Agnosticism

The commonest confusion about agnosticism is to picture belief as a single line running from theist through agnostic to atheist, with agnosticism as the moderate middle. That gets the geometry wrong, and almost every argument about the two words traces back to it.

There are two questions, not one, and they are independent. The first is about belief: do you believe that a god exists? Answering yes makes you a theist, and anything short of yes makes you an atheist in the broad sense. The second is about knowledge: do you claim to know? Claiming to know makes you gnostic on the point, declining to claim it makes you agnostic. Because the questions are separate, all four combinations exist and are held by real people.

An agnostic theist believes without claiming certainty, which describes a very large number of religious people and is close to what the apophatic strands of Christianity, Judaism and Islam positively insist on: that God exceeds what human knowledge can capture. A gnostic theist claims certain knowledge that God exists. An agnostic atheist does not claim to know and does not believe, which is probably the most populous position among people who use either label. A gnostic atheist claims to know that no god exists, and is rarer than either side's rhetoric suggests.

Bertrand Russell put the practical upshot better than anyone. Before a philosophical audience he would call himself an agnostic, because he could not disprove the God of the philosophers any more than he could disprove the gods of Homer; before a general audience he would say atheist, because in ordinary speech that conveyed his actual expectations. The two labels were describing different things about the same man.

A consequence follows for anyone reading survey data: questionnaires that force a single choice between agnostic and atheist are collapsing two axes into one, and their figures cannot simply be added together or compared across countries.

Agnosticism comes in a weaker and a stronger version, and they are not small variations on each other.

The weak or temporary form holds that whether any god exists is not currently known. It leaves open that evidence or argument might one day settle it, and it is compatible with active inquiry. The strong or permanent form holds that the question cannot be settled by any human being at all, because its object lies outside the reach of experience and no possible observation would bear on it. On the strong view the shortage is not of evidence but of any procedure that could produce evidence. Thomas Henry Huxley's own statements sit nearer the strong end for a particular class of questions, though he framed his position as a method rather than a doctrine and resisted turning it into a creed.

The strong form has philosophical ancestry. Kant argued that reason cannot reach past possible experience to prove or disprove the existence of God, and that the traditional proofs fail for that structural reason rather than through any fixable defect. Herbert Spencer, writing in the decade in which the word was coined, built a system around the Unknowable, an ultimate reality of which nothing determinate can be said.

There is a further step that agnosticism itself does not take. Theological noncognitivism, sometimes called ignosticism and usually credited to the rabbi Sherwin Wine, holds that the question is not merely unanswerable but defective: that the term God, as ordinarily offered, is too under-specified for the sentence God exists to carry a truth value at all. On that view the agnostic has already conceded too much by treating the proposition as a genuine one awaiting evidence, and the right response is to ask for a definition before anything else. Whether this is a deeper form of agnosticism or a rejection of it is itself argued about.

Agnosticism is attacked from both directions, and the attacks are the quickest route into what it does and does not claim.

Believers argue that suspension is not the neutral resting place it presents itself as. William James put the case in The Will to Believe (1896): where an option is genuine, forced and momentous, refusing to decide is itself a decision, taken under exactly the same uncertainty as either alternative, and it forecloses whatever the believing life might have shown. Pascal's wager is the older and cruder version of the same complaint. Both hold that in practice one lives as though one had answered, whatever one says.

Atheists argue the opposite way. Dawkins's seven-point spectrum in The God Delusion is the standard modern statement: he places himself at six rather than seven, technically an agnostic, while insisting the working position is atheism, because nobody suspends judgement in this way about other unfalsifiable claims. The sharper form of the complaint is sociological rather than philosophical, that agnostic has become, in much ordinary use, a way of avoiding the social cost attached to the word atheist rather than a distinct epistemic position. That charge is contested, and agnostics reject it.

The agnostic reply is that both criticisms assume the question behaves like other questions. If the strong form is right and no evidence could bear on the matter, then confident denial is as unwarranted as confident assertion, and treating the two as asymmetrical begs precisely what is at issue.

There is a separate difficulty for anyone reading numbers. In modern surveys agnostic frequently records something closer to unsure, undecided or has not thought about it than the considered epistemology Huxley meant by it. The label and the position it named have drifted apart, and the drift makes survey shares a poor proxy for how many people actually hold the philosophical view.