In response to:

Billions and Billions of Demons from the January 9, 1997 issue

To the Editors:

Richard Lewontin’s review of Carl Sagan’s The Demon-Haunted World [NYR, January 9] raises many of the right questions about the over-confident rhetoric of many scientists. Anyone who believes Sagan’s claims about the ultimate hopes for “scientific method” should study Lewontin’s underminings carefully.

It is not enough, however, to say that Sagan “has opened the wrong envelope.” Lewontin himself has not quite opened the right one. He is right in claiming that Sagan gives no clue about how scientific method might prove its own claims to superiority. But he falls into a shopworn and destructive dichotomy when he says, “The case for the scientific method should itself be ‘scientific’ and not merely rhetorical” (my italics). The “case” can never be “scientific” in the sense that Lewontin and Sagan imply. Lewontin does acknowledge that scientists inescapably rely on “rhetorical” proofs (authority, tradition) for most of what they care about; they depend on theoretical assumptions unprovable by hard science, and their promises are often absurdly overblown.

But he mistakenly suggests that such rhetorical reliance is a fault that the right kind of scientist—no doubt his kind—could escape. Since rhetoric is for him always on the non-rational side, it can have nothing to do with the genuinely scientific. He thus underplays two facts: that everyone, in all human circumstances, relies on rhetoric, and that some forms of rhetoric are thoroughly rational (for example, most of his arguments in his review).

Lewontin concludes that “we do not know [my italics] how to provide” “the power to discover the truth.” He’s right, if by “know” he means “able to provide experimental or statistical or mathematical proof, for every rational belief.” But we can distinguish strong rhetoric (for example, “indubitable” arguments that “hard evidence should count”) from the obviously dubious or just plain silly. From roughly mid-century on, beginning with the still underestimated works of Kenneth Burke, we have had a flood of serious inquiry into the many genuine distinctions in degrees of rationality obscured by that polarizing of “science,” the rational, and “mere rhetoric,” the irrational.

Lewontin also needs to do some rhetorical analysis of his God-terms, “matter” and “materialism.” “No matter how counter-intuitive,” he says, “no matter how mystifying to the uninitiated,” Sagan’s and his faith in materialism is “absolute, for we cannot allow a Divine Foot in the door.” Surely he knows that “matter” is by now an utterly ambiguous and controversial notion. Ever since Einstein “equated” mass, energy, and the speed of light, “matter” has snuck further and further under vast carpets of waves and patterns that, if still somehow “matter,” ain’t what matter used to be. It’s often much more like what people have traditionally thought of as spirit, or even “the power of the Lord.” Our presses are now flooded with books by genuine scientists grappling with what at one time would have been thought questions reserved for theologians.

In short, while I wish everyone shared Sagan’s and Lewontin’s rejection of kookie faiths which are easily undermined with hard evidence, such as belief in Martian invaders and spoon-benders, Lewontin’s case against Sagan’s naiveté would be a lot stronger if he acknowledged what we rhetorical theorists have been claiming for more than two millennia: about most of the issues that we care about, hard science can teach us nothing, but that does not mean that when science fails what is left is irrationality. What is left are a broad range of more or less rational arguments, to be tested not in the laboratory but in the courts of reasonable discussion.

The brilliant, learned Lewontin might well start his re-education by settling into a careful reading—or re-reading?—of Aristotle’s Rhetoric. Then he could do a book tracing the non-scientific rhetorics, defensible and indefensible, of various prominent scientists. Such a book would join the fine work of economist Donald McCloskey on economists’ non-economic, non-scientific arguments, in his book The Rhetoric of Economics.

Wayne C. Booth
The University of Chicago
Chicago, Illinois

To the Editors:

In “Billions and Billions of Demons,” Richard Lewontin has directed some misleading assertions and unwarranted criticism at genetics research in particular and science in general. In continuing his crusade against genetics as it is studied and applied in the context of what he has previously termed “bourgeois science” and now refers to as “elite culture,” Lewontin implies that genetics has made no significant contributions to medicine. With regard to cancer treatment, he writes: “The realization of the role played by DNA has had absolutely no consequence for either therapy or prevention….” And on “diseases” in general: “We do not yet have a single case of a prevention or cure arising from a knowledge of DNA sequences….”

Statements such as these, especially by a geneticist, are surely disingenuous, concealing significant facts behind the terms “therapy,” “prevention,” and “cure.” There is no magic bullet but Lewontin certainly knows that the ravages of Tay-Sachs disease, a catastrophic genetic disorder which invariably results in deterioration and death within the first few years of life and which is disproportionately common among Ashkenazic Jews and French-Canadians, can be and is being avoided through genetic screening and counseling. Not only have such births been avoided, but normal births have resulted from the reassurances received by couples who would otherwise have been reluctant to have children. More than forty genetic disorders have been correlated with high-risk ethnic groups, and in several conditions genetic screening and counseling can indicate interventions that can at least moderate disease. “Prevention” and “cure” are clearly not the only avenues along which knowledge of DNA sequences has justified the research. And this says nothing about applications in the fields of agricultural science.


Lewontin’s targets are not limited to genetics; he also takes aim at the entire scientific enterprise which he now sees, dialectically, in terms of a “confrontation between elite culture and popular culture.” Creationist opposition to teaching evolutionary biology is blamed on “the elite culture [that] was now extending its domination by attacking the control that families had maintained over the ideological formation of their children.” He even finds fault with the open, self-critical tradition of science, comparing it unfavorably with Hasidic scholasticism: “If [one] really wants to hear serious disputation about the nature of the universe, [one] should leave the academic precincts…and spend a few minutes in an Orthodox study house in Brooklyn.”

I wonder if it is noticeable in a few minutes that Hasidic culture is no less elite than science, and much more sexist—learned males in yeshivas; learned males and females in universities. And while he was singing the praises of the Brooklyn Hasidim he might have mentioned that, until 1984, their community was wracked by Tay-Sachs disease. Then, through the concerted actions of neurologists at a local hospital, the United Jewish Appeal of New York, and a Hasidic group, Dor Yeshorim, a program of genetic screening, counseling, and traditional matchmaking reduced the incidence of Tay-Sachs disease in the community by nearly 80 percent.* Some of those skilled dialecticians owe their peace of mind to the knowledge of DNA sequencing that spared them from the horrible tragedy of parenting a Tay-Sachs child.

Lewontin’s antipathy to the culture of scientific research is misconceived and misguided. In attempting to discredit what he and many scholars and scientists in the elite culture see as an imperfect society he is directing his fire at its most commendable institution, an institution which in fact functions in accordance with essentially the same norms across various social systems. It is unfortunate that a scientist of Lewontin’s caliber continues, in the name of world betterment, to minimize the achievements of science and fellow scientists. Perhaps he will heed the appeal of a common sailor in the film The Battleship Potemkin, a product of popular culture: “Lower your guns, you’re shooting at your brothers.”

Harold Dorn
Stevens Institute of Technology
Hoboken, New Jersey

To The Editors:

Richard Lewontin asserts that the discovery of the role played by DNA in cancer has had “absolutely no consequence for either therapy or prevention.” The same cannot be said for several other diseases. For example, the use of the DNA-based Polymerase Chain Reaction (PCR) now allows us to detect Cytomegalovirus infections of the nervous system in AIDS patients early enough to begin treatment that can prevent major disability. Genetic testing for inherited nervous system diseases like Huntington’s chorea allows detection before a patient has shown symptoms, and, with any luck, before the mutation has been passed on to a successive generation. Recombinant DNA technology allows the production of large quantities of biological agents ranging from insulin to growth factors and interferons, which show promise in the treatment of diseases like renal cell carcinoma and Lou Gehrig’s disease. These advances would be impossible without the type of basic research into the role of genes in disease that Lewontin disparages.

Additionally, Lewontin criticizes the attempt to sequence the human genome by stating that if he knows “the DNA sequence of a gene [he has] no hint about the function of a protein specified by that gene, or how it enters into an organism’s biology.” This is untrue. By comparing the sequence of a newly discovered gene or protein to those of proteins whose function has already been elucidated by more traditional (and arduous) methods, one can make educated guesses about the function of the new gene. Using computerized databases containing thousands of sequences, these guesses usually turn out to be correct, and we are spared the effort and expense of having to repeat functional studies on every gene that is discovered. Given the vast number of genes whose sequences and functions are currently known, large tracts of the human genome sequence will be interpretable solely by sequence comparison.


Dr. Lewontin correctly argues that the claims of scientists about the societal value of their work product should be examined skeptically. However, I believe that even a cursory analysis of the advances of biological science in the last twenty years shows that we are better off because of the massive government funding of molecular biology.

Richard A. Bernstein, M.D., Ph.D.
Department of Medicine
Mount Sinai Hospital
New York City

To the Editors:

…Professor Lewontin makes important points about the impediments placed by social structures upon increasing scientific literacy among the public. However, he also appears to hold misconceptions about the nature of progress in molecular cancer research that illustrate why he believes scientists are guilty of the perceived deceptions. He suggests that researchers wasted time and money early on studying elusive oncogenic viruses and, frustrated by lack of success, turned their attention to human genes without giving the viruses due “comeuppance.” While it is true that oncogenic viruses seem to be directly involved in only a minority of cancers, it was the study of oncogenic animal viruses that first pointed to the human genes themselves as the true culprits.

In fact, the work of J. Michael Bishop and Harold Varmus (for which they were awarded a Nobel Prize) demonstrated that the cancer genes of certain acutely oncogenic animal viruses were mutated copies of genes originally derived from their host organisms. One type of cancer gene, designated ras, was originally found in a rat virus. The human equivalent of this gene is mutated in a high percentage of human cancers. There are valid scientific reasons to predict that drugs (some of which are in clinical trials) capable of disrupting the function of the protein encoded by the ras gene will be useful therapeutics in the future. A wealth of knowledge about the function of ras proteins has made such drug development possible, but this information would not have been generated if the discovery of cancer-related mutations in ras genes had not occurred first. Thus, there has been a direct continuum of progress from the study of oncogenic viruses to current drug development. This line of inquiry therefore supports the notion, derided by Professor Lewontin, that knowledge of the DNA sequence of defective genes will be beneficial in therapeutic intervention, at least in cancer.

Robert S. Krauss, Ph.D.
Assistant Professor
Department of Biochemistry
Mount Sinai Hospital
New York City

Richard Lewontin replies:

What madness must have possessed me when I wrote “mere rhetoric”? I can only plead that while working on my review I was immersed in Plato who, as Wayne Booth knows, had a very low opinion of rhetoric. Indeed, he rejected its usual characterization as an art (tékhne), regarding it as only a practical skill or knack (tribé) like cooking and cosmetics, a form of gratification and flattery. I realize that modern views of rhetoric depend on Aristotle’s encyclopedic treatment in his “Art of Rhetoric” (what the eminent Victorian classical scholar, Richard Jebb, thought of as “one of the driest books in the world”). Aristotle like Plato thought that “the whole business of Rhetoric was to influence opinion” and he also agreed that one ought to lead people only to right opinion, so he deftly swallowed Plato by including within rhetoric both a recitation of the evidence and the logical structure of argument based on the facts.
Matters of style are treated by Aristotle only last and as a vehicle for the truth. But this neat anatomization of the subject, and the subordination of style to content that flows from Aristotle’s moral view of the function of rhetoric, seem to me to miss an essential reality about argument. Our susceptibility to the resonance of language itself often leads us to accept what we otherwise would not believe. Although every moral and political conviction I have speaks against it, I am nearly driven by Milton’s insidious poetry to believe that it is “Better to reign in Hell than serve in Heaven.” For a pedagogical Aristotle this would be a particularly fine example of an “antithesis” in the “periodic” mode dissected in Book III, Part 9: the parallel but opposing structure, “Better to…, than to…,” the contrast of functionally linked “reign” and “serve,” the final antithesis of “Heaven” and “Hell,” two words beginning with the same sound but with differ-ent endings. For me, however, it is an almost irresistibly seductive flight of English speech, a Siren music that can be withstood only by lashing myself to the mast of my convictions. Many times more dangerous than “mere” rhetoric, artful language creates its own logic.

Booth’s gloss on my lament that we do not know how to provide people with the power to discover truth is off the mark. It has nothing to do with dividing all claims into the rational and irrational, into mathematically proven and just plain silly. Only the most simple-minded and philosophically naive scientist, of whom there are many, thinks that science is characterized entirely by hard inference and mathematical proofs based on indisputable data. I put a great deal of weight, as did Aristotle, on arguments about what is probable (in the everyday sense) and on arguments by example. Indeed the entire science of statistics is designed to cope with the ambiguity of most scientific evidence, and my professor, Theodosius Dobzhansky, the most eminent experimental evolutionist of his day, used to say that “statistics is a way of making bad data look good.” The problem of the power to discern truth lies precisely in knowing how to evaluate, even roughly, the ambiguous knowledge of the world that is produced by so much of science. It is a problem because that evaluation requires an acquaintance with an immense penumbra of fact and theory that surrounds any particular observation and which provides its context. Unfortunately, there is not world enough, or time.

It is true, as Booth says, that matter ain’t what it used to be, but then neither is materialism. Einstein’s equation of matter and energy makes the materialist’s life easier, not harder, because it brings under the aegis of elementary physical explanation phenomena that would otherwise be in the realms of mystery. By “materialism” we mean the claim that all the motions and states of the physical (including the biological) universe form a closed world of causation, solely under the influence of a small number of known measurable forces. As I pointed out, this claim requires us to accept that these forces have some counterintuitive properties, our intuitions having been formed by the experience of our gross senses. The fact that “matter” can be turned into “energy” is one of the less difficult ones for us to imagine because our ordinary experience seems to tell us (quite erroneously) that this happens every time we burn a piece of paper. And if any doubt remains that the conversion of mass into energy can be accomplished at will by the work of human hands, a trip to Hiroshima will settle the matter.

A transparent example of the attempt to make the mere aroma of words do the work that neither the evidence nor logic can is offered by Harold Dorn’s letter. It is filled with evocative terms like “bourgeois,” “dialectically,” “in the name of world betterment,” unnamed “social systems,” finally topped off with “The Battleship Potemkin.” We have passed from Einstein to Eisenstein. I am apparently a member of the recently invented but undefined “academic left,” devotees of a passé Marxism that “takes aim at the entire scientific enterprise” because of an “antipathy to the culture of scientific research.” Yet none of this appears, even by implication, in my review. So he says that in my essay the rise of creationist activism is “blamed on” what I used to call “bourgeois” but now describe as “elite” culture. This is rather like saying that I “blame” the abolitionist movement for Secession and the Civil War. I did not invent the notion of the confrontation between elite and popular culture nor the identification of an elite culture by poor farmers and miners. Dorn seems simply ignorant of the well-established and uncontroversial history of American social movements of which populism is a recurrent variety. He might begin his education in this matter by reading Richard Hofstadter’s famous thirty-year-old work, Anti-Intellectualism in American Life. He may then understand why the characterization of intellectuals as “an effete corps of impudent snobs” came from the mouth of Spiro Agnew.

I did, indeed, write that Talmudic scholars were more aggressive in their pilpul than scientists, but it is Dorn, not I, who regards that as “singing the praises” of the former. For me, at least, it is the content and not the contention that counts.

Dorn raises one substantive issue also mentioned en passant in Dr. Bernstein’s letter. They point to the undoubted power that a knowledge of DNA sequences has provided to genetic counseling and to the identification, in utero, of potentially debilitated fetuses. There is no question that prenatal diagnosis and selective abortion have reduced the sum of human misery and that this is a positive consequence of molecular biology. But I wrote of cure and disease prevention. To conflate, as Dorn does, the prevention of disease with the prevention of lives that will involve disease, is to traduce completely the meaning of preventive medicine. It would lead to the grotesque claim that the National Socialists did more to “prevent” future generations of Tay-Sachs sufferers than all the efforts of science to date. Genetic counseling and selective abortion are substitutes for disease prevention and cure. Their use is a sign, not of the success of molecular biology in producing a scientific medicine, but of its failure.

Dr. Bernstein adds an interesting variant, of which I was unaware, to the uses of DNA sequences for screening purposes, namely to detect a class of virus infections in AIDS patients. Neither this, however, nor the use of recombinant DNA technology to help make commercial quantities of the drugs he mentions, have anything to do with “basic research into the role of genes in disease.” A revealing shift in the grounds of the argument has taken place here, a shift that is characteristic of the polemic on molecular biology. The claim in my review was quite specific. I asserted that the DNA sequence of a gene whose altered form is a cause of disease does not contain the information we need either to cure or prevent the disease. Further, that we have no case where a knowledge of the gene sequence has led to a treatment. No counterexample is offered by those who are made unhappy by this assertion, but instead all sorts of other uses for DNA sequencing are mentioned, as if DNA sequencing as a technique and molecular biology as a science were under total attack.

But this attack is nonexistent. I make a living as a molecular geneticist. My laboratory has expensive machinery for sequencing DNA, an operation we perform every day of the week, because the knowledge about genes that DNA sequencing and genetic engineering have produced represents the single most important advance in population and evolutionary genetics in the history of these sciences. A similar claim can be made about the elucidation of the network of molecular signals in the early development of organisms. And then there are all the commercial uses of DNA technology in industrial and agricultural production. But none of this has changed that fact that Tay-Sachs disease is still incurable, that cancer can be treated only by a gross attack on the cancerous tissue, and that no one can tell us how they will use a knowledge of the relevant DNA sequences to cure or prevent these or any other disorders in living human beings.

The other way in which Dr. Bernstein’s letter shifts the ground, although imperceptibly to the lay reader, is his use of the word “function.” It is certainly true that in relatively few cases genes may contain short stretches of DNA that have been seen in other genes, “motifs” that suggest that the protein coded by the gene is imported into the cell nucleus or that it becomes associated with the DNA of other genes as a kind of signaling device or that it is likely to be associated with cell membranes in some way. But these properties are a long way from what we mean by “function” in the physiological and developmental sense. Whether the product of such a gene is related to memory, blood pressure, cancer, or the color of my eyes, whether it influences the shape of a cell or whether the cell divides or dies, cannot be judged from the presence of such motifs. An example is a large class of proteins, the G-protein receptors, that can be seen from their sequences to be associated with cell membranes, but whose functions turn out to include color vision, olfaction, immune response, blood clotting, and the reproductive cycle.

It is also true that many genes have arisen in evolution by the duplication of other genes and then have diverged in DNA sequence from their parental forms. Sometimes, when this divergence has not been too great we may guess the function of one of the genes from a previous knowledge of the function of its duplicate. But often we cannot, even when the similarity of sequence of the two genes is obvious. A colleague of mine has discovered a new gene that has clearly been made from a large piece of a different gene, one that codes for an enzyme that splits alcohol. The product of the new gene, however, has a totally different function, as yet undetermined. Even worse, genes may diverge from each other in evolution until finally they have no sequence resemblance at all, yet the protein they code retains exactly the same function. The most famous case is an enzyme, lysozyme, which has totally diverged in sequence between yeast and chickens but is known, from looking at organisms of various intermediate degrees of evolutionary relationship, to be part of a single evolutionary sequence. Despite the total sequence divergence, the three-dimensional folded structure of the lysozyme protein is remarkably similar in all organisms and this, presumably, accounts for the conservation of function.

This folding cannot be predicted from the DNA or protein sequence. Indeed, a major barrier to getting useful knowledge from gene sequences is that even the most sophisticated theories of the chemical physics of large molecules do a bad job of predicting the folding of proteins from their amino acid sequences. The notion that with a big enough computer and the complete genome sequence of an organism we will know what all that DNA is being used for belongs just after astrology in Sagan’s list.

The ground has shifted yet further in Dr. Krauss’s letter. A byproduct of viral oncology may indeed have been the discovery of cancer-inducing gene mutations. It would be extraordinary if all the money and work put into cancer virology had been a total waste, and I never suggested that it was. Once again, a specific criticism has been perceived as a general nihilism. Whatever value may have come secondarily from the concentration on viruses as the cause of cancer, the much-proclaimed program to get rid of cancer by treating it as an infectious disease wasted away, died, and was buried in a remote corner of the cemetery without a public funeral. The heirs simply took the cash from the estate and invested it in another enterprise.

No one can deny that whatever information can be obtained about one bit of the physical world has ramifications for the knowledge and manipulation of other parts. Often these ramifications are unexpected. Even more often, especially in those particularly complicated parts of the physical world constituted by living organisms, the knowledge that is accessible is separated from what we really want to understand by very hard questions whose answers we do not know how to obtain. At one time or another in their lives most scientists realize the extent of this separation but they also perceive, probably correctly, that public faith and concrete investment in science depend on a connection between knowledge and result that is much simpler. So, they are led to make extravagant claims for the operation of science, both for the objectivity and compelling power of a formulaic “scientific method,” and for the direct applicability of elementary knowledge to problems of human welfare. When challenged, they throw up an obfuscating cloud of quite interesting and sometimes even quite useful results of scientific investigation, in the hope that no one will notice that the original problem has not been solved, or that it has, but by a pathway quite unrelated to what they have been doing.

Certain cells of the body, the T lymphocytes, are part of the immune reaction to cancer and are active in their metabolism when in contact with cancerous tissue. It has recently been reported that, by genetic engineering, it has been possible to make T lymphocytes that secrete a cell toxin that will kill other cells invoking this immune response. Thus, “killer” lymphocytes will specifically kill the tumor cells even in very small tumors, but not harm normal tissue. It might even be effective against viral and immune disorders. If this trick actually works clinically it will be hailed, quite correctly, as a triumph of molecular biology in the war on cancer. The wonderful irony is that cancer will have been cured using molecular biology, but without knowing anything new about basic cell biology, genetics, and the causes of cancer. The method is simply a very clever extension of what we already do to treat cancer, poison it.

This Issue

March 6, 1997