The Extended Phenotype
Chapter 4 opens the case for the prosecution against the selfish organism, and begins to hint at the second aspect of the Necker Cube. Chapter 5 opens the case for the ‘replicator’ as the fundamental unit of natural selection. Chapter 6 returns to the individual organism and shows how neither it, nor any other major candidate except the small genetic fragment, qualifies as a true replicator. Rather, the individual organism should be thought of as a ‘vehicle’ for replicators. Chapter 7 is a digression on research methodology. Chapter 8 raises some awkward anomalies for the selfish organism, and Chapter 9 continues the theme. Chapter 10 discusses various notions of ‘individual fitness’, and concludes that they are confusing, and probably dispensable.
Chapters 11, 12 and 13 are the heart of the book. They develop, by gradual degrees, the idea of the extended phenotype itself, the second face of the Necker Cube. Finally, in Chapter 14, we turn back with refreshed curiosity to the individual organism and ask why, after all, it is such a salient level in the hierarchy of life.
2 Genetic Determinism and Gene Selectionism
Long after his death, tenacious rumours persisted that Adolf Hitler had been seen alive and well in South America, or in Denmark, and for years a surprising number of people with no love for the man only reluctantly accepted that he was dead (Trevor-Roper 1972). In the First World War a story that a hundred thousand Russian troops had been seen landing in Scotland ‘with snow on their boots’ became widely current, apparently because of the memorable vividness of that snow (Taylor 1963). In our own time myths such as that of computers persistently sending householders electricity bills for a million pounds (Evans 1979), or of well-heeled welfare-soroungers with two expensive cars parked outside their government-subsidized council houses, are familiar to the point of cliché. There are some falsehoods, or half-truths, that seem to engender in us an active desire to believe them and pass them on even if we find them unpleasant, maybe in part, perversely, because we find them unpleasant.
Computers and electronic ‘chips’ provoke more than their fair share of such myth-making, perhaps because computer technology advances at a speed which is literally frightening. I know an old person who has it on good authority that ‘chips’ are usurping human functions to the extent not only of ‘driving tractors’ but even of ‘fertilizing women’. Genes, as I shall show, are the source of what may be an even larger mythology than computers. Imagine the result of combining these two powerful myths, the gene myth and the computer myth! I believe that I may have inadvertently achieved some such unfortunate synthesis in the minds of a few readers of my previous book, and the result was comic misunderstanding. Happily, such misunderstanding was not widespread, but it is worth trying to avoid a repeat of it here, and that is one purpose of the present chapter. I shall expose the myth of genetic determinism, and explain why it is necessary to use language that can be unfortunately misunderstood as genetic determinism.
A reviewer of Wilson’s (1978) On Human Nature, wrote: ‘…although he does not go as far as Richard Dawkins (The Selfish Gene …) in proposing sex-linked genes for “philandering”, for Wilson human males have a genetic tendency towards polygyny, females towards constancy (don’t blame your mates for sleeping around, ladies, it’s not their fault they are genetically programmed). Genetic determinism constantly creeps in at the back door’ (Rose 1978). The reviewer’s clear implication is that the authors he is criticizing believe in the existence of genes that force human males to be irremediable philanderers who cannot therefore be blamed for marital infidelity. The reader is left with the impression that those authors are protagonists in the ‘nature or nurture’ debate, and, moreover, died-in-the-wool hereditarians with male chauvinist leanings.
In fact my original passage about ‘philanderer males’ was not about humans. It was a simple mathematical model of some unspecified animal (not that it matters, I had a bird in mind). It was not explicitly (see below) a model of genes, and if it had been about genes they would have been sex-limited, not sex-linked! It was a model of ‘strategies’ in the sense of Maynard Smith (1974). The ‘philanderer’ strategy was postulated, not as the way males behave, but as one of two hypothetical alternatives, the other being the ‘faithful’ strategy. The purpose of this very simple model was to illustrate the kinds of conditions under which philandering might be favoured by natural selection, and the kinds of conditions under which faithfulness might be favoured. There was no presumption that philandering was more likely in males than faithfulness. Indeed, the particular run of the simulation that I published culminated in a mixed male population in which faithfulness slightly predominated (Dawkins 1976a, p. 165, although see Schuster & Sigmund 1981). There is not just one misunderstanding in Rose’s remarks, but multiple compounded misunderstanding. There is a wanton eagerness to misunderstand. It bears the stamp of snow-covered Russian jackboots, of little black microchips marching to usurp the male role and steal our tractor-drivers’ jobs. It is a manifestation of a powerful myth, in this case the great gene myth.
The gene myth is epitomized in Rose’s parenthetic little joke about ladies not blaming their mates for sleeping around. It is the myth of ‘genetic determinism’. Evidently, for Rose, genetic determinism is determinism in the full philosophical sense of irreversible inevitability. He assumes that the existence of a gene ‘for’ X implies that X cannot be escaped. In the words of another critic of ‘genetic determinism’, Gould (1978, p. 238), ‘If we are programmed to be what we are, then these traits are ineluctable. We may, at best, channel them, but we cannot change them either by will, education, or culture.’
The validity of the determinist point of view and, separately, its bearing on an individual’s moral responsibility for his actions, has been debated by philosophers and theologians for centuries past, and no doubt will be for centuries to come. I suspect that both Rose and Gould are determinists in that they believe in a physical, materialistic basis for all our actions. So am I. We would also probably all three agree that human nervous systems are so complex that in practice we can forget about determinism and behave as if we had free will. Neurones may be amplifiers of fundamentally indeterminate physical events. The only point I wish to make is that, whatever view one takes on the question of determinism, the insertion of the word ‘genetic’ is not going to make any difference. If you are a full-blooded determinist you will believe that all your actions are predetermined by physical causes in the past, and you may or may not also believe that you therefore cannot be held responsible for your sexual infidelities. But, be that as it may, what difference can it possibly make whether some of those physical causes are genetic? Why are genetic determinants thought to be any more ineluctable, or blame-absolving, than ‘environmental’ ones?
The belief that genes are somehow super-deterministic, in comparison with environmental causes, is a myth of extraordinary tenacity, and it can give rise to real emotional distress. I was only dimly aware of this until it was movingly brought home to me in a question session at a meeting of the American Association for the Advancement of Science in 1978. A young woman asked the lecturer, a prominent ‘sociobiologist’, whether there was any evidence for genetic sex differences in human psychology. I hardly heard the lecturer’s answer, so astonished was I by the emotion with which the question was put. The woman seemed to set great store by the answer and was almost in tears. After a moment of genuine and innocent bafflement the explanation hit me. Something or somebody, certainly not the eminent sociobiologist himself, had misled her into thinking that genetic determination is for keeps; she seriously believed that a ‘yes’ answer to her question would, if correct, condemn her as a female individual to a life of feminine pursuits, chained to the nursery and the kitchen sink. But if, unlike most of us, she is a determinist in that strong Calvinistic sense, she should be equally upset whether the causal factors concerned are genetic or ‘environmental’.
What does it ever mean to say that something determines something? Philosophers, possibly with justificati
on, make heavy weather of the concept of causation, but to a working biologist causation is a rather simple statistical concept. Operationally we can never demonstrate that a particular observed event C caused a particular result R, although it will often be judged highly likely. What biologists in practice usually do is to establish statistically that events of class R reliably follow events of class C. They need a number of paired instances of the two classes of events in order to do so: one anecdote is not enough.
Even the observation that R events reliably tend to follow C events after a relatively fixed time interval provides only a working hypothesis that C events cause R events. The hypothesis is confirmed, within the limits of the statistical method, only if the C events are delivered by an experimenter rather than simply noted by an observer, and are still reliably followed by R events. It is not necessary that every C should be followed by an R, nor that every R should be preceded by a C (who has not had to contend with arguments such as ‘smoking cannot cause lung cancer, because I knew a non-smoker who died of it, and a heavy smoker who is still going strong at ninety’?). Statistical methods are designed to help us assess, to any specified level of probabilistic confidence, whether the results we obtain really indicate a causal relationship.
If, then, it were true that the possession of a Y chromosome had a causal influence on, say, musical ability or fondness for knitting, what would this mean? It would mean that, in some specified population and in some specified environment, an observer in possession of information about an individual’s sex would be able to make a statistically more accurate prediction as to the person’s musical ability than an observer ignorant of the person’s sex. The emphasis is on the word ‘statistically’, and let us throw in an ‘other things being equal’ for good measure. The observer might be provided with some additional information, say on the person’s education or upbringing, which would lead him to revise, or even reverse, his prediction based on sex. If females are statistically more likely than males to enjoy knitting, this does not mean that all females enjoy knitting, nor even that a majority do.
It is also fully compatible with the view that the reason females enjoy knitting is that society brings them up to enjoy knitting. If society systematically trains children without penises to knit and play with dolls, and trains children with penises to play with guns and toy soldiers, any resulting differences in male and female preferences are strictly speaking genetically determined differences! They are determined, through the medium of societal custom, by the fact of possession or non-possession of a penis, and that is determined (in a normal environment and in the absence of ingenious plastic surgery or hormone therapy) by sex chromosomes.
Obviously, on this view, if we experimentally brought up a sample of boys to play with dolls and a sample of girls to play with guns, we would expect easily to reverse the normal preferences. This might be an interesting experiment to do, for the result just might turn out to be that girls still prefer dolls and boys still prefer guns. If so, this might tell us something about the tenacity, in the face of a particular environmental manipulation, of a genetic difference. But all genetic causes have to work in the context of an environment of some kind. If a genetic sex difference makes itself felt through the medium of a sex-biased education system, it is still a genetic difference. If it makes itself felt through some other medium, such that manipulations of the education system do not perturb it, it is, in principle, no more and no less a genetic difference than in the former, education-sensitive case: no doubt some other environmental manipulation could be found which did perturb it.
Human psychological attributes vary along almost as many dimensions as psychologists can measure. It is difficult in practice (Kempthorne 1978), but in principle we could partition this variation among such putative causal factors as age, height, years of education, type of education classified in many different ways, number of siblings, birth order, colour of mother’s eyes, father’s skill in shoeing horses, and, of course, sex chromosomes. We could also examine two-way and multi-way interactions between such factors. For present purposes the important point is that the variance we seek to explain will have many causes, which interact in complex ways. Undoubtedly genetic variance is a significant cause of much phenotypic variance in observed populations, but its effects may be overridden, modified, enhanced or reversed by other causes. Genes may modify the effects of other genes, and may modify the effects of the environment. Environmental events, both internal and external, may modify the effects of genes, and may modify the effects of other environmental events.
People seem to have little difficulty in accepting the modifiability of ‘environmental’ effects on human development. If a child has had bad teaching in mathematics, it is accepted that the resulting deficiency can be remedied by extra good teaching the following year. But any suggestion that the child’s mathematical deficiency might have a genetic origin is likely to be greeted with something approaching despair: if it is in the genes ‘it is written’, it is ‘determined’ and nothing can be done about it; you might as well give up attempting to teach the child mathematics. This is pernicious rubbish on an almost astrological scale. Genetic causes and environmental causes are in principle no different from each other. Some influences of both types may be hard to reverse; others may be easy to reverse. Some may be usually hard to reverse but easy if the right agent is applied. The important point is that there is no general reason for expecting genetic influences to be any more irreversible than environmental ones.
What did genes do to deserve their sinister, juggernaut-like reputation? Why do we not make a similar bogey out of, say, nursery education or confirmation classes? Why are genes thought to be so much more fixed and inescapable in their effects than television, nuns, or books? Don’t blame your mates for sleeping around, ladies, it’s not their fault they have been inflamed by pornographic literature! The alleged Jesuit boast, ‘Give me the child for his first seven years, and I’ll give you the man’, may have some truth in it. Educational, or other cultural influences may, in some circumstances, be just as unmodifiable and irreversible as genes and ‘stars’ are popularly thought to be.
I suppose part of the reason genes have become deterministic bogeys is a confusion resulting from the well-known fact of the non-inheritance of acquired characteristics. Before this century it was widely believed that experience and other acquisitions of an individual’s lifetime were somehow imprinted on the hereditary substance and transmitted to the children. The abandoning of this belief, and its replacement by Weismann’s doctrine of the continuity of the germ plasm, and its molecular counterpart the ‘central dogma’, is one of the great achievements of modern biology. If we steep ourselves in the implications of Weismannian orthodoxy, there really does seem to be something juggernaut-like and inexorable about genes. They march through generations, influencing the form and behaviour of a succession of mortal bodies, but, except for rare and non-specific mutagenic effects, they are never influenced by the experience or environment of those bodies. The genes in me came from my four grandparents; they flowed straight through my parents to me, and nothing that my parents achieved, acquired, learned or experienced had any effect on those genes as they flowed through. Perhaps there is something a little sinister about that. But, however inexorable and undeviating the genes may be as they march down the generations, the nature of their phenotypic effects on the bodies they flow through is by no means inexorable and undeviating. If I am homozygous for a gene G, nothing save mutation can prevent my passing G on to all my children. So much is inexorable. But whether or not I, or my children, show the phenotypic effect normally associated with possession of G may depend very much on how we are brought up, what diet or education we experience, and what other genes we happen to possess. So, of the two effects that genes have on the world—manufacturing copies of themselves, and influencing phenotypes—the first is inflexible apart from the rare possibility of mutation; the second may be exceedingly flexible. I think a confusio
n between evolution and development is, then, partly responsible for the myth of genetic determinism.
But there is another myth complicating matters, and I have already mentioned it at the beginning of this chapter. The computer myth is almost as deep-seated in the modern mind as the gene myth. Notice that both passages I quoted contain the word ‘programmed’. Thus Rose sarcastically absolved promiscuous men from blame because they are genetically programmed. Gould says that if we are programmed to be what we are then these traits are ineluctable. And it is true that we ordinarily use the word programmed to indicate unthinking inflexibility, the antithesis of freedom of action. Computers and ‘robots’ are, by repute, notoriously inflexible, carrying out instructions to the letter, even if the consequences are obviously absurd. Why else would they send out those famous million pound bills that everybody’s friend’s friend’s cousin’s acquaintance keeps receiving? I had forgotten the great computer myth, as well as the great gene myth, or I would have been more careful when I myself wrote of genes swarming ‘inside gigantic lumbering robots …’, and of ourselves as ‘survival machines—robot vehicles blindly programmed to preserve the selfish molecules known as genes’ (Dawkins 1976a). These passages have been triumphantly quoted, and requoted apparently from secondary and even tertiary sources, as examples of rabid genetic determinism (e.g. ‘Nabi’ 1981). I am not apologizing for using the language of robotics. I would use it again without hesitation. But I now realize that it is necessary to give more explanation.
From 13 years’ experience of teaching it, I know that a main problem with the ‘selfish-gene survival machine’ way of looking at natural selection is a particular risk of misunderstanding. The metaphor of the intelligent gene reckoning up how best to ensure its own survival (Hamilton 1972) is a powerful and illuminating one. But it is all too easy to get carried away, and allow hypothetical genes cognitive wisdom and foresight in planning their ‘strategy’. At least three out of twelve misunderstandings of kin selection (Dawkins 1979a) are directly attributable to this basic error. Time and again, non-biologists have tried to justify a form of group selection to me by, in effect, imputing foresight to genes: ‘The long-term interests of a gene require the continued existence of the species; therefore shouldn’t you expect adaptations to prevent species extinction, even at the expense of short-term individual reproductive success?’ It was in an attempt to forestall errors like this that I used the language of automation and robotics, and used the word ‘blindly’ in referring to genetic programming. But it is, of course, the genes that are blind, not the animals they program. Nervous systems, like man-made computers, can be sufficiently complex to show intelligence and foresight.