The Presumption of Atheism and the Burden of Proof

The Presumption of Atheism

Abstract

On this view, the rational starting point is the absence of theistic belief, and the theist bears the burden of proof. Its two roots are W. K. Clifford's ethics of belief — that believing without sufficient evidence is not merely mistaken but wrong — and Antony Flew's procedural proposal that debate about God should begin from a "presumption of atheism," where "atheism" is understood as the mere lack of theistic belief rather than the denial of God. Combine an evidentialist rule with an absence-of-belief definition, and disbelief becomes the position one occupies by default, cost-free, until positive evidence for God is produced.

Formal statement

1. It is wrong (epistemically, and per Clifford morally) to believe any proposition on insufficient evidence.
2. "Atheism" is best construed, for the purpose of setting the terms of debate, as the state of lacking the belief that God exists — not as the belief that God does not exist.
3. One who merely lacks a belief asserts no proposition and so incurs no burden of proof.
4. There is (the atheist claims) no sufficient evidence for the existence of God.
5. Therefore the presumption lies with atheism: absent sufficient evidence for theism, withholding belief — indeed disbelieving — is the rational default, and the burden rests entirely on the theist.

Key evidence / textual basis

Clifford's principle is stated by the SEP in its baldest form: "It is wrong always, everywhere, and for anyone, to believe anything upon insufficient evidence" (SEP Epistemology of Religion §2). The encyclopedia notes that Clifford's "use of moral vocabulary conveys well the overriding character of the reasons epistemology is said to provide" (SEP §2) — that is, if evidentialism is the correct epistemology and religious belief fails its test, one has an overriding reason to give the belief up. The parallel SEP entry on pragmatic arguments records the same rule and adds Clifford's rationale: the harm of credulous belief is social, since a community that loses "the habit of testing things and inquiring into them… must sink back into savagery" (SEP Pragmatic Arguments §3).

Flew's contribution is to attach a definition to this rule. The SEP records that defining "atheist" as one who merely lacks the belief that God exists "was famously proposed by the philosopher Antony Flew and arguably played a role in his (1972) defense of an alleged presumption of 'atheism'" (SEP Atheism and Agnosticism §1). The strategic point of the absence-of-belief definition is exactly that a person who merely lacks a belief seems to owe no argument for lacking it — so, on this construal, the theist alone must produce evidence. The SEP further notes that this "presumption of atheism" rests on evidentialism: the defense of it "may well be implicit in Flew's famous 'The Presumption of Atheism' (1972)" and turns on an Ockhamist principle "that, in the absence of evidence for the existence of things of kind X, belief in Xs is not reasonable" (SEP Epistemology of Religion §3).

An empiricist epistemology of the kind that makes such a presumption feel natural is available in our corpus in Bertrand Russell's Problems of Philosophy. Russell holds that "all our knowledge of truths is infected with some degree of doubt" (Russell 1912, ch. XIII), and that beyond self-evident truths what we have is "probable opinion" to be proportioned to its grounds — a temper of mind congenial to the demand that existence-claims earn their keep. (Russell's own china-teapot version of the burden-of-proof argument is not in this text; his general evidentialist epistemology is.)

Scripturally, this view stands in tension with Rom 1:19-20 (bib), which claims that "what can be known about God is plain," so that unbelief is not an innocent default but a suppression of available knowledge — a claim the presumption-of-atheism advocate must reject or reinterpret.

Leading proponents

• Antony Flew — "The Presumption of Atheism" (1972); proposed the absence-of-belief definition that shifts the burden onto the theist. (Primary text copyright / not in corpus; represented here via SEP.) Flew later, and famously, abandoned atheism, but the 1972 procedural argument remains the locus classicus of the burden-shifting move.
• W. K. Clifford — "The Ethics of Belief" (1877); the evidentialist rule that supplies the presumption its normative force. (Primary text not in corpus; quoted via SEP.)

Strongest counter-arguments

The most-cited objection targets premise 2. The SEP argues that the absence-of-belief definition "fails as an umbrella term" and generates paradoxes: on it, "babies, cats, and rocks" count as atheists, and — more seriously — a person who lacks belief in God while judging "the truth of theism… more probable than its falsity" would be counted an atheist, which "is counterintuitive in the extreme" (SEP Atheism and Agnosticism §1). The encyclopedia contends that in philosophy "atheism" is standardly and best understood as the proposition that God does not exist, symmetric with theism, precisely so it can "be true or false and can be the conclusion of an argument" (SEP §1). If that is right, atheism is an assertion, and premise 3 collapses.

A second objection, from Scott Shalkowski, is an ad hominem tu quoque that the SEP reports approvingly: defenders of the presumption "tend in fact to be atheists not agnostics, yet a careful examination… of the examples used to support Ockham's Razor show that either they are ones in which there is independent evidence for denying the existence of Xs or ones in which suspense of judgement seems to be the appropriate response, not denial" (SEP Epistemology of Religion §3). In other words, the Ockhamist principle, honestly applied, yields agnosticism, not atheism — so it cannot underwrite a presumption of atheism specifically. The SEP's verdict: "critics of religious belief are not intellectual superiors correcting the mistakes of their inferiors, but disputants" (SEP §3).

Responses

Defenders reply that Flew's definition need not win as the only legitimate meaning of "atheism"; it needs only to be a legitimate way of framing a debate. The SEP concedes that Flew's is "certainly a legitimate definition in the sense that it reports how a significant number of people use the term" (SEP Atheism and Agnosticism §1). And the most sophisticated heir of the presumption — the "low priors argument" — sidesteps the definitional quarrel entirely. It grants the agnostic that the evidence may be "ambiguous or absent," yet argues that "theism starts out with a very low probability before taking into account any evidence," so that ambiguous evidence leaves theism "very probably false" (SEP §6.2). This, the SEP says, "implicitly addresses" the burden-of-proof question "in a much more sophisticated and promising way" than "bad analogies to Santa Claus, flying spaghetti monsters, and Bertrand Russell's… famous china teapot" (SEP §6.2). Whether theism really has a low intrinsic probability is, of course, exactly what a theist will contest.

Assessment

Assessment: Live — the crude burden-shifting version (mere lack of belief incurs no burden) is under sustained pressure from the symmetry objection and from Shalkowski's point that Ockham's razor yields agnosticism rather than atheism. But the underlying evidentialist intuition, refined into the low-priors argument, remains a serious and much-discussed position. The presumption of atheism is weakest as a definitional trick and strongest as a claim about prior probability.

Evidentialist Parity / Theistic Burden-Sharing

Abstract

This theistic view accepts the evidentialist arena but denies that the theist plays in it alone. If "atheism" is the proposition that God does not exist — the standard philosophical usage — then it is a substantive claim about ultimate reality, and its assertor bears a burden of proof symmetric with the theist's. The believer, William Lane Craig among them, then argues that this burden is dischargeable: natural theology (cosmological, teleological, moral arguments) supplies the positive evidence the evidentialist demands, so the theist meets the standard rather than evading it.

Formal statement

1. In philosophy, "atheism" standardly denotes the proposition that God does not exist, symmetric with theism (the proposition that God exists).
2. Whoever asserts a proposition bears a burden of proof for it.
3. Therefore both the theist and the (standardly-defined) atheist bear burdens; neither position is the cost-free default.
4. The theist can discharge this burden by giving a natural theology — deductively valid or cumulatively probable arguments from premises the interlocutor accepts.
5. Therefore the demand for evidence, far from favoring atheism, is one the theist can and should meet head-on.

Key evidence / textual basis

The parity claim rests on the SEP's central argument for the standard definition: because philosophers treat "theism" as a proposition that "is true or false" and can be "the conclusion of an argument," it is best to define "atheism" analogously, "as the contradictory of theism" (SEP Atheism and Agnosticism §1). On that construal atheism is not a restful absence of belief but a positive denial, and "It proposes positive disbelief rather than mere suspension of belief" (SEP §1, quoting Rowe). Symmetric propositions carry symmetric burdens.

The second half of the view — that the theist's burden is dischargeable — is exactly the project the SEP labels natural theology: "Theistic philosophers may grant evidentialism and even grant its hegemony, but defend their religious convictions by providing the case which evidentialists demand" (SEP Epistemology of Religion §4). The encyclopedia distinguishes two strategies here: demonstration — "a deductively valid argument from premises which are themselves justifiably held with full belief" — and probabilistic cumulative case, "ones which are probable, either in the sense of having weight but being inconclusive or in the sense of having a mathematical probability assigned to them" (SEP §4). Notably, the SEP's own bibliography lists Craig's 1979 Kalam Cosmological Argument as an instance of the demonstrative project (SEP §4) — a first-party textual anchor for Craig's role, since his own volume is not in our corpus. The concrete discharging arguments are developed in our The Kalam Cosmological Argument and related natural-theology articles.

Scripturally this view is at home with Rom 1:19-20 (bib) — the claim that God's "invisible attributes… have been clearly perceived… in the things that have been made" is, in effect, a scriptural natural theology, asserting that the evidence is there to be discharged into, not absent.

Leading proponents

• William Lane Craig — insists the atheist who says "there is no God" makes a knowledge-claim requiring justification, and undertakes to meet the theist's own burden through natural theology (Kalam, fine-tuning, moral, resurrection arguments). His Kalam Cosmological Argument (1979) is cited by the SEP as a demonstrative natural-theology project (SEP §4). (Craig's monographs on the burden-of-proof question itself are not in our corpus; his role here is anchored via the SEP citation and the companion Kalam article.)

Strongest counter-arguments

The sharpest reply is that parity is illusory because the two propositions differ in prior probability. Even granting that "God does not exist" is a genuine proposition, the low-priors advocate argues that it starts with a high intrinsic probability and theism with a low one, so "ambiguous or absent evidence" leaves atheism "very probably true" (SEP Atheism and Agnosticism §6.2). Symmetry of logical form does not entail symmetry of epistemic starting point. A second objection: even if the theist's arguments have weight, "even a highly probable argument differs from a demonstration in that the former is vulnerable to probabilistic counter-arguments," so that "a probabilistic version of the Argument from Evil might subsequently reduce the probability" the natural theologian has built up (SEP Epistemology of Religion §5). Discharging the burden is thus never final; it invites the counter-case developed in The Evidential Problem of Evil and Divine Hiddenness.

Responses

The parity advocate answers the prior-probability objection by contesting the alleged low prior: the claim that theism has a low intrinsic probability is itself a substantive, defensible-or-refutable thesis (the SEP flags that Le Poidevin "never shows that the intrinsic probability of a proposition depends only on its specificity," and cites Swinburne against him — SEP Atheism and Agnosticism §4). To the "never final" objection, the response is that vulnerability to counter-argument is a feature of all empirical rationality, not a special defect of theism — the same holds of any scientific theory — so it does not restore a presumption against theism; it merely keeps the case open. Whether the cumulative theistic case in fact outweighs the evidential problem of evil is where the debate genuinely lives.

Assessment

Assessment: Live — the definitional half of the view (standard-definition atheism bears a burden) is well supported by the SEP's own analysis and is arguably the majority position among philosophers of religion. The dischargeability half depends entirely on whether the first-order natural-theology arguments succeed, which is contested article-by-article across this wiki.

Reformed Epistemology

Abstract

Where the first two views argue within the evidentialist arena, Alvin Plantinga's Reformed epistemology challenges the arena itself. Belief in God, on this view, can be "properly basic" — warranted directly, not on the basis of argument, much as beliefs about the past or the external world are — provided it is appropriately grounded and defended against objections. If so, the evidentialist premise on which both the presumption of atheism and the parity view depend (that belief needs propositional evidence) is not a neutral law of reason but a contestable, and possibly self-undermining, epistemological thesis. The burden-of-proof question is thereby reframed: the theist need not first win an evidential contest to be rational.

Formal statement

1. A belief is warranted if it is produced by cognitive faculties functioning properly, in an appropriate environment, aimed at truth — whether or not it is inferred from other beliefs.
2. Belief in God can be produced in this way (e.g., grounded in religious experiences of awe, gratitude, contrition, or a sensus divinitatis), and so can be "properly basic."
3. Properly basic beliefs need no propositional evidence to be rational; they need only be grounded and defended against known defeaters.
4. Therefore the evidentialist demand — that theistic belief is irrational unless supported by sufficient argument — is false as a universal principle.
5. Moreover, evidentialism may be self-referentially inconsistent, since "there is no evidence for evidentialism"; so the principle that grounds the whole burden-of-proof dispute cannot simply be assumed.

Key evidence / textual basis

The SEP states the core claim directly: "Reformed epistemologists, such as Wolterstorff (1976) and Plantinga (1983) reject evidentialism, arguing that beliefs are warranted without Enlightenment-approved evidence provided they are (a) grounded, and (b) defended against known objections" (SEP Epistemology of Religion §6). Grounding is glossed on Plantinga's own terms: "S's belief that p is grounded in event E if (a) in the circumstances E caused S to believe that p, and (b) S's coming to believe that p was a case of proper functioning" (SEP §6). The self-referential objection to evidentialism is likewise reported as Plantinga's: "evidentialism is self-referentially inconsistent for there is no evidence for evidentialism" (SEP §3).

The companion entry situates the same move in the burden-of-proof debate. Confronting Le Poidevin's premise that the total evidence is ambiguous, the SEP notes that "So-called 'Reformed epistemologists' (e.g., Plantinga 2000) might challenge the second premise… on the grounds that many beliefs about God, like many beliefs about the past, are 'properly basic'… and so are, in effect, a part of the total evidence" (SEP Atheism and Agnosticism §4). In other words, if the sensus divinitatis delivers God-beliefs, then those beliefs are themselves data — and the presumption that the theist begins with nothing to show is denied at the root.

Scripturally the view is often paired with Rom 1:19-20 (bib) (the knowledge of God as immediate and universal) and with Heb 11:6 (bib) — "without faith it is impossible to please him, for whoever would draw near to God must believe that he exists" — which frames belief less as the terminus of an argument than as the posture in which God is known. Ps 14:1 (bib), "The fool says in his heart, 'There is no God,'" reads denial as a moral-cognitive failure rather than a neutral default, cohering with the Reformed diagnosis of a disordered sensus divinitatis.

Leading proponents

• Alvin Plantinga — "Reason and Belief in God" (1983); Warranted Christian Belief (2000). Argues theistic belief can be properly basic and warranted apart from argument, and presses the self-referential objection against evidentialism. (Primary texts copyright / not in corpus; represented via SEP.)
• Nicholas Wolterstorff — Reason within the Bounds of Religion (1976); co-founder of the Reformed-epistemology program (cited via SEP; profile pending {{PROFILE-PENDING}}).

Strongest counter-arguments

The SEP presses two serious objections. First, the analogy to memory and perception is weak: the alleged sensus divinitatis "operates at most sporadically and far from universally," "can easily be resisted," and "the existence of the beliefs it is supposed to produce can easily be explained without supposing that the faculty exists at all" (SEP Atheism and Agnosticism §4). If the faculty's outputs are dispensable, treating them as basic data looks question-begging. Second, and symmetrically, Reformed epistemology appears to license atheistic proper basicality just as readily: Jerome Gellman's "experience of godlessness," "occasioned by… the evils that surround us," "would seem to ground atheism in the same way that the experience of forgiveness can ground theism" (SEP Epistemology of Religion §6). The framework may thus produce a "stand-off between theists and de-bunkers of religion" (SEP §6) rather than a decision. A further worry from Swinburne: Plantinga's later work "largely ignores the question of justification, or reasonableness" — whether the beliefs are "probable relative to total evidence" (SEP §6).

Responses

Reformed epistemologists reply that the "weak analogy" charge assumes what it must prove — namely that the sensus divinitatis does not exist; but "if God exist it is" the product of proper functioning, and "if atheist naturalism is correct then theism would not be" (SEP Epistemology of Religion §6). As Andrew Moon observes in the SEP's telling, from that stand-off "we may draw the conclusion that the de-bunker has failed to undercut religious belief" (SEP §6) — a defensive success, even if not an offensive one. The parity-with-atheism worry (Gellman) is met by the second Reformed requirement: warranted basic beliefs must be defended against known objections, and the theist claims the experience of forgiveness survives defeaters that the experience of godlessness does not — though the SEP treats this as unresolved.

A cousin of Reformed epistemology deserves a place here: the Jamesian pragmatic reply to Clifford. Where Plantinga denies that evidence is required, William James denies that withholding is mandatory when evidence is silent. James's target is precisely "the agnostic imperative (withhold belief whenever the evidence is insufficient)" (SEP Pragmatic Arguments §3). His governing thought: "a rule of thinking which would absolutely prevent me from acknowledging certain kinds of truth if those kinds of truth were really there, would be an irrational rule" (SEP §3, quoting James 1896). Clifford's rule, on James's analysis, embodies one strategy — "avoid error at all costs" — while an equally rational strategy is to "seek truth by any means available, even at the risk of error" (SEP §3). When an option is living, momentous, and forced — a "genuine option" that the evidence leaves open — James holds it may be "decidable on passional grounds" (SEP §3). Crucially, James "is not arguing against conforming one's belief to the evidence, whenever there's a preponderance of evidence" (SEP §3); his quarrel is only with Clifford's prohibition on believing when evidence is silent. This connects to Heb 11:6 (bib): faith as a warranted response under evidential ambiguity, not a violation of intellectual duty. Bertrand Russell stands on the other side — for the empiricist, belief should track "probable opinion" and its grounds (Russell 1912, ch. XIII), and the Jamesian license to believe past the evidence looks like wishful thinking dressed as epistemology.

Assessment

Assessment: Strong — as a defensive thesis (theistic belief can be rational without a completed natural theology, and evidentialism cannot simply be assumed) Reformed epistemology is widely regarded as having reset the burden-of-proof debate; the SEP grants that it forces the evidentialist to defend, rather than presuppose, the demand for evidence. As an offensive thesis it is weaker: the framework's apparent symmetry (godlessness grounding atheism as forgiveness grounds theism) means it tends to a stand-off rather than a demonstration, and Swinburne's charge that it sidelines the question of all-things-considered probability remains live.