How the Mind Works by Steven Pinker
How the Mind Works
by Steven Pinker
Norton. 606 pp. $29.95
Nothing is more fascinating than how we think, why we think that way, and how our thinking affects our actions. In this book Steven Pinker, a professor in the department of brain and cognitive science at MIT, takes on the general task of explaining the process, and in so doing he finds something to say about almost every propensity of the mind at which a stick could be shaken: why we see colors as we do, why we are disgusted by worms, what attracts us in the opposite sex, why we are given to emotions, why we are or are not promiscuous, why mothers sometimes kill their newborns, and on and on.
Pinker is only forty-three, but he has been publishing professional and popular articles and books since 1979, and enjoys a reputation as something of an enfant terrible. It is easy to see why from How the Mind Works, a long and loosely organized book written in a style that blends vernacular expressions and witticisms (often Jewish) with sophisticated ideas presented in fairly technical language. The book is both fun and, deliberately, outrageous; that is no doubt why it commanded a very large advance, why it was made a selection of the Book-of-the-Month Club, why it has been widely and controversially reviewed, and why its ideas are likely to be influential. Unfortunately, those ideas happen to be wrong.
Pinker’s thesis, in brief, is that the human mind is a computational device, rigged by evolutionary forces to benefit the individual in the competition for survival and for dominance in the gene pool. In How the Mind Works he begins by showing how the computational model helps us to interpret data gleaned from various experiments related to human vision and to early learning in children. Then he goes on to the juicier topics, arguing that all sorts of human behavior, from emotions to sex roles, can be explained as the results of computational operations disciplined over the course of human history by the implacable workings of Darwinian natural selection.
A few examples will help give a sense not only of Pinker’s approach but of the scope and purpose of his book. A great deal of space is devoted to a discussion of the mechanics of vision, drawn in large part from the work of the late David Marr. We learn, for example, about the remarkable ability of the mind always to see colors “correctly,” even under unfavorable circumstances. A snowball, for instance, looks white even in dim light, while a chunk of coal in bright light looks black, though in fact more light may be bouncing off the coal than off the snow. Why do we see the “right” colors? Because our mental program is set up for that very purpose: it is more helpful for us to keep the colors right than to record, like a light meter, how bright an object may happen to be. Similarly, through shading and perspective, a two-dimensional drawing on a page, shown to people in experiments, appears to them as a folded piece of cardboard—that is, to possess depth. Why? Because our beneficial program is set up for 3-D. And so forth.
From experiments with children, Pinker informs us, we can infer that the mind categorizes visual and other sense data in preformed “boxes.” Evolution has created these boxes for us to fit our experiences into, even though we ourselves may have had no previous encounter with a given subject matter or set of behaviors. The mind is not a blank tablet; evolution has written on it before we open our eyes. If our natural expectations sometimes prove to be wrong—children expect objects to be cohesive, so when they are shown a hand picking up what looks like an object, but then part of the “object” stays behind, they register surprise—those expectations nevertheless remain useful, at least as a beginning point. Evolution proceeds not according to truth, but according to success.
This lesson sets us up for Pinker’s revelations concerning how the mind works in human experience. To our ordinary way of thinking, for instance, something is disgusting because it is, well, disgusting; it turns our stomachs. But Pinker has a different explanation: disgust is actually nature’s way of supplying a useful primer in microbiology, warning us away from (say) animal parts and excrescences that rot and decay as well as from the creatures that might touch or eat them. This evolutionary effort to keep us healthy may not work perfectly, but it is better than either no response at all or other responses that nature might have randomly tried before.
Why do sex roles differ? Because, Pinker again explains evolutionarily, although both men and women have the same impulse to maximize their contribution to the pool of genes, they differ in the respective investments they make in their off spring. Men are promiscuous because to them the energy cost of maximizing their gene input is small, whereas for women that cost is much higher, pregnancy and child care being great consumers of time and strength. It is evolution that commands men to go after nubile (hence fertile) young women—and as many of them as possible—just as evolution commands women to seek protective men who will stay around, men who are rich, powerful, and often older. Men’s minds are programmed to be excited and enticed by the sight of nude young women; hence the popularity of Playboy. For women, rape is a constant peril, and male nudity a threat; Playgirl, Pinker reports, is successful not with women but with gay men.
How can Pinker be so sure about all these dispositions of the mind, especially since (unlike the case of vision) he does not have experiments to rely on? He uses the method of “reverse engineering”—a term that customarily defines the process whereby a competitor’s product is purchased and taken apart so as to determine how it works and what possibly new techniques were employed to achieve its designated end. Pinker’s use of the term is metaphorical. He arrives at his analysis of human conduct through an enlightened surmise of the objectives of natural selection—that is, what a thing is “for,” how it contributes toward success in the grand Darwinian scheme—and then works backward to how and why it does what it does.
Are Pinker’s enlightened surmises plausible? Some are. Is his reasoning not circular? It is; unlike reverse engineers in industry, he proceeds not from a known end (the product) but from an assumed end—Darwinian survival—and his reconstructions of the means toward that end are themselves often hypothetical. But the problem in How the Mind Works goes even deeper than this. The real problem is Pinker’s reliance on the computational model, which, even on its own terms, is seriously flawed.
The computational model says that all mental workings are in every respect a function of the brain’s manipulation of squiggles or bits called symbols. In Pinker’s psychology, the question of what, if anything, those symbols symbolize does not arise. As he describes the brain’s activity, bits of matter symbolizing things bump into other bits of matter also symbolizing things, and when “eventually the bits of matter constituting a symbol bump into bits of matter connected to the muscles, behavior happens.”
Let us be clear about this. It would be a misunderstanding of Pinker to say that, for him, symbolizing is a deliberate undertaking, or that symbols stand for concepts, ideas, or images. We might think that the phrase, “She betrayed me,” is something a man might say out of a belief that his wife had lied to him or undertaken compromising liaisons with another man—basing himself, that is, on a common understanding of what lies are, and what deceit is. But Pinker tells a different story—one in which, apart from its activity with squiggles and bits, there is nothing left over for the “mind” to do. Light energy and sound vibrations impinge on the peripheral sense organs, and when the patterns from them fit a certain program-configuration in the brain, muscles are stimulated in such a way that the sounds, “I hate you, Jane,” are heard echoing down the hall: a jealousy subroutine has been set in motion.
What is being suggested here is—again on its own terms—preposterous. No one has ever proposed a way to analyze billions of possible sense inputs under a single name or description like Pinker’s jealousy program. Such a program would have to take in sense data of all sorts from visual and auditory sources. Light, for instance, would come from all angles and in different illuminations to form an object—OK, we have a program for that. But then it would have to be processed as “my wife”—for this there is no evidence of a program—and the next stages get more difficult still. What, by light and sound bits only, is deceit? The odd mood change? Disturbances of the imagination? How has nature established—rather, preestablished—a program so flexible that it can adjust to furtive glances, or to coded messages in love letters, telephone calls, or e-mail in English, Urdu, and Tagalog?
An analogy comes to mind. My word-processor has a program to check my use of English grammar. It will tell me when “which” should be used, and when “that” is called for instead. Of course it is quite inflexible, its corrective message appearing on the screen only when certain definite patterns of O’s and l’s occur under certain specifiable conditions. It is precisely the predictability of the patterning that makes the grammar check possible. But in jealousy, there is little that takes predictable form; jealousy is tied to individuals and their personalities, to the structure of a society, to historical epochs, to the state of technology. What kinds of muscle control will the program contain that, in a manner both laid down by evolution and adapting to circumstance, will take the form of crying, shouting, sneaking around, moping, hitting, or killing? And when will the muscles be activated, and against whom?
Pinker might reply that the task, though formidable, is accomplished stepwise. First the data are put into “thing” categories (woman); then the things are put into “person” and “relationship” categories (my wife); then the relevant activities are put into behavior categories (lying, moodiness, absence from home); and finally they are assembled into classifications (infidelity) and possible reactions (crying, beating). But this just begs the question, since on Pinker’s own understanding we have no way of categorizing bits of matter (or electrical charges in neurons) as beliefs and actions. The computational device may make coal look black, but it does not do mental things—or, if it does, nothing we yet know allows us to say that it does.
According to Pinker, “betrayal,” “jealousy,” “love” are just sounds or marks we make when the bits have bounced in certain ways: a process, something that happens to us (for our own good, or the good of the species). If that is so, then Pinker himself cannot claim to know anything about jealousy or love as we conceive of them, but only as bumping symbolic bits rumbling and chuffing around the dark closets of the brainpan. Yet Pinker writes about jealousy in the ordinary sense as if he thought he were telling us something about it, and as if what he had to say were intelligible. Maybe he has been taken in, along with the rest of us.
To these difficulties with Pinker’s conception of our mental modules—that is, the difficulties of how very elaborate they would have to be, even if we were to accept that they exist—must be added others. The mind, Pinker urges, is a product of the brain’s design. From its propensity toward language acquisition to its propensity to prefer slim, young blondes, it is a collection of computational regulations laid down in all their physical manifestations by the codons of the DNA molecule. Turning, then, to the molecular vehicle of our behavior, we must pose an important question: is that marvelous double helix, the Bosporus through which all our genetic accomplishments must be squeezed, capable of containing so much information?
Pinker must assume the answer to this question is yes, for he never hesitates to ascribe any habit, disposition, attitude, desire, or intelligence to some capacity of that wondrous object. In this, however, he is mistaken. There is no basis for believing that the genome contains enough inherited and/or mutated capacity to create the hominid whom Pinker describes. If that capacity is not there—and, on his telling, it can be nowhere else—then his lectures and remonstrations remain unsupported by his story of how genes are selected through random trial and error.
The genetic burden, we know today, is carried by the DNA-RNA molecules. Though their behavior cannot be adequately covered even in a big book, much less in a review, it is safe to generalize on the way the so-called genetic code works. In human beings the DNA consists of about three billion base pairs of nucleotides. There are four kinds of these nucleotides, which in groups of three make a codon. The codons lie next to each other in a line along each side of the two spines of the DNA molecule. A wonderful process unzips the two sides of the double helix, and one side is read and copied in relatively short segments. Those segments—groups of a few to many hundreds of codons—constitute a gene.
It is possible to estimate the number of genes contained by the DNA molecule by throwing out of our calculation those codons (called introns) that do not do anything—that is, that are not part of a gene segment—and then averaging the gene size and dividing the result into the number of codons (called exons) functioning as genes. The calculation goes like this. Of the approximately three billion base pairs in the human genome, about 90 percent, surprisingly, are junk: they do not function as parts of genes. That leaves 30 million (10 percent of three billion). Using an average gene size of 300 nucleotides, or 100 codons, we arrive at a total of 100,000 genes in the DNA molecule. Some estimates go as low as 60,000, but none goes higher than 100,000.
So: can 100,000 genes build a human being in all its details, from structural parts such as bones and internal organs to functional parts such as heart and respiratory rates, walking and running coordination, language acquisition, logical inference, and emotions? The central dogma of microbiology is that a single gene codes for a single protein; that is all, no more and no less. (That is not strictly true, but the generalization is close enough.) Genes do not code for mating rituals, for fight or flight responses, or even for body parts or anything that would count as a feature of an organism.
Since no single gene makes any trait of an organism, a bacterium, a plant, or a Pinker, how many together would be required to make some trait, how many a Pinker? The answer is profoundly unknown. The DNA molecule in human beings is divided into 46 chromosomes (actually 47, the “y” making the 47th), and every day we read that some position on some chromosome is associated with a genetic disorder or propensity to disorder, from cancer to Down’s syndrome. But no part or group of parts of any chromosome is known to be responsible for the construction of a liver or a kneecap. We do not have the slightest idea of how two or more genes combine to make anything, whether a single cell or a hair follicle. There are regulatory genes that have something or other to do with combinations of genes, but little is known of them, and nothing is known about whole-component construction by means of such regulatory genes.
The logic of genetic part-building is not only unknown, but unimagined. How can genes that individually code for proteins and only for proteins do anything more than that by being combined in the serial order of the DNA grouping? The combination would have to be the consequence either of several nucleotide genes making a super-code or of several proteins built in an order that in turn amounts to a code. Even if there is a supercode, there is still reason to doubt that enough information would be carried by it to accommodate all the work of inheritance ascribed to the DNA molecule.
How many genes, for instance, are necessary to make a kidney, with its various component and subcomponent structures, their means of organization, and the hormonal controls dictated and coordinated by the brain? Might we assume hundreds? How many for the eye? Thousands? More? How many to prepare the brain for language acquisition? Millions, billions? It is impossible even to speculate, since we have not the slightest idea of what is entailed in language from the point of view of a speaker/ thinker any more than we have an idea of how genes might go together to provide this capacity. Not only are answers absent, a question that might be productive of an answer has not yet been formulated.
Pinker simply proceeds on the assumption that human mental abilities are attributable to genetic inheritance, modified by adaptation. This or that gene, he says, was selected because it provided this or that advantage to the individual or to the species. He does not notice that in the case of almost any feature for which we may specify an advantage, there will be many more that would have had to go along in a sort of symphony of common achievement. Upright posture required modifications from our forebears’ stooped and ambling carriage that involved bones, leg muscles, the spine, and the circulatory system; an instrument of use, the hand, once the new posture was achieved; and, perhaps most importantly, an inclination or willingness to adopt the distinctive stance. All of these and more must go together to fit a being to stand, walk, and run on two legs. How many coordinated and more or less simultaneous code changes have had to take place, and from which prior functional entities, in order to account genetically for disgust, romantic love, or musical taste?
Science is silent.
Pinker’s Entire effort, in short, is wildly fanciful, a work of fiction based not on biological science but on a cartoon genetics, little smiling faces zapping suddenly into convenient organs and operations. There may indeed be some natural process at work in constructing an organism or a species by means of instructions laid down in the DNA double helix. But if so, it has not yet been revealed. I suspect that 100 years will not be sufficient to reveal it, and scientists will have plenty of microbiological research to keep them busy in the interim.
But one thing is plain: the explanation finally arrived at will in all probability bear no resemblance whatsoever to the facile, stands-to-reason kind of Darwinism advanced by Steven Pinker. It is a pity that he has been invited to pronounce as if knowingly on subjects of intense and mighty concern to us as we desperately struggle to find the right path in life. For science as yet has no guidance to give us, and Pinker hasn’t a clue.