QuestionsDoes God exist?The celestial teapot: who must prove it?
Does God exist?AtheismThe burden of proof

The celestial teapot: who must prove it?

Bertrand Russell 1872 · 1970Clarify thought with logic, doubt dogmas.
See profile
What it claims

There is no reason to believe in God. Doubting is not presumption: it is for the believer to prove.

Why

Whoever claims something exists must prove it, and no proof of God's existence holds up.

The same idea through history
Carl Sagan1995The invisible dragon in the garage: a claim that no test can refute is worth nothing as knowledge. It is the same burden of proof as Russell's teapot, with another animal.
In short

"The celestial teapot: who must prove it?" is an argument by Bertrand Russell (Is There a God?, 1952 (published 1997)) on the question "Does God exist?" (answer: atheism). Its claim: There is no reason to believe in God. Doubting is not presumption: it is for the believer to prove. According to AskPhilosophy's logic check (Z3 solver), the argument is valid only if one grants a premise the author leaves unstated (P4). AskPhilosophy presents 4 objections, 4 of them with a documented reply, and the original text behind each premise, checked word for word.

Is There a God?, 1952 (published 1997)
Valid with an implicit premise

If I say that a small teapot orbits the Sun between the Earth and Mars, too small to be seen through a telescope, nobody can prove me wrong. Should we therefore believe it? Russell uses this image for a simple question: when a claim can be neither proved nor disproved, who must bring the proof? The argument does not conclude that God does not exist. It concludes that we have no reason to believe in God, which corresponds here to the atheist position ("no good reason to believe in one"). Russell himself called himself an agnostic in the strict sense.

P1It is for whoever claims that something exists to prove it, not for the sceptic to prove that it does not. Without proof, we have no reason to believe it, even if nobody can disprove it.
P2What counts as proof is an argument or evidence. That a belief is ancient, widespread, taught to children every Sunday, does not make it any more reasonable.
P3Yet none of the proofs of God's existence survives examination: not the first cause, not design, not the moral arguments. There is "no shred of evidence" of a purpose in the universe.
P4 · implicit premiseAs far as proof is concerned, "God exists" is a claim of existence like any other: the same rule applies to it as to the teapot.
Therefore there is no reason to believe in God. Doubting is not presumption: it is for the believer to prove.

click a premise to reject it

If you accept every premise, logic forces you to accept the conclusion. So the debate is about the premises.
Checked by a logic solver: the premises cannot all be true while the conclusion is false.
Without the implicit premise: Without P4, the computer finds a case where all the other premises are true, no proof exists, and yet there is a reason to believe: if "God exists" is not an ordinary claim of existence, the teapot rule does not apply to it.
Valid with P4, which Russell does not state: he moves from the teapot to religious dogmas as if they were the same case. That is where the strongest objections strike: God is not an object lost in space, and some philosophers think one can believe in him without proof, as we believe without proof that the past existed. Note also what the argument does not say: it does not prove that God does not exist. Russell grants this himself ("my position is agnostic").
Analogy. A classmate claims an invisible, silent, odourless dragon lives in his garage. You cannot prove him wrong. But that is not your job: as long as he brings nothing, you have no reason to believe him. And if the whole town had believed it for centuries, that would change nothing about the proof.

Objections

4

Each objection is a conversation: click to see who replies and where the debate stands.

A
A common objection: "absence of evidence is not evidence of absence" attacks premise 1

Having no proof that something exists is not having proof that it does not. The sceptic must say whether he claims he can prove that God does not exist; otherwise he is no further on than the believer. This is the question Father Copleston puts to him at the very start of their 1948 debate: are you an agnostic or an atheist?

Russell replies
Russell
Russell grants it at once: no, he would not say that God's non-existence can be proved; his position is agnostic.
The debate goes on (our summary) The believer would answer that the admission changes everything: if God's existence and non-existence can both not be proved, the teapot does not show that doubting is more reasonable than believing, only that both sides must bring reasons.
Where the debate stands (our summary)Russell and his critics agree on one point: it cannot be proved that God does not exist. The disagreement is about what follows: for Russell, the burden of proof stays with the believer; for his critics, this admission puts believer and sceptic on an equal footing, each having to bring reasons.
F
Father Copleston and the Thomists 1948 · attacks premise 3

It is false that belief in God is without proof, like the teapot. There is a serious metaphysical proof: contingent things do not contain the reason for their existence; so there must be a being that contains it in itself and cannot not exist.

Russell replies
Russell
Nothing allows us to suppose that the world, besides each of its parts, must have a cause. Russell sees no ground for this assumption, and asks to be given one. (The full debate is in the next argument: "The universe is just there".)
The debate goes on (our summary) Copleston at once gives the ground asked for: either the series of things has a cause, or it is sufficient to itself; but it cannot be sufficient to itself, since it is nothing more than its members and each of them depends on something else. For him, then, belief in God does have a proof, and the comparison with the teapot fails.
Where the debate stands (our summary)Core disagreement: is the argument from contingency a proof of God's existence? Copleston says it is; Russell refuses to move from the parts of the world to the world as a whole. The full debate is in the argument "The universe is just there".
William James (The Will to Believe) 1896 · attacks premise 1

Demanding proof before believing is not always reasonable. When the question is vital, when we cannot avoid choosing and the evidence does not settle it, we have the right to let the heart speak: refusing to believe for fear of being wrong is also a choice, and a wager. Russell, for his part, sums up James's argument differently: we ought to believe in God because, otherwise, we shall not behave well.

Russell replies
Russell
At best this argument shows that the belief is useful, not that it is true. Yet a genuine believer wants to know whether God really exists. Belief in Father Christmas is useful in the nursery, but adults do not conclude that he is real.
The debate goes on (our summary) James would reply that Russell is aiming at a version that is not his: he does not say that usefulness proves truth, but that when reason cannot decide we have the right to choose to believe, and that refusing to choose is already a choice.
Where the debate stands (our summary)Core disagreement: may one believe without proof when the question is vital and the evidence does not settle it? James says yes; Russell answers that a belief's usefulness does not prove it true. Open point: no widely accepted answer.
Alvin Plantinga ("Reason and Belief in God") 1983 · attacks premise 4

The rule "no proof, no reason to believe" does not hold for all our beliefs. We believe without argument that the past existed or that other people have minds, and we are right to. Belief in God can be of this kind, "properly basic": it is not a hypothesis about one more object in the universe, like the teapot, and it is not irrational for lack of proof.

a defender replies
summary: their text is not yet in our corpus
No reply possible: objection raised in 1983, after the author's death (1970).
Herman Philipse defender of the argument · 2012
Summary: for Philipse, Plantinga's model is plausible only if God exists: it reassures the believer who is already convinced, but does not exempt from reasons someone who asks whether their faith is true. And Plantinga gives no independent evidence that his "sense of the divine" is more reliable than that of Muslims. A belief that claims to be true must therefore be backed by arguments, like any other claim; the most promising route, in his view, is Swinburne's cumulative case, which he then judges insufficient.
Where the debate stands (our summary)Core disagreement: does the rule "no proof, no reason to believe" apply to belief in God? Plantinga holds it can be properly basic, without argument; Philipse holds that a belief claiming to be true must, like any other, be backed by reasons. Open point: no consensus.
What today's voices say
Defend it
Ricky Gervais'You say there's a god. I say can you prove that? You say no. I say I don't believe you then.'
Richard DawkinsHe used Russell's teapot in A Devil's Chaplain (2003) and The God Delusion (2006) to show that belief and disbelief do not deserve equal respect.
In the same family · The burden of proof
Do not claim to know what you do not knowThomas Henry Huxley · 1869 · debate over 3 levels
Does God exist?Agnosticism 4