The Argument From the Idea

Around 1078, Anselm of Canterbury tried to prove that God exists using nothing but the definition of the word.

The argument is short enough to state completely.

God is that than which nothing greater can be conceived. Now suppose God exists only in the understanding, as an idea. Something that exists in reality is greater than something that exists only as an idea. So we could conceive of something greater than God, namely that same being existing in reality. But that contradicts the definition, since nothing greater than God can be conceived. Therefore God exists in reality.

Whatever else is true of it, notice what this argument does not use. No observation of the world. No evidence of any kind. No claim about design, or causation, or miracles, or testimony. If it worked, it would settle the largest question anybody has ever asked from an armchair, using only the meaning of a term.

That is why it has fascinated people for nine hundred years, including people with no religious stake in the answer at all.

The first objection arrived almost immediately, from a monk named Gaunilo, and it is still the best-known one.

Gaunilo said: run the same argument for an island. Consider the greatest conceivable island, more perfect than any other. An island that exists is greater than one that does not. So the greatest conceivable island must exist.

It does not. So either the form of the argument is broken, or something distinguishes God from islands in a way that lets the argument run in one case and not the other.

Defenders of Anselm say exactly that: an island has no intrinsic maximum, since you can always add another palm tree, whereas the concept of a greatest conceivable being is not open-ended in that way. Whether this rescues the argument or restates the problem is still discussed, which tells you something.

The deepest objection came from Kant, six centuries later, and it is directed at a hidden assumption.

Existence, Kant said, is not a predicate. When you list a thing's properties, saying that it exists adds nothing to the description. A hundred real coins do not contain a penny more than a hundred imagined coins; they differ in whether there are any, not in what they are like. So existence cannot be one of the properties that makes something the greatest conceivable thing, because it is not that kind of property at all.

That is generally taken as decisive, and the argument keeps coming back anyway, in modal logic and in contemporary metaphysics, in forms Anselm would not recognise.

That pattern is worth noticing independently of whether you care about the conclusion. An argument that everybody can feel is wrong, and that nobody can straightforwardly refute, is usually pointing at something genuinely unresolved.

Here what it is pointing at is the relationship between what we can coherently define and what there is. We do not have a settled account of that, and until we do, arguments of this shape will keep finding new clothes.