America, founded in secularism as a beacon of eighteenth century enlightenment, is becoming the victim of religious politics, a circumstance that would have horrified the Founding Fathers. The political ascendancy today values embryonic cells over adult people. It obsesses about gay marriage, ahead of genuinely important issues that actually make a difference to the world. It gains crucial electoral support from a religious constituency whose grip on reality is so tenuous that they expect to be 'raptured' up to heaven, leaving their clothes as empty as their minds. More extreme specimens actually long for a world war, which they identify as the 'Armageddon' that is to presage the Second Coming. Sam Harris, in his new short book, Letter to a Christian Nation, hits the bull's-eye as usual:

It is, therefore, not an exaggeration to say that if the city of New York were suddenly replaced by a ball of fire, some significant percentage of the American population would see a silver-lining in the subsequent mushroom cloud, as it would suggest to them that the best thing that is ever going to happen was about to happen: the return of Christ . . .Imagine the consequences if any significant component of the U.S. government actually believed that the world was about to end and that its ending would be glorious. The fact that nearly half of the American population apparently believes this, purely on the basis of religious dogma, should be considered a moral and ¬intellectual emergency.

Does Bush check the Rapture Index daily, as Reagan did his stars? We don't know, but would anyone be surprised?

My scientific colleagues have additional reasons to declare emergency. Ignorant and absolutist attacks on stem cell research are just the tip of an iceberg. What we have here is nothing less than a global assault on rationality, and the Enlightenment values that inspired the founding of this first and greatest of secular republics. Science education - and hence the whole future of science in this country - is under threat. Temporarily beaten back in a Pennsylvania court, the 'breathtaking inanity' (Judge John Jones's immortal phrase) of 'intelligent design' continually flares up in local bush-fires. Dowsing them is a time-consuming but important responsibility, and scientists are finally being jolted out of their complacency. For years they quietly got on with their science, lamentably underestimating the creationists who, being neither competent nor interested in science, attended to the serious political business of subverting local school boards. Scientists, and intellectuals generally, are now waking up to the threat from the American Taliban.

Scientists divide into two schools of thought over the best tactics with which to face the threat. The Neville Chamberlain 'appeasement' school focuses on the battle for evolution. Consequently, its members identify fundamentalism as the enemy, and they bend over backwards to appease 'moderate' or 'sensible' religion (not a difficult task, for bishops and theologians despise fundamentalists as much as scientists do). Scientists of the Winston Churchill school, by contrast, see the fight for evolution as only one battle in a larger war: a looming war between supernaturalism on the one side and rationality on the other. For them, bishops and theologians belong with creationists in the supernatural camp, and are not to be appeased.

The Chamberlain school accuses Churchillians of rocking the boat to the point of muddying the waters. The philosopher of science Michael Ruse wrote:

We who love science must realize that the enemy of our enemies is our friend. Too often evolutionists spend time insulting would-be allies. This is especially true of secular evolutionists. Atheists spend more time running down sympathetic Christians than they do countering ¬creationists. When John Paul II wrote a letter endorsing Darwinism, Richard Dawkins's response was simply that the pope was a hypocrite, that he could not be genuine about science and that Dawkins himself simply preferred an honest fundamentalist.

A recent article in the New York Times by Cornelia Dean quotes the astronomer Owen Gingerich as saying that, by simultaneously advocating evolution and atheism, 'Dr Dawkins "probably single-handedly makes more converts to intelligent design than any of the leading intelligent design theorists".' This is not the first, not the second, not even the third time this plonkingly witless point has been made (and more than one reply has aptly cited Uncle Remus: "Oh please please Brer Fox, don't throw me in that awful briar patch").

Chamberlainites are apt to quote the late Stephen Jay Gould's 'NOMA' - 'non-overlapping magisteria'. Gould claimed that science and true religion never come into conflict because they exist in completely separate dimensions of discourse:

To say it for all my colleagues and for the umpteenth millionth time (from college bull sessions to learned treatises): science simply cannot (by its legitimate methods) adjudicate the issue of God's possible superintendence of nature. We neither affirm nor deny it; we simply can't comment on it as scientists.

This sounds terrific, right up until you give it a moment's thought. You then realize that the presence of a creative deity in the universe is clearly a scientific hypothesis. Indeed, it is hard to imagine a more momentous hypothesis in all of science. A universe with a god would be a completely different kind of universe from one without, and it would be a scientific difference. God could clinch the matter in his favour at any moment by staging a spectacular demonstration of his powers, one that would satisfy the exacting standards of science. Even the infamous Templeton Foundation recognized that God is a scientific hypothesis - by funding double-blind trials to test whether remote prayer would speed the recovery of heart patients. It didn't, of course, although a control group who knew they had been prayed for tended to get worse (how about a class action suit against the Templeton Foundation?) Despite such well-financed efforts, no evidence for God's existence has yet appeared.

To see the disingenuous hypocrisy of religious people who embrace NOMA, imagine that forensic archeologists, by some unlikely set of circumstances, discovered DNA evidence demonstrating that Jesus was born of a virgin mother and had no father. If NOMA enthusiasts were sincere, they should dismiss the archeologists' DNA out of hand: "Irrelevant. Scientific evidence has no bearing on theological questions. Wrong magisterium." Does anyone seriously imagine that they would say anything remotely like that? You can bet your boots that not just the fundamentalists but every professor of theology and every bishop in the land would trumpet the archeological evidence to the skies.

Either Jesus had a father or he didn't. The question is a scientific one, and scientific evidence, if any were available, would be used to settle it. The same is true of any miracle - and the deliberate and intentional creation of the universe would have to have been the mother and father of all miracles. Either it happened or it didn't. It is a fact, one way or the other, and in our state of uncertainty we can put a probability on it - an estimate that may change as more information comes in. Humanity's best estimate of the probability of divine creation dropped steeply in 1859 when The Origin of Species was published, and it has declined steadily during the subsequent decades, as evolution consolidated itself from plausible theory in the nineteenth century to established fact today.

The Chamberlain tactic of snuggling up to 'sensible' religion, in order to present a united front against ('intelligent design') creationists, is fine if your central concern is the battle for evolution. That is a valid central concern, and I salute those who press it, such as Eugenie Scott in Evolution versus Creationism. But if you are concerned with the stupendous scientific question of whether the universe was created by a supernatural intelligence or not, the lines are drawn completely differently. On this larger issue, fundamentalists are united with 'moderate' religion on one side, and I find myself on the other.

Of course, this all presupposes that the God we are talking about is a personal intelligence such as Yahweh, Allah, Baal, Wotan, Zeus or Lord Krishna. If, by 'God', you mean love, nature, goodness, the universe, the laws of physics, the spirit of humanity, or Planck's constant, none of the above applies. An American student asked her professor whether he had a view about me. 'Sure,' he replied. 'He's positive science is incompatible with religion, but he waxes ecstatic about nature and the universe. To me, that is ¬religion!' Well, if that's what you choose to mean by religion, fine, that makes me a religious man. But if your God is a being who designs universes, listens to prayers, forgives sins, wreaks miracles, reads your thoughts, cares about your welfare and raises you from the dead, you are unlikely to be satisfied. As the distinguished American physicist Steven Weinberg said, "If you want to say that 'God is energy,' then you can find God in a lump of coal." But don't expect congregations to flock to your church.

When Einstein said 'Did God have a choice in creating the Universe?' he meant 'Could the universe have begun in more than one way?' 'God does not play dice' was Einstein's poetic way of doubting Heisenberg's indeterminacy principle. Einstein was famously irritated when theists misunderstood him to mean a personal God. But what did he expect? The hunger to misunderstand should have been palpable to him. 'Religious' physicists usually turn out to be so only in the Einsteinian sense: they are atheists of a poetic disposition. So am I. But, given the widespread yearning for that great misunderstanding, deliberately to confuse Einsteinian pantheism with supernatural religion is an act of intellectual high treason.

Accepting, then, that the God Hypothesis is a proper scientific hypothesis whose truth or falsehood is hidden from us only by lack of evidence, what should be our best estimate of the probability that God exists, given the evidence now available? Pretty low I think, and here's why.

First, most of the traditional arguments for God's existence, from Aquinas on, are easily demolished. Several of them, such as the First Cause argument, work by setting up an infinite regress which God is wheeled out to terminate. But we are never told why God is magically able to terminate regresses while needing no explanation himself. To be sure, we do need some kind of explanation for the origin of all things. Physicists and cosmologists are hard at work on the problem. But whatever the answer - a random quantum fluctuation or a Hawking/Penrose singularity or whatever we end up calling it - it will be simple. Complex, statistically improbable things, by definition, don't just happen; they demand an explanation in their own right. They are impotent to terminate regresses, in a way that simple things are not. The first cause cannot have been an intelligence - let alone an intelligence that answers prayers and enjoys being worshipped. Intelligent, creative, complex, statistically improbable things come late into the universe, as the product of evolution or some other process of gradual escalation from simple beginnings. They come late into the universe and therefore cannot be responsible for designing it.

Another of Aquinas' efforts, the Argument from Degree, is worth spelling out, for it epitomises the characteristic flabbiness of theological reasoning. We notice degrees of, say, goodness or temperature, and we measure them, Aquinas said, by reference to a maximum:

Now the maximum in any genus is the cause of all in that genus, as fire, which is the maximum of heat, is the cause of all hot things . . . Therefore, there must also be something which is to all beings the cause of their being, goodness, and every other perfection; and this we call God.

That's an argument? You might as well say that people vary in smelliness but we can make the judgment only by reference to a perfect maximum of conceivable smelliness. Therefore there must exist a pre-eminently peerless stinker, and we call him God. Or substitute any dimension of comparison you like, and derive an equivalently fatuous conclusion. That's theology.

The only one of the traditional arguments for God that is widely used today is the teleological argument, sometimes called the Argument from Design although - since the name begs the question of its validity - it should better be called the Argument for Design. It is the familiar 'watchmaker' argument, which is surely one of the most superficially plausible bad arguments ever discovered - and it is rediscovered by just about everybody until they are taught the logical fallacy and Darwin's brilliant alternative.

In the familiar world of human artifacts, complicated things that look designed are designed. To naïve observers, it seems to follow that similarly complicated things in the natural world that look designed - things like eyes and hearts - are designed too. It isn't just an argument by analogy. There is a semblance of statistical reasoning here too - fallacious, but carrying an illusion of plausibility. If you randomly scramble the fragments of an eye or a leg or a heart a million times, you'd be lucky to hit even one combination that could see, walk or pump. This demonstrates that such devices could not have been put together by chance. And of course, no sensible scientist ever said they could. Lamentably, the scientific education of most British and American students omits all mention of Darwinism, and therefore the only alternative to chance that most people can imagine is design.

Even before Darwin's time, the illogicality was glaring: how could it ever have been a good idea to postulate, in explanation for the existence of improbable things, a designer who would have to be even more improbable? The entire argument is a logical non-starter, as David Hume realized before Darwin was born. What Hume didn't know was the supremely elegant alternative to both chance and design that Darwin was to give us. Natural selection is so stunningly powerful and elegant, it not only explains the whole of life, it raises our consciousness and boosts our confidence in science's future ability to explain everything else.

Natural selection is not just an alternative to chance. It is the only ultimate alternative ever suggested. Design is a workable explanation for organized complexity only in the short term. It is not an ultimate explanation, because designers themselves demand an explanation. If, as Francis Crick and Leslie Orgel once playfully speculated, life on this planet was deliberately seeded by a payload of bacteria in the nose cone of a rocket, we still need an explanation for the intelligent aliens who dispatched the rocket. Ultimately they must have evolved by gradual degrees from simpler beginnings. Only evolution, or some kind of gradualistic 'crane' (to use Daniel Dennett's neat term), is capable of terminating the regress. Natural selection is an anti-chance process, which gradually builds up complexity, step by tiny step. The end product of this ratcheting process is an eye, or a heart, or a brain - a device whose improbable complexity is utterly baffling until you spot the gentle ramp that leads up to it.

Whether my conjecture is right that evolution is the only explanation for life in the universe, there is no doubt that it is the explanation for life on this planet. Evolution is a fact, and it is among the more secure facts known to science. But it had to get started somehow. Natural selection cannot work its wonders until certain minimal conditions are in place, of which the most important is an accurate system of replication - DNA, or something that works like DNA.

The origin of life on this planet - which means the origin of the first self-replicating molecule - is hard to study, because it (probably) only happened once, 4 billion years ago and under very different conditions from those with which we are familiar. We may never know how it happened. Unlike the ordinary evolutionary events that followed, it must have been a genuinely very improbable - in the sense of unpredictable - event: too improbable, perhaps, for chemists to reproduce it in the laboratory or even devise a plausible theory for what happened. This weirdly paradoxical conclusion - that a chemical account of the origin of life, in order to be plausible, has to be implausible - would follow if it were the case that life is extremely rare in the universe. And indeed we have never encountered any hint of extraterrestrial life, not even by radio - the circumstance that prompted Enrico Fermi's cry: "Where is everybody?"

Suppose life's origin on a planet took place through a hugely improbable stroke of luck, so improbable that it happens on only one in a billion planets. The National Science Foundation would laugh at any chemist whose proposed research had only a one in a hundred chance of succeeding, let alone one in a billion. Yet, given that there are at least a billion billion planets in the universe, even such absurdly low odds as these will yield life on a billion planets. And - this is where the famous anthropic principle comes in - Earth has to be one of them, because here we are.

If you set out in a spaceship to find the one planet in the galaxy that has life, the odds against your finding it would be so great that the task would be indistinguishable, in practice, from impossible. But if you are alive (as you manifestly are if you are about to step into a spaceship) you needn't bother to go looking for that one planet because, by definition, you are already standing on it. The anthropic principle really is rather elegant. By the way, I don't actually think the origin of life was as improbable as all that. I think the galaxy has plenty of islands of life dotted about, even if the islands are too spaced out for any one to hope for a meeting with any other. My point is only that, given the number of planets in the universe, the origin of life could in theory be as lucky as a blindfolded golfer scoring a hole in one. The beauty of the anthropic principle is that, even in the teeth of such stupefying odds against, it still gives us a perfectly satisfying explanation for life's presence on our own planet.

The anthropic principle is usually applied not to planets but to universes. Physicists have suggested that the laws and constants of physics are too good - as if the universe were set up to favour our eventual evolution. It is as though there were, say, half a dozen dials representing the major constants of physics. Each of the dials could in principle be tuned to any of a wide range of values. Almost all of these knob-twiddlings would yield a universe in which life would be impossible. Some universes would fizzle out within the first picosecond. Others would contain no elements heavier than hydrogen and helium. In yet others, matter would never condense into stars (and you need stars in order to forge the elements of chemistry and hence life). You can estimate the very low odds against the six knobs all just happening to be correctly tuned, and conclude that a divine knob-twiddler must have been at work. But, as we have already seen, that explanation is vacuous because it begs the biggest question of all. The divine knob twiddler would himself have to have been at least as improbable as the settings of his knobs.

Again, the anthropic principle delivers its devastatingly neat solution. Physicists already have reason to suspect that our universe - everything we can see - is only one universe among perhaps billions. Some theorists postulate a multiverse of foam, where the universe we know is just one bubble. Each bubble has its own laws and constants. Our familiar laws of physics are parochial bylaws. Of all the universes in the foam, only a minority has what it takes to generate life. And, with anthropic hindsight, we obviously have to be sitting in a member of that minority, because, well, here we are, aren't we? As physicists have said, it is no accident that we see stars in our sky, for a universe without stars would also lack the chemical elements necessary for life. There may be universes whose skies have no stars: but they also have no inhabitants to notice the lack. Similarly, it is no accident that we see a rich diversity of living species: for an evolutionary process that is capable of yielding a species that can see things and reflect on them cannot help producing lots of other species at the same time. The reflective species must be surrounded by an ecosystem, as it must be surrounded by stars.

The anthropic principle entitles us to postulate a massive dose of luck in accounting for the existence of life on our planet. But there are limits. We are allowed one stroke of luck for the origin of evolution, and perhaps for a couple of other unique events like the origin of the eukaryotic cell and the origin of consciousness. But that's the end of our entitlement to large-scale luck. We emphatically cannot invoke major strokes of luck to account for the illusion of design that glows from each of the billion species of living creature that have ever lived on Earth. The evolution of life is a general and continuing process, producing essentially the same result in all species, however different the details.

Contrary to what is sometimes alleged, evolution is a predictive science. If you pick any hitherto unstudied species and subject it to minute scrutiny, any evolutionist will confidently predict that each individual will be observed to do everything in its power, in the particular way of the species - plant, herbivore, carnivore, nectivore or whatever it is - to survive and propagate the DNA that rides inside it. We won't be around long enough to test the prediction but we can say, with great confidence, that if a comet strikes Earth and wipes out the mammals, a new fauna will rise to fill their shoes, just as the mammals filled those of the dinosaurs 65 million years ago. And the range of parts played by the new cast of life's drama will be similar in broad outline, though not in detail, to the roles played by the mammals, and the dinosaurs before them, and the mammal-like reptiles before the dinosaurs. The same rules are predictably being followed, in millions of species all over the globe, and for hundreds of millions of years. Such a general observation requires an entirely different explanatory principle from the anthropic principle that explains one-off events like the origin of life, or the origin of the universe, by luck. That entirely different principle is natural selection.

We explain our existence by a combination of the anthropic principle and Darwin's principle of natural selection. That combination provides a complete and deeply satisfying explanation for everything that we see and know. Not only is the god hypothesis unnecessary. It is spectacularly unparsimonious. Not only do we need no God to explain the universe and life. God stands out in the universe as the most glaring of all superfluous sore thumbs. We cannot, of course, disprove God, just as we can't disprove Thor, fairies, leprechauns and the Flying Spaghetti Monster. But, like those other fantasies that we can't disprove, we can say that God is very very improbable.