Some are bizarre: the theory that culture, ritual and religion developed as a by-product of concealed ovulation, synchronised menstruation and the resulting sexual mystification of men fascinates Kohn, though not, I think, to the point where he entirely believes in it. But all the accounts have one feature in common which casts an odd light on the relations between religion and science. They all, one way or another, propose that that a religious imagination may have been necessary for us to become human.
This is not an argument from classification. It doesn't say we were not human before we were religious, or artistic: on the contrary, it says we were religious before we were human, and without the emotions necessary for ritual and art we would never have developed fully human intelligence. It is bound up with an argument about the costs and benefits of language, and the role of communications in animal life, which is subtle, and profoundly pessimistic.
Communication, in one influential theory, originated by Richard Dawkins and John Krebs, is essentially to be understood as an effort by an animal to manipulate its fellows. But if the point of your communication is to get me to do what you want, then what's in it for me? The central problem of language, from a socio-biological point of view, is not why the capacity to speak it evolved, but how we got the capacity to understand it, when it is so often advantageous to other people to tell us lies.
The answer is that language must have been built on pre-existing mechanisms of trust. It's worthwhile for an animal to trust another if lying would be just as expensive as telling the truth. And of course, this calculus persists into our modern lives: journalists are supposed to ask themselves at every turn "Why is this lying bastard lying to me?" and to accept the truth of a statement only if the speaker gains nothing by deceiving us. But at this point the argument takes a sharp twist. For we do not in fact go through life asking all the time why everyone is lying to us. The emotional strain of treating everyone like a salesman is just too great, except for a few psychopaths and novelists.
We are adapted, then, to trust people, or at least not to distrust all of them all of the time. If this were not the case, and we did not have an emotional predisposition to trust people, language, and so human intelligence, could never have developed. But, since it is the basic principle of sociobiology that emotions are by and large adaptive, it follows that our ancestors must have been largely trustworthy, and have peopled their imaginations with trustworthy beings.
The trick to these deductions is rather like reading an insurance policy. Insurance companies are really just very successful bookies, who only offer bets which they will, if they take enough of them, be bound to win. So if someone is willing to insure you for something at less than ruinous odds, you can be fairly certain it won't happen. There are no guarantees, of course. But in the long run, the insurance company will always win. Similarly, if there are deep-lying patterns of emotion and behaviour in humans, we can be sure that though they will sometimes lead to great error, they will usually pay off. We find it agreeable to believe each other because we find it agreeable to tell the truth.
But if you accept this view of our emotions as fundamental to our intellectual development it follows that we are inescapably animals who put an emotional cast on the world. We may not see it through rose-tinted spectacles, but we certainly see it organised and given form by emotion. This I take to be the fundamental attitude from which all religious experiences and belief systems spring: that the world means something; that it has a natural emotional charge. That, in turn, is why it makes sense to speak of some forms of atheism as religious, because they seem to see the meaning of the world in the non-existence of God.
Nor is there any reason to suppose mankind will grow out if this. If you look back at the high doctrines of progressive atheism current in the Twenties and Thirties - perhaps the best example would be Olaf Stapledon's majestic science fiction novel Last and First Men, what has dated is the hope, now easily identifiable as religious, that we would grow out of our stupid religions. And it's hard to see how any sociobiologist could ever expect the religious emotions to fade away. If emotions are, just like insurance policies, a container of statistical knowledge about the way the world works, then we could not lose our tendency to project them on the world without losing all our higher faculties.
Of course, emotions don't feel like abstract statistical knowledge at all. To see how they can work like that, it's worth while reading the book to discover how the urge to show off when chipping a flint axe may have been one of the really essential things that made our ancestors human and show, by twisting routes led to a world which people change by chipping larger, more elaborate theories.