Foxes in the Show

Foreword to a new edition of George Williams’ Adaptation and Natural Selection (2018).

The neo-Darwinian synthesis of the 1930s and 1940s was a collective Anglo-American achievement, defined by a recognizable ‘canon’ of seminal books, those of Fisher, Haldane, Mayr, Dobzhansky, Simpson and others. Julian Huxley’s Evolution: the modern synthesis bequeathed its title to the whole movement although for its theoretical content it doesn’t stand out. If I were asked to nominate one book from the second half of the twentieth century that deserves to take an honoured place alongside the ‘canon’ of the 1930s and 1940s, I would choose Adaptation and Natural Selection by George C. Williams. On opening it I have the feeling of being ushered into the presence of a penetrating and outstanding mind, the same feeling I get, indeed, from reading The Genetical Theory of Natural Selection, although Williams, unlike Fisher, was no mathematician. In George Williams we have an author of immense learning and incisive critical intelligence, who thought deeply about every aspect of evolution and ecology. Williams not only enlarged the synthesis, he exposed with great clarity where many of its followers had gone astray, even in some cases the original authors themselves. This is a book that every serious student of biology must read, a book that irrevocably changes the way we look at life. Throughout my career as an Oxford tutor, I obviously recommended many books to my students. But I think this was the only one I insisted that all should read.

Here’s a list of major mistakes a student is likely to make before reading this book, but will not make afterwards. ‘Mutations are an adaptation to speed up evolution.’ ‘Dominance hierarchies are an adaptation to make sure the strongest individuals reproduce.’ ‘Territoriality is an adaptation to space the species out and beneficially limit the population.’ ‘Sex ratios are optimized to make the best use of species resources.’ ‘Death from old age is an adaptation to clear superannuated individuals out of the way and make room for the young.’ ‘Natural selection favours species that resist extinction.’ ‘Species parcel out niches for the benefit of a balanced ecosystem.’ ‘Predators hunt “prudently,” taking care not to deplete prey that they are going to need in the future.’ ‘Individuals limit their reproduction to avoid overpopulation.’

‘Adaptation’ is the first word in the title and the book is largely a plea for a proper, scientific study of adaptation – a scientific teleonomy, to adopt Pittendrigh’s term as advocated by Williams. But Williams is the last person who could justly be tarred as a naive ‘adaptationist’. This pejorative was given wide currency by Gould and Lewontin in their overrated ‘spandrels paper’ of 1979. It denotes those who assume without evidence that everything an animal is or does must be an adaptation. Unfortunately, their critique of adaptationism has been misunderstood, not least by some philosophers such as the late Jerry Fodor,[1] as a critique of the very idea of adaptation itself.

A ‘spandrel’ is a non-adaptive by-product. The name comes from the gaps between gothic arches which are a necessary but non-functional by-product of the functionally important arches themselves. Long before the word was introduced into biology, Williams, a leading advocate of adaptation as a proper subject for scientific study, gave an incisive critique of what would later be called spandrels. His vivid example, which regularly grabbed the attention of my Oxford students, was a fox repeatedly running along its own tracks in the snow. Its paws increasingly flattened the snow, which made each successive journey easier and faster. But it would be wrong to say the fox’s paws were adapted to flatten snow. They can’t help flattening snow. This particular beneficial effect is a by-product. Williams summed up the message pithily: adaptation is an ‘onerous concept’.

If I might paraphrase the Anglican marriage service in a way that Williams might not, any attribution of adaptation should not be entered into unadvisedly or lightly; but reverently, discreetly, advisedly, soberly and in the fear of Occam’s Razor. You must first assure yourself that you could, if called upon to do so, translate your adaptation theory back into the rigorous terms of neo-Darwinism. The ‘adaptation’ you postulate must not just be ‘beneficial’ in some vague, panglossian sense. You must clearly set out, and be prepared to defend, a strictly Darwinian pathway to the evolution of the alleged adaptation. The ‘benefit’ must accrue at the proper level in the hierarchy of life, which is the unit of Darwinian natural selection. And the proper level, for Williams as for me, is that of the individual genes responsible for the putative adaptation.

The term ‘panglossian’ was introduced into biology by J. B. S. Haldane, one of the architects of the synthesis. His star pupil John Maynard Smith reported that Haldane proposed three ‘theorems’ to satirize errors in scientific thinking.

Aunt Jobiska’s Theorem (from Edward Lear): ‘It’s a fact the whole world knows.’

The Bellman’s Theorem (from Lewis Carroll): ‘What I tell you three times is true.’

Pangloss’s Theorem (from Voltaire and applying especially to biology): ‘All is for the best in the best of all possible worlds.’

My second paragraph above was a list of panglossian errors frequently perpetrated by professors and students alike (including my own undergraduate self). Adaptations cannot be just ‘good’. It is not enough that they convey ‘benefit’. They have to be good for some entity that has been naturally selected precisely because of benefit to itself. And that entity, as Williams powerfully argues, will normally be the gene. I’m fond of a Williams bon mot from his later book, Natural Selection: ‘A gene pool is an imperfect record of a running average of selection pressures over a long period of time in an area often much larger than individual dispersal distances.’ But why the gene? And ‘gene’ in what sense? Williams’ rationale was so clear and irrefutable, I’m inclined to quote it in full but you only have to turn to page 23, the ‘Socrates paragraph’, which also grabbed my Oxford students by the collar when they read it. Here’s the central point.

With Socrates’ death, not only did his phenotype disappear, but also his genotype . . . Socrates’ genes may be with us yet, but not his genotype because meiosis and recombination destroy genotypes as surely as death . . . It is only the meiotically dissociated fragments of the genotype that are transmitted in sexual reproduction, and those fragments are further fragmented by meiosis in the next generation. If there is an ultimately indivisible fragment it is, by definition, ‘the gene’ that is treated in the abstract discussions of population geneticists.

That last sentence is the answer to my second question, ‘Gene, in what sense?’ I summed up the Williams answer a decade later when I jokingly wrote that The Selfish Gene might better have been called The slightly selfish big bit of chromosome and the even more selfish little bit of chromosome. It could also have been called The Cooperative Gene, and here lies the answer to perhaps the commonest criticism of the ‘gene’s-eye view’ of natural selection. There is no simple, atomistic, one-to-one mapping between single genes and units of phenotype. Most genes have effects in many parts of the body, and most phenotypic features are influenced by many genes: how then, the critics bleat, can ‘the gene’ be the unit of natural selection? The objection is easily answered and Williams dispatches it with characteristic aplomb:

No matter how functionally dependent a gene may be, and no matter how complicated its interactions with other genes and environmental factors, it must always be true that a given gene substitution will have an arithmetic mean effect on fitness in any population. (p. 57)

Williams is eloquent on the idea that the other genes in the genome (which in the long run means in the population gene pool) constitute the main environment in which a gene operates – the ‘background’ against which it is naturally selected. The fallacy (a sadly common one) is to assume that a coadapted gene complex is necessarily selected as a unit. Rather, each gene in the complex is selected individually for its compatibility with the other genes in the complex, which are in turn being selected for the very same compatibility.

Return for a moment to Williams’ picturesque example of the fox in the snow. I think he’d have accepted the following reservation to his ‘spandrel’ or by-product lesson. Natural selection actually could favour an adaptive broadening of fox paws for the function of flattening snow. But only if the resulting path benefited the fox itself (and its family) alone, rather than foxes in general. It might, for example, be confined to the individual fox’s own territory. This brings me to the central core of the book, which is Williams’ critique of ‘group selection’. This is as needed today as it was in 1966, for group selectionism won’t lie down. With its magnetic allure, perhaps politically or even aesthetically motivated, group selectionism keeps coming back for more, in ways that, I can’t resist confessing, remind me of Monty Python’s Black Knight.[2]

Williams admits that natural selection could theoretically choose among groups. He just doesn’t think it’s important in practice. His meaning of group selection includes what he later called ‘clade selection’. A hypothetical example might be a tendency for within-species natural selection (which Williams calls ‘organic selection’) to favour larger-sized individuals while at the same time whole species of smaller individuals are less likely to go extinct (‘biotic selection’). Some authors espouse a different form of group selection in which altruistic or cooperative behaviour of individuals, or indeed a tendency to live in groups, is thought to be favoured because it benefits the group. Williams declines to invoke group selection where the phenomena are more parsimoniously explained by kin selection (Hamilton’s seminal papers had just appeared) or reciprocation (Trivers’ clever theorizing lay in the future, but Williams anticipates the basic idea). As for living in groups, there are, of course, numerous ways in which individuals benefit: huddling for warmth, safety in numbers when predators strike, the ‘many eyes effect’ when spotting opportunities, aerodynamic or hydrodynamic facilitation in flocking birds or schooling fish, ‘non-zero-sum games’ in bringing down large prey, etc. Indeed, all these examples are nowadays often handled by game theory models in which individuals maximize their own benefit in the context of other individuals maximizing their own benefit. Group benefit plays absolutely no part in such models. Incidentally, Williams has a prescient anticipation of evolutionary game theory, in a slightly different context. Williams revisited and updated his critique of group selection in his 1992 book Natural Selection, but I’ll say no more about it here.

In the final chapter of Adaptation and Natural Selection, where he lays out his programme for a scientific teleonomy, Williams quotes William Paley’s Natural Theology on the vertebrate eye. His purpose is to illustrate the self-evident ‘design’ of living creatures, an immensely powerful illusion of design, which pervades some (though not all – that would be adaptationism) biological entities such as the eye. The complex, statistically improbable juxtaposition of mutually suited functionally cooperating parts – precision-focusing lens, precision-adjusting iris diaphragm, retina with millions of colour-coding photocells, optic nerve trunk cable to the brain: such phenomena (and they are legion in all parts of all animals and plants) can only be explained if the principles of chemistry and physics are supplemented by ‘the one additional postulate of natural selection and its consequence, adaptation’. Philosophers and others who don’t see the glaring need for natural selection (or divine creation as Paley would have it) must simply be ignorant of the relevant beautiful facts. Have they never seen a David Attenborough film? Or looked down a microscope at a cell? Or contemplated their own hand?

Williams urges us to take seriously the need for a special kind of explanation of adaptation, but to pay cautious attention to the precise mechanism of natural selection and the level in the hierarchy of life where it acts. It is his contention that genic selection occupies the appropriate level. Group selection is a theoretical possibility, but it lacks the power to build up Paleyesque complexity: Darwin’s ‘organs of extreme perfection and complexity’, organs which, for Hume, ‘ravish into admiration all men who have ever contemplated them’. We marvel at complex organs that give individuals the power to see, birds to fly, bats to echolocate, dogs to smell, cheetahs to sprint. There are no complex organs that give species, or groups, or ecosystems, the power to do anything. Those larger groupings of individuals are just not the kind of entity that has complex ‘organs’ or, indeed, adaptations of any kind. What groups do is a consequence, a by-product indeed, of what their component individuals do.

George Williams, a tall, imposing, Abe Lincolnesque figure, quiet, kind, thoughtful, modest, made major research contributions to solving outstanding problems in evolutionary biology – really big problems like the evolution of sex, of senescence, of life-history strategies. He was a pioneer of the up-and-coming but still undervalued subject of Darwinian medicine. His Natural Selection: domains, levels and challenges was an important successor to this book. But I think Adaptation and Natural Selection is his outstanding achievement. When I re-read it before writing this foreword I expected to find passages that needed critical updating or even deleting. I failed. It can still be recommended to today’s students without reserve. Not for its historical interest like some books of the synthesis, but because this fifty-year-old book is still biologically illuminating, wise and – as far as I can judge – correct.

[1] Daniel Dennett, personal communication.

[2] Played by John Cleese, wearing full armour, in their film Monty Python and the Holy Grail.