule-based system provided much of the basis of the information-processing model of human cognition.

The Computer Metaphor and Information-Processing Models

The invention of computers, and their increasing availability in academic circles in the 1950s and 1960s, had the indirect effect of changing many psychologists' beliefs about mental processes and whether they could be studied objectively. Computers were able to carry out tasks, such as arithmetic and problem solving, that require cognition (i.e., mental processes) when humans carry them out. Computer scientists used terms like information processing to describe what happens when a computer carries out a program. Programs specify the series of internal states that a computer undergoes between presentation of some input data and production of some output as it carries out some task, such as adding two numbers. A diagram of the processing components of a typical computer is presented in Figure 1.3. Let us say that the computer is running a program that gives you the phone number of a person whose name you enter using the keyboard. The computer's memory contains a database, listing people by name and phone number. When you type in a name, it is placed in the central processor, where the program uses it as input. The program takes the name and attempts to match it to items within the database in memory. If there is a successful match, the phone number is transferred from memory to the central processor, where it is then produced as output—either as a number on the screen, as printed output, or as spoken output.

Figure 1.3 Information-processing components of a computer.

Thus, in carrying out this simple task, the computer goes through a series of internal states (e.g., taking in information, scanning memory, transferring information from memory to the central processor, etc.). In theory, at least, those internal states could be specified, as long as you knew the design of the program that was running and the data that the program was using.

These facts led many to believe that we could study and interpret human mental processes as analogous to a computer carrying out a task by running through a program. Researchers in computer science, most notably Alan Newell and Herbert Simon (e.g., Newell, Shaw, & Simon, 1958; Newell & Simon, 1972), proposed that humans could also be described as information-processing devices, similar in some important respects to computers, although obviously made of different sorts of stuff. It is not that cognitive scientists believe that we have silicon chips in our heads; the analogy of the mind as a computer is at a functional level, or at the level of the software. Newell and Simon proposed that one should conceive of a human carrying out any task that involved cognition (i.e., thinking) as if he or she were a computer carrying out a program. If human mental processes are similar to the series of internal states produced as a computer carries out a program, then one should be able, in principle, to devise computer programs that mimic or simulate human thinking.

The concept of modularity, just introduced in the discussion of Chomsky's analysis of language as a cognitive module, is also illustrated in computers. Computers are not general, all-purpose machines. If you buy a new computer, you will be asked what types of software you want—word processing, a graphics package, statistical software, and so on. These will be loaded separately. Likewise, psychologists typically study cognitive skills in isolation from each other; it is assumed that the programmed rules for human mathematical computations are distinct from those for constructing sentences, or for imagining the face of a friend. The modularity seen in computer information-processing leads to the expectation that human cognition might also exhibit modularity. Although there have been challenges to the information-processing approach to cognition (such as parallel distributed processing models, which will be discussed later in this and other chapters), it has proved a useful framework for thinking about how people carry out various cognitive tasks.

Study of Cognition in Humans

Tolman's (1932) research on maze learning in rats, which led to the concept of a cognitive map, was an early attempt to study cognitive phenomena in lower organisms using objective methods. However, many psychologists interested in human cognition felt that Tolman's research was limited in its applicability to more complex cognitive processes that humans could carry out. To the new generation of cognitive psychologists (e.g., Miller, 1956; Miller, Galanter, & Pribram, 1960), human thinking was more than internal stimuli and responses. One early influential cognitive psychologist was Allan Paivio (e.g., 1971; 2006), a researcher interested in the effects of visual imagery on memory. An example of a study similar to those carried out by Paivio is given in Box 1.2; please carry out the demonstration before reading further.

Box 1.2: Demonstration of Memory for Word Pairs

Please get a pencil and paper before reading further. Below is a list of pairs of words. Each pair is followed by one of two words: REPEAT or IMAGE. If the word pair is followed by REPEAT, then repeat the pair five times to yourself, and then rate how hard it was to pronounce the pair, on a scale from 1 (very easy) to 5 (very hard). Write the rating in the space after the pair. If the word pair is followed by IMAGE, then take about five seconds to form an image of the words in the pair interacting, and then rate how hard it was to do that on a scale of 1 (easy) to 5 (hard). Write that rating in the blank. When you have finished, continue with the Test for Word Pairs.

Test for Word Pairs

Here are the first words from each pair in the top of the box. Without looking at the pairs, try to recall the second word and write it down. Then go back and check whether you recalled more of the words from the imagery pairs or the repetition pairs.

diamond –

sauce –

beggar –

factory –

marriage –

cattle –

money –

gem –

street –

hotel –

In a series of studies, Paivio (1971) demonstrated that people recalled information more easily when they used imagery as the basis for learning. Chances are that you, too, remembered more words in the imagery pairs than the repetition pairs in the demonstration in Box 1.2. In his research, Paivio first asked people to rate, on a 1–7 scale, how easily they could think of an image for the meaning of many concrete (e.g., bell) or abstract (e.g., independence) words (Paivio, Yuille, & Madigan, 1968). He then showed that words that had been ranked high in imageability (e.g., car) were recalled more easily than words that were low in concreteness (such as idea; for a review of many studies, see Paivio, 1971). Paivio proposed that participants could easily form images when they studied concrete words, thereby creating a dual code—both visual and verbal—for these words, and thus making them easier to recall. Paivio's research was important because it was an attempt to deal directly with cognitive processes and to use mentalistic concepts—in this case imagery—as a component part of the explanation of complex behavior.

The Cognitive Revolution: Summary

By the early 1960s, many changes had taken place in psychology. There were criticisms raised about the adequacy of S–R analyses of behavior, both from within psychology and from outside. There were also a number of researchers who had embarked on research programs directed toward the understanding of cognitive processes. Several of these programs originated in Europe (e.g., Bartlett, the Gestalt psychologists), but there was also interest in human cognition among U.S. psychologists (e.g., Paivio, 1971) and linguists (Chomsky, 1959, 1965). In addition, the advent of computers provided a concrete example of a physical system that carried out processes that resembled human cognition. This raised the analogy of the mind as a computer, and the possibility that humans and computers were similar at a functional level. These streams of research came together in the 1960s to form the new discipline of cognitive psychology.

The New Cognitive Psychology

Publication of the book Cognitive Psychology, by Ulric Neisser (1967), was evidence that the new cognitive viewpoint in psychology had become a dominant paradigm within psychological research. This book had several important effects, one of which was giving a name to the new developments. Neisser also used the computer metaphor to organize the presentation of material concerning human functioning. The organization of Neisser's book followed information as it worked its way into the organism, as outlined in Figure 1.4. According to his analysis, information passed through a series of stages, from perceptual processes to memory, from which it could be recalled when needed. Imagine the cognitive processes involved when we meet an old acquaintance, John, on the street. The first stage involves registering the parts or features of the stimulus, for example, the lines, angles, and curves of the stimulus, out of which a representation of John is constructed. The next stage might be recognizing that a visual object has been presented. The object would then be classified as a person, then as our friend John specifically, and finally stored in memory as a recent encounter with John. The information could then be used as needed, for example, as the basis for affirmative answer if someone asked, Did you see John today?

Figure 1.4 Information-processing model; bottom-up processing.

The coverage in Neisser's book was most heavily concentrated at the perceptual end of the information-processing sequence, such as pattern recognition and attention. Less than 10% of Neisser's book was concerned with the higher mental processes, (e.g., memory, concept formation, and problem solving). Neisser acknowledged this lack of balance, and commented that at the time not very much was known about the higher processes. This book, on the other hand, will have about two-thirds of its pages devoted to the higher processes. This is because we have learned much about these topics over the years since Neisser's pioneering book was published. Another change from Neisser's book to the present one is an increase in emphasis on the role of knowledge in even the lower-level or perceptual processes. Neisser did discuss the role of knowledge in memory and perception, but we will place much more emphasis on the role of knowledge in all our cognitive functioning. That is why our book begins with memory, as information we have already stored in memory influences even lower-level processes such as perception.

An example of an information-processing model of human cognition that follows a serial path—outlining a series of stages in processing, as discussed by Neisser—is shown in Figure 1.5. The model deals with the visual processing of letters. The first stage of processing involves analysis of the letter into its important parts, or features, which are then used to identify the letter. When recognizing an A, for example, the features activated would be two slanted lines (/ and \) adjoined at the top, and a horizontal line (−). Once the features of the input have been identified, the bundle of features would first be identified as a physical object, and then identified as a specific letter. Finally, the results of this analysis can be stored in memory (I saw an A).

Figure 1.5 Letter-recognition model.

Posner, Boies, Eichelman, and Taylor (1969) conducted research using Donders's subtractive method to specify the stages that took place as people processed linguistic symbols, such as letters. The basic procedure involved presentation of pairs of upper- and/or lowercase letters to the participant, such as AA, aa, aA, or AB. In one condition, the participants were instructed to press one of two buttons, corresponding to whether two letters were physically identical (such as AA, aa) or physically different (Aa or AB). In a second condition, the judgment was based on whether the letters had the same name (AA, aa, and Aa) or different names (AB). People were faster to judge that physically identical letters are the same (AA, aa; average response time = 859 msec) than to judge that upper- and lowercase letters are (Aa; average response time = 955 msec).

The results of the Posner et al. (1969) study (as depicted in Figure 1.6) indicate that the first stage of letter recognition is based on the visual form of the letters; the second stage involves recall of the letter name from memory. On the basis of these results, Posner and his colleagues concluded that information about letters is processed through several stages, with each stage becoming more removed from the physical stimulus. Each stage thus utilized a different code to make the judgments: first a visual code (where AA has the advantage over Aa), then a name code (where AA and Aa judgments are equal). Thus, by carefully controlling the properties of the stimuli and the judgment that the participants were asked to