# Evidentialism ## Intro Evidentialism is the view that whether a belief is justified depends on one thing only: how well it fits the evidence the person actually has. Not on whether it makes them happy, not on whether it is useful, not on whether it turns out true by luck. Just the evidence. Stated that plainly, most Christians agree with it. Scripture tells people to test things, to examine claims, to be ready with reasons. The Bereans get praised for checking what they were told against the text. The trouble starts when evidentialism gets a second job. W.K. Clifford, writing in 1877, turned it into a moral rule: it is *wrong*, he said, always and everywhere and for anyone, to believe anything on insufficient evidence. That word "wrong" is doing a lot of work. It converts a claim about justification into an accusation, and that accusation became the backbone of the modern case against religious belief. When someone says "faith is believing without evidence, and that is irresponsible," they are quoting Clifford whether they know it or not. Two replies matter. William James answered in 1896 that in certain choices, ones that are forced, live, and momentous, Clifford's rule is not a neutral requirement of reason but one policy among others, a decision to risk missing truth rather than risk error. Alvin Plantinga answered much later, and more damagingly, that the objection quietly assumes a theory of knowledge that cannot pass its own test. That last point is the one to remember. Clifford's principle is not itself believed on sufficient evidence. It is not self-evident, no experiment supports it, and no argument establishes it. By its own standard, believing it is wrong. Quick reply line: *"Clifford's rule says it's wrong to believe anything without sufficient evidence. What's the sufficient evidence for Clifford's rule? It fails its own test, which is why almost no epistemologist holds it in that strong form anymore."* ## In full **Evidentialism** names three related but genuinely distinct things, and most confused arguments in this area come from sliding between them. 1. **Evidentialism as a thesis about epistemic justification.** The technical position, most carefully developed by **Richard Feldman and Earl Conee** ("Evidentialism," *Philosophical Studies*, 1985; expanded in *Evidentialism*, Oxford, 2004): the epistemic justification of a belief is determined wholly by the quality of the believer's evidence for it. This is a mainstream, respectable position with no anti-religious content whatever. 2. **Evidentialism as an ethics of belief.** The moralized version from **W.K. Clifford**'s "The Ethics of Belief" (1877), on which believing beyond one's evidence is not merely unjustified but *wrong*. This supplies the normative force that the [Argument from Epistemic Normativity](/codex/argument-from-epistemic-normativity/) examines, and it is the version that does the accusing. 3. **The evidentialist objection to religious belief.** The applied claim that belief in God is unjustified because the evidence is insufficient, so the believer is in violation of an epistemic duty. This is [Alvin Plantinga](/codex/alvin-plantinga/)'s target across the *Warrant* trilogy, and it is the one with the structural defect. Confusingly, **"evidentialism" also names a school of Christian apologetic method** (Gary Habermas, John Warwick Montgomery, Josh McDowell), which argues for Christianity from historical and empirical evidence and stands opposed to [Presuppositionalism](/codex/presuppositionalism/) on method. That sense has nothing to do with senses 1 to 3 beyond the shared root, and the codex's own [Minimal Facts Argument](/codex/minimal-facts-argument/) belongs to it. Do not let an opponent equivocate between "evidentialism the objection" and "evidentialism the apologetic method." ## Clifford and James **Clifford's principle**, from "The Ethics of Belief" (1877): it is wrong always, everywhere, and for anyone, to believe anything upon insufficient evidence. His illustration is a shipowner who talks himself into believing his vessel seaworthy, sends it out, and drowns its passengers. Clifford's verdict is that the owner is guilty even if the ship had happened to survive, because the guilt attaches to how the belief was formed rather than to the outcome. The example is well chosen and the intuition is real. Where Clifford overreaches is in generalising from a case with obvious third-party stakes to *every* belief held by *anyone*. **James's reply**, in *The Will to Believe* (1896), grants that Clifford is right in most cases and denies that he is right in all. James identifies a class of options that are **live** (both branches are genuine possibilities for this person), **forced** (there is no neutral ground; declining to choose is itself a choice), and **momentous** (the stakes are significant and the opportunity may not repeat). In such cases, he argues, Clifford's rule amounts to a policy of *risking the loss of truth in order to avoid the risk of error*, which is a preference rather than a neutral requirement. The religious question, James claims, is exactly such an option. The codex treats the live use of this exchange at [Agnostic Retreat Defeater](/codex/agnostic-retreat-defeater/), where the modern "suspension of belief is the neutral default" move is shown to be Clifford's principle smuggled in as though it were free. ## Plantinga's critique The most damaging objection is structural rather than rhetorical, and it runs in two steps. **Step one: the objection presupposes classical foundationalism.** To say belief in God is unjustified because unsupported by evidence, one needs an account of which beliefs may be held *without* evidential support, since not every belief can be supported by another on pain of regress. See [Foundationalism](/codex/foundationalism/). The classical answer, running from Descartes through the modern evidentialist tradition, is that a belief is properly basic only if it is self-evident, incorrigible, or evident to the senses. **Step two: that criterion fails its own test.** The proposition *a belief is properly basic only if it is self-evident, incorrigible, or evident to the senses* is not itself self-evident, not incorrigible, and not evident to the senses. Nor is there a non-question-begging argument for it from beliefs that are. So by its own standard it may not be believed. See [Self-refutation](/codex/self-refutation/). Plantinga's constructive alternative is that belief in God can be **properly basic** without being groundless, arising in appropriate circumstances through faculties functioning as designed, which is the core of [Reformed Epistemology](/codex/reformed-epistemology/) and of the proper-function account of warrant in *Warrant and Proper Function* (1993). The same self-reference problem hits Clifford's principle directly, and it is worth stating separately because it is the version an opponent will actually have quoted. Clifford's rule is a philosophical claim about the ethics of belief. There is no experiment for it, it is not self-evident, and the argument from the shipowner establishes at most that recklessness with others' lives is culpable, not that all under-evidenced belief anywhere is wrong. Held to its own standard, believing it is wrong. ## What the Christian should concede A great deal, and conceding it early is both honest and tactically strong. - **Sense 1 is largely correct and Christians should say so.** Belief ought to track evidence. Scripture commands examination rather than credulity: [1 Thessalonians 5:21](/codex/1-thessalonians-5-21/) on testing everything, [Acts 17:11](/codex/acts-17-11/) on the Bereans checking the claim against the text, [1 Peter 3:15](/codex/1-peter-3-15/) on being ready with a reasoned defence, [Proverbs 18:17](/codex/proverbs-18-17/) on hearing the other side. A tradition that produced [Aquinas](/codex/thomas-aquinas/) and the entire evidential apologetic project is not embarrassed by the demand for evidence. - **Clifford's shipowner is genuinely culpable.** Do not defend motivated reasoning in order to protect faith. The Christian objection is to Clifford's *scope*, not to his intuition about that case. - **"Faith" in Scripture is not belief without evidence.** *Pistis* is better rendered trust, and trust in a person is normally grounded in evidence about that person. See [Faith is Belief Without Evidence Objection Defeater](/codex/faith-is-belief-without-evidence-objection-defeater/) and [Faith and Reason](/codex/faith-and-reason/). What the Christian denies is that the evidentialist objection has the resources to convict, given that its central principle cannot satisfy itself, and given that on naturalism there is no clear account of what makes the duty binding in the first place. That second point is the [Argument from Epistemic Normativity](/codex/argument-from-epistemic-normativity/). ## Positions compared | Position | Justification depends on | Status of the norm | Religious belief | |---|---|---|---| | **Feldman and Conee evidentialism** | fit with the believer's total evidence | epistemic, not explicitly moral | undetermined; depends on what the evidence shows | | **Clifford's ethics of belief** | fit with sufficient evidence | moral duty, categorical | unjustified and blameworthy where evidence is thin | | **James's pragmatism** | evidence, plus permissible passional choice in forced live momentous options | one policy among others | permissible in the religious case | | **[Reformed Epistemology](/codex/reformed-epistemology/)** | proper function of faculties in a congenial environment | warrant replaces duty-talk | can be properly basic, not requiring argument | | **Naturalized epistemology** | causal facts about belief formation | descriptive, norm-talk deemphasised | not a normative question at all. See [Naturalized Epistemology](/codex/naturalized-epistemology/) | ## See also - [Argument from Epistemic Normativity](/codex/argument-from-epistemic-normativity/), the dilemma that presses on where Clifford's "wrong" gets its authority - [Reformed Epistemology](/codex/reformed-epistemology/), Plantinga's alternative account of warrant and proper basicality - [Naturalized Epistemology](/codex/naturalized-epistemology/), the descriptive turn away from the normative project - [W.V.O. Quine](/codex/w-v-o-quine/), whose 1969 essay drove that turn - [Foundationalism](/codex/foundationalism/), the regress problem the evidentialist objection presupposes an answer to - [Self-refutation](/codex/self-refutation/), the pattern Clifford's principle and classical foundationalism both fall into - [Faith is Belief Without Evidence Objection Defeater](/codex/faith-is-belief-without-evidence-objection-defeater/), the popular form of the objection - [Agnostic Retreat Defeater](/codex/agnostic-retreat-defeater/), where the Clifford and James dispute is deployed against the neutral-default claim - [Faith and Reason](/codex/faith-and-reason/), the wider relationship - [Minimal Facts Argument](/codex/minimal-facts-argument/), the codex's own evidential case, which belongs to apologetic evidentialism - [Epistemic Standards for Theism](/codex/epistemic-standards-for-theism/), what standards it is fair to hold theistic claims to - [Alvin Plantinga](/codex/alvin-plantinga/), [David Hume](/codex/david-hume/), [Bertrand Russell](/codex/bertrand-russell/), principal figures - [Epistemology](/codex/epistemology/), [Christian Theories of Knowledge](/codex/christian-theories-of-knowledge/), domain hubs
## Common questions this page answers **Q: What is evidentialism?** The view that a belief's justification is determined entirely by how well it fits the evidence the believer has. Richard Feldman and Earl Conee developed the careful modern version. Confusingly the word names three other things too: W.K. Clifford's moral rule that believing on insufficient evidence is *wrong*, the applied objection that religious belief violates that rule, and a school of Christian apologetics that argues from historical evidence. Most muddled arguments in this area come from sliding between those senses. **Q: Is it wrong to believe something without evidence?** Clifford said yes, always and everywhere and for anyone, in "The Ethics of Belief" (1877). Two problems. First, his own principle is not believed on sufficient evidence, since it is not self-evident, not verifiable by any experiment, and not established by his shipowner example, which shows only that recklessness with other people's lives is culpable. It fails its own test. Second, William James argued in 1896 that for choices which are forced, live and momentous, Clifford's rule is a policy of risking loss of truth to avoid risk of error, which is a preference rather than a neutral requirement. **Q: Do Christians reject evidentialism?** Not the core of it, and they should say so plainly. Scripture commands testing claims rather than swallowing them: [1 Thessalonians 5:21](/codex/1-thessalonians-5-21/), [Acts 17:11](/codex/acts-17-11/) on the Bereans, [1 Peter 3:15](/codex/1-peter-3-15/) on giving a reasoned defence. There is an entire school of Christian apologetics that is *called* evidentialism, arguing for the resurrection from historical evidence. What Christians deny is that the evidentialist objection to belief in God can convict, since its governing principle cannot satisfy itself. **Q: What is Plantinga's objection to evidentialism?** That the evidentialist objection to religious belief presupposes classical foundationalism, on which a belief may be held without support only if it is self-evident, incorrigible, or evident to the senses. But that criterion is itself none of those three, and there is no non-circular argument for it. So it cannot be believed by its own lights. Plantinga's alternative is that belief in God can be properly basic, arising through faculties functioning as designed. See [Reformed Epistemology](/codex/reformed-epistemology/). **Q: What is the difference between evidentialism and presuppositionalism in apologetics?** They are rival methods, not rival epistemologies. Evidential apologetics (Gary Habermas, John Warwick Montgomery, Josh McDowell) argues from shared historical and empirical evidence toward Christian conclusions, as the codex does at [Minimal Facts Argument](/codex/minimal-facts-argument/). [Presuppositionalism](/codex/presuppositionalism/) holds that no evidence is neutral and argues instead that the unbeliever's own reasoning presupposes the Christian worldview. Note the naming collision: this sense of "evidentialism" has nothing to do with Clifford's objection to religious belief.