Interview with Richard Dawkins


Professor Dawkins, your books contain both discussions of theory and descriptions of extraordinary plants and animals, but which of these aspects was it that first brought you to biology?


It was not so much the wonderful natural history but the philosophical questions, such as how biology may help us to answer questions like “Why are we all here?” An interest in natural history for me came afterwards.


Your first, highly influential book, The Selfish Gene, was published in 1976. In a much-quoted passage from the book, you say that we are “survival machines” for our genes. Your use of terms such as “machines” and “lumbering robots” has caused negative reactions in some quarters.


When one uses language like that to make a certain point, it is often misunderstood to be making a different point. The first misunderstanding is that machines or robots are simple. When I used the analogy of the “lumbering robots” I had just returned from a wonderful conference on Artificial Intelligence which had greatly inspired me, so the last thing on my mind was crude simplicity! By using the analogy, I wanted to convey the idea that the body is a mechanism that is pre-programmed, in this case by the genes rather than by human programmers. We do not think of a robot as having subjective awareness, or subjective deliberation; we think of it as being programmed with however much complexity it needs to cope with all the eventualities that may be thrown at it. The idea of genes as purposeful, conscious entities was the last thing I wanted to suggest. (It has been taken that way.) The way the genes work is to act as programmers, pre-programming robot machines which are very complicated and which do take highly flexible action for the benefit of their programmers. The flexibility comes in the machine, not in the programmers, which are just lengths of DNA.


Talking about robots and machines takes us to the idea of digitization. In Climbing Mount Improbable, you say that elephant DNA is like virus DNA in being fundamentally a “Duplicate Me” computer program, but with a large digression—the elephant itself. Presumably, since the discovery of the structure of DNA in 1953, the whole idea of digitization and the development of computers have profoundly influenced thinking in evolutionary biology.


Yes, post-1953 molecular biology is all digital. If you look at a journal of molecular biology, it is just like a journal of computer science, differing only in the details. Huge masses of digital data that comprise the genome of this animal or that plant are presented, with chunks of DNA being cut, copied, pasted, transposed—just like computer data. That is one point. The comment about the elephant and the virus is a further point. The analogy there is, in the first place, to a computer virus—a piece of computer code which says “copy me” and which may incidentally say something spiteful like “erase all the data off the hard disk”, but most importantly it says “copy me”. It may do some clever things to get itself copied because people take steps to avoid being infected, so ingenious programmers develop more and more effective viruses and worms to get around the defences. The elephant is an enormously elaborate get-around. It is still a “copy me” program. Bacterial DNA copies itself by the most direct way. Given that most of the DNA in the world is bacterial DNA, all the obvious ways of getting copied are filled. There are a few less obvious, indirect ways, like making an elephant, making a kangaroo, or making an oak tree, and they are glorious and bewildering in their complexity. But fundamentally what they are saying is “copy me”.


You have also used computers to simulate evolution, with your Biomorph program, described in The Blind Watchmaker. Computers have been effective at showing how rapidly organs such as the eye can develop through natural selection, but people are now using simulated computer worlds as substrates to develop artificial life. What are your thoughts on this?


When I developed Biomorph, a long time ago now, in 1986, it was purely didactic. It was to show what surprising power there is in the simple random-mutation-followed-by-non-random-survival algorithm. In this case, I used the human eye as the selecting agent. Other people have carried on and produced more beautiful and more elaborate artificial selection programs. Artificial life is more interesting still, because people are trying to simulate other aspects of life—not just to evolve entities that appeal to a human selector, but to develop artificial organisms that survive and evolve, in a certain circumscribed sense, in an environment, a world defined in the computer model. Such models usually involve an artificial physics, an ecology with scarce resources, and a genetics. There is then evolution towards adaptations to survive in that world. Of course, nobody has yet produced a computer world that is even remotely as complicated as the real world, but each model is a partial approximation, and each tells you something different. I think, therefore, that artificial life is a very illuminating field of study. To me, its value lies in stimulating human thought. Rather than being a model of exactly how life has arisen on this planet, it is a way of freeing our minds to think about what in general has to be true wherever life is found. In advance of going to other planets, artificial life is a mind-expanding way to think about that.


You refer to life on other planets. Life arose, or only needed to arise once, some 4 billion years ago, on the Earth. We have now found many extrasolar planets, and we are even beginning to be able to identify elements in their atmospheres. How common do you think life might be in the universe?


My gut feeling is, much the same as that of other people, that life is probably not unique to this planet. I suspect that it is moderately rare; it could be so rare that, because the universe is so large, the islands of life may be so separated from each other that no form of life ever encounters any other, which would be rather sad. In The Blind Watchmaker I made a sort of logical case that it is possible that life on this planet is unique. My point was that it is a false argument to say that because life has arisen here, it cannot be all that improbable, that therefore there must be life elsewhere. If the origin of life really were gigantically improbable, so that it has only arisen on one planet in the universe, then that planet is this one, because here we are talking about it. By the same token, if life has only arisen once, we should not be in the least surprised if chemists fail to recreate the conditions for the origin of life—that is, of self-replicating molecules. It would be better to approach the question the other way round, by asking chemists how improbable the origin of life is, based on chemical reasoning.


The only other piece of data that we have to go on, which is weak but it is there, is that we have not apparently been visited, neither physically nor by radio. We can therefore use that as an additional statistical argument to say that if life were above a certain threshold of commonness, then by now we should expect to have been visited, if only by radio communication. Here we must also consider the probability of going from life at all to life complicated enough to invent space travel or invent radio. Mark Ridley, the evolutionary biologist, suggests in his book Mendel’s Demon that origins of life are common, but complex life is rare. If you have a rationale for thinking that, then you could argue that the universe is teeming with life, but not with intelligent life, so it is not surprising that we have not yet made contact.


Coming back to our planet, which is teeming with life, in The Extended Phenotype, the most scientifically sophisticated of your books, you talk about the “long reach of the gene”. How long is the reach of the gene?


This is a sort of verbal or conceptual trick. The phenotype of the gene [the physical effects it produces] is normally thought of as being limited to the body in which the gene sits. That is because we are thoroughly accustomed to the idea that life comes in discrete packages called bodies. We have the idea that a gene’s phenotypic expression has a definite and limited range. The idea of the extended phenotype, or the long reach of the gene, is that the phenotypic effect of the gene can reach outside the body in which it sits. The way I approach it is by a verbal argument which gradually leads the reader further and further away from familiar territory.


The first step is to consider artefacts made by animals, such as birds’ nests. A bird’s nest, especially a beautiful and well-fashioned nest like that of a weaver bird, is very clearly a Darwinian adaptation—it is fashioned as a biological object to do a biological job. There are no genes in that nest. The genes which help to fashion it, that were naturally selected to perfect that adaptation, are in weaver bird bodies. The adaptation is a product of natural selection on genes that are not part of it. The phenotype is an extended phenotype of weaver bird genes.


The next step is to consider phenotypic expression in other bodies, for example, in the case of parasites that live inside their hosts, which can manipulate the body of the host for their own benefit. Flukes in ants cause ants to change their behaviour in such a way that the ants are more likely to be eaten by sheep, their next host. The change in the ants’ behaviour is a phenotypic expression of fluke genes. The justification for saying that is that it is a Darwinian adaptation, and Darwinian adaptations can only arise through natural selection of genes.


We can then consider the phenotypic effects of genes of parasites that do not live inside the bodies of their hosts—the effect of a baby cuckoo on its foster parent, for example. The foster parent’s behaviour is, by the same Darwinian argument, a phenotypic expression of cuckoo genes. Now we have reached the point where genes in one body can reach out to a long-distant phenotypic expression. A nightingale singing in the wood is influencing other nightingales, changing their hormonal state, causing their ovaries to grow. This is phenotypic expression in a female nightingale of a male nightingale’s genes. The long reach of the gene can be hundreds of yards.


Clearly, genes also affect human behaviour and emotions. But we come again, here, to an area that has been widely discussed and has been controversial. What, in your view, is the role of Darwinism when it comes to human culture and society?


That should really be separated into two quite different levels. Darwinism working on genes produces the brains that we all have, without which there would be no culture. There is, then, ordinary gene-level Darwinism going on, but that is seldom helpful in explaining differences in human culture. When we marvel at the differences in cultures in New Guinea, South America, and so on, it is very unlikely that looking for genetic differences will illuminate those cultural differences, which from a naive Darwinian point of view is a bit of a pity. To help understand these, we should, instead, focus on the similarities among genetically programmed brains that are susceptible to cultural influences. What all people have in common is a tendency to develop cultures, which from a biological point of view means things like the tendency to imitate your parents and peers. Given that brains are set up in such a way, you would expect that in different parts of the world, different cultures would arise, because of arbitrary local differences that become reinforced. That is a slightly more helpful way of using Darwinism to account for culture, but here we are not accounting for differences directly but indirectly.


There might also be a kind of non-genetic Darwinism at work, applying to the things being imitated in ways analogous to genes, for example beliefs, copied habits, or copied dress fashions that might have high survival value. Here, survival value emphatically does not mean ability to help the individual survive. As in the gene’s-eye view, it means ability to help itself survive. If there is a particular belief or a particular fashion in clothes which has what it takes to survive—if, for example, it is compatible with some other aspect of the ethos of the local culture—then, just as genes tend to be favoured if they are compatible with other genes already in the gene pool, leading to co-adapted gene complexes, so co-adapted complexes of cultural units of inheritance can result. There is an analogy of these so-called “memeplexes” to co-adapted gene complexes, but we are not talking about genes at all now.


In River Out of Eden, you have used the metaphor of the digitized river of life, split into some 3 billion branches to represent species, branching out when species divide, the waters then becoming totally separate. We are now in a situation technologically where we can pick up water from one river and put it into another, something that does not happen, by definition, in the natural world—in other words, taking a gene from one species and putting it in another. What do you feel are the risks and benefits of such activity?


This is very new. Genetic engineering makes it possible to mix gene pools of different species, and there are famous examples such as the antifreeze gene in Arctic fish that has been imported into plant genomes to prevent frost damage. From an academic point of view, this is extremely interesting. It illustrates the point that genes are equivalent to subroutines in a computer language, and a subroutine for doing a particular job like antifreeze will work in other programs. It is eloquent testimony to the universality of the genetic code that it is possible to take subroutines from one organism and insert them in a radically different kind of organism. It is potentially enormously valuable, because so much R&D work has been done already by natural selection. Because natural selection on fish has done the R&D work on antifreeze it saves an enormous amount of trouble simply to copy and paste the subroutine into the tomato from the fish.


What about harm? There is an enormous amount of hysterical hype about this, which shows itself in words like “Frankenfoods”. This is a gross misunderstanding of what is going on. Without understanding that genes are just subroutines written in the same computer language whichever organism they come from, people can have the idea that a gene that comes from a fish must in some sense be a “fishy” gene, that somehow you are contaminating the tomatoes with fishiness. You are not, just as you would not be contaminating a financial spreadsheet with rocket science because the programmer borrowed the square-root subroutine from a rocket science program.


Are there dangers? There can be dangers in anything new, and it is always important of course to test everything. But I do not think there is any general reason to believe that importing genes from widely separated gene pools is, as a general rule, likely to be any more dangerous than any of the other things we do, when we try any new medicines, for example.


Is there, then, any general misgiving we ought to have about importing genes? Possibly. I can illustrate by an analogy. When British colonists to New Zealand released animals from the European fauna into the wild, they contaminated the unique fauna of those islands irrevocably. Another example is the American grey squirrel, which has become rampant across Britain. Nowadays there would be all hell to pay if anyone did such a thing, because our consciousness has been raised. Yet at the time, people had no idea they were doing something irresponsible. Conceivably, our descendants could look back on our first fumbling attempts to import genes from one gene pool into another and say “How could they do it?” I’m not sure quite what it would mess up, but then, the 19th-century colonists of New Zealand did not know what they were messing up.


But unlike subroutines, genes don’t generally act discretely, do they? They do operate in teams…


That is another point. Yes, they do act as teams. It is not obvious that the antifreeze gene would have worked in a tomato because it might have required other fish genes in order to have its effect. That is a matter of finding out whether it does and if necessary introducing the whole team. And I suppose it is also possible that the [isolated] subroutine, when put into a foreign gene pool, could have some bad effects.


In Unweaving the Rainbow, your latest book, the central theme is how much looking at the world with a scientific eye, developing a scientific understanding, enriches one’s experience of that world. We are supposed to be living in a scientific age now, and certainly we use a huge amount of technology that has been derived from our scientific understanding. How do you account for the continued need for your books—for a lack of understanding of evolution beyond immediate scientific circles? Is it that evolutionary theory, because you have to take a different perspective, is actually quite counterintuitive and hence difficult to grasp?


There are particular difficulties in understanding evolution. I do not think they are great. I think they are clearly less than the difficulties in understanding quantum theory or relativity. Quite apart from the difficulty of quantum theory, there is also the suspension of common sense that you have to undergo. You have to force yourself to accept highly counterintuitive beliefs in order to grasp quantum theory even partly. Evolution has a bit of that. You have to suspend your prejudice in favour of seeing things happen on the sort of timescale that the human brain is used to and accept that things can change on a much greater timescale. Geology has the same problem. But that is not a major qualitative problem for the mind to grasp.


There is also the negative influence of alternative dogmas which many people are brought up with. In some cases, in order to understand the scientific worldview, they have actively to overthrow the worldview with which they were brought up, and that can be quite painful. The sheer magnitude of what evolution has achieved, the beauty, the complexity, the elegance of life, is so colossal that it can be hard to get your mind around the idea that there is one simple explanation for it all. Again I suppose that comes down to the time involved. Anybody can, without much difficulty, understand how one species can change into another species that is a tiny bit different, with slightly longer legs or darker hair, or whatever. But to put all those little steps together, one after another, cumulatively, to produce the change from a bacterium to a human is something that challenges naive common sense—again, not such a blatant defiance as quantum theory, but it is still a bit difficult. Lewis Wolpert, the distinguished embryologist, wrote a book called The Unnatural Nature of Science where he in a way contradicts Huxley’s famous statement that science is nothing but organized common sense. Wolpert says in fact that it is not common sense at all, that it is inimical to common sense.


It certainly might explain why science is regarded as difficult at school.


I think there are two sorts of difficulty. It is difficult in the sense of being hard work, which is different from being counterintuitive. Wolpert has a nice example of counterintuitive fact that you can work out mathematically, which is very intriguing. He points out that every time you drink a glass of water, the number of molecules in the glass of water is so huge compared to the number of glasses of water there are in the world, in all the water in all the seas and lakes and rivers, that the chances are very high that one molecule in the glass of water you are drinking passed through the bladder of Oliver Cromwell! (You can insert any name.) That is a nice example of a counterintuitive fact that, once you understand it, you can work out. But that is not hard in the sense that one says doing science at school is hard—it is intriguing and interesting. And I think if people were told more of that sort of thing, they might be more keen to do science.


BIBLIOGRAPHY
Dawkins, Richard, The Selfish Gene, New Edn. Oxford University Press, 1989. Dawkins, Richard, The Extended Phenotype, New Edn. Oxford University Press, 1999. Dawkins, Richard, The Blind Watchmaker, New Edn. Penguin Books, 2000. Dawkins, Richard, River Out of Eden. Science Masters Series, Orion, 1996. Dawkins, Richard, and Lalla Ward (illustrator), Climbing Mount Improbable. Penguin Books, 1997. Dawkins, Richard, Unweaving the Rainbow, New Edn. Penguin Books, 1999. Ridley, Mark, Mendel’s Demon. Orion Books, 2001. Wolpert, Lewis, The Unnatural Nature of Science. Faber and Faber, 1993.
Microsoft ® Encarta ® 2009. © 1993-2008 Microsoft Corporation. All rights reserved.