In an interview he gave not long before he died, P G Wodehouse was asked whether he had any religious beliefs. His reply - "It's frightfully hard to say" - has always seemed to me to display an insight into the ambiguities of belief that is commonly lacking when philosophers talk about religion. Typically, philosophers take it for granted that religions are systems of belief, and condemn them for failing to meet standards of proof that are applied in other areas of human life, above all in science. Contemporary philosophy has not advanced very far beyond the views of the late-Victorian anthropologist J G Frazer, who, in his once hugely influential study of myth, The Golden Bough (1890), portrayed religion as a form of magical thinking.
In the spirit of 19th-century positivism, Frazer assumed that myths were primitive scientific theories - a highly reductive view that later cultural anthropologists have abandoned. Yet the idea of religion as a magical belief system remains the default position of most contemporary philosophers, and it is evident on every page of Daniel Dennett's Breaking the Spell.
Dennett is best known for his militant atheism, and, like other evangelists of unbelief, he views the world through the conceptual grid of western monotheism. His view of religion itself proves this; he defines it as a social system "whose participants avow belief in a supernatural agent or agents whose approval is to be sought". This may be commonplace as a perception of religion, but it is also highly parochial. One cannot make a sharp distinction between natural processes and supernatural agents unless one presupposes a view of the world something like that presented in the biblical creation story, and the distinction is not found in most of the world's religions. For example, in animism - which must rank as the oldest and most universal religion - spirits are seen as part of the natural world.
More fundamentally, it is a mistake to assume that belief is the core of religion. This may seem self-evident to many philosophers, but in fact belief is not very important in most religions. Even within Christianity there are traditions, such as Eastern Orthodoxy, in which it has never been central. For the majority of humankind, religion has always been about practice rather than belief. In fixating on the belief-content of religion, Dennett emulates Christianity at its most rationalistic and dogmatic. Pascal knew better, and understood that faith is not so much the basis of the religious life as a derivative from it. Dennett mocks those who say that life without faith has no meaning as "believers in belief". Yet he displays a zealous faith in unbelief that is far more inimical to doubt, and there is more scepticism in a single line of the Pensees than in the whole of Dennett's leaden tome.
Breaking the Spell approaches its subject with a relentless, simple-minded cleverness that precludes anything like profundity, and much of it seems designed to demonstrate the author's intellectual ingenuity rather than to advance the reader's understanding. (For example, there is a curious appendix on Kim Philby in which Dennett maintains that Philby's career is a real-life instance of W V Quine's principle of radical indeterminacy.) When Dennett delivers on the promise of the book - a naturalistic explanation of religion - the result is embarrassingly naive. The explanation turns out to be a variation on Richard Dawkins's theory of memes - units of information whose competition somehow explains the development of thought. One problem with memes is that, unlike genes, they are not identifiable physical structures. Ideas are elusive things - think of the ways in which artistic styles emerge and develop. It shows a sorry lack of cultural understanding to imagine that the baroque, say, can be reduced to a few simple structures.
In a postscript, Dennett defends memes against the criticism that they lack the clear identity of genes, but the real objection is that it is not a theory at all, as it fails to identify anything like a mechanism of cultural evolution. This is hardly surprising, given that there is nothing in the history of ideas that resembles natural selection in biology. Some ideas seem to be more contagious than others, but those which prevail are often the ones that have power on their side. Pagan religion did not disappear from the ancient world because it lost out in competition with non-pagan memes but because, following the conversion of Constantine, it was repressed. Like other evolutionist ideologies, the theory of memes passes over the role of power in history.
The appeal of the theory is that it reduces the fertile chaos of human thought to objects that can be manipulated, and seems to open up the prospect of memetic engineering - consciously directing the intellectual evolution of the species by disseminating some memes and discouraging others. In previous books Dennett has hinted that human evolution could be directed in this way, with his own ideas helping to guide the process, but happily the possibilities of memetic engineering are rather limited. Ideas can be suppressed, but they cannot be controlled. They have too many unexpected consequences, and always slip out of the hands of their authors. Like history as a whole, the history of ideas will always be partly a matter of chance. An attempt to defeat this contingency, the theory of memes is at bottom an expression of magical thinking and as remote from genuine science as "intelligent design".
Unlike Dennett, Lewis Wolpert writes as a practising scientist, and Six Impossible Things Before Breakfast is a robust defence of materialism that contains much interesting information. Relying heavily on evidence from the advancing science of evolutionary psychology, Wolpert interprets religion as a type of adaptive behaviour in which our beliefs are shaped by our practical needs. Like Dennett, he seems ignorant of the vast range of religious traditions in which belief is peripheral. Again, he thinks of religion as having to do with supernatural phenomena, writing naively: "Religion is concerned with the supernatural, and this involves forces and causes beyond our normal experience of nature." After a long discussion which shows that he sees it in Frazer-like terms as the prototype of irrationality, Wolpert concludes that religion is "deeply rooted in our biology".
I am sure he is right, but it is not supernatural belief that is hard-wired in humans: it is the need for myth, and it fuels secular belief as much as traditional religion. The idea that humanity is moving to a higher state of intellectual development is pervasive in modern culture, but belied by the facts. In the 1930s, many secular thinkers believed a new type of human being was emerging in the Soviet Union, and today there are many who believe that the advance of science will rid the world of irrationality.
This humanist faith in progress is a myth no different in kind from the stories that are repeated in churches and temples. Myths are not primitive scientific theories that belong in the infancy of the species. They are symbolic narratives that give meaning to the lives of those who accept them. The chief difference between religious and secular believers is that, while the former have long known their myths to be extremely questionable, the latter imagine their own to be literally true.
Breaking the Spell and Six Impossible Things Before Breakfast present themselves as examples of freethinking. Actually they are secular sermons, which will be of interest chiefly to anxious humanists seeking to boost their sagging faith. If it is the liber-ating air of sceptical doubt you want, you are better off reading Pascal - or P G Wodehouse.
John Gray is the author of Al-Qaeda and What It Means to Be Modern (Faber & Faber)