The Blind Watchmaker

  But of all the systems of classification that could be dreamed up, there is one unique system, unique in the sense that words like ‘correct’ and ‘incorrect’, ‘true’ and ‘false’ can be applied to it with perfect agreement given perfect information. That unique system is the system based on evolutionary relationships. To avoid confusion I shall give this system the name that biologists give to its strictest form: cladistic taxonomy.

  In cladistic taxonomy, the ultimate criterion for grouping organisms together is closeness of cousinship or, in other words, relative recency of common ancestry. Birds, for instance, are distinguished from non-birds by the fact that they are all descended from a common ancestor, which is not an ancestor of any non-bird. Mammals are all descended from a common ancestor, which is not an ancestor of any non-mammal. Birds and mammals have a more remote common ancestor, which they share with lots of other animals like snakes and lizards and tuataras. The animals descended from this common ancestor are all called amniotes. So, birds and mammals are amniotes. ‘Reptiles’ is not a true taxonomic term, according to cladists, because it is defined by exception: all amniotes except birds and mammals. In other words, the most recent common ancestor of all ‘reptiles’ (snakes, turtles, etc.) is also ancestral to some non-‘reptiles’, namely birds and mammals.

  Within mammals, rats and mice share a recent common ancestor with each other; leopards and lions share a recent common ancestor with each other; so do chimpanzees and humans with each other. Closely related animals are animals that share a recent common ancestor. More distantly related animals share an earlier common ancestor. Very distantly related animals, like people and slugs, share a very early common ancestor. Organisms can never be totally unrelated to one another, since it is all but certain that life as we know it originated only once on earth.

  True cladistic taxonomy is strictly hierarchical, an expression which I shall use to mean that it can be represented as a tree whose branches always diverge and never converge again. In my view (some schools of taxonomists, that we shall discuss later, would disagree), it is strictly hierarchical not because hierarchical classification is convenient, like a librarian’s classification, nor because everything in the world falls naturally into a hierarchical pattern, but simply because the pattern of evolutionary descent is hierarchical. Once the tree of life has branched beyond a certain minimum distance (basically the bounds of the species), the branches never ever come together again (there may be very rare exceptions, as in the origin of the eukaryotic cell mentioned in Chapter 7). Birds and mammals are descended from a common ancestor, but they are now separate branches of the evolutionary tree, and they will never come together again: there will never be a hybrid between a bird and a mammal. A group of organisms that has this property of all being descended from a common ancestor, which is not an ancestor of any non-member of the group, is called a clade, after the Greek for a tree branch.

  Another way of representing this idea of strict hierarchy is in terms of ‘perfect nesting’. We write the names of any set of animals on a large sheet of paper and draw rings round related sets. For example, rat and mouse would be united in a small ring indicating that they are close cousins, with a recent common ancestor. Guinea-pig and capybara would be united with each other in another small ring. The rat/mouse ring and the guinea-pig/capybara ring would, in turn, be united with each other (and beavers and porcupines and squirrels and lots of other animals) in a larger ring labelled with its own name, rodents. Inner rings are said to be ‘nested’ inside larger, outer rings. Somewhere else on the paper, lion and tiger would be united with one another in a small ring. This ring would be included with others in a ring labelled cats. Cats, dogs, weasels, bears, etc. would all be united, in a series of rings within rings, in a single large ring labelled carnivores. The rodent ring and the carnivore ring would then take part in a more global series of rings within rings in a very large ring labelled mammals.

  The important thing about this system of rings within rings is that it is perfectly nested. Never, not on a single solitary occasion, will the rings that we draw intersect each other. Take any two overlapping rings, and it will always be true to say that one lies wholly inside the other. The area enclosed by the inner one is always totally enclosed by the outer one: there are never any partial overlaps. This property of perfect taxonomic nesting is not exhibited by books, languages, soil types, or schools of thought in philosophy. If a librarian draws a ring round the biology books and another ring round the theology books, he will find that the two rings overlap. In the zone of overlap are books with titles like ‘Biology and Christian Belief’.

  On the face of it, we might expect the classification of languages to exhibit the property of perfect nesting. Languages, as we saw in Chapter 8, evolve in a rather animal-like way. Languages that have recently diverged from a common ancestor, like Swedish, Norwegian and Danish, are much more similar to each other than they are to languages that diverged longer ago, like Icelandic. But languages don’t only diverge, they also merge. Modern English is a hybrid between Germanic and Romance languages that had diverged much earlier, and English would therefore not fit neatly in any hierarchical nesting diagram. The rings that enclosed English would be found to intersect, to overlap partially. Biological classificatory rings never intersect in this way, because biological evolution above the species level is always divergent.

  Returning to the library example, no librarian can entirely avoid the problem of intermediates or overlaps. It is no use housing the biology and theology sections next door to each other and putting intermediate books in the corridor between them; for what then do we do with books that are intermediate between biology and chemistry, between physics and theology, history and theology, history and biology? I think I am right in saying that the problem of intermediates is inescapably, inherently a part of all taxonomic systems other than that which springs from evolutionary biology. Speaking personally, it is a problem that gives me almost physical discomfort when I am attempting the modest filing tasks that arise in my professional life: shelving my own books, and reprints of scientific papers that colleagues (with the kindest of intentions send me; filing administrative papers; old letters, and so on. Whatever categories one adopts for a filing system, there are always awkward items that don’t fit, and the uncomfortable indecision leads me, I am sorry to say, to leave odd papers out on the table, sometimes for years at a time until it is safe to throw them away. Often one has unsatisfactory recourse to a ‘miscellaneous’ category, a category which, once initiated, has a menacing tendency to grow. I sometimes wonder whether librarians, and keepers of all museums except biological museums, are particularly prone to ulcers.

  In the taxonomy of living creatures these filing problems do not arise. There are no ‘miscellaneous’ animals. As long as we stay above the level of the species, and as long as we study only modern animals (or animals in any given time slice: see below) there are no awkward intermediates. If an animal appears to be an awkward intermediate, say it seems to be exactly intermediate between a mammal and a bird, an evolutionist can be confident that it must definitely be one or the other. The appearance of intermediacy must be an illusion. The unlucky librarian can take no such reassurance. It is perfectly possible for a book to belong simultaneously in both the history and the biology departments. Cladistically inclined biologists never indulge in any librarians’ arguments over whether it is ‘convenient’ to classify whales as mammals or as fish, or as intermediate between mammals and fish. The only argument we have is a factual one. In this case, as it happens, the facts lead all modern biologists to the same conclusion. Whales are mammals and not fish, and they are not, even to a tiny degree, intermediate. They are no closer to fish than humans are, or duck-billed platypuses, or any other mammals.

  Indeed, it is important to understand that all mammals — humans, whales, duck-billed platypuses, and the rest — are exactly equally close to fish, since all mammals are linked to fish via the same common ancestor. Th
e myth that mammals, for instance, form a ladder or ‘scale’, with ‘lower’ ones being closer to fish than ‘higher’ ones, is a piece of snobbery that owes nothing to evolution. It is an ancient, pre-evolutionary notion, sometimes called the ‘great chain of being’, which should have been destroyed by evolution but which was, mysteriously, absorbed into the way many people thought about evolution.

  At this point I cannot resist drawing attention to the irony in the challenge that creationists are fond of hurling at evolutionists: ‘Produce your intermediates. If evolution were true, there should be animals that are half way between a cat and a dog, or between a frog and an elephant. But has anyone ever seen a frelephant?’ I have been sent creationist pamphlets that attempt to ridicule evolution with drawings of grotesque chimeras, horse hindquarters grafted to a dog’s front end, for instance. The authors seem to imagine that evolutionists should expect such intermediate animals to exist. This not only misses the point, it is the precise antithesis of the point. One of the strongest expectations the theory of evolution gives us is that intermediates of this kind should not exist. This is the burden of my comparison between animals and library books.

  The taxonomy of evolved living beings, then, has the unique property of providing perfect agreement in a world of perfect information. That is what I meant by saying that words like ‘true’ and ‘false’ could be applied to claims in cladistic taxonomy, though not to claims in any librarian’s taxonomy. We must make two qualifications. First, in the real world we don’t have perfect information. Biologists may disagree with one another over the facts of ancestry, and the disputes may be difficult to settle because of imperfect information — not enough fossils, say. I shall return to this. Second, a different kind of problem arises if we have too many fossils. The neat and clear-cut discreteness of classification is liable to evaporate if we try to include all animals that have ever lived, rather than just modern animals. This is because, however distant from each other two modern animals may be — say they are a bird and a mammal — they did, once upon a time, have a common ancestor. If we are faced with trying to fit that ancestor into our modern classification, we may have problems.

  The moment we start to consider extinct animals, it is no longer true that there are no intermediates. On the contrary, we now have to contend with potentially continuous series of intermediates. The distinction between modern birds, and modern non-birds like mammals, is a clear-cut one only because the intermediates converging backwards on the common ancestor are all dead. To make the point most forcibly, think again of a hypothetically ‘kind’ nature, providing us with a complete fossil record; with a fossil of every animal that ever lived. When I introduced this fantasy in the previous chapter, I mentioned that from one point of view nature would actually be being unkind. I was thinking then of the toil of studying and describing all the fossils, but we now come to another aspect of that paradoxical unkindness. A complete fossil record would make it very difficult to classify animals into discrete nameable groups. If we had a complete fossil record, we should have to give up discrete names and resort to some mathematical or graphical notation of sliding scales. The human mind far prefers discrete names, so in one sense it is just as well that the fossil record is poor.

  If we consider all animals that have ever lived instead of just modern animals, such words as ‘human’ and ‘bird’ become just as blurred and unclear at the edges as words like ‘tall’ and ‘fat’. Zoologists can argue unresolvably over whether a particular fossil is, or is not, a bird. Indeed they often do argue this very question over the famous fossil Archaeopteryx. It turns out that if ‘bird/non-bird’ is a clearer distinction than ‘tall/short’, it is only because in the bird/non-bird case the awkward intermediates are all dead. If a curiously selective plague came along and killed all people of intermediate height, ‘tall’ and ‘short’ would come to have just as precise a meaning as ‘bird’ or ‘mammal’.

  It isn’t just zoological classification that is saved from awkward ambiguity only by the convenient fact that most intermediates are now extinct. The same is true of human ethics and law. Our legal and moral systems are deeply species-bound. The director of a zoo is legally entitled to ‘put down’ a chimpanzee that is surplus to requirements, while any suggestion that he might ‘put down’ a redundant keeper or ticket-seller would be greeted with howls of incredulous outrage. The chimpanzee is the property of the zoo. Humans are nowadays not supposed to be anybody’s property, yet the rationale for discriminating against chimpanzees in this way is seldom spelled out, and I doubt if there is a defensible rationale at all. Such is the breathtaking speciesism of our Christian-inspired attitudes, the abortion of a single human zygote (most of them are destined to be spontaneously aborted anyway) can arouse more moral solicitude and righteous indignation than the vivisection of any number of intelligent adult chimpanzees! I have heard decent, liberal scientists, who had no intention of actually cutting up live chimpanzees, nevertheless passionately defending their right to do so if they chose, without interference from the law. Such people are often the first to bristle at the smallest infringement of human rights. The only reason we can be comfortable with such a double standard is that the intermediates between humans and chimps are all dead.

  The last common ancestor of humans and chimps lived perhaps as recently as five million years ago, definitely more recently than the common ancestor of chimps and orang-utans, and perhaps 30 million years more recently than the common ancestor of chimps and monkeys. Chimpanzees and we share more than 99 per cent of our genes. If, in various forgotten islands around the world, survivors of all intermediates back to the chimp/human common ancestor were discovered, who can doubt that our laws and our moral conventions would be profoundly affected, especially as there would presumably be some interbreeding along the spectrum? Either the whole spectrum would have to be granted full human rights (Votes for Chimps), or there would have to be an elaborate apartheid-like system of discriminatory laws, with courts deciding whether particular individuals were legally ‘chimps’ or legally ‘humans’; and people would fret about their daughter’s desire to marry one of ‘them’. I suppose the world is already too well explored for us to hope that this chastening fantasy will ever come true. But anybody who thinks that there is something obvious and self-evident about human ‘rights’ should reflect that it is just sheer luck that these embarrassing intermediates happen not to have survived. Alternatively, maybe if chimpanzees hadn’t been discovered until today they would now be seen as the embarrassing intermediates.

  Readers of the previous chapter may remark that the whole argument, that categories become blurred if we don’t stick to contemporary animals, assumes that evolution goes at a constant speed, rather than being punctuated. The more our view of evolution approaches the extreme of smooth, continuous change, the more pessimistic shall we be about the very possibility of applying such words as bird or non-bird, human or non-human, to all animals that ever lived. An extreme saltationist could believe that there really was a first human, whose mutant brain was twice the size of his father’s brain and that of his chimp-like brother.

  The advocates of punctuated equilibrium are for the most part not, as we have seen, true saltationists. Nevertheless, to them the problem of the ambiguity of names is bound to seem less severe than it will on a more continuous view. The naming problem would arise even for punctuationists if literally every animal that had ever lived was preserved as a fossil, because the punctuationists are really gradualists when we come right down to detail. But, since they assume that we are particularly unlikely to find fossils documenting short periods of rapid transition, while being particularly likely to find fossils documenting the long periods of stasis, the ‘naming problem’ will be less severe on a punctuationist view than on a nonpunctuationist view of evolution.

  It is for this reason that the punctuationists, especially Niles Eldredge, make a big point of treating ‘the species’ as a real ‘entity’. To a non-punctuationi
st, ‘the species’ is definable only because the awkward intermediates are dead. An extreme anti-punctuationist, taking a long view of the entirety of evolutionary history, cannot see ‘the species’ as a discrete entity at all. He can see only a smeary continuum. On his view a species never has a clearly defined beginning, and it only sometimes has a clearly defined end (extinction); often a species does not end decisively but turns gradually into a new species. A punctuationist, on the other hand, sees a species as coming into existence at a particular time (strictly there is a transition period with a duration of tens of thousands of years, but this duration is short by geological standards). Moreover, he sees a species as having a definite, or at least rapidly accomplished, end, not a gradual fading into a new species. Since most of the life of a species, on the punctuationist view, is spent in unchanging stasis, and since a species has a discrete beginning and end, it follows that, to a punctuationist, a species can be said to have a definite, measurable ‘life span’. The non-punctuationist would not see a species as having a ‘life span’ like an individual organism. The extreme punctuationist sees ‘the species’ as a discrete entity that really deserves its own name. The extreme anti-punctuationist sees ‘the species’ as an arbitrary stretch of a continuously flowing river, with no particular reason to draw lines delimiting its beginning and end.

  In a punctuationist book on the history of a group of animals, say the history of the horses over the past 30 million years, the characters in the drama may all be species rather than individual organisms, because the punctuationist author thinks of species as real ‘things’, with their own discrete identity. Species will suddenly arrive on the scene, and as suddenly they will disappear, replaced by successor species. It will be a history of successions, as one species gives way to another. But if an anti-punctuationist writes the same history, he will use species names only as a vague convenience. When he looks longitudinally through time, he ceases to see species as discrete entities. The real actors in his drama will be individual organisms in shifting populations. In his book it will be individual animals that give way to descendant individual animals, not species that give way to species. It is not surprising, then, that punctuations tend to believe in a kind of natural selection at the species level, which they regard as analogous to Darwinian selection at the ordinary individual level. Non-punctuationists, on the other hand, are likely to see natural selection as working at no higher level than the individual organism. The idea of ‘species selection’ has less appeal for them, because they do not think of species as entities with a discrete existence through geological time.