Science in the Soul

by Richard Dawkins

  Darwinian selection, too, optimizes within economic limits and can be said to have values in the same sense. John Maynard Smith said: ‘If there were no constraints on what is possible, the best phenotype would live for ever, would be impregnable to predators, would lay eggs at an infinite rate, and so on.’

  Nicholas Humphrey continues the argument with another analogy from engineering.

  Henry Ford, it is said,*26 commissioned a survey of the car scrap yards of America to find out if there were parts of the Model T Ford which never failed. His inspectors came back with reports of almost every kind of breakdown: axles, brakes, pistons – all were liable to go wrong. But they drew attention to one notable exception: the kingpins of the scrapped cars invariably had years of life left in them. With ruthless logic Ford concluded that the kingpins on the Model T were too good for their job and ordered that in future they should be made to an inferior specification…Nature is surely at least as careful an economist as Henry Ford.

  Humphrey applied his lesson to the evolution of intelligence, but it can equally be applied to bones or anything else. Let us commission a survey of the corpses of gibbons, to see if there are some bones that never fail. We find that every bone in the body breaks at one time or another, but with one notable exception: let’s (rather implausibly) say the femur (thigh bone) is never known to break. Henry Ford would be in no doubt. In future, the femur must be made to an inferior specification.

  Natural selection would agree. Individuals with slightly thinner femurs, who diverted the material saved into some other purpose, say building up other bones and making them less likely to break, would survive better. Or females might take the calcium shaved off the thickness of the femur and put it into milk, thereby improving the survival of their offspring – and with them the genes for making the economy.

  In a machine or an animal the (simplified) ideal is that all the parts should wear out simultaneously. If there is one part that consistently has years of life left in it after the others have worn out, it is overdesigned. Materials that went into building it up should, instead, be diverted to other parts. If there is one part that consistently wears out before anything else, it is underdesigned. It should be built up, using materials taken away from other parts. Natural selection will tend to uphold an equilibration rule: ‘Rob from strong bones to pay weak ones, until all are of equal strength.’

  The reason this is an oversimplification is that not all the bits of an animal or a machine are equally important. That’s why inflight entertainment systems go wrong thankfully more often than rudders or jet engines. A gibbon might be able to afford a broken femur better than a broken humerus. Its way of life depends upon ‘brachiation’ (swinging through the trees by its arms). A gibbon with a broken leg might survive to have another child. A gibbon with a broken arm probably wouldn’t. So the equilibration rule I mentioned has to be tempered: ‘Rob from strong bones to pay weak ones, until you have equalized the risks to your survival accruing from breakages in all parts of your skeleton.’

  But who is the ‘you’ admonished in the equilibration rule? It certainly isn’t an individual gibbon, who is not, we assume, capable of making compensatory adjustments to its own bones. The ‘you’ is an abstraction. You can think of it as a lineage of gibbons in ancestor/descendant relation to one another, represented by the genes that they share. As the lineage progresses, ancestors whose genes make the right adjustments survive to leave descendants who inherit those correctly equilibrating genes. The genes that we see in the world tend to be the ones that get the equilibration right, because they have survived through a long line of successful ancestors who have not suffered the breakage of underdesigned, or the waste of overdesigned, bones.

  So much for bones. Now we need to establish, in Darwinian terms, what values are doing for animals and plants. Where bones stiffen limbs, what do values do for their possessors? By values, I am now going to mean the criteria, in the brain, by which animals choose how to behave.

  The majority of things in the universe don’t actively strive for anything. They just are. I am concerned with the minority that do strive for things, entities that appear to work towards some end and then stop when they’ve achieved it. This minority I shall call value-driven. Some of them are animals and plants, some man-made machines.

  Thermostats, heat-seeking Sidewinder missiles, and numerous physiological systems in animals and plants are controlled by negative feedback. There is a target value which is defined in the system. Discrepancies from the target value are sensed and fed back into the system, causing it to change its state in the direction of reducing the discrepancy.

  Other value-seeking systems improve with experience. From the point of view of defining values in learning systems, the key concept is reinforcement. Reinforcers are positive (‘rewards’) or negative (‘punishments’). Rewards are states of the world which, when encountered, cause an animal to repeat whatever it recently did. Punishments are states of the world which, when encountered, cause an animal to avoid repeating whatever it recently did.

  The stimuli that animals treat as rewards and punishments can be seen as values. Psychologists make a further distinction between primary and secondary reinforcers (both rewards and punishments). Chimpanzees learn to work for food as a primary reward, but they will also learn to work for the equivalent of money – secondary rewards – plastic tokens which they have previously learned to stuff into a slot machine to get food.

  Some psychological theorists have argued that there is only one primary built-in reward (‘drive reduction’ or ‘need reduction’), upon which all others are built. Others, such as Konrad Lorenz, the grand old man of ethology,*27 argued that Darwinian natural selection has built-in complicated rewarding mechanisms, specified differently and in detail for each species to fit its unique way of life.

  Perhaps the most elaborately detailed examples of primary values come from bird song. Different species develop their songs in different ways. The American song sparrow is a fascinating mixture. Young birds brought up completely alone end up singing normal song sparrow song. So, unlike, say, bullfinches, they don’t learn by imitation. But they do learn. Young song sparrows teach themselves to sing by babbling at random and repeating those fragments that match a built-in template. The template is a genetically specified preconception of what a song sparrow ought to sound like. You could say that the information is built in by the genes, but into the sensory part of the brain. It has to be transferred to the motor side by learning. And the sensation specified by the template, by definition, is a reward: the bird repeats actions that deliver it. But, as rewards go, it is very elaborate and precisely specified in detail.

  It is examples like this that stimulated Lorenz to use the colourful phrase ‘innate schoolmarm’ (or ‘innate teaching mechanism’) in his lengthy attempts to resolve the ancient dispute over nativism versus environmentalism. His point was that, however important learning is, there has to be innate guidance of what we shall learn. In particular, each species needs to be supplied with its own specifications of what to treat as rewarding, what punishing. Primary values, Lorenz was saying, have to come from Darwinian natural selection.

  Given enough time we should be able to breed, by artificial selection, a race of animals that enjoy pain and hate pleasure. Of course, by the animals’ newly evolved definition this statement is oxymoronic. I’ll rephrase it. By artificial selection we could reverse the previous definitions of pleasure and pain.*28

  The animals so modified will be less well equipped to survive than their wild ancestors. Wild ancestors have been naturally selected to enjoy those stimuli most likely to improve their survival, and to treat as painful those stimuli most likely, statistically, to kill them. Injury to the body, puncturing skin, breaking bones: all are perceived as painful, for good Darwinian reasons. Our artificially selected animals will enjoy having their skin pierced, will actively seek to break their own bones, will bask in a temperature so hot or so cold as to endanger th
eir survival.

  Similar artificial selection would work with humans. Not only could you breed for tastes, you could breed for callousness, sympathy, loyalty, slothfulness, piety, meanness or the Protestant work ethic. This is a less radical claim than it sounds, for genes don’t fix behaviour deterministically, they only contribute quantitatively to statistical tendencies. Nor, as we saw when discussing the values of science, does it imply a single gene for each of these complicated things, any more than the feasibility of breeding racehorses implies a single gene for speed. In the absence of artificial breeding, our own values are presumably influenced by natural selection under conditions that prevailed in the Pleistocene epoch in Africa.

  Humans are unique in many ways. Perhaps our most obviously unique feature is language. Whereas eyes have evolved between forty and sixty times independently around the animal kingdom,*29 language has evolved only once.*30 It seems learned but there is strong genetic supervision of the learning process. The particular language that we speak is learned, but the tendency to learn language rather than just any old thing is inherited and evolved specifically in our human line. We also inherit evolved rules for grammar. The exact readout of these rules varies from language to language, but their deep structure is laid down by the genes, and is presumably evolved by natural selection just as surely as our lusts and our bones. Evidence is good that the brain contains a language ‘module’, a computational mechanism that actively seeks to learn language and actively uses grammatical rules to structure it.

  According to the young and thriving discipline of evolutionary psychology, the language learning module stands as exemplar of a whole set of inherited, special-purpose computational modules. We might expect modules for sex and reproduction, for analysing kinship (important for doling out altruism and avoiding dysgenic incest), for counting debts and policing obligations, for judging fairness and natural justice, perhaps for throwing projectiles accurately towards a distant target, and for classifying useful animals and plants. These modules will presumably be mediated by specific, built-in values.*31

  If we turn Darwinian eyes on our modern, civilized selves and our predilections – our aesthetic values, our capacity for pleasure – it is important to wear sophisticated spectacles. Do not ask how a middle manager’s ambitions for a bigger desk and a softer office carpet benefit his selfish genes. Ask, instead, how these urban partialities might stem from a mental module which was selected to do something else, in a different place and time. For office carpet, perhaps (and I mean perhaps) read soft and warm animal skins whose possession betokened hunting success. The whole art of applying Darwinian thinking to modern, domesticated humanity is to discern the correct rewriting rules. Take your question about the foibles of civilized, urban humanity and rewrite it back half a million years and out onto the African plains.

  Evolutionary psychologists have coined the term environment of evolutionary adaptedness, or EEA, for that set of conditions in which our wild ancestors evolved. There’s a lot that we don’t know about the EEA; the fossil record is limited. Some of what we guess about it comes, through a kind of reverse engineering, from examining ourselves and trying to work out the kind of environment to which our attributes would have been well adapted.

  We know that the EEA was located in Africa; probably, though not certainly, scrubby savannah. It is plausible that our ancestors lived in these conditions as hunter-gatherers, perhaps in something like the way modern hunter-gatherer tribes live in the Kalahari but, at least in earlier periods, with a less developed technology. We know that fire was tamed more than a million years ago by Homo erectus, the species that was probably our immediate predecessor in evolution. It is controversial when our ancestors dispersed out of Africa. We know that there were Homo erectus in Asia a million years ago, but many believe that nobody today is descended from those early migrants and that all surviving humans are the descendants of a second, more recent exodus of Homo sapiens out of Africa.*32

  Whenever the exodus, there has evidently been time for humans to adapt to non-African conditions. Arctic humans are different from tropical. We northerners have lost the black pigmentation that our African ancestors presumably had. There has been time for biochemistries to diverge in response to diet. Some peoples – perhaps those with herding traditions – retain into adulthood the ability to digest milk. In other peoples, only children can digest it; the adults suffer from the condition known as lactose intolerance. Presumably the differences have evolved by natural selection in different culturally determined environments. If natural selection has had time to shape our bodies and our biochemistry since some of us left Africa, it should also have had time to shape our brains and our values. So we needn’t necessarily pay total heed to specifically African aspects of the EEA. Nevertheless, the genus Homo has spent at least nine-tenths of its time in Africa, and the hominins have spent 99 per cent of their time in Africa, so, insofar as our values are inherited from our ancestors, we might still expect a substantial African influence.

  Various researchers, most notably Gordon Orians of the University of Washington, have examined aesthetic preferences for various kinds of landscapes. What kinds of environments do we seek to recreate in our gardens? These workers try to relate the sorts of places we find attractive to the sorts of places that our wild ancestors would have encountered as nomads wandering from camp site to camp site in the EEA. For example, we might be expected to like trees of the genus Acacia or other trees that resemble them. We might prefer landscapes in which the trees were low and dotted about, rather than deep forest landscapes, or deserts, both of which might carry threatening messages for us.

  There seem to be some grounds for suspicion of this kind of work. Less justified would be a general scepticism that anything so complex or highflown as preference for a landscape could possibly be programmed into the genes. On the contrary, there is nothing intrinsically implausible about such values being inherited. Once again, a sexual parallel comes to mind. The act of sexual intercourse, if we contemplate it dispassionately, is pretty bizarre. The idea that there could be genes ‘for’ enjoying this preposterously unlikely act of rhythmic insertion and withdrawal might strike us as implausible in the extreme. But it is inescapable if we accept that sexual desire has evolved by Darwinian selection. Darwinian selection can’t work if there are no genes to select. And if we can inherit genes for enjoying penile insertion, there is nothing inherently implausible in the idea that we could inherit genes for admiring certain landscapes, enjoying certain kinds of music, hating the taste of mangoes or anything else.

  Fear of heights, manifesting itself in vertigo and in the common dreams of falling, might well be natural in species that spend a good deal of their time up trees, as our ancestors did. Fear of spiders, snakes and scorpions might with benefit be built into any African species. If you have a nightmare about snakes, it is just possible that you are dreaming not about symbolic phalluses but actually about snakes. Biologists have often noted that phobic reactions are commonly exhibited towards spiders and snakes, almost never to electric bulb sockets and motor cars. Yet in our temperate and urban world snakes and spiders no longer constitute a source of danger, while electric sockets and cars are potentially lethal.

  It is notoriously hard to persuade drivers to slow down in fog, or refrain from tailgating at high speed. The economist Armen Alchian ingeniously suggested abolishing seatbelts and compulsorily fixing a sharp spear to all cars, pointing from the centre of the steering wheel straight at the driver’s heart. I think I’d find it persuasive, whether for atavistic reasons I don’t know. Also persuasive is the following calculation. If a car travelling at 80 miles per hour is abruptly slammed to a halt, this is equivalent to hitting the ground after falling from a tall building. In other words, when you are driving fast, it is as if you were hanging from the top of a high-rise building by a rope sufficiently thin that its probability of breaking is equal to the probability that the driver in front of you will do something really stupid
. I know almost nobody who could happily sit on a windowsill up a skyscraper, few who unequivocally enjoy bungee-jumping. Yet almost everybody happily travels at high speed along motorways, even if they clearly understand in a cerebral way the danger they are in. I think it is quite plausible that we are genetically programmed to be afraid of heights and sharp points, but that we have to learn (and are not very good at) being afraid of travelling at high speeds.

  Social habits that are universal among all peoples, such as laughing, smiling, weeping, religion, and a statistical tendency to avoid incest, are likely to have been present in our common ancestors too. Hans Hass and Irenäus Eibl-Eibesfeldt travelled the world clandestinely filming people’s facial expressions, and concluded that there are cross-cultural universals in styles of flirting, of threatening, and in a fairly complicated repertoire of facial expressions. They filmed one child born blind, whose smile and other expressions of emotion were normal although she had never seen another face.

  Children notoriously have a highly developed sense of natural justice, and ‘not fair’ is one of the first expressions to spring to the lips of a disgruntled child. This does not, of course, show that a sense of fairness is universally built in by the genes, but some might consider it suggestive in the same kind of way as the smile of the child born blind. It would be tidy if different cultures, the world over, shared the same ideas of natural justice. But there are some disconcerting differences. Most people attending this lecture would think it unjust to punish an individual for the crimes of his grandfather. Yet there are cultures where the transgenerational vendetta is taken for granted and is presumably regarded as naturally just.*33 This perhaps suggests that, at least in detail, our sense of natural justice is pretty flexible and variable.