My impression is that the fire-breathing atheists about whom we hear so much – the celebrated quartet of Richard Dawkins, Christopher Hitchens, Sam Harris, and Dan Dennett – think of religious commitments in terms of mistaken or at least hopelessly improbable and therefore irrational ontology. Believers think that something exists, but the overwhelmingly probable truth is that it does not. I may be wrong that this is what they think, but whether they do so or not, I am sure others do. Yet this interpretation of the issue is itself mistaken, and indeed doubly so. It is mistaken, or perhaps it would be better to say unimaginative or off-key to think of religious frames of mind primarily in terms of belief. And it is similarly misdirected to think that the belief component primarily concerns the existence of anything. I read both these insights into Hume's celebrated works on the philosophy of religion, but their lesson seems never to have been properly absorbed. Even Wittgenstein, I think, although he came close and certainly flirted with more adequate views, failed to take their measure properly, although I will be unable to substantiate that opinion here.

To make both my claims plausible, we can contrast a religious frame of mind with a simple case of mistaken ontology. Bertrand Russell's comparison with the empirically absurd belief that there exists a teapot flying in its own extraterrestrial orbit around the sun is a convenient example. The first obvious point is that nobody would think of belief in such a thing as itself religious. It has none of the hallmarks. It has nothing to do with the conduct of life, with ethics, with the formation of congregations, with ritual, with the sense of anything as sacred, with consolation, hope, despair, or the many other emotional and social clothings of religion. It is simply a daft secular belief.

But now suppose some of those clothings arrive. The teapot is important. The texts describing it are sacred. Out of its spout issue instructions for living. There are proper ways to show respect for it, cemented into services and rituals. There are sacred sites, and taboos associated with it, and priests who interpret its sayings, or who alone are authorized to lift its hidden lid. It is anathema to mock it, or them. Many people assert that without the belief in the teapot their lives would be meaningless. And so on and so on. The teapot has now become an object of veneration – a religious object.

Now my two claims can come further into focus. The first is that the religious clothing that arrived was not primarily a matter of belief. It was a matter of practical dispositions or stances towards things. It was partly like having a favourite poem in your head, or a preference for one kind of music or another. It was partly as well a commitment to some practices and some permissions and prohibitions: an immersion in a ‘way of life’.

The second is that ontology has actually dropped out of the picture. It simply does not matter any more whether there is such a teapot or not. A mythical teapot can perform the religious function of being a focus for emotions, attitudes, and social practices just as well as an actual teapot – better, in fact, since mythical teapots are not the objects of science. It will be no part of any religious practice to ask whether the teapot is pink or blue, how much it weighs, how wide it is, any more than we ask the same kinds of question about Odysseus or Krishna (Wittgenstein remarked that people who talk of God's all-seeing eye do not talk about his eyebrows). Teapot theologians would, rightly, be horrified at the idea of sending a spaceship to intercept the teapot: it is no part of the way of life that empirical confirmation or disconfirmation of that kind is remotely appropriate.

Hume said that religionists are in a ‘somewhat unaccountable’ state of mind, somewhere between belief and disbelief. We can now see what he may have meant. The story about the teapot is serious – as serious as birth, death, marriage, or any of the other aspects of living that the texts and rituals and ordained forms regulate and symbolize. It will not do to mock these things. On the other hand it is empirically disengaged, or if we like, empirically frivolous, for nothing empirical is relevant to its function as an intentional object, a focus for all the emotions and practices which the religious service of the teapot entails. Just as a child can be afraid of the Jabberwocky without really having any conception of it (‘it seems to fill my head with ideas’, said Alice, ‘but I do not know what they are’), so the religionist can venerate the teapot without any conception of it beyond a vague and changing kaleidoscope of imaginings.

So what we are really dealing with is not ontology and belief, but anthropology and ways of life. I certainly do not wish to imply that this puts the new atheists out of business. There may be actual beliefs involved in the mix, such as historical beliefs about this and that, and they may be improbable and irrational, like creationist beliefs about the geological timescale. Much more importantly, there may be much to dislike about contemporary ways of life that find Jesus of Nazareth as their focus, or Mohammed or Krishna or Buddha, or L. Ron Hubbard. There may be authoritative Rawlsian principles attempting to show that it is wrong for religionists to bring their faith-based moralities into the public sphere, although I rather doubt it myself. Keeping religion out of the public square is a political problem, not an intellectual one.

But the term ‘atheist’, so carefully avoided by Hume himself, is no longer appropriate. It implies that there is a definite ontological belief that some people have and others do not, and I have been arguing that this mislocates the issue. The term ‘agnostic’ is no better, since it has the same implication of a definite ontological question, only one to which we do not know the answer.

So my recommendation is that we leave any traditional philosophy of religion exactly where Hume left it at the end of the Dialogues. I fear that this recommendation may be more gratefully received by many theologians and practitioners, who are themselves suspicious of what is sometimes called ‘onto-theology’, than by the new atheists. But in the spirit of annoying all parties, I continue. Instead we should pick up Hume's Natural History of Religion, or its successors in the works of Kant, Feuerbach, Marx or Durkheim. And if we want to wean people away from their myths, or the particular colouration their myths have taken at particular times and places, then we must do what Nietzsche did at the end of the nineteenth century, which was to recognize moral corruption when we find it, and then to rail, preach, inveigh, fulminate, or thunder against it. Of course, whether we find more of it inside churches, synagogues, mosques and temples, or outside them, may itself be a difficult matter to resolve.