# Burden of Proof ## Intro The burden of proof is the obligation to support what you have asserted. In its original home, a courtroom, it is assigned by rule: the prosecution bears it, the defendant does not, and the standard is set in advance. Outside a courtroom there is no rule assigning it, which is why so many arguments about God turn into arguments about who has to go first. The standard atheist position is that the burden falls entirely on the theist. The theist asserts that a God exists; the atheist merely declines to accept the assertion; and one does not have to prove a negative. Antony Flew gave this its classic form as the **presumption of atheism** (1972), arguing that debate should begin from atheism as the default position, with theism required to earn its way out. Three replies dismantle the position without the exchange ever becoming a standoff. **First, the burden follows the assertion, not the label.** Anyone who makes a claim owes support for it. If an atheist says only that he is unpersuaded, he has said something about his own mental state and has taken no position anyone could argue with. If he says God does not exist, that is a claim about reality with the same standing as its denial, and it carries a burden. Many people move between the two definitions mid-conversation, holding the strong position and defending the weak one. **Second, the shared burden is the honest arrangement.** Two people offering competing accounts of why anything exists, why the universe is ordered, why consciousness and moral obligation are features of the world, are both explaining. Neither is a bystander. The productive framing is not "prove it to me" but "here is my account of these features of reality, and here is yours; which explains more with less." **Third, the no-default rule cuts both ways.** There is no neutral starting position on the existence of God, because the alternative to theism is not a blank but a rival metaphysics. Atheism carries commitments about the origin of the universe, the status of moral facts, and the reliability of reason. Those commitments are not free. Quick reply line: *"I will carry my burden and I am glad to go first. But be clear which position you are defending. If you are saying God does not exist, you have made a claim about reality and you owe an argument. If you are only saying you personally are not convinced, that is a fact about you, and there is nothing there for me to engage."* ## In full The burden of proof, *onus probandi*, is the dialectical obligation to provide justification for a proposition one advances. The formal maxim is *ei incumbit probatio qui dicit, non qui negat*, the proof lies on the one who asserts, not on the one who denies, and understanding what that maxim does and does not license is most of the argument. The maxim is sound as a rule of legal procedure, where the stakes of a wrongful conviction justify an asymmetric standard set in advance. It does not transfer cleanly to metaphysics, because in a metaphysical dispute both parties typically hold positive views, and there is no institution assigning roles. ## The two definitions of atheism Nearly every dispute here reduces to an equivocation between two positions: | Position | Content | Burden | |---|---|---| | **Strong or positive atheism** | God does not exist | Full burden, this is a claim about reality | | **Weak or negative atheism** | I lack belief in God | No burden, but also no position | The weak definition purchases immunity at the cost of content. A lack of belief is a psychological report. It is true of rocks and infants, it cannot be argued with, and it cannot be argued *for*. The moment someone offers reasons why one *should* lack belief, arguments from evil, from divine hiddenness, from the sufficiency of naturalistic explanation, they have moved to the strong position, because reasons are offered for propositions and not for mental states. The practical move is to fix the definition early and hold it. See [Atheism is a Belief](/codex/atheism-is-a-belief/) for the developed engagement. ## Can you prove a negative? The claim that one cannot is stated far more often than it is examined, and it is false as a general principle. - Negatives are proven routinely. There is no largest prime number, and Euclid proved it. There is no elephant in this room, and looking establishes it. - The real principle is narrower: a **universal** negative over an unsurveyable domain is difficult to establish exhaustively. That is a point about the size of the search space, not about negation. - It does not apply to the God of classical theism at all, because that God is not proposed as an object hidden somewhere in the universe that a search might have missed. He is proposed as the explanation of the existence of anything whatever. A claim of that kind is refuted by showing incoherence or by showing that the explanatory work is done better otherwise, which is precisely what the arguments from evil and from naturalistic sufficiency attempt. Those are arguments, which is the point: the atheist has them, and having them means having a burden. ## Russell's teapot and its relatives Bertrand Russell's teapot, invisible dragons, and the flying spaghetti monster are all versions of one move: God is grouped with entities that are unfalsifiable, arbitrary and unmotivated, so that the burden of disproof is felt to be absurd. The analogy fails on the feature that matters. A teapot orbiting the sun is a contingent physical object, arbitrarily specified, explaining nothing. God as classical theism proposes him is a necessary being invoked to explain features of reality that need explaining: that a contingent universe exists at all, that it is intelligible and finely structured, that consciousness and rational thought occur, that moral obligation is experienced as binding. See [Cosmological Arguments](/codex/cosmological-arguments/), [Moral Argument](/codex/moral-argument/) and [Argument from the Reliability of Reason](/codex/argument-from-the-reliability-of-reason/). Whether those arguments succeed is exactly what the debate is about. But an entity proposed as the terminus of explanation is not in the same category as an entity proposed for no reason at all, and the analogy assumes the arguments have already failed rather than showing it. ## Extraordinary claims Carl Sagan's formulation, that extraordinary claims require extraordinary evidence, adapted from Laplace. It is not unreasonable, and it is worth granting rather than fighting, because the interesting question is what makes a claim extraordinary. If it means low prior probability, the standard is Bayesian and unobjectionable, and the dispute becomes about what the correct priors are. That is a real argument and one worth having. If it means violating what we already believe, the principle is conservative in a way that would have blocked most scientific revolutions, and it invites the reply that the extraordinary claim in the room may be that a finely structured, intelligible universe containing conscious moral agents arose without explanation. ## Practical guidance - **Accept the burden and go first.** Refusing looks evasive and forfeits the initiative. The theist has the arguments; using them is better than negotiating about who must. - **Get the definition on the record early.** One question does it: *"Are you claiming God does not exist, or reporting that you are not convinced?"* - **When reasons appear, name the shift.** The moment an argument from evil is offered, the interlocutor has assumed a burden, and it should be noted without hostility. - **Do not accept the interrogation format.** A structure in which one side only ever asks and never defends is not a debate. Ask what account they give of the same features of reality. - **Distinguish burden from standard.** Even granting the burden entirely, nothing follows about how high the bar is. Demanding certainty is a separate move and an unreasonable one, since no interesting claim about the world meets it. - **Note the related epistemological question.** Whether belief in God requires argument at all is contested, and [Reformed Epistemology](/codex/reformed-epistemology/) argues it can be properly basic. That is a different line from the one on this page and should not be mixed into it mid-exchange. ## See also - [Atheism is a Belief](/codex/atheism-is-a-belief/), the debate-prep engagement with the definition question - [Epistemology](/codex/epistemology/), the parent discipline - [Evidentialism](/codex/evidentialism/), the position that belief requires proportioned evidence - [Reformed Epistemology](/codex/reformed-epistemology/), the alternative account of warranted belief - [Argument from Ignorance](/codex/argument-from-ignorance/), the fallacy adjacent to burden disputes - [Antony Flew](/codex/antony-flew/), author of the presumption of atheism, and later a theist - [Agnosticism](/codex/agnosticism/), the position that declines both claims - [Cosmological Arguments](/codex/cosmological-arguments/), the explanatory work the teapot analogy ignores - [Moral Argument](/codex/moral-argument/), another feature requiring explanation - [Argument from the Reliability of Reason](/codex/argument-from-the-reliability-of-reason/), the naturalist's own burden
## Common questions this page answers **Q: Who has the burden of proof in a debate about God?** Whoever makes a claim. If the theist asserts that God exists, he owes an argument. If the atheist asserts that God does not exist, he owes one equally, because that is also a claim about reality. If the atheist says only that he is not personally convinced, he has reported a mental state rather than taken a position, which carries no burden but also offers nothing to discuss. **Q: Isn't atheism just the lack of belief, so it needs no defence?** That definition avoids a burden by giving up all content. A lack of belief is a psychological fact, true of rocks and infants, and nothing can be argued for or against it. The moment someone offers reasons why one should lack belief, such as the problem of evil or divine hiddenness, they have moved to the claim that God does not exist, because reasons support propositions and not mental states. **Q: Can you prove a negative?** Yes, routinely. Euclid proved there is no largest prime. The real principle is narrower: a universal negative across an enormous unsurveyable domain is hard to establish exhaustively. That has no application to classical theism, which does not propose God as an object hidden somewhere in the universe but as the explanation for there being a universe at all. **Q: What is wrong with Russell's teapot analogy?** It compares God to an entity proposed for no reason and explaining nothing. Classical theism proposes God as the explanation of features of reality that require explanation: that a contingent universe exists, that it is intelligible and finely structured, that consciousness and moral obligation occur. Whether those arguments succeed is the debate; the analogy assumes they have already failed rather than showing it. **Q: Do extraordinary claims require extraordinary evidence?** Granting it is more useful than fighting it, because everything then depends on what makes a claim extraordinary. If it means low prior probability, the standard is unobjectionable and the dispute becomes about what the correct priors are. If it means conflicting with current belief, the principle would have blocked most scientific revolutions, and it invites the question of whether a fine-tuned universe containing conscious moral agents is itself the extraordinary claim. **Q: How do I respond when an atheist refuses to defend any position?** Ask one question: are you claiming God does not exist, or reporting that you are not convinced? Then hold the answer. If it is the first, ask for the argument. If it is the second, point out that nothing has been asserted for you to engage, and ask what account they give of the universe's existence, its intelligibility, and moral obligation. A format where one side only asks and never answers is not a debate.