Third, the 'central theorem' may substitute for 'genes' the more general term 'replicators'. The fact that religion is ubiquitous probably means that it has worked to the benefit of something, but it may not be us or our genes. It may be to the benefit of only the religious ideas themselves, to the extent that they behave in a somewhat gene-like way, as replicators. I shall deal with this below, under the heading 'Tread softly, because you tread on my memes'. Meanwhile, I press on with more traditional interpretations of Darwinism, in which 'benefit' is assumed to mean benefit to individual survival and reproduction.

Hunter-gatherer peoples such as Australian aboriginal tribes presumably live in something like the way our distant ancestors did. The New Zealand/Australian philosopher of science Kim Sterelny points up a dramatic contrast in their lives. On the one hand aboriginals are superb survivors under conditions that test their practical skills to the uttermost. But, Sterelny goes on, intelligent as our species might be, we are perversely intelligent. The very same peoples who are so savvy about the natural world and how to survive in it simultaneously clutter their minds with beliefs that are palpably false and for which the word 'useless' is a generous understatement. Sterelny himself is familiar with aboriginal peoples of Papua New Guinea. They survive under arduous conditions where food is hard to come by, by dint of 'a legendarily accurate understanding of their biological environment. But they combine this understanding with deep and destructive obsessions about female menstrual pollution and about witchcraft. Many of the local cultures are tormented by fears of witchcraft and magic, and by the violence that accompanies those fears.' Sterelny challenges us to explain 'how we can be simultaneously so smart and so dumb'.⁷⁶

Though the details differ across the world, no known culture lacks some version of the time-consuming, wealth-consuming, hostility-provoking rituals, the anti-factual, counter-productive fantasies of religion. Some educated individuals may have abandoned religion, but all were brought up in a religious culture from which they usually had to make a conscious decision to depart. The old Northern Ireland joke, 'Yes, but are you a Protestant atheist or a Catholic atheist?', is spiked with bitter truth. Religious behaviour can be called a human universal in the same way as heterosexual behaviour can. Both generalizations allow individual exceptions, but all those exceptions understand only too well the rule from which they have departed. Universal features of a species demand a Darwinian explanation.

Obviously, there is no difficulty in explaining the Darwinian advantage of sexual behaviour. It is about making babies, even on those occasions where contraception or homosexuality seems to belie it. But what about religious behaviour? Why do humans fast, kneel, genuflect, self-flagellate, nod maniacally towards a wall, crusade, or otherwise indulge in costly practices that can consume life and, in extreme cases, terminate it?

There is a little evidence that religious belief protects people from stress-related diseases. The evidence is not strong, but it would not be surprising if it were true, for the same kind of reason as faith- healing might turn out to work in a few cases. I wish it were not necessary to add that such beneficial effects in no way boost the truth value of religion's claims. In George Bernard Shaw's words, 'The fact that a believer is happier than a skeptic is no more to the point than the fact that a drunken man is happier than a sober one.'

Part of what a doctor can give a patient is consolation and reassurance. This is not to be dismissed out of hand. My doctor doesn't literally practise faith-healing by laying on of hands. But many's the time I've been instantly 'cured' of some minor ailment by a reassuring voice from an intelligent face surmounting a stethoscope. The placebo effect is well documented and not even very mysterious. Dummy pills, with no pharmacological activity at all, demonstrably improve health. That is why double-blind drug trials must use placebos as controls. It's why homoeopathic remedies appear to work, even though they are so dilute that they have the same amount of active ingredient as the placebo control — zero molecules. Incidentally, an unfortunate by-product of the encroachment by lawyers on doctors' territory is that doctors are now afraid to prescribe placebos in normal practice. Or bureaucracy may oblige them to identify the placebo in written notes to which the patient has access, which of course defeats the object. Homoeopaths may be achieving relative success because they, unlike orthodox practitioners, are still allowed to administer placebos — under another name. They also have more time to devote to talking and simply being kind to the patient. In the early part of its long history, moreover, homoeopathy's reputation was inadvertently enhanced by the fact that its remedies did nothing at all — by contrast with orthodox medical practices, such as bloodletting, which did active harm.

Is religion a placebo that prolongs life by reducing stress? Possibly, although the theory must run a gauntlet of sceptics who point out the many circumstances in which religion causes rather than relieves stress. It is hard to believe, for example, that health is improved by the semi-permanent state of morbid guilt suffered by a Roman Catholic possessed of normal human frailty and less than normal intelligence. Perhaps it is unfair to single out the Catholics. The American comedian Cathy Ladman observes that 'All religions are the same: religion is basically guilt, with different holidays.' In any case, I find the placebo theory unworthy of the massively pervasive worldwide phenomenon of religion. I don't think the reason we have religion is that it reduced the stress levels of our ancestors. That's not a big enough theory for the job, although it may have played a subsidiary role. Religion is a large phenomenon and it needs a large theory to explain it.

Other theories miss the point of Darwinian explanations altogether. I'm talking about suggestions like 'religion satisfies our curiosity about the universe and our place in it', or 'religion is consoling'. There may be some psychological truth here, as we shall see in Chapter 10, but neither is in itself a Darwinian explanation. As Steven Pinker pointedly said of the consolation theory, in How the Mind Works: 'it only raises the question of why a mind would evolve to find comfort in beliefs it can plainly see are false. A freezing person finds no comfort in believing he is warm; a person face-to-face with a lion is not put at ease by the conviction that it is a rabbit.' At the very least, the consolation theory needs to be translated into Darwinian terms, and that is harder than you might think. Psychological explanations to the effect that people find some belief agreeable or disagreeable are proximate, not ultimate, explanations.

Darwinians make much of this distinction between proximate and ultimate. The proximate explanation for the explosion in the cylinder of an internal combustion engine invokes the sparking plug. The ultimate explanation concerns the purpose for which the explosion was designed: to impel a piston from the cylinder, thereby turning a crankshaft. The proximate cause of religion might be hyperactivity in a particular node of the brain. I shall not pursue the neurological idea of a 'god centre' in the brain because I am not concerned here with proximate questions. That is not to belittle them. I recommend Michael Shermer's How We Believe: The Search for God in an Age of Science for a succinct discussion, which includes the suggestion by Michael Persinger and others that visionary religious experiences are related to temporal lobe epilepsy.