The Argument from Improbability
In what he considers to be the central argument of his book The God Delusion (TGD), Richard Dawkins claims to have shown that ‘God almost certainly does not exist’ (TGD, p. 158). His argument is based on probability and indeed he refers to it as the argument from 
improbability, which he thinks ‘demonstrates that God, though not technically disprovable, is very very improbable indeed’ (TGD, p. 109).
The argument is also known as the ultimate Boeing 747 gambit. The cosmologist Fred Hoyle claimed that the probability of natural processes producing living organisms is no greater than the probability of a hurricane assembling a Boeing 747 in a scrapyard. This illustration is often used to show that just as a Boeing 747 points to a designer so too does the existence of life. Dawkins attempts to turn this argument on its head. According to Dawkins, God is even more complex and so he claims that God’s existence is even more improbable than a Boeing 747 coming about by chance. He puts it like this, ‘However statistically improbable the entity you seek to explain by invoking a designer, the designer himself has got to be at least as improbable. God is the Ultimate Boeing 747’ (TGD, p. 114).
How does this relate to the presumption of atheism considered in Part 1 and the insufficient evidence objection in Part 2? One difficulty here is in understanding exactly what Dawkins’ argument is supposed to show. In using probability in this way, his claim is ambiguous because when we talk about probabilities we need to specify what background evidence or knowledge has been taken into account. The probability that I will assign to my friend Tom’s having won the lottery in a given week will be very low given only the background evidence that he buys a ticket every week. But if I see him on TV receiving the winning cheque, then based on this evidence, I can be very confident that he has in fact won. The probability of his winning is very low based on only background evidence, but very high based on specific evidence in its favour.
For this reason, Dawkins’ argument cannot be understood as claiming that God’s existence is very improbable irrespective of what evidence is available since surely there could, at least in principle, be evidence that would overturn any initial probability and make it very likely that God exists. The most reasonable way to interpret Dawkins’ argument is as an argument to show that God’s existence is improbable before any specific evidence in favour of God’s existence is taken into account. No doubt Dawkins thinks that God’s existence is also improbable once all the evidence is taken into account, but his argument (even if successful) doesn’t establish that. He does provide other reasons for reaching that conclusion. Many people, including many atheists, find his reasons unconvincing, but that is not the topic of the discussion here. Whatever the merits of his other arguments, his argument from improbability cannot establish both that God’s existence is improbable before any specific evidence is taken into account and that God’s existence is still improbable after relevant evidence is taken into account. That would be double counting.
To put all of this another way, while Dawkins would affirm the insufficient evidence objection to God’s existence, his argument from improbability is much more plausibly understood as offering an argument in favour of the presumption of atheism, or perhaps better, a strong version of it. The idea is that God, if he existed would be very complex, and hence very improbable unless very convincing evidence is forthcoming. Atheism, if Dawkins’ argument works, should be the default position and the burden of proof would lie with the theist.
As I argued in Part 1, the presumption of atheism is not an obvious truth. If it is to be accepted, a good argument in its favour is needed. While some philosophers have offered arguments of this kind (for a response see my article ‘Probability and the Presumption of Atheism’), at a popular level atheists often simply presuppose the presumption of atheism. That is, they simply assume that the burden of proof lies with the theist. Some make silly comparisons between God and Santa Claus, others just think it is obvious that God is so strange or so different from our experience that his existence can be virtually ruled out without argument, and hence a great burden of proof lies with the theist. Now, while Dawkins does at times advocate this kind of approach, with the argument from improbability he does, to his credit, offer an argument for the presumption of atheism.
However, while Dawkins is to be congratulated for offering an argument, there are excellent reasons to think that his argument fails. The key problem lies with his attempt to link complexity and probability. First, let’s think about what he means by complexity. Roughly speaking, the kind of complexity he has in mind is that which is found in an entity that is composed of a variety of separable parts arranged in a highly specific manner so that it is able to function. Clearly, a Boeing 747 is a good example. But is God complex in this sense? Perhaps God is complex in some sense of the word, but God isn’t composed of separable parts and so isn’t complex in Dawkins’ sense. This means that Dawkins’ argument can’t be applied to God. Some atheists will say that provided God is complex in some sense, then God’s existence is highly improbable, but Dawkins’ argument doesn’t establish this conclusion since it only applies to a particular notion of complexity (one that isn’t possessed by God). Atheists who wish to continue to claim that God’s existence is highly improbable based on complexity will need an alternative argument; theists can’t simply be expected to accept their claim as an obvious truth.
Secondly, and more importantly, how is complexity to be linked with improbability? As Dawkins puts it, ‘The argument from improbability states that complex things could not have come about by chance’ (TGD, p. 114). That makes sense. Is the existence of a Boeing 747 improbable? No, what is improbable is that it could have come about by chance. Clearly, however, God would not have come about by chance. This means that even if God were complex in Dawkins’ sense, it wouldn’t follow that God’s existence is improbable, only that God is unlikely to have come about by chance. This will hardly worry the theist!
In summary, there is no reason to think that God would have the relevant kind of complexity, and such complexity wouldn’t lead to improbability anyway. For more on this topic, see There’s Probably No God – a response to Richard Dawkins or for a more detailed response see Darwin, Design and Dawkins’ Dilemma.
We’ve considered three popular atheist arguments: the presumption of atheism, the insufficient evidence objection, and the argument from improbability. These arguments often crop up in discussions about God. So common are they that in many cases they, or something like them, are not presented as arguments at all, but simply assumed as part of the discussion. This is more often the case for the first two since it is often assumed that the burden of proof lies with the theist or asserted that there is not enough evidence to believe in God. In fact, it’s often asserted that there is no evidence at all. The argument from improbability might not be stated in a precise way, but again it is often simply assumed that God must be very improbable because God is very complex or far removed from our experience. Or it might be thought that there is some good technical reason based on probability theory that shows that God almost certainly does not exist.
As we have seen, however, none of these arguments is convincing. Those who wish to defend belief in God need to point out why this is so and be able to identify when such ideas, whether consciously or not, might prevent people from taking the case for God seriously.