bigideas

There’s been a bit of agnosticism in the air recently, and it reminded me of a popular argument against the existence of God. I think it’s a bad argument, and I’d like to try to set out why. For our purposes here, “God” can stand for any supernatural something at all — I’ll try to stay out of theology.

Non-Existence By Default

The argument goes like this. There are two ways you can convince me that something exists: using logic (“analytic” reasoning) or using science (“synthetic” reasoning). Neither has been in any way successful in convincing me that God exists. Therefore, it’s reasonable to believe that God doesn’t exist.

This argument works, insofar as it does, because of a hidden premise, which is something like this: “By default, nothing exists”. This is an odd premise, but it’s served us pretty well. By default, I don’t believe in God. Now, you can’t present me with either an analytic or a synthetic reason to change my default position. Therefore, I (still) don’t believe in God.

Notice that this is an argument about belief, not existence. Someone who accepts this line of reasoning does not become an atheist — one who believes there is no God — but an agnostic who acts like an atheist. This person hasn’t argued that God doesn’t exist but that no grounds can be given for believing that God does exist.

Consider Tom, who murders John. Tom hides his crime very well. Nobody but Tom knows he did it. Now, there is a principal of law (and charity) enshrined in the phrase “innocent until proven guilty” and so Jane is absolutely right to believe that Tom did not murder John. That doesn’t mean that in fact he didn’t do it.

Reasonableness

Perhaps this principle can be extended to a claim about existence, rather than knowledge, but surely only if the request for “evidence” (analytic or synthetic) is reasonable. By this I mean that if God did exist, we should reasonably be able to provide it. I’m not sure why we would think this. Many analytic attempts (mostly ontological arguments) have been tried and failed. What’s more, since God is by definition supernatural it’s not clear how science, which studies nature, might gather any kind of pertinent evidence. Note that I’m not talking about specific beliefs of specific religions here, which in some cases are scientifically testable, but the general prospect of the existence of a supernatural something at all.

This is why Dawkins’s argument about fairies works to support afairyism, but isn’t a perfect analogy for atheism. If little winged ladies lived at the bottom of some people’s gardens, you can bet we’d know about it by now. It’s possible that they’ve hidden really really well, but it’s honestly not likely. Until you show me some evidence to the contrary, I’ll stick with my default position that fairies don’t exist. (It’s worth noting in passing that, here as in almost all matters of science, “evidence” almost always means “evidence that people I trust have seen some evidence”).

The requirement of reasonableness might be illustrated as follows. Imagine Tom, John and Jane live in a country run entirely by mathematicians. Their whole culture is built on analytic thinking — only water-tight logical proofs are considered to have any real force. Well, then it would hardly matter how careless Tom had been in disguising the murder, since any number of outlandish explanations for his innocence might be put forward. If you require that Tom is innocent until analytically proven guilty then he’ll get off scot-free, as will every other criminal. This is why mathematicians are rarely given anything important to do.

By analogy, in questions of the existence of supernatural beings it may be that neither analytic nor synthetic arguments are reasonable things to ask for. This is what many religious believers will tell you: that they’ve had a personal experience, they can’t explain it to you and so on. And if that’s the case, then perhaps — just perhaps — the generally useful principle of non-existence-by-default needs to be adjusted a bit. I’m not sure whether it does or doesn’t, and I’ve no idea how, but until we’ve properly thought that through we can’t dine out on the principle as if we had.

[Philosophers: I appreciate I've simplified the ideas of "analytic" and "synthetic" rather drastically in order to make my point. Anyone who wants to find out more about these ideas could do worse than read the Stanford article.]