Genes, Memes, & Minds

As an evolutionary biologist, I am used to being misunderstood by philosophers. Even my favorite philosopher, Karl Popper, although he later repented, argued for many years that evolution theory is metaphysics rather than science. It is therefore a pleasure to meet a philosopher who understands what Darwinism is about, and approves of it.

Dennett goes well beyond biology. He sees Darwinism as a corrosive acid, capable of dissolving our earlier belief and forcing a reconsideration of much of sociology and philosophy. Although modestly written, this is not a modest book. Dennett argues that, if we understand Darwin’s dangerous idea, we are forced to reject or modify much of our current intellectual baggage—for example, the ideas of Stephen Jay Gould, Noam Chomsky, Jerry Fodor, John Searle, E.O. Wilson, and Roger Penrose. As it happens, he is not the first to see how farreaching are the effects of Darwin’s idea. Darwin himself wrote, “He who understands baboon would do more toward metaphysics than Locke.” It is a remarkable fact that these words were written in a private notebook when Darwin was still seeking a theory of evolution.

Dennett’s central thesis is that evolution by natural selection is an algorithmic process. An algorithm is defined in the OED as “a procedure or set of rules for calculation and problem-solving.” The rules must be so simple and precise that it does not matter whether they are carried out by a machine or an intelligent agent; the results will be the same. He emphasizes three features of an algorithmic process. First, “substrate neutrality”: arithmetic can be performed with pencil and paper, a calculator made of gear wheels or transistors, or even, as was hilariously demonstrated at an open day at my son’s school, jets of water. It is the logic that matters, not the material substrate. Second, mindlessness: each step in the process is so simple that it can be carried out by an idiot or a cogwhell. Third, guaranteed results: whatever it is that an algorithm does, it does every time (although, as Dennett emphasizes, an algorithm can incorporate random processes, and so can generate unpredictable results).

How can it be that such a mindless process can generate such wonderful results? In particular, how can it have produced us? Before Darwin, it was the accepted opinion of both philosophers and biologists that the complex adaptations of living things implied an intelligent designer. The essence of Darwin’s dangerous idea is that adaptations can arise by natural selection, without need of intelligence: that is, they can be the products of an algorithmic process. Dennett repeatedly uses the analogy of “cranes” and “skyhooks.” These are both devices for lifting things—in evolution, for generating increasingly complex designs—but of very different kinds. A crane is a structure or process which is itself the product of the natural selection of replicating entities, but which, once it has arisen, makes it possible for still more complex structures to evolve.

Two examples will make the point clearer. The first populations of replicating entities lacked sex: that is, there was no way in which different replicators could unite to form a new individual. Once sex did arise, it greatly accelerated the evolutionary process. Sex is, in Dennett’s terminology, a crane. Sex did not arise because it would accelerate evolution in the future: natural selection does not have foresight. Indeed, there is still controversy among evolutionary biologists about how and why sex did originate, although I think that a plausible account is now possible.

To give a second example of a crane, many animals learn by experience. The brain is as much a product of natural selection as the liver or the kidney. In the simplest animals it serves mainly to generate fixed responses to external stimuli. Once, however, the connectivity of the neurons in an individual can be modified by experience, an animal will alter its behavior in a useful way as a result of that experience. It is also possible to program a computer so that the strength of the internal connections alter with experience; such computers are surprisingly good at learning. As B.F. Skinner pointed out, trial-and-error learning is an exact analogue of evolution by natural selection. Dennett agrees with Skinner about this, although rejecting much else that he said. The important point is that the brain is, in Dennett’s words, a crane. It evolved by natural selection, but, once evolved, it made the evolution of further complexity possible.

Dennett’s view of evolution, then, is one of cranes building cranes building cranes, each new crane arising by an essentially mindless process of selection. I fully agree with this view. Indeed, my recent book with Eörs Szathmáry, The Major Transitions in Evolution,¹ is an account of this succession of cranes, starting with the origin of the first replicators and the genetic code, and ending (so far) with the origin of human language. The problem is to explain how each new crane arose by a process of selection, without miracles, or “skyhooks.” Skyhooks are a stark contrast. They are mysterious lifting devices, whose origin cannot be explained. It is Dennett’s thesis that we must eschew skyhooks and make do with cranes.

In addition to the Darwinian notion of natural selection, there are two other sources of Dennett’s view of the world. The first is a set of ideas that includes computing science, artificial intelligence, and cognitive science: it is from here that he acquired his conviction that algorithmic processes can generate mind-like activities. For example, computers, following simple rules, can design railway timetables or play chess, tasks once thought to require the operation of the human mind. The second is the gene’s-eye view of evolution pioneered by G.C. Williams and Richard Dawkins. According to this view, evolution is a necessary consequence of the existence of replicating entities; in biology, those entities are genes, but the principle holds for any kind of replicators. I have thought for some time that Dawkins and Williams have made a more fruitful contribution to philosophy than most philosophers, and I am pleased to see this opinion so generously recognized.

Dennett uses one other general idea which, perhaps because of my personal history, I find particularly appealing. This is the idea of “reverse engineering.” Usually, engineers start with a function they wish to perform, and design a structure to do the job. Biologists often find themselves confronted by a structure and ask themselves what function it was selected to perform. Harvey’s discovery that the heart is a pump was an early triumph of such reverse engineering.

As it happens, engineers occasionally do reverse engineering. In 1945, when I was working in an aircraft design office, the Royal Aircraft Establishment at Farnborough put on an exhibition of recently captured German equipment. A friend and I spent two fascinating days looking at the equipment and asking ourselves, “Why did they do that?” Both the V1 flying bomb and the V2 rocket were on show, and we spent some time trying to figure out how they worked. I seem to remember that the V2 rocket had a gyroscope puzzlingly connected to the fuel supply to the motors; surely, one would think, it should be connected to the guidance system. I leave it as an exercise to any readers who fancy themselves as reverse engineers to work out why, if my memory is correct, it was connected to the fuel supply. Since I became a biologist, I have spent most of my time asking questions like that.

Of course, when thinking about the V2 rocket I was thinking about a product of human design, whereas, a few years later, when I was thinking about the shapes of mammalian teeth, I was asking why mammals were better at chewing, and so left more descendants. But this difference had no effect on the way I thought about the two problems. Indeed, I have become increasingly convinced that there is no way of telling the difference between an evolved organism and an artifact designed by an intelligent being. Thus imagine that the first spacemen to land on Mars are met by an object which appears to have sense organs (eyes, ears) and organs of locomotion (legs, wings). How will they know whether it is an evolved organism, or a robot designed by an evolved organism? Only, I think, by finding out where it came from, and perhaps not even then.

Dennett suggests that criticisms of the neo-Darwinist synthesis come, in the main, from those who are reluctant to believe that they are the product of an algorithmic process and who lust after skyhooks. First among these, he suggests, is Stephen Jay Gould. Gould occupies a rather curious position, particularly on his side of the Atlantic. Because of the excellence of his essays, he has come to be seen by non-biologists as the preeminent evolutionary theorist. In contrast, the evolutionary biologists with whom I have discussed his work tend to see him as a man whose ideas are so confused as to be hardly worth bothering with, but as one who should not be publicly criticized because he is at least on our side against the creationists. All this would not matter, were it not that he is giving non-biologists a largely false picture of the state of evolutionary theory.

There are, Dennett suggests, three main aspects of Gould’s thought which reveal a wish to escape from Darwin’s algorithmic grip. The first is his critique, with Richard Lewontin, of the “adaptationist paradigm.” I have some responsibility for this critique. As organizer of a symposium in London on adaptation, I invited Lewontin, as a well-known critic of naive adaptationist arguments, to contribute. Lewontin, for reasons that, as an exaircraft engineer, I well understand, dislikes flying, and suggested that he write a joint paper with Gould, which Gould would present. The result was the now-famous paper entitled “Spandrels of San Marco.” Its thesis is that many structures in the animal world are not adapted for any function, but, like the spandrels of San Marco, are accidental and unselected consequences of something else. Further, they argued, many adaptive explanations are “Just So Stories,” unsupported by evidence.

By and large, I think their paper had a healthy effect. There are plenty of bad adaptive stories: we can all laugh at the suggestion that flamingos are pink because it camouflages them against the sunset. Their critique forced us to clean up our act and to provide evidence for our stories. But adaptationism remains the core of biological thinking. Confronted with feathers, or eyes, or ribosomes, we cannot not ask what they are for. It would be no more plausible to suppose that they are accidental and non-selected byproducts of something else than it would be to suppose that the gyroscope in the V2 rocket was connected as it was because some German fitter made a mistake.

I cannot resist being autobiographical. My own first scientific papers, written while I was turning myself from an engineer into a biologist, concerned animal mechanics. Why do primitive flying animals have long tails? Why is the jaw articulation in carnivores level with the tooth row, but in most herbivores above it? Why are the gaits of mammals as they are? Back in 1950, it was hard to get such papers accepted for publication. Editors did not like functional explanations. I had particular difficulties with my paper on gaits, which may have been the first genuine “optimization” paper in biology. To find the optimal gait, the one that would minimize energy expenditure at a given speed, I wrote down a differential equation: (dw/dj=o). The editor, offended by this attempt to sully a respectable biological journal with mathematics, asked “Why don’t you cancel the d’s?” I fear that this story will amuse only some readers, but I want to get it on record.

Since those early days, I have spent much time studying structures and processes which might be thought to raise difficulties for an adaptive view of life—for example, aging, sex, ritualized behavior. As Theodosius Dobzhansky put it, none of these processes makes any sense except in the light of evolution. Essentially Dennett is right: if there are philosophers out there who think that biology can do without functional explanation, they are just wrong.

Two other Gouldian themes, punctuated equilibria, and the non-repeatability of evolution, can be dealt with more briefly. The tale of punctuated equilibria is an odd one. Its factual basis, commonly reported by paleontologists, is that lineages often change very little for millions of years, and then change rather rapidly. When the idea was first put forward by Gould and Niles Eldredge, it was presented as just what one would expect to see if the orthodox view, that species often arise by rapid evolution in small peripheral populations, is indeed accurate. If only they had left the argument there! Their paper would then have been seen as a useful extension of the picture given in Tempo and Mode in Evolution by George Gaylord Simpson, which was the Darwinian orthodoxy when I was a student. Sometimes, however, Gould appears to be saying that the changes, when they occurred, were not the result of natural selection, but of some other process—genetic revolutions, “hopeful monsters” (large mutational changes), or what you will. Since “sudden” in the fossil record means thousands of generations, there is no reason whatever for supposing any such thing.

The non-repeatability of evolution—the idea that if evolution were to happen again from the same starting point, it would not repeat itself—is true, but not new: it is what most scientists have always thought. But what, Dennett asks, is the significance of these various reservations—anti-adaptationism, punctuated equilibria, non-repeatability? The answer, he suggests, is that Gould is trying to escape from an algorithmic explanation of life. Dennett may be right.

The natural selection of replicators—essentially, of nucleic acid molecules—may explain the evolution of animals and plants, but what about humans? We, surely, are more than the product of our genes. Indeed we are, admits Dennett, but it does not follow that we are anything other than the products of an algorithmic process. At this point, he embraces Dawkins’s notion of a meme. A meme can be anything from the limerick about the young man of Belgrave (mutated in the US, I’m told, to a young fellow called Dave) to the doctrine of the Trinity. A meme is an idea that can lodge in a person’s mind, and can be transmitted, in print or by word of mouth, to other minds. In other words, it is a replicator. What is peculiar about humans is that they can hold ideas in their heads, and transmit them to others: they provide an environment in which a new kind of replicator, memes, can evolve. The human mind is another example of a crane. It evolved by natural selection, without need for an intelligent designer. Once evolved, however, it provides a medium in which a new kind of evolution by natural selection can occur, involving a new kind of replicator, the meme.

My uneasiness with the notion of memes arises because we do not know the rules whereby they are transmitted. A science of population genetics is possible because the laws of transmission—Mendel’s laws—are known. Dennett would agree that no comparable science of memetics is as yet possible. His point is a philosophical rather than a scientific one. We see humans as the joint products of their genes and their memes—indeed, what else could they possibly be?—even if we have no predictive science of meme change. Once a human mind capable of harboring memes evolved, a new kind of evolution, cultural evolution, became possible, more rapid by far than genetic evolution.

What is needed for the harboring and transmission of memes? Essentially, it is language. The past thirty years has seen a debate on the nature of language. For Skinner, the ability to learn a language was just an aspect of our general learning ability. For Chomsky and his students, it is a special faculty, both in the sense of being peculiar to humans and of being peculiar to language. Dennett accepts, and I agree, that this argument has been won by Chomsky: there is indeed a special “language organ” that enables children to learn to talk. I see Chomsky, and I think Dennett would agree, as one of the half-dozen commanding intellects of this century.

I therefore find Chomsky’s views on evolution completely baffling. If the ability to learn a language is innate, it is genetically programmed, and must have evolved. But Chomsky refuses to think about how this might have happened. For example, in 1988 he wrote, “In the case of such systems as language or wings, it is not easy even to imagine a course of selection that might have given rise to them.” This is typical of his remarks on evolution. There is, in fact, no difficulty in imagining how wings might have evolved. Language is difficult because it leaves no fossils; it has evolved just once (unlike wings, which have evolved at least four times); and there is an enormous gap between the best that apes, whales, or parrots can do and what almost all humans can do.

Happily, some Chomskian linguists, notably Steven Pinker, are taking up the challenge.² It is not hard to think of functional intermediates between ape language and human language, but it is hard to decide what were the actual intermediates. Perhaps more interestingly, new kinds of organs—and the language organ is certainly new—do not usually arise from nothing, but as modifications of preexisting organs with different functions. Teeth are modified scales, legs are modified fins, and, after complex transformations, ears are modified parts of the lateral line organs of fish. What was the language organ doing before it acquired its present function?

The question may not be unanswerable. The best chance may lie in genetic analysis. The Canadian linguist Myrna Gopnik has identified one human gene which, if mutated, causes a limited but specific deficiency in grammatical competence.³ There is a real chance that genetic analysis will, in time, reveal the nature and origin of the human language organ, just as it is already revealing how animal form appears during embryological development.

Why does Chomsky not wish to think about evolution? Dennett, who is as puzzled as I am, has an interesting idea. Chomsky, he suggests, would readily accept an explanation of linguistic competence in terms of some general physical law, but not in terms of messy, ad hoc, contingent engineering design, which is the best that natural selection can do. If so, he is not alone in his taste for general, elegant explanations. My friend Brian Goodwin, the developmental biologist, cannot bear the idea that the explanation for development may be a series of ad hoc contrivances, and another friend, the Japanese evolution theorist Motoo Kimura, sadly now dead, once rejected an idea of mine with the words, “It is possible, but it would be so inelegant.” But I fear that the world is inelegant. There is a lesson which Chomsky’s students, if not the great man himself, will have to learn. Science is a unity. Biology cannot ignore chemistry, much as I wish it could; for the same reason, linguistics cannot ignore biology.

One last point before leaving language. In 1989, Pinker and his student Paul Bloom presented a paper at MIT entitled “Natural Language and Natural Selection,” which argued that the origin of the language organ, like that of other organs, must be explained in Darwinian terms. Dennett tells us that it was the “level of hostility and ignorance about evolution that was unabashedly expressed by eminent cognitive scientists on that occasion” that persuaded him that he could no longer put off writing the present book. I am delighted that he was provoked.

Dennett is critical of sociobiology, or at least of its application to humans. He accuses it of “greedy reductionism,” of trying to reduce human behavior too directly to biology. Our behavior is affected by memes as well as by genes, and the attempt to explain it, as biologists explain much of animal behavior, as a direct adaptation ensuring gene transmission is therefore unjustified. However, he is more sympathetic to the new wave of sociobiology represented by evolutionary psychology. The most interesting claim made by evolutionary psychologists is that the mind contains specialized modules that evolved to perform particular tasks.

This is obviously true of that part of the brain concerned with analyzing visual input, and, if Chomsky is right, it is true of language. It has been proposed that there are also modules concerned, for example, with the detection of cheating and with the identification and classification of living organisms. Of course, even if there are such modules, they cannot be completely isolated. In science, as in other fields, progress often depends on seeing analogies between apparently different processes. For example, my own main contribution has been to see the analogies between human games and the things that spiders, trees, and even viruses do. This would not be possible if the mind consisted of isolated modules. Although he is attracted by the notion of modularity, Dennett warns that the mere fact that humans in different societies behave in similar ways cannot be taken as evidence of genetic determination. People may simply be doing what any intelligent being would do in the circumstances: making a forced move in design space, to use his terminology.

A potentially serious challenge to his position is posed by an argument, put forward by Roger Penrose and others, that Gödel’s theorem can be used to show that human intelligence cannot be algorithmic. The argument goes as follows—Gödel proved that there exist, within any non-contradictory mathematical system, some true statements that cannot be proved. Yet human mathematicians can intuit the truth of some such statements. Since anything that can be proved can, in principle, be proved algorithmically, it follows that humans can do something that algorithms cannot. Dennett replies that, although there is no algorithm that can prove a given statement to be true, there may well be algorithms that can suggest statements that are very probably true. Humans, perhaps, use algorithms of the latter kind. Their intuited mathematical truths may just be very good guesses. By analogy, a computer programmed to play chess cannot, with certainty, find the best possible move, but it does find very good moves. Dennett’s argument on this point should be read with care. I am not sure I have understood it correctly, but I like it, partly because I cannot see what else human intelligence could be, other than algorithmic, and partly, perhaps, because while I am rather good at having mathematical intuitions, I have learned that they are sometimes wrong.

Dennett’s last topic is the evolution of morality. Here it is important to distinguish two questions: “How could humans come to have a sense of right and wrong?” and “What is right and what is wrong?” I do not think the first question is all that difficult. I would expect any intelligent organism that lives in groups to evolve an ability to hold beliefs about right behavior, and to be influenced in those beliefs by myth and ritual. We do not only have beliefs: we make contracts. It is worth asking what cognitive equipment is needed to make a contract. At the very least, it requires language and a “theory of mind”: that is, we must be able to perceive other people as beings like ourselves, with minds like ours. Both these qualities are probably unique to humans.

But is there any way in which we can decide, with certainty, which actions are right? Dennett’s view, which I share, is that there is not, unless you hold that some book, for example the Bible, is the word of God, and that human beings are here to do God’s bidding. If a person is simply the product of his or her genetic makeup and environmental history, including all the ideas that he or she has assimilated, there is simply no source whence absolute morality could come. Of course, this does not exempt us from making moral judgments: it only means that we cannot be sure that we are right.

At the start of his book, Dennett says that he aims to persuade us not only that the world is free of skyhooks, but that we can live happily in such a world. In the last two chapters, he tries to deliver on this second promise. He is surprisingly successful. In essence, he says that, no matter how mindless the processes of evolution may be, they have, in fact, produced a world of astonishing diversity and beauty, which we can enjoy, and ought to protect. It is a conclusion that echoes the final words of the The Origin of Species: “…from so simple a beginning, endless forms most beautiful and most wonderful have been, and are being evolved.”