What does it mean to be agnostic?

The short answer

An agnostic is someone who holds that the existence of God is unknown — and, in the strongest version of the position, unknowable. Agnostics don’t claim to know whether a god exists. They sit with the uncertainty rather than resolving it through faith in either direction.

Agnosticism is often confused with atheism, but they answer different questions. Atheism is about belief: “I don’t believe in God.” Agnosticism is about knowledge: “I don’t know if God exists.” The distinction matters because the two questions have different logical structures. You can be both at once. Many people are agnostic atheists — they don’t believe in God and also think the question cannot be settled definitively. Others are agnostic theists: they believe, but admit their belief is not grounded in certainty.

Where the word actually comes from

The term was coined in 1869 by Thomas Henry Huxley, the English biologist who earned the nickname “Darwin’s Bulldog” for his defense of evolution. Huxley was attending a meeting of the Metaphysical Society in London, a discussion club whose members ranged from Catholic theologians to materialists. Everyone there had a label — Christian, pantheist, positivist — and Huxley found himself without one that fit.

He later explained that he coined “agnostic” as a counterpart to “gnostic” — those who claim to know the deeper truths. A gnostic asserts knowledge of the ultimate; an agnostic declines to assert it. Huxley was not claiming to disbelieve in God. He was claiming something stronger and more specific: that on the evidence then available, the honest answer to the question was “I don’t know,” and that pretending otherwise — in either direction — was a violation of intellectual integrity.

In his 1889 essay Agnosticism, Huxley put it this way: “Positively the principle may be expressed: In matters of the intellect, follow your reason as far as it will take you, without regard to any other consideration. And negatively: In matters of the intellect, do not pretend that conclusions are certain which are not demonstrated or demonstrable.” For Huxley, agnosticism was not a timid middle position. It was a rule — an epistemic discipline — that applied to every question, not just the question of God.

Weak versus strong agnosticism

Contemporary philosophers usually split agnosticism into two versions. Weak agnosticism(sometimes called empirical or temporary agnosticism) holds that the existence of God is currently unknown but is in principle the kind of question that could have an answer. A weak agnostic is waiting for better evidence. They are saying: “I don’t know yet, but I am open to learning.”

Strong agnosticism (also called strict or permanent agnosticism) goes further. It holds that the existence of God is in principle unknowable — not just currently unknown. On this view, the question is not the kind of thing finite human minds could ever settle, because the object of inquiry is outside the domain where ordinary methods of evidence and reasoning apply. Huxley himself held a version of strong agnosticism about ultimate metaphysical questions while remaining a weak agnostic about many specific empirical ones.

The distinction matters in practice because the two positions generate different responses to new evidence. A weak agnostic could, in principle, be convinced either way by a sufficiently strong argument or experience. A strong agnostic is making a claim about the limits of human knowledge itself, and cannot be moved by ordinary evidence alone — only by a deeper argument about what kind of thing the God question is.

The Dawkins spectrum

In The God Delusion (2006), Richard Dawkins offered a seven-point scale for how confident someone is about the existence of God. The scale is useful because it treats the question as one of probability rather than certainty, and it makes clear that “agnostic” is not a single point but a range.

1. Strong theist.“I do not believe. I know” — 100 percent certain God exists.
2. De facto theist.Very high probability but short of certainty. “I don’t know for sure, but I strongly believe in God and live my life on the assumption that he is there.”
3. Leaning theist. Higher than 50 percent but not very high. Inclined to believe, open to uncertainty.
4. Completely impartial.Exactly 50/50. “God’s existence and non-existence are equally probable.”
5. Leaning atheist. Lower than 50 percent but not very low. Inclined to doubt, but not certain.
6. De facto atheist.Very low probability but short of zero. “I cannot know for certain, but I think God is very improbable, and I live my life on the assumption that he is not there.”
7. Strong atheist.“I know there is no God.”

Dawkins himself identified as a 6 — a de facto atheist — and argued that honest 7s are rare because nobody can be 100 percent certain of a negative metaphysical claim. The scale helps because it dissolves the fake binary between “believer” and “atheist.” Most real positions sit somewhere between 2 and 6, and the move from one number to the next is often gradual and evidence-sensitive rather than a sudden conversion.

Clifford’s ethics of belief

The philosophical backbone of serious agnosticism is an essay by the Victorian mathematician William Kingdon Clifford called The Ethics of Belief (1877). Clifford opened with a parable: a shipowner prepares to send an emigrant vessel to sea. He knows the ship is old and has had trouble before. He ought to inspect it, or at least have it inspected. Instead, he talks himself into confidence — God watches over those who set out in good faith, the ship has crossed before, worrying about it would be ungenerous. He sends the ship. It sinks with all hands.

Clifford’s conclusion is sharp: “It is wrong, always, everywhere, and for anyone, to believe anything upon insufficient evidence.” The shipowner is morally culpable not because he was mistaken, but because he let himself believe what he had no right to believe. Belief on insufficient evidence is not just an intellectual failing; it is a kind of ethical negligence. When you form a belief without checking it, you pass it on to others — through speech, through example, through action — and you bear some responsibility for where it takes them.

This is the ethical weight behind rigorous agnosticism. The agnostic is not saying “I can’t decide” in a shrug. The agnostic is saying: the stakes of this question are too high to pretend I know more than I do. Believing what I cannot justify is a kind of fraud, first on myself and then on everyone I talk to. Until I have a good reason, I owe it to the truth — and to the people I’ll influence — to say “I don’t know.”

Russell’s teapot

Bertrand Russell is usually classified as an atheist, but he argued for something that looks very much like strong agnosticism in a famous 1952 essay. Suppose, he wrote, that he claimed there was a china teapot in elliptical orbit between the Earth and Mars, too small to be detected by any existing telescope. No one could prove him wrong. But the fact that no one could disprove the teapot would not obligate anyone to take the claim seriously. The burden of proof, Russell insisted, is on the person making the positive claim, not on the person who refuses to grant it.

This is the deeper point that distinguishes serious agnosticism from wishy-washy fence-sitting. Saying “I don’t know” is not symmetric. It is not the same as saying “both positions are equally plausible.” The agnostic can perfectly well hold that belief in God is less warranted than disbelief, while still declining to claim certainty either way. Russell did exactly that. So did Huxley. The agnostic position is structurally closer to atheism than most outsiders realize, because it treats the burden of proof the same way.

Is agnosticism a cop-out?

Some confident atheists — and some confident theists — argue that agnosticism is an evasion. The theist version: you’re just refusing to commit. The atheist version: you’re giving unwarranted credibility to claims that have no evidence behind them. Both charges miss what the serious agnostic is actually doing.

The serious agnostic is not saying that every claim about God deserves equal respect. The serious agnostic is making a specific epistemic claim: that the question at issue is one where the evidence is inadequate to support confident knowledge. Calling this a cop-out is like accusing a scientist of being evasive for reporting a result with wide confidence intervals. The confidence intervals are not a failure of nerve. They are a representation of what the evidence actually supports.

Where the “cop-out” charge sometimes lands is against casual agnosticism — the person who says “I don’t know” and then also never thinks about the question, never reads the arguments, never examines their default commitments. Huxley, Russell, and Clifford would all have considered that dishonest for a different reason: it treats a live intellectual question as something you can set aside. Their agnosticism was active, not passive. It was a discipline of continuously interrogating what you believe and why.

A steelman of the theistic reply

The strongest theistic response to agnosticism accepts the epistemic humility and argues that agnosticism, honestly held, is unstable. The argument runs like this: you cannot actually live as an agnostic. Every day you make decisions that presuppose some answer to the ultimate questions — decisions about how to spend your time, how to raise your children, whether to forgive, whether to hope. You cannot suspend judgment and still act. Pascal made a version of this point in the Wager: the decision to bet is forced.

Contemporary theologians push further. The Oxford philosopher Basil Mitchell argued that the agnostic is claiming a godlike view from nowhere that they haven’t earned — pretending to stand outside all commitments and judge which ones are warranted from that neutral vantage point. But no one actually stands there. Everyone enters the question with prior commitments, cultural inheritances, and emotional stakes. The agnostic who thinks they have escaped faith has just put their faith somewhere else.

The reply is worth taking seriously because it points at something real: there is no default position on the God question from which the answers flow. The agnostic has to reason in the same fog everyone else does, with the same finite tools. What the serious agnostic can say in response is that this is precisely why the intellectual discipline matters — if no one has a view from nowhere, it is even more important to be explicit about what you are assuming and why, and to hold your conclusions at the confidence level your evidence actually supports. Honesty in the fog is still possible. Pretending the fog isn’t there is not.

How many people identify as agnostic?

Globally, somewhere between 450 and 500 million people identify as either atheist or agnostic, and the category is growing. The exact numbers are difficult to measure — agnostics are often lumped with other non-religious groups in surveys, and in many countries social or family pressure suppresses open identification. But the direction is consistent. Non-religious identification is the fastest-growing category in almost every wealthy democracy.

A 2006 Financial Times / Harris poll captured one snapshot of the rates across major Western countries:

These numbers have grown in most countries since 2006, and the trend is accelerating fastest among people under 40. The United States — historically an outlier among developed nations for its high levels of religious affiliation — has seen the “religiously unaffiliated” category grow from around 8 percent in 1990 to roughly 30 percent today. Most of that growth is agnostic and “nothing in particular,” not confident atheism.

Agnosticism in a deconverting life

For people leaving religion, agnosticism is usually the first stable ground they find. The certainty of faith dissolves, but certainty on the other side doesn’t immediately replace it. For a while — sometimes for years — the honest answer is “I don’t know anymore.” That is not a failed deconversion. It is the default starting point for anyone who is not willing to simply swap one unexamined certainty for another.

Many people pass through agnosticism toward confident atheism over time, as they find that the arguments for specific religious claims do not survive scrutiny and that the natural world does not seem to require a divine explanation. Others stay put. Both are defensible landing spots. What matters — in the Huxley, Clifford, Russell tradition — is not which label you wear but whether you can give an honest account of how confident you are and why.

Sources

1. Thomas Henry Huxley, “Agnosticism” (1889), in Collected Essays, Volume V.
2. William Kingdon Clifford, “The Ethics of Belief” (1877), Contemporary Review.
3. Bertrand Russell, “Is There a God?” (1952), commissioned by Illustrated Magazine.
4. Richard Dawkins, The God Delusion (2006), chapter 2.
5. “Religion: Year in Review 2010: Worldwide Adherents of All Religions”. Encyclopædia Britannica, 2010.
6. “Religious Views and Beliefs Vary Greatly by Country”. Financial Times / Harris Interactive, December 20, 2006.