I found this review of Breaking the Spell from two years ago while looking for something else. It was in the Guardian at the time, but it still makes sense, at least to me.
It is hard being an atheist with a sense of proportion. No one in this country, will persecute you; and it’s not really very hard to disbelieve in God, but the temptation to strike attitudes in front of the universe persists even in people who are about to spend 450 pages arguing that the universe is not the sort of thing that might be impressed. Thus, Daniel Dennett writes early in this book: "I for one am not in awe of your faith. I am appalled by your arrogance, by your unreasoning certainty that you have all the answers." and he’s not talking about Richard Dawkins. He goes on "I wonder if any believers in the End Times will have the intellectual honesty and courage to read this book through"
Intellectual honesty and courage are not the only qualities required to reach the far end of the book. In his preface, Dennett does remark that every foreign reader who saw drafts of the book complained of its American bias. His defence is that it’s aimed at an American audience, since it is American fundamentalism which most threatens what he values about his own society. So after the preliminary pep-talk to the choir, he gives a very forceful and lucid account of the reasons why we need to study religious behaviour as a human phenomenon: apparently this programme comes as a tremendous shock to those Americans who have never heard of David Hume, William James, or Terry Pratchett.
This is followed by an excellent and clear summary of the state of some new-ish scientific research into the psychology of religious belief. If you want to naturalise religion, as Dennett does, and to show that it is a human activity arising form the normal workings of nature, then you need to discover what parts of our evolved human nature it appeals to. There is in fact quite a lot of psychological research into our capacity to believe in ghosts, spirits, and other things for which there is no experimental warrant. The anthropologists Pascal Boyer and Scott Atran have both written very interestingly on the subject, and Dennett summarises and credits their work in a way that should do much to promote it.
Dennett understands there are vast differences between primitive or animist religions and the sophisticated beliefs of a modern Jesuit. This isn’t always clear in Boyer or Atran, who would like their explanations to cover all forms of religious activity. But Dennett sees that religious feelings are modified by the social structures in which they are expressed and that there is an important difference between believing in a crocodile god, who lives under the mountain five miles away, and subscribing to the doctrine of the Trinity.
A simple psychological account won’t do here, any more than psychology can explain economics, or sociology; psychology may explain how we will react to our social environments, but it can’t explain or predict how these environments will arise and change.
Another great advantage to Dennett’s book is that he sees that religious belief is not really propositional. The next president of the USA could be a man who believes that America was peopled by one of the lost tribes of Israel, and later extensively visited by Jesus in pre-Columbian times; that it makes sense to baptise your dead ancestors, and that all these truths were revealed, on tablets of gold, by a being called the Angel Moroni, to a farmhand in upstate New York. Mitt Romney, the Republican governor of Massachusetts, and a serious presidential candidate, is a Mormon. Richard Dawkins might regard Romney’s professed beliefs as evidence of simple insanity. Dennett sees that their status is more complicated and interesting than that. He understands that modern religions derive their coherence precisely from the fact a creed is a statement of belonging as much as of belief.
So he doesn’t skirt the complications of theorising about religion: he sees the difficulties, and he marches bravely into the swamp and then – about half way through the book – at exactly the point where we’re wondering how to reach firm ground, he stops, inflates a hot air balloon that’s labelled "memes", climbs into it, and floats away.
Memes are familiar to readers of Dennett’s earlier work. They are a ideas, words, tunes, strategies, catchphrases – anything that people can copy, or appear to copy from one another. In every case where the word is used, it can be replaced by one of these other terms with a corresponding gain in precision and explanatory power. If we look at Governor Romney’s Mormon beliefs, they are essentially a way of marking what tribes he belongs to. In that very important sense, they are arbitrary signifiers, like words. They mean what their indented audience takes them to mean.
And if you’re really trying to produce a naturalistic account of religion, "memes" distract from the worrying and frightening questions. People like Dennett and Dawkins who pride themselves on their tough-minded, ruthless, reductionist approach to biology never seem to apply this kind of reasoning to human society. Why should we expect religions to behave for the benefit of professors in Cambridge, or Oxford, or even for the benefit of humanity?
If we are going to be atheists, and to regard religions as human constructions serving human ends, we should not shrink from the idea that these ends are likely to be sometimes inimical to other humans outside the group. For all the rhetoric about the wickedness of religious belief, I don’t think Dennett takes this idea very seriously.
Religions are one of the ways in which humans understand, and create their own societies. Thus they are essential to warlike societies as much as to peaceful ones. The urgent question isn’t whether religion provokes warlike behaviour. It is whether warlike behaviour benefits those who carry it out, for if it does, religions will surely find ways to justify it.
Few of us in this culture now are in favour of fanaticism; but it is obviously possible to be a fanatical atheist, so it turns out to be fanaticism that’s the problem, not religion. More profoundly, a scientific or evolutionary analysis of fanaticism might ask what use it was to fanatics and the answer is clearly that sometimes it was very useful indeed – at least to their surviving relatives and to their tribe. This may be difficult for us to see because the myth we learnt was that fanaticism was a substitute for high technology. Fanatics were the guys galloping towards the machine guns, and reasoned, logical, scientific people sat behind the machine guns and calmly mowed them down. But that may have been a nineteenth century aberration, like the belief that the enlightenment must spread simply because our beliefs are true.