To the Editor:

Arthur B. Cody’s argument in his review of Steven Pinker’s How the Mind Works [Books in Review, March] is a parody of sloppy New Age thinking. Pinker’s use of evolutionary theory in explaining human behavior, Mr. Cody asserts, is “wildly fanciful, a work of fiction.” Why? Because evolution consists of genetic change, and “we cannot map the passage from genetic structure to human behavior.” Why not? Because genes “do not code for mating rituals, for fight or flight responses.” Rather, Mr. Cody informs us,

a single gene codes for a single protein; that is all, no more and no less. . . . [H]ow many [genes] would be required to make some trait . . .? The answer is profoundly unknown.

And he concludes that

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 been revealed.

The reader should understand that this is not an attack on Steven Pinker alone or even on evolutionary psychology—it is an attack on all evolutionary theory. The basic evolutionary model states that genotype affects phenotype and, through natural selection, phenotype affects genotype. The result is the evolution of both genotype and phenotype. But if Mr. Cody is correct, we cannot argue that genotype affects phenotype, except in that very tiny number of cases in which we know the precise mechanism through which this occurs. Darwin (who not only did not know about DNA but did not even know about genes) must have been “wildly fanciful” in his analysis of Galapagos fauna. To Mr. Cody, the neo-Darwinian synthesis in evolutionary biology, as widely accepted a theory as any in modern science, is simply a “work of fiction.”

If Mr. Cody understands the scientific method, he gives no evidence of it here. And though supposedly a philosopher, he appears never to have heard of even the most elementary distinctions such as those among material, formal, final, and efficient causes (Aristotle). For example: we knew that incest reduces the viability of offspring long before we knew how it does so, and we knew that smoking reduces life expectancy long before we understood the mechanism involved. Similarly, we know in considerable detail and with precise and predictive models that evolution affects the behavior of organisms, but in spite of advances in biochemistry over the past two decades, there is much we still do not know about the causal mechanisms leading from genotype to phenotype. Nor is it certain that we will ever be able to chart the pathways from genes to behavior, since complex biological systems may well have emergent properties that cannot be predicted from their underlying constituent parts. But this does not compromise behavioral ecology in the slightest.

Steven Pinker argues that human beings have brain structures that affect behavior in specific ways, that these brain structures are a product of the genetic constitution of human beings, and that this genetic constitution is the product of biological evolution. These arguments are all true, I believe, and none depends on having any detailed knowledge of developmental biochemistry. Pinker may be wrong in whole or in part in his specific assertions, but Mr. Cody’s arguments are quite beside the point, and his review completely uninformative.

Herbert Gintis
University of Massachusetts
Amherst, Massachusetts

_____________

To the Editor:

Having read several laudatory reviews of Steven Pinker’s book on the mind, I found it a pleasure to read the insightful one by Arthur B. Cody.

When all else fails, it is the common practice of popular writing in biology to invoke the tautology of natural selection to account for complex properties of organisms. As Mr. Cody points out so well, genetics has limited explanatory power for making a cell, much less an organism.

However, it is possible to provide some answers to his questions of how many genes are involved in making an eye. Experiments indicate that mutations in some two-thirds of the genes of a fruit fly affect the development of the eye. That means roughly 10,000 genes are involved. If a similar fraction of the genome were involved in the eye development of humans, the number of “eye” genes would be increased five to tenfold. The permutations and combinations of variations from mutations among such large numbers of genes rise to numbers too large to be handled by any conceivable computer in predicting the developmental outcome of the functions of those genes. Similar numbers of genes are likely to contribute to the development of other organs.

To say that development is determined by natural selection is a meaningless statement because there is no set of experiments that can prove it. As Niels Bohr said in a paraphrase of Wittgenstein, “Science must pass over in silence what it cannot define by experiment.” It will take a quantum advance in biological thought before we can cope with such problems.

Harry Rubin
Department of Molecular and
Cell Biology
University of California
Berkeley, California

_____________

Arthur B. Cody writes:

Herbert Gintis, a noted economist who has also ably defended Darwinism for many years, maintains that there are few if any explanations in science that are supported by a complete understanding of the processes at work. In this I agree with him. He also reminds us of what Darwin accomplished without any knowledge of genetics; but we are not in a similar position today. It is not that our present knowledge is limited, as Mr. Gintis states, but rather the reverse: we know a lot, and what we know goes against the cornerstone of the neo-Darwinian theory of evolution, the claim that variation in the phenotype is the result of random changes in the copious genotype.

We know with certainty that the human genome consists of no more than 100,000 genes, and we know that genes express proteins, not body plans. We also have reason to believe that the fertilized egg through which inheritance moves contains important molecules other than DNA, but none is a candidate for the implementation of the informational program ascribed to the DNA molecule. The DNA helix alone is simply not big enough to provide all the information required to build an organism—including, among other things, every regulatory system; every accommodation to injury and sickness; every system of fluid, muscular, and electrical movement; every complex organ from kidney to brain; and now, according to Steven Pinker and others, every mental disposition and cognitive dimension.

From this, however, I do not conclude that Jean-Baptiste Lamarck, who maintained that evolution progresses by the inheritance of acquired traits, was right, though it might appear that only directed evolution gets us through the numbers; instead, I recommend patience. For we still do not know how evolution actually works. Natural selection is plausible, and, as such, it not only permits but encourages teleological speculation disguised as scientific reasoning. Plausibility quiets the nerves but it is not proof. It may turn out, and I suspect it will, that the “quantum advance in biological thought” that Harry Rubin anticipates will reveal the principles of evolution to be no less unique and implausible in biology than those of quantum theory have proved to be in physics.

Mr. Rubin, a member of the National Academy of Science, has had a long and distinguished career in cell biology and immunology. I am happy indeed to have his support and to be informed of the experiments to which he refers. Those experiments lead to the astonishing conclusion that two-thirds of a fruit fly’s genes “affect the development” of the eye. Obviously, two-thirds of the fly’s DNA information is not devoted exclusively to the eye; instead, the implication is that there are many systems so intimately related to the eye that mutations in them bring about changes in the eye as well.

This feature of organic development adds enormously to the task natural selection is called upon to perform. The standard thesis insists that all mutations in the genotype occur randomly; any other suggestion introduces purpose or design into the process, and that is strictly forbidden. But if the development from one eye stage to the next requires a random search of all related systems, the numbers involved in such a search would be truly staggering.

To be more specific consider first the number of possible proteins in the design space. There are 20 amino acids that can be strung together in groups of 50 to 5,000 in forming a protein. That means a base 20 raised to the 50th power, plus 20 to the 51st power, and so on to 20 to the 5,000th power. (The number 10 raised to the 100th power, an insignificantly little number by comparison, is said to count all the subatomic particles in the universe.) Now combine that number of possible proteins with tens of thousands like it in the systems simultaneously affected by any change in the eye and you see what Mr. Rubin means by a resultant number no conceivable computer could handle.

Consider then Steven Pinker’s supposition that all the habits, attitudes, evaluations, predilections, and so forth of the mind are laid down in the brain through a process of natural selection perfected in a few million years while our ancestors were hunting and gathering. The random process of natural selection working on the genotype for so formidable an achievement seems not merely inadequate but altogether infeasible.

Now to a correction. In my review, the numbers I used in going from the human DNA molecule to the human genome were bollixed, even though the conclusion remains the same—i.e., the total number of human genes lies between an estimated low of 60,000 and a high of 100,000. The correct reasoning goes as follows: begin with six billion base pairs in the genome that include inputs from both parents; reduce that to three billion as the resultant collection for the individual; reduce that by 90 percent of junk DNA; reduce that by, say, 90 percent for the introns and other codons not used in protein expression (though used in numerous other ways)—and you get 30 million. Divide 30 million by an average of 300 nucleotides (or 100 codons), and the final number is 100,000. This calculation is meant only to illustrate the relationship between the total DNA molecule and the number of genes in the human genome; it serves no other purpose.

_____________