Outgrowing God

by Richard Dawkins

  Here’s a slightly more technical way of putting the point. Humans have a tendency to believe in agency. What is agency? Well, an agent is a thing that deliberately does something for a purpose. When the wind rustles the long grass, there is no agency. Wind is not an agent. A lion is an agent. A lion is an agent whose purpose is to eat you. It will modify its behaviour in sophisticated ways in order to catch you, and work energetically and flexibly to thwart your efforts to escape. It’s worth being scared of agency. But it can be a waste of time and effort, because the suspected agent may be something like the wind. The more dangerous your life tends to be on average, the more the balance should shift towards seeing agents everywhere and therefore sometimes believing falsehoods.

  Nowadays we mostly no longer have to be scared of lions or sabretooths. But even modern humans can be scared of the dark. Children are scared of bogeymen. Adults are scared of muggers and burglars. Alone in bed at night, you hear a noise. It could be the wind. It could be the timbers of the old house, creaking as they settle. But it could be an armed burglar. Maybe nothing so specific as a burglar. As far as you are concerned, you fear an unnamed agent, as opposed to a non-agent like the wind or a creaking beam. The fear of agents, even if irrational, even if inappropriate on this particular occasion, may lurk within us from our ancestral past. My colleague Dr Andy Thomson put it like this in his book Why We Believe in God(s): we are likely to mistake a shadow for a burglar; we are unlikely to mistake a burglar for a shadow. We have a bias towards seeing agents, even when there aren’t any. And religion is all about seeing agency all around us.

  Our ancestors’ religions were ‘animistic’: they saw agents everywhere they looked, and often they called them gods. This is how the Greek gods started out, as is clear from Stephen Fry’s lovely book Mythos. All over the world there were river gods and thunder gods, sea gods and moon gods, fire gods and sun gods, gods – or perhaps demons – of the dark forest. The sun was a god, an agent who had to be wooed and placated with prayers and sacrifices, otherwise he might decide not to rise tomorrow. The fire was a god who needed feeding or he’d go out. Thunder was a god – what else but a god could account for such a terrifying noise? The weather was so unpredictable, yet so important to life, it was natural to think agents were behind its changing moods. Surely there must be a way to end the terrible drought? A really big sacrifice to the rain god might do it. A terrible storm just wrecked our house. Perhaps we didn’t heap sufficient praise on the storm gods and they were angry.

  Yahweh evolved in people’s minds to become the one God of the Jews, and eventually of Christians and Muslims too. Before that he was a ‘storm god’, one of many gods of the Canaanite peoples from whom the Jews sprang. Other Bronze Age Canaanite gods who were originally worshipped alongside Yahweh included Baal the fertility god and El, the chief god, and his wife the goddess Asherah. According to some scholars of religious history, Yahweh was later merged in people’s minds with El and Asherah to eventually become the one and only God of the Jews. So Bronze Age animism came to be pared down to Iron Age monotheism. Later, Christianity and Islam adopted the God of the Jews. And later still the storm god of the Canaanites evolved further sophistication and became the hero of books on theology by learned professors at Oxford and Harvard.

  I suggested that people made sacrifices to the weather gods in the hope of breaking a drought. But why would they think it might help? The human brain is a pattern-seeker. Natural selection has built into our brains a tendency to notice patterns such as sequences: what follows what. We notice that thunder follows lightning, rain follows after grey clouds gather, crops don’t grow if there is no rain. But ‘what follows what’ is complicated. ‘What follows what’ turns out to mean not ‘what always follows what’ but ‘what sometimes follows what’. Pregnancy follows sexual intercourse, but only sometimes.

  Often we think we notice a pattern when there really isn’t one. Often we fail to notice a pattern when there really is one. The mathematicians known as statisticians distinguish two ways of getting things wrong when we try to recognize these patterns. They call them false positives and false negatives. A false positive is thinking you see a pattern when there isn’t one. Superstition is a common type of false positive error. A false negative is failing to notice a pattern when there really is one. There’s a real pattern between being bitten by a mosquito and getting malaria. But it doesn’t invariably follow, and nobody picked up on it until Sir Ronald Ross in 1897. There’s no real pattern between a black cat crossing your path and subsequent misfortune. But many superstitious people have believed that particular false positive.

  Last year we prayed to the rain gods and it then rained.

  Surely that pattern must have meant something?

  No, it was meaningless. A false positive. It was going to rain anyway. But it’s hard to shake off the superstition.

  The child was ill with a fever. We sacrificed a goat to the gods and the child got better. So we’d better sacrifice a goat the next time somebody gets a high fever.

  The immune system often cures people of malaria anyway. But try telling that to a superstitious person who is convinced that sacrificing a goat did the trick.

  Even if you notice an unvaryingly repeated pattern – something follows something else reliably, every single time – it doesn’t prove that the earlier event caused the later one. The church clock in the village of Runton Acorn always strikes the hour shortly before the clock in the neighbouring village of Runton Parva. But does the Runton Acorn clock cause the Runton Parva clock to strike? Observation alone can’t settle the question. Not even repeated observation. The only sure way to demonstrate cause is an experiment. You have to manipulate the situation. Climb up into the Runton Acorn tower and stop the clock. Does the Runton Parva clock then fail to strike? Then experimentally set the Runton Acorn clock ten minutes fast. Does the Runton Parva clock still strike just after it? Of course you have to repeat the experiment a respectable number of times to rule out chance – random luck.

  It takes a sophisticated, perhaps rather nerdish mind to do proper experiments to test whether an apparent pattern is really there. You’d have to be very nerdish indeed to bother to do the church clocks experiment. And if the question is whether a noise really is a lion, the experimental approach could be fatal. No wonder our ancestors resorted to superstition instead.

  A famous experimental psychologist called B. F. Skinner demonstrated superstition in pigeons. His pigeons ‘noticed’ patterns that were not really there: false positives. Each of eight pigeons was placed in a separate box called a ‘Skinner box’. Each box had an electrically operated feeding apparatus which could deliver food to the hungry pigeons. Normally, Skinner boxes are wired up to deliver food only when the bird does something, like peck a switch in the wall of the box. But Skinner did something different for this particular experiment. He severed the connection between the feeding apparatus and the birds’ behaviour. Nothing that the birds did had any effect on whether they got fed. Food was delivered into the box sporadically, regardless of what the bird did. Or, indeed, if it did nothing.

  The result was fascinating. Six of the eight birds developed superstitious habits of various kinds. One bird walked round and round in an anti-clockwise direction, making two or three turns between rewards. We could say it had a superstitious belief that turning anti-clockwise caused the food to come. A second bird repeatedly thrust its head into one of the upper corners of the box. It ‘thought’ that that was what persuaded the feeding apparatus to deliver. Two other birds developed a ‘pendulum’ habit with the head. They thrust the head fast to left or right, then brought it back more slowly. Another bird’s superstitious habit was to toss the head upwards, as though throwing some non-existent object up in the air. And the sixth bird directed pecking movements towards the floor, without ever hitting the floor.

  Skinner called it superstitious behaviour and I think he was right to do so. What must have h
appened is this. A bird just happened to make a particular movement, say thrusting its head up into the corner, immediately before the feeding apparatus clunked into action. The bird ‘thought’ (not necessarily consciously) it was its head movement that had caused the food to arrive. So it did it again. And, as it happened, that was just the right time for the next food consignment to arrive. Each bird learned a different superstitious habit, repeating whatever it happened to do before food arrived by chance. And that, it seems likely, is how our ancestors developed the habit of, say, praying, or sacrificing a goat, to cure a child of a fever. The other resemblance between Skinner’s pigeons and humans is that, in different parts of the world, local peoples develop different superstitious beliefs. Just like the six different pigeons in their ‘local’ Skinner boxes.

  Gamblers, too, whether at the roulette wheel or a one-armed bandit, are rewarded at random, whatever they do. A gambler thinks he notices that he has more luck when he wears his ‘lucky shirt’. Or he once prayed for luck and promptly won the jackpot. Just like Skinner’s pigeons, he does it again. Never wins the jackpot again but can’t rid himself of the habit of praying. You can’t influence the probability that a slot machine will deliver the jackpot. Or that the ball on a roulette wheel will land where you want it to. Yet gamblers from Monte Carlo to Las Vegas are riddled with superstitious beliefs that they can.

  Long ago, before computers had screens, they printed things out on a teleprinter instead. Once, while working in my university computer room, I watched a student who was desperately impatient for the computer to respond. He repeatedly rapped his knuckles on the teleprinter, although he must really have known it couldn’t possibly persuade the computer to hurry up. Maybe he had once done it just before the computer happened to spew out its results anyway and never rid himself of the superstitious habit. Like Skinner’s pigeons.

  Let’s suppose that in a time of drought our ancestors took it into their heads to sacrifice to the rain god. Every day. And eventually the rain came. Maybe a lot of sacrifices were necessary before the rain god – as they thought – was persuaded. The superstitious people never tried the experiment of not sacrificing to the rain god, to see if rain would come anyway. That’s what a scientist would do. But our ancestors weren’t scientists. And they didn’t dare risk not sacrificing to the rain god.

  Of course, I’m speculating. But I think it’s plausible. It’s exactly the kind of thing many tribespeople do, to this day. And Skinner’s experiments were not speculation. They really happened. Nor is it speculation that human gamblers trust in lucky numbers, lucky charms and prayers. People tend to pray, or develop superstitious habits, whenever there’s uncertainty as to what will happen (what we call ‘chance’ or ‘luck’) and we want a particular outcome. Superstition in itself probably didn’t help our ancestors to survive. But a general tendency to look for patterns in the world – making an effort to notice when events tend to be followed by important other events – probably did. And superstition was a byproduct of this. As with the zebras balancing the risk of being eaten against the risk of not eating enough, human pattern-seekers had to strike a balance between two risks: the risk of noticing a pattern when there isn’t one (superstitious false positive) and the risk of failing to notice a pattern when there is one (false negative). A tendency to notice patterns was favoured by natural selection. Superstition and religious belief were a byproduct of that tendency.

  Now here’s another line of thought. Our earliest human ancestors lived in a dangerous place, the African savanna. There were poisonous snakes, scorpions, spiders and centipedes underfoot. There were pythons and leopards lurking in the trees, lions behind bushes, crocodiles in the river. Adults knew of these dangers but children needed to be told. Parents would surely have warned their children, just as parents in modern cities warn their children to look left and right before crossing a road. Natural selection would have favoured parents who warned their children. And natural selection would have favoured genes that built into child brains a tendency to believe their parents.

  That much is easy to understand. Now for the tricky part. If adults ever gave children bad advice alongside good advice, the child brain would have no way to distinguish bad advice from good. If the child brain were capable of making that distinction, the adult advice wouldn’t be necessary anyway. The child would just know, for instance, that snakes are dangerous. The whole point is that if children already knew, parents wouldn’t need to tell them. So if, for some reason, a parent were to give a child useless advice – like ‘You have to pray five times a day’ – the child would have no way of knowing that it was useless. Natural selection simply builds into the child brain the rule ‘Believe whatever your parents tell you’. And that rule will come into force even when ‘what your parents tell you’ is actually silly or untrue. Or just based on a pigeon-like superstition.

  But, you are probably asking, why should a parent tell a child something silly or untrue? Well, the parents themselves were once children. They were once given advice by their own parents. They too had no way of judging which advice was good and which advice was useless or bad. Advice, whether good or bad, gets passed on to the next generation. As for how it got started in the first place, pigeon-like superstition was probably part of the story. As the generations went by, the useless or superstitious advice got modified, amplified, by the same Chinese Whispers effect we saw at work in Chapters 2 and 3. In different parts of the world, different advice would get passed on. Which is exactly what we notice has happened, when we look around the world.

  Of course some intelligent children, when they grow up, look at the evidence and succeed in breaking away from bad or useless advice from previous generations – grow out of it. Think about the title of this book. But that doesn’t always happen, and I believe this partly explains how religions get started and why they persist. It’s a sort of byproduct theory. Useless or superstitious beliefs, like the need to pray five times a day, or the need to sacrifice a goat to cure malaria, get passed on as a byproduct of sensible beliefs – or rather, as a byproduct of child brains being shaped by natural selection to believe parents, teachers, priests and other elders. And that is favoured by natural selection, because much of what elders tell children is sensible.

  The byproduct theory is a truly Darwinian explanation for religious beliefs. True Darwinian explanations are all about genes becoming more numerous in a population. There are other kinds of explanation which look a bit like Darwinian explanations but aren’t really. For instance, whole groups or nations might survive better because of their religion. And this means the religion itself survives. Suppose two nations have different religions. One has a warlike god, like Yahweh/Allah. Or like the warlike gods of the Vikings. The priests of such gods preach the virtues of courage in battle. They teach, perhaps, that a warrior who dies a martyr’s death will go straight to a special martyrs’ heaven. Or will go straight to Valhalla. They might even promise beautiful virgins in heaven to those men who die fighting for the tribal god (do you, like me, feel sorry for the poor virgins?). The other nation has a peaceful god or gods. Their priests don’t advocate war. They don’t preach heavenly bliss for those who die fighting. Maybe they don’t preach any kind of heaven at all. All other things being equal, which nation will have the bravest warriors? Which nation is more likely to conquer the other? And therefore, which of the two religions is most likely to spread? The question answers itself. It is a matter of history that the spread of Islam, from Arabia throughout the Middle East and the Indian subcontinent, was due to military conquest. And the same goes for the spread of Christianity by the Spanish conquerors in South and Central America.

  There are other possible ways in which religions might help nations or tribes, as well as in warfare. It’s been suggested – I think quite plausibly – that a shared religion, and shared myths, rituals and traditions, helped societies to bond together and cooperate in ways that benefited everybody in them. It may seem daft to
pray for rain, since modern science knows that praying for rain can’t affect the weather. But what if coming together in a rhythmic rain dance helps promote solidarity and cooperation in the tribe? It’s worth a thought, and respected colleagues have given it one.*1 Another possible non-Darwinian reason for the flourishing of religion is that kings and priests exploited the faith of their people as a means of dominating their societies. Yet another (and actually this one is close to being truly Darwinian) is the theory that ideas themselves – I’ve called them ‘memes’, to distinguish them from genes – including religious ideas, compete against rival memes in a gene-like way, to become more numerous in minds. There’s no space here to explore these various theories; I just mention them to give you an idea of the kinds of debates being pursued. But now I need to move on.

  In Chapter 6, I promised I’d return to the question of why natural selection favours niceness – at least, a limited form of niceness, which might serve as a kind of evolutionary basis for morality, a sense of what is good, and the desirability of doing good things. But I must say first that I think the changes in morality that I talked about in Chapter 6 are more important. Natural selection may build into our brains the basis for a limited amount of niceness. But it builds in the basis for nastiness too. As so often, there’s a balance. What has happened in history is that the balance has shifted. In the nice direction, as we saw in Chapter 6.

  So, what is the evolutionary basis for niceness? In Chapter 8 we saw that evolution is all about successful genes becoming more frequent in the gene pool (that’s what successful means). Genes that equip individuals to run faster (though not so fast that their legs break like a racehorse’s) become more numerous. Genes that make moths, lizards and frogs harder to see against tree bark become more numerous. Genes that make parents care for their children become more numerous, because copies of the very same genes survive in the bodies of the children cared for. So, being nice to your own children is a no-brainer, as far as natural selection is concerned.