Breaking the Spell: Religion as a natural phenomenon, by Daniel C Dennett

To the dismay and bewilderment of secularists, a couple of centuries of scientific reason have not produced the changes they expected in the climate of belief. Some varieties of religion have dried up in the light of reason, but the global effect has been increased and ominous turbulence. Even the godless now fear what may come down from the skies. This sense of crisis underlies the programme that the philosopher Daniel Dennett seeks to introduce in his new book. Having failed to neutralise religion by indirect influence, scientific reason must be brought directly to bear on it. The benefits and costs of religious belief must be weighed scientifically; scientists must develop accounts of how religion arises from the way the mind works.

While Dennett sees an urgent need to analyse religion as a natural phenomenon, he sees little point in pursuing the question of whether it is a supernatural one as well. He is already a convinced atheist, and a self-proclaimed "bright" - a term intended to do for non-believers what "gay" has done for homosexuals, combining positive connotations with a sense of assertiveness and commitment. Believers, he suggests, might like to call themselves "supers", a similarly positive tag that refers to the supernatural.

They are unlikely to welcome his offering, or to mistake it for a genuine token of respect. Dennett is happy to let his disdain for religion show through the framework of inquiry, insistent that it is disinterested. His opening sally compares the Word of God to a parasite that infects ants' brains. That is the view of religion articulated by Richard Dawkins, who sees faiths as "viruses of the mind".

Though Dennett is evidently in sympathy with his English comrade, he differs crucially in strategy. Dawkins takes a polemical position and wages frank war on God. Dennett's engine of rational inquiry insists that hypotheses must remain on the table until settled according to scientific procedure.

He proposes that religious ideas may be parasitic, harming the hosts whose minds they occupy, or may have mutually beneficial relationships with their host, or may be "along for the ride". Deciding between these possibilities is a goal of the research programme he urges.

To some ears this may sound like overweening scientism, a vain belief in science as a superior form of religion. But the impression left by Dennett's strategic compromise between sentiment and inquiry is closer to managerial politics than any grand vision of enlightenment. His studiedly open set of options resembles one of those consultation exercises that local authorities like to conduct, whose wording offers clues to which policy the authority has already decided to pursue.

The book's title looks like a clue. "Breaking the spell" suggests the removal of an external influence, probably a malign one. Dennett argues convincingly that the human readiness to believe in supernatural entities may have arisen from the evolved capacity to understand that other individuals have intentions.

This understanding provides a basis for empathy and a defence against deception as well as a means to anticipate others' behaviour. These capacities proved so valuable that people came to assume that intentions lay behind all events, and conjured up supernatural beings to account for intentions which could not be ascribed to humans or animals. These entities, or rather ideas of them, have evolved and multiplied in human minds.

In the course of their evolution they have imposed immense costs on their hosts. As Dennett points out, "To an evolutionist, rituals stand out like peacocks in a sunlit glade". To an evolutionist, the extravagance of a peacock's fan insists it must confer benefits that outweigh its costs. To an evolutionary psychologist, the universal extravagance of religious rituals, with their costs in time, resources, pain and privation, should suggest as vividly as a mandrill's bottom that religion may be adaptive. It might also suggest that, as costly signals are hard to fake and tend to be reliable, these advantages may relate to the establishment of trust, turning groups into communities.

But here, as throughout Breaking the Spell, the revving of a powerful analytical engine is followed by the sound of it slipping out of gear. Dennett suggests that rituals may be a means of ensuring that the ideas behind them are reproduced faithfully. The possibility that rituals benefit the individuals performing them disappears behind the focus on what may be in it for the ideas themselves.

Dennett is not alone among Darwinians in his reluctance to explore the possibility that religion is adaptive. Only a few evolutionary thinkers have taken up the study of religion, and evolutionary psychology has failed to incorporate religion with the same enthusiasm it has applied to other universal forms of human behaviour, despite its keen interest in the evolution of morality.

Dennett's reference to anthropologists' and historians' "blinkered perspectives" smacks of pots and kettles. Towards the end, he does declare that "scientists have much to learn from the historians and cultural anthropologists", but by that stage it evokes the distinction he draws earlier between what people profess and what they actually believe.

This absence may also reflect Dennett's intellectual preference, as a philosopher, for abstraction rather than detail. The minutiae of anthropologists' field observations or the archaeological record do not detain him on his way to his discussion of religion today, where past form can be dismissed. Even if religion was the foundation of human society, you could say the same about hunting and gathering, and we're not going back to that. People nowadays can live moral lives and sustain decent communities without religious faith, although it seems only fair to acknowledge that these moral frameworks are secularised versions of earlier religious ones.

Dennett insists that a sense of reverence for the sacred is not a necessary qualification for the analysis of religion. But the project does require passion, which is why Dawkins's essayistic voice is convincing even if not found persuasive. Breaking the Spell reads like an exercise in fulfilling a disagreeable obligation. It takes a hundred pages to justify itself, drops interesting ideas no sooner than it has picked them up, and we're still waiting for it to hit its stride when we run into the appendices. The style is familiar but the effect very different from a book such as Darwin's Dangerous Idea, where Dennett's voice is filled by something he really believes in.

To make real progress his project must engage with religion both as a natural phenomenon and a social one. It is in the latter terms that we can best make sense of why a populist Christian movement arose in America, insisting that the earth is just a few thousand years old, a good century after educated Britons had come to accept that the planet is far older than the Bible suggests. Scientists trying to comprehend religion don't just need to learn from historians and anthropologists; they need to learn how to make their fellow scholars into allies.

Marek Kohn's A Reason for Everything is published by Faber