We have at least quantitatively improved our estimate of one previously shrouded term of the Drake Equation. This permits a significant, if still moderate, easing of our agnosticism about the final value yielded by the equation. We must still be agnostic about life on other worlds — but a little bit less agnostic, because we are just that bit less ignorant. Science can chip away at agnosticism, in a way that Huxley bent over backwards to deny for the special case of God. I am arguing that, notwithstanding the polite abstinence of Huxley, Gould and many others, the God question is not in principle and forever outside the remit of science. As with the nature of the stars, contra Comte, and as with the likelihood of life in orbit around them, science can make at least probabilistic inroads into the territory of agnosticism.

My definition of the God Hypothesis included the words 'superhuman' and 'supernatural'. To clarify the difference, imagine that a SETI radio telescope actually did pick up a signal from outer space which showed, unequivocally, that we are not alone. It is a non-trivial question, by the way, what kind of signal would convince us of its intelligent origin. A good approach is to turn the question around. What should we intelligently do in order to advertise our presence to extraterrestrial listeners? Rhythmic pulses wouldn't do it. Jocelyn Bell Burnell, the radio astronomer who first discovered the pulsar in 1967, was moved by the precision of its 1.33-second periodicity to name it, tongue in cheek, the LGM (Little Green Men) signal. She later found a second pulsar, elsewhere in the sky and of different periodicity, which pretty much disposed of the LGM hypothesis. Metronomic rhythms can be generated by many non-intelligent phenomena, from swaying branches to dripping water, from time lags in self-regulating feedback loops to spinning and orbiting celestial bodies. More than a thousand pulsars have now been found in our galaxy, and it is generally accepted that each one is a spinning neutron star emitting radio energy that sweeps around like a lighthouse beam. It is amazing to think of a star rotating on a timescale of seconds (imagine if each of our days lasted 1.33 seconds instead of 24 hours), but just about everything we know of neutron stars is amazing. The point is that the pulsar phenomenon is now understood as a product of simple physics, not intelligence.

Nothing simply rhythmic, then, would announce our intelligent presence to the waiting universe. Prime numbers are often mentioned as the recipe of choice, since it is difficult to think of a purely physical process that could generate them. Whether by detecting prime numbers or by some other means, imagine that SETI does come up with unequivocal evidence of extraterrestrial intelligence, followed, perhaps, by a massive transmission of knowledge and wisdom, along the science-fiction lines of Fred Hoyle's A for Andromeda or Carl Sagan's Contact. How should we respond? A pardonable reaction would be something akin to worship, for any civilization capable of broadcasting a signal over such an immense distance is likely to be greatly superior to ours. Even if that civilization is not more advanced than ours at the time of transmission, the enormous distance between us entitles us to calculate that they must be millennia ahead of us by the time the message reaches us (unless they have driven themselves extinct, which is not unlikely).

Whether we ever get to know about them or not, there are very probably alien civilizations that are superhuman, to the point of being god-like in ways that exceed anything a theologian could possibly imagine. Their technical achievements would seem as supernatural to us as ours would seem to a Dark Age peasant transported to the twenty-first century. Imagine his response to a laptop computer, a mobile telephone, a hydrogen bomb or a jumbo jet. As Arthur C. Clarke put it, in his Third Law: 'Any sufficiently advanced technology is indistinguishable from magic.' The miracles wrought by our technology would have seemed to the ancients no less remarkable than the tales of Moses parting the waters, or Jesus walking upon them. The aliens of our SETI signal would be to us like gods, just as missionaries were treated as gods (and exploited the undeserved honour to the hilt) when they turned up in Stone Age cultures bearing guns, telescopes, matches, and almanacs predicting eclipses to the second.

In what sense, then, would the most advanced SETI aliens not be gods? In what sense would they be superhuman but not supernatural? In a very important sense, which goes to the heart of this book. The crucial difference between gods and god-like extraterrestrials lies not in their properties but in their provenance. Entities that are complex enough to be intelligent are products of an evolutionary process. No matter how god-like they may seem when we encounter them, they didn't start that way. Science-fiction authors, such as Daniel F. Galouye in Counterfeit World, have even suggested (and I cannot think how to disprove it) that we live in a computer simulation, set up by some vastly superior civilization. But the simulators themselves would have to come from somewhere. The laws of probability forbid all notions of their spontaneously appearing without simpler antecedents. They probably owe their existence to a (perhaps unfamiliar) version of Darwinian evolution: some sort of cumulatively ratcheting 'crane' as opposed to 'skyhook', to use Daniel Dennett's terminology.⁴⁵ Skyhooks — including all gods — are magic spells. They do no bona fide explanatory work and demand more explanation than they provide. Cranes are explanatory devices that actually do explain. Natural selection is the champion crane of all time. It has lifted life from primeval simplicity to the dizzy heights of complexity, beauty and apparent design that dazzle us today. This will be a dominant theme of Chapter 4, 'Why there almost certainly is no God'. But first, before proceeding with my main reason for actively disbelieving in God's existence, I have a responsibility to dispose of the positive arguments for belief that have been offered through history.

The five 'proofs' asserted by Thomas Aquinas in the thirteenth century don't prove anything, and are easily — though I hesitate to say so, given his eminence — exposed as vacuous. The first three are just different ways of saying the same thing, and they can be considered together. All involve an infinite regress — the answer to a question raises a prior question, and so on ad infinitum.

All three of these arguments rely upon the idea of a regress and invoke God to terminate it. They make the entirely unwarranted assumption that God himself is immune to the regress. Even if we allow the dubious luxury of arbitrarily conjuring up a terminator to an infinite regress and giving it a name, simply because we need one, there is absolutely no reason to endow that terminator with any of the properties normally ascribed to God: omnipotence, omniscience, goodness, creativity of design, to say nothing of such human attributes as listening to prayers, forgiving sins and reading innermost thoughts. Incidentally, it has not escaped the notice of logicians that omniscience and omnipotence are mutually incompatible. If God is omniscient, he must already know how he is going to intervene to change the course of history using his omnipotence. But that means he can't change his mind about his intervention, which means he is not omnipotent. Karen Owens has captured this witty little paradox in equally engaging verse: