The Story of Psychology

by Morton Hunt

  But the advocates of propositional representation have equally good grounds for their view. They contend that images cannot convey such relationships as “has,” “causes,” and “rhymes with,” or represent categories and abstract concepts. Herbert Simon and William Chase found that chess masters could reproduce an entire board position after viewing it for just a few seconds—but only if it was a true board position in an actual game. If it was a random arrangement of the pieces, they could not. The implication: The masters’ memory was not visual but was based on the geometrical relations—the attack and defense move potentials—of the pieces. Finally, information in computer programs is stored in propositional form, and if computability is a good model of cognition, it stands to reason that the mind stores information similarly.49

  Quite reasonably, a third position has been taken for some time by many theorists: There are several types of representation—propositions, mental models, and images, each encoding information at a different level of abstraction. Finally, a fourth position is that different types of mental imagery use different brain networks: imagery involving spatial relations (as in imaginary rotation of an object) relies on a network in the parietal lobes, while imagery involving high-resolution shapes relies on a network in the occipital lobes.50 (Even if true, that position doesn’t help us understand how the masses of neuronal impulses arriving by either network get to be “seen” by us as mental images.)

  Schemas: In 1932 the English psychologist Frederic Bartlett told subjects folk tales from non-Western sources and then asked them to recall the tales. They remembered the stories inaccurately, inadvertently filling in gaps, modifying events so as to provide reasons for what happened, and omitting details that made no sense to Western minds. Bartlett concluded that “remembering is not the re-excitation of innumerable fixed, lifeless, and fragmentary traces” but “an imaginative reconstruction, or construction” based on our own organized mass of experiences. He called that organized mass “schemata”; others prefer the anglicized version “schemas.”51

  Bartlett’s idea has been revived and elaborated in recent years. Schemas—also known as “frames” and “scripts”—are now thought of as packages of integrated information on various topics, retained in memory, on which we rely to interpret the allusive and fragmentary information that ordinary conversation—and even most narrative writing—consists of. In 1978, David Rumelhart, then of the University of California in San Diego, reported on experiments in which he read stories, sentence by sentence, to his subjects to see how and when they formed a clear idea of what the stories were about. When, for instance, they heard this: “I was brought into a large white room, and my eyes began to blink because the bright light hurt them,” some 80 percent at once assumed they were hearing either a hospital or interrogation scene, and supplied a wealth of information to the few words they had heard. If the next sentence or two contradicted this supposition, they changed it and filled out the story anew from a different schema.52

  Much other recent work on schemas and a related type of information package known as a “script” has firmly established that it is by drawing on our expectations and organized knowledge structures that we understand and interpret—or often misinterpret—what we hear, read, and experience. Memory, in sum, is not only an information register to be consulted as needed but a program that directs our thinking.53

  Forgetting: Many studies have explored why we forget some things and not others, and what can be done to improve memory, particularly in the elderly, most of whom undergo some degree of nonpathological memory impairment. (Normal age-related memory problems can often be ameliorated by mnemonic and other training. There is also the possibility that not far down the road a pharmacological treatment, rebalancing altered neurotransmitter output, will be found.)54

  Much interesting research has concerned not the total loss of particular memories but the forgetting of important details and their replacement by new material. Our legal system relies heavily on the assumption that if we remember an event at all, we remember it as it was. The courts and many psychotherapists also believe that forgotten material retrieved through hypnosis is a true record of what happened. But psychotherapists have also long had evidence that we alter memories to make them more acceptable to the ego, and for many years Elizabeth Loftus has been amassing experimental evidence showing that the memory of a startling or traumatic event can be distorted by the trauma itself, that the memory of an event can be slanted by a skillful attorney’s loaded questions, that we graft new information on to the memory of an event as time goes by and have no way of retrieving the original version, and that hypnosis sometimes retrieves deeply buried memories—and sometimes manufactured ones.55 (In 2005, Loftus won a $200,000 Gravemeyer Award for her research on false recollections.)

  Nearly all of us, however, feel sure that certain events are indelibly and accurately burned into our memories. Recollections of such experiences as first hearing about the assassination of President Kennedy or the explosion of the Challenger space shuttle are known to psychologists as “flashbulb memories,” because they are vivid tableaus that seem utterly unforgettable. Ulric Neisser and an associate, Nicole Harsch, seized an extraordinary opportunity to study the phenomenon. The day after the Challenger disaster (which occurred on January 28, 1986), they asked a large sample of undergraduates to record, in detail, how they had heard the news of the explosion. Two and a half years later, those respondents who could still be reached were asked to answer a questionnaire about the event and six months later were interviewed.

  Over a third of the students’ recollections about time, place, who told them, and so on were dead wrong, as judged by their 1986 reports, and nearly another quarter were partly wrong. When the subjects were shown their original statements, Harsch and Neisser reported, “Many were quite upset by the discrepancies with their present memories… Interestingly, many continued to prefer their current 1989 recall to the version in the original 1986 record.” Where did the errors come from? Harsch and Neisser call them “narrative reconstructions” of the type described by Bartlett in 1932.56

  Sometimes, even in the fast-developing cognitive revolution, plus ça change…

  Language

  Scientists infer natural laws from specimens, events, natural phenomena, and experimental findings of one sort or another. The comparable raw materials for cognitive scientists are thoughts, but the neural discharges, or brainwaves, that indicate thought, though they can be seen by EEG (electroencephalography), reveal nothing of the content. Gestures, facial expressions, mathematical or artistic symbols, and demonstrations (as in sports training) convey thoughts, but within a very narrow range. The principal form in which thinking can be observed is language, which has accordingly been called “the window to the mind.”

  Or, one might also say, the spoor of thought, since language not only conveys thoughts but in its structure bears traces of how our minds work. The study of thought processes as revealed by these traces is the province of psycholinguistics. (Linguistics, an older discipline, deals with the characteristics of language itself.)

  An example of such a trace: Small children tend to treat irregular verbs and nouns as if they were regular (“Doggy runned away,” “Dat baby has two toofs”). But they have not heard adults make such mistakes, and therefore are not imitating them. The errors, psycholinguists say, show that children recognize such regularities in adult speech as adding “ed” to make a simple past tense, “s” or “es” to make a plural noun, and take these to be applicable to all verbs and nouns (the tendency is called “overregularization”)—evidence that the human mind spontaneously generalizes from examples, then applies the generalization to new cases. Cognitive psychologists long had two different hypotheses about how this takes place: one, that regular past tense forms are generated by a rule and irregular ones retrieved from memory; and two, that both forms are generated by a single system and differ only in their reliance on sound and semantics. An fMRI-based study
has just settled the issue: Brain area activation is the same for both regular and irregular verbs, thus confirming the single-system hypothesis.57

  The acquisition of verb tenses is only one of a number of marks left by the thinking process that psycholinguists have found in language. They are not peculiar to English; analogous phenomena can be found in every language and seem to be characteristic of human thought. Human languages appear to be governed by the same universal principles and constraints.

  The universality does not, of course, involve grammar and vocabulary; in those respects English, Swahili, and Basque, for instance, have nothing in common. Yet children who grow up hearing each of those languages recognize, without being taught, the difference between singular and plural forms of a noun, the verb forms that denote present and past, and so on, and construct for themselves the rules governing their language. Similarly, they intuit the basic rules governing word order and are able to construct simple declarative sentences made up of words in the proper sequence. No child of English-speaking parents ever says, “Milk more some want I,” nor does a child of parents who speak any other language get its basic word order wrong.

  Psychology had little relationship to linguistics before midcentury, but in the dawn of the cognitive revolution some cognitive psychologists and linguists dimly saw that new developments in each of their disciplines called for explanations by the other one. For instance, certain new theories of the linguists about how grammar works implied that the mind, when dealing with concepts, performs complex manipulations that are not accounted for by behaviorist psychology. In 1953, a number of psychologists and linguists held a conference at Cornell University, discussed their areas of common interest, and adopted the term “psycholinguistics” to designate the study of the psychology of language.

  Psycholinguistics was still a little-known discipline when, four years later, a twenty-nine-year-old member of the Harvard Society of Fellows published a monograph that thrust the subject into the limelight. The theory proposed in that monograph has been called one of the two most important developments in psychology in that era (the other being artificial intelligence).58 Its author was Noam Chomsky, some of whose ideas we heard about earlier.

  Chomsky, a shaggy-haired, bespectacled, rumpled genius—the Central Casting stereotype of an intellectual—very nearly did not become a psycholinguist.59 He grew up during the Depression years in the radical Jewish community of New York; his father, however, was a distinguished Hebrew scholar, and even as a youth Chomsky picked up some knowledge of the structure of the Semitic languages and some idea of what linguistics was about. These two themes, radical politics and language, have dominated his life ever since. His work in linguistics, the basis of his renown in cognitive psychology, came about when he met Zellig Harris, a professor of linguistics at the University of Pennsylvania.

  Harris, who got him excited about linguistics, was trying to develop a system based on behaviorist principles—a system that could account for language patterns without reference to meaning. But his scheme was flawed, and for some years Chomsky labored diligently in the effort to make it work. When he could not, he abandoned Harris’s theory and within two years had developed his own. It is ironic that Chomsky is a leftist; the central thesis of his theory, advanced in his monograph Syntactic Structures, is that certain aspects of linguistic knowledge and ability are innate, not learned, a doctrine that leftists, liberals, and behaviorist-trained psychologists considered mentalistic and reactionary.

  The child, Chomsky maintained, makes sense of heard speech and acquires language not by means of the grammar of the language (“surface grammar,” in his terminology) but by an inherent ability to recognize deep-lying syntactical relationships among the component phrases of the heard sentence—what he calls the “deep structure” of the underlying connections. As evidence he points to the ease with which children understand what is meant when one form of sentence is transformed into another—when, for example, a declarative statement is reworded as a question—and make such transformations themselves. If surface grammar were what children relied on, they would extract incorrect rules for transforming sentences. From instances like these:

  The man is tall.

  Is the man tall?

  they would derive the rule: Start at the beginning, move on to the first appearance of “is” or another verb, and shift that verb to the front. But the rule is too simple; it fails as soon as one confronts a sentence like:

  The man who is tall is in the room—

  where the rule would lead them to say

  Is the man who tall is in the room?

  But children never make that mistake. They make trivial ones like “toofs” but not substantive ones; they sense the relationships among the elements of the thought—its syntactic constituents or “phrase structures.” It is by means of this knowledge of “universal grammar” that children make sense of what they hear and effortlessly construct correct sentences they have never heard.

  When and how do they come by a knowledge of universal grammar and deep structure? Chomsky’s answer perfectly expresses the revolution against the behaviorist doctrine that the newborn’s mind is a tabula rasa. Somewhere in the brain, he maintains, is a specialized neural structure—he calls it the Language Acquisition Device, or L.A.D.—that is genetically wired to recognize the ways in which the things and actions represented by noun phrases and verb phrases are related to one another as agent, action, and object.

  Chomsky and the many psycholinguists who adopted his view or developed their own versions of it set out in new form the ancient question, banned during the behaviorist era, of whether knowledge exists in the mind before experience. Their answer: While language itself is learned, the brain is so constructed that children spontaneously extract the rules of speech from what they hear without being taught those rules and, making only minor errors, use them when constructing sentences.

  Though usually serious and intense, Chomsky is certainly capable of wit. To illustrate the deep relationships among the components of a sentence, he concocted a completely absurd one that has become famous: “Colorless green ideas sleep furiously.” Although totally nonsensical, it feels very different to the reader from an equally nonsensical rearrangement of the words: “Ideas furiously green colorless sleep.” Anyone familiar with English finds the first version somehow comfortable—it almost seems to mean something—while the second is uncomfortable gibberish. The reason is that the first version obeys the rules of both surface grammar and deep structure; the second does not.

  Chomsky’s theory touched off fierce controversy, largely because of its innatism, although he did not posit inborn ideas but only the inborn capacity to experience language in useful ways. Some critics, rejecting the hypothesis of an L.A.D., agreed that the ability to acquire language is innate but said that it is a byproduct of general intellectual abilities. Others to whom the theory of an innate L.A.D. is unacceptable keep finding grounds on which to reject it. One such ground, for instance, is that genetically transmitted organs are subject to variations. If so, some children should have abnormal L.A.D.’s and be deficient in some areas of language comprehension, but there seems to be no evidence of that.60

  Aside from the controversy, for half a century psycholinguists and cognitive psychologists have been gathering evidence that shows how language relates to thought and reveals thought processes. Some patiently observe the errors and self-corrections children make in learning language, some analyze language games, some study developmental language disorders like dyslexia and acquired language disorders produced by brain injuries, and some conduct reaction-time experiments. An instance of the last: Herbert Clark and others have found that when subjects are shown a simple pattern, such as a star above a plus sign, and alongside it either a true affirmation (“Star is above plus”) or a true denial (“Star is not below plus”), it takes them two tenths of a second longer to say the denial is true than to say the affirmation is true. We seem to be programmed to
think more easily about what is than about what is not, and we have to turn negative sentences into positive ones in order to deal with them.61

  Today, many psycholinguists, as a result of their research, give more credit to environmental influences in language acquisition than Chomsky does. They stress, for instance, the informal language training provided by “motherese,” the special way mothers (and some fathers) talk to small children. Nevertheless, while many psycholinguists question details of Chomsky’s L.A.D. theory (which he himself has much qualified and modified over the years), most agree that human beings have a genetically determined ability to understand and acquire any language.

  Psycholinguists have also explored other important questions about the relation of language to thought. Do we always or only sometimes think in words? Is thought possible without words? Do the words of our native language shape or limit our thinking? The issues have been much debated and much studied. A few highlights:

  —The linguist Benjamin Whorf theorized in 1957 that thought is molded by the syntax and vocabulary of one’s native language, and offered cross-cultural evidence to prove his point. One of his examples was that the Hopi Indian language does not distinguish, at least not as we do, between past, present, and future (a rare exception to a nearly universal rule). Instead, a Hopi speaker indicates through inflections whether he or she is talking about an event that actually happened, one that is expected to happen, or about such events in general. Whorf and his followers accordingly maintained that the language we use shapes or influences what we see and think.62