The Blind Watchmaker

by Richard Dawkins

  They usually begin by measuring everything they can about their animals. You have to be a bit clever about how you interpret these measurements, but I shan’t go into that. The end result is that the measurements are all combined together to produce an index of resemblance (or, its opposite, an index of difference) between each animal and each other animal. If you wish, you can actually visualize the animals as clouds of points in space. Rats, mice, hamsters, etc. would all be found in one part of the space. Far away in another part of the space would be another little cloud, consisting of lions, tigers, leopards, cheetahs, etc. The distance between any two points in the space is a measure of how closely the two animals resemble each other, when large numbers of their attributes are combined together. The distance between lion and tiger is small. So is the distance between rat and mouse. But the distance between rat and tiger, or mouse and lion, is large. The combining together of attributes is usually done with the aid of a computer. The space that these animals are sitting in is superficially a bit like Biomorph Land, but the ‘distances’ reflect bodily resemblances rather than genetic resemblances.

  Having calculated an index of average resemblance (or distance) between each animal and each other animal, the computer is next programmed to scan the set of distances/resemblances and to try to fit them into a hierarchical clustering pattern. Unfortunately there is a lot of controversy about exactly which calculation method should be used to look for clusters. There is no one obviously correct method, and the methods don’t all give the same answer. Worse, it is possible that some of these computer methods are over-‘eager’ to ‘see’ hierarchically arranged clusters within clusters, even if they aren’t really there. The school of distance measures, or ‘numerical taxonomists’, has become a bit unfashionable lately. My view is that the unfashionableness is a temporary phase, as fashions often are, and that this kind of ‘numerical taxonomy’ is by no means easily to be written off. I expect a comeback.

  The other school of pure-pattern measurers are the ones that call themselves transformed cladists, for reasons of history as we have seen. It is from within this group that the ‘nastiness’ mainly emanates. I shall not follow the usual practice of tracing their historical origins from within the ranks of true cladists. In their underlying philosophy, so-called transformed cladists have more in common with the other school of pure-pattern measurers, the ones often called ‘pheneticists’ or ‘numerical taxonomists’, whom I have just discussed under the title of average-distance measurers. What these share with each other is an antipathy to dragging evolution into the practice of taxonomy, although this does not necessarily betoken any hostility to the idea of evolution itself.

  What the transformed cladists share with true cladists is many of their methods in practice. Both think, right from the start, in terms of bifurcating trees. And both pick out certain kinds of characteristics as taxonomically important, other kinds of characteristics as taxonomically worthless. They differ with respect to the rationale that they give to this discrimination. Like average-distance measurers, transformed cladists are not out to discover family trees. They are looking for trees of pure resemblance. They agree with the average-distance measurers to leave open the question of whether the pattern of resemblance reflects evolutionary history. But unlike the distance measurers, who, at least in theory, are prepared to let Nature tell them whether she is actually hierarchically organized, the transformed cladists assume that she is. It is an axiom, an article of faith with them, that things are to be classified into branching hierarchies (or, equivalently, into nested nests). Because the branching tree has nothing to do with evolution, it need not necessarily be applied to living things. The methods of transformed cladistics can, according to their advocates, be used for classifying not just animals and plants but stones, planets, library books and Bronze Age pots. In other words they would not subscribe to the point I made with my library comparison, that evolution is the only sound basis for a uniquely hierarchical classification.

  The average-distance measures, as we saw, measure how far each animal is from each other animal, where ‘far’ means ‘does not resemble’ and ‘near’ means ‘resembles’. Only then, after calculating a sort of summed average index of resemblance, do they start trying to interpret their results in terms of a branching, cluster-within-clustery hierarchy or ‘tree’ diagram. The transformed cladists, however, like the true cladists that they once were, bring in clustery, branchy thinking right at the outset. Like true cladists, they would begin, at least in principle, by writing down all possible bifurcating trees, and then choosing the best.

  But what are they actually talking about when they consider each possible ‘tree’, and what do they mean by the best? What hypothetical state of the world does each tree correspond to? To a true cladist, a follower of W. Hennig, the answer is very clear. Each of the 15 possible trees uniting four animals represents a possible family tree. Of all the 15 conceivable family trees uniting four animals, one and only one must be the correct one. The history of the animals’ ancestors really did happen, in the world. There are 15 possible histories, if we make the assumption that all branchings are two-way branchings. Fourteen of those possible histories must be wrong. Only one can be right; can correspond to the way the history actually happened. Of all the 135,135 possible family trees culminating in 8 animals, 135,134 must be wrong. Only one represents historical truth. It may not be easy to be sure which one is the correct one, but the true cladist can at least be sure that not more than one is correct.

  But what do the 15 (or 135,135, or whatever it is) possible trees, and the one correct tree, correspond to in the nonevolutionary world of the transformed cladist? The answer, as my colleague and former student Mark Ridley has pointed out in his Evolution and Classification, is nothing very much. The transformed cladist refuses to allow the concept of ancestry to enter his considerations. ‘Ancestor’, to him, is a dirty word. But on the other hand he insists that classification must be branching hierarchy. So, if the 15 (or 135,135) possible hierarchical trees are not trees of ancestral history, what on earth are they? There is nothing for it but to appeal to ancient philosophy for some woolly, idealistic notion that the world just is organized hierarchically; some notion that everything in the world has its ‘opposite’, its mystical ying or yang. It never gets much more concrete than that. It certainly is not possible, in the nonevolutionary world of the transformed cladist, to make strong and clear statements such as ‘only one out of the 945 possible trees uniting 6 animals can be right; all the rest must be wrong’.

  Why is ancestor a dirty word to cladists? It is not (I hope) that they think that there never were any ancestors. It is rather that they have decided that ancestors have no place in taxonomy. This is a defensible position as far as the day-to-day practice of taxonomy is concerned. No cladist actually draws flesh and blood ancestors on family trees, though traditional evolutionary taxonomists sometimes do. Cladists, of all stripes, treat all relationships between real, observed animals as cousinships, as a matter of form. This is perfectly sensible. What is not sensible is to carry this over into a taboo against the very concept of ancestors, against the use of the language of ancestry in providing the fundamental justification for adopting the hierarchically branching tree as the basis for your taxonomy.

  I have left till last the oddest aspect of the transformed cladism school of taxonomy. Not content with a perfectly sensible belief that there is something to be said for leaving evolutionary and ancestral assumptions out of the practice of taxonomy, a belief that they share with pheneticist ‘distance measures’, some transformed cladists have gone right over the top and concluded that there must be something wrong with evolution itself. The fact is almost too bizarre to credit, but some of the leading ‘transformed cladists’ profess an actual hostility to the idea of evolution itself, especially the Darwinian theory of evolution. Two of them, G. Nelson and N. Platnick from the American Museum of Natural History in New York, have gone so far as to write th
at ‘Darwinism … is, in short, a theory that has been put to the test and found false’. I should love to know what this ‘test’ is and, even more, I should love to know by what alternative theory Nelson and Platnick would explain the phenomena that Darwinism explains, especially adaptive complexity.

  It isn’t that any transformed cladists are themselves fundamentalist creationists. My own interpretation is that they enjoy an exaggerated idea of the importance of taxonomy in biology. They have decided, perhaps rightly, that they can do taxonomy better if they forget about evolution, and especially if they never use the concept of the ancestor in thinking about taxonomy. In the same way, a student of, say, nerve cells, might decide that he is not aided by thinking about evolution. The nerve specialist agrees that his nerve cells are the products of evolution, but he does not need to use this fact in his research. He needs to know a lot about physics and chemistry, but he believes that Darwinism is irrelevant to his day-to-day research on nerve impulses. This is a defensible position. But you can’t reasonably say that, because you don’t need to use a particular theory in the day to day practice of your particular branch of science, therefore that theory is false. You will only say this if you have a remarkably grandiose estimation of the importance of your own branch of science.

  Even then, it isn’t logical. A physicist certainly doesn’t need Darwinism in order to do physics. He might think that biology is a trivial subject compared with physics. It would follow from this that, in his opinion, Darwinism is of trivial importance to science. But he could not sensibly conclude from this that it is therefore false! Yet this is essentially what some of the leaders of the school of transformed cladistics seem to have done. ‘False’, note well, is precisely the word Nelson and Platnick used. Needless to say, their words have been picked up by the sensitive microphones that I mentioned in the previous chapter, and the result has been considerable publicity. They have earned themselves a place of honour in fundamentalist, creationist literature. When a leading transformed cladist came to give a guest lecture in my university recently, he drew a bigger crowd than any other guest lecturer that year! It isn’t hard to see why.

  There is no doubt at all that remarks like ‘Darwinism … is a theory that has been put to the test and found false’, coming from established biologists on the staff of a respected national museum, will be meat and drink to creationists and others who actively have an interest in perpetrating falsehoods. This is the only reason I have troubled my readers with the topic of transformed cladism at all. As Mark Ridley more mildly said, in a review of the book in which Nelson and Platnick made that remark about Darwinism being false, Who would have guessed that all that they really meant was that ancestral species are tricky to represent in cladistic classification? Of course it is difficult to pin down the precise identity of ancestors, and there is a good case for not even trying to do so. But to make statements that encourage others to conclude that there never were any ancestors is to debauch language and betray truth.

  Now I’d better go out and dig the garden, or something.

  CHAPTER 11

  Doomed rivals

  No serious biologist doubts the fact that evolution has happened, nor that all living creatures are cousins of one another. Some biologists, however, have had doubts about Darwin’s particular theory of how evolution happened. Sometimes this turns out to be just an argument about words. The theory of punctuated evolution, for instance, can be represented as anti-Darwinian. As I argued in Chapter 9, however, it is really a minor variety of Darwinism, and does not belong in any chapter about rival theories. But there are other theories that are most definitely not versions of Darwinism, theories that go flatly against the very spirit of Darwinism. These rival theories are the subject of this chapter. They include various versions of what is called Lamarckism; also other points of view such as ‘neutralism’, ‘mutationism’ and creationism which have, from time to time, been advanced as alternatives to Darwinian selection.

  The obvious way to decide between rival theories is to examine the evidence. Lamarckian types of theory, for instance, are traditionally rejected — and rightly so — because no good evidence for them has ever been found (not for want of energetic trying, in some cases by zealots prepared to fake evidence). In this chapter I shall take a different tack, largely because so many other books have examined the evidence and concluded in favour of Darwinism. Instead of examining the evidence for and against rival theories, I shall adopt a more armchair approach. My argument will be that Darwinism is the only known theory that is in principle capable of explaining certain aspects of life. If I am right it means that, even if there were no actual evidence in favour of the Darwinian theory (there is, of course) we should still be justified in preferring it over all rival theories.

  One way in which to dramatize this point is to make a prediction. I predict that, if a form of life is ever discovered in another part of the universe, however outlandish and weirdly alien that form of life may be in detail, it will be found to resemble life on Earth in one key respect: it will have evolved by some kind of Darwinian natural selection. Unfortunately, this is a prediction that we shall, in all probability, not be able to test in our lifetimes, but it remains a way of dramatizing an important truth about life on our own planet. The Darwinian theory is in principle capable of explaining life. No other theory that has ever been suggested is in principle capable of explaining life. I shall demonstrate this by discussing all known rival theories, not the evidence for or against them, but their adequacy, in principle, as explanations for life.

  First, I must specify what it means to ‘explain’ life. There are, of course, many properties of living things that we could list, and some of them might be explicable by rival theories. Many facts about the distribution of protein molecules, as we have seen, may be due to neutral genetic mutations rather than Darwinian selection. There is one particular property of living things, however, that I want to single out as explicable only by Darwinian selection. This property is the one that has been the recurring topic of this book: adaptive complexity. Living organisms are well fitted to survive and reproduce in their environments, in ways too numerous and statistically improbable to have come about in a single chance blow. Following Paley, I have used the example of the eye. Two or three of an eye’s well-designed’ features could, conceivably, have come about in a single lucky accident. It is the sheer number of interlocking parts, all well adapted to seeing and well adapted to each other, that demands a special kind of explanation beyond mere chance. The Darwinian explanation, of course, involves chance too, in the form of mutation. But the chance is filtered cumulatively by selection, step by step, over many generations. Other chapters have shown that this theory is capable of providing a satisfying explanation for adaptive complexity. In this chapter I shall argue that all other known theories are not capable of so doing.

  First, let us take Darwinism’s most prominent historical rival, Lamarckism. When Lamarckism was first proposed in the early nineteenth century, it was not as a rival to Darwinism, because Darwinism had not yet been thought of. The Chevalier de Lamarck was ahead of his time. He was one of those eighteenth-century intellectuals who argued in favour of evolution. In this he was right, and he would deserve to be honoured for this alone, along with Charles Darwin’s grandfather Erasmus and others. Lamarck also offered the best theory of the mechanism of evolution that anyone could come up with at the time, but there is no reason to suppose that, if the Darwinian theory of mechanism had been around at the time, he would have rejected it. It was not around, and it is Lamarck’s misfortune that, at least in the English-speaking world, his name has become a label for an error — his theory of the mechanism of evolution — rather than for his correct belief in the fact that evolution has occurred. This is not a history book, and I shall not attempt a scholarly dissection of exactly what Lamarck himself said. There was a dose of mysticism in Lamarck’s actual words — for instance, he had a strong belief in progress up what many people,
even today, think of as the ladder of life; and he spoke of animals striving as if they, in some sense, consciously wanted to evolve. I shall extract from Lamarckism those non-mystical elements which, at least at first sight, seem to have a sporting chance of offering a real alternative to Darwinism. These elements, the only ones adopted by modern ‘neo-Lamarckians’, are basically two: the inheritance of acquired characteristics, and the principle of use and disuse.

  The principle of use and disuse states that those parts of an organism’s body that are used grow larger. Those parts that are not used tend to wither away. It is an observed fact that when you exercise particular muscles they grow; muscles that are never used shrink. By examining a man’s body we can tell which muscles he uses and which he does not. We may even be able to guess his profession or his recreation. Enthusiasts of the ‘body-building’ cult make use of the principle of use and disuse to ‘build’ their bodies, almost like a piece of sculpture, into whatever unnatural shape is demanded by fashion in this peculiar minority culture. Muscles are not the only parts of the body that respond to use in this kind of way. Walk barefoot, and you acquire tougher skin on your soles. It is easy to tell a farmer from a bank clerk by looking at their hands alone. The farmer’s hands are horny, toughened by long exposure to rough work. If the clerk’s hands are horny at all, it amounts only to a little callus on the writing finger.

  The principle of use and disuse enables animals to become better at the job of surviving in their world, progressively better during their own lifetime as a result of living in that world. Humans, through direct exposure to sunlight, or lack of it, develop a skin colour which equips them better to survive in the particular local conditions. Too much sunlight is dangerous. Enthusiastic sun-bathers with very fair skins are susceptible to skin cancer. Too little sunlight, on the other hand, leads to vitamin-D deficiency and rickets, sometimes seen in hereditarily black children living in Scandinavia. The brown pigment melanin, which is synthesized under the influence of sunlight, makes a screen to protect the underlying tissues from the harmful effects of further sunlight. If a suntanned person moves to a less sunny climate the melanin disappears, and the body is able to benefit from what little sun there is. This can be represented as an instance of the principle of use and disuse: skin goes brown when it is ‘used’, and fades to white when it is not ‘used’. Some tropical races, of course, inherit a thick screen of melanin whether or not they are exposed to sunlight as individuals.