Clifford's Principle¶

An ethics-of-belief norm holding that accepting or retaining a belief without sufficient evidence is morally wrong, together with a duty not to disregard relevant counterevidence.

Version: v1 · 2026-09-28 · History
Domain-specific #: 8481
Domain group: Humanities
Origin domain: Philosophy
Subdomains: Ethics of Belief, Epistemology → Philosophy
Aliases: Cliffordian evidentialism, Clifford's evidentialist principle

Core Idea¶

Clifford's principle turns evidential responsibility into a moral demand. It assesses not only whether a proposition is true, but whether an agent was entitled to believe it on the evidence available and sought. Its stringent formulation applies to anyone and any belief rather than reserving scrutiny for public or high-stakes claims.

The companion duty not to ignore relevant evidence gives the norm a procedural side. Accumulating congenial reasons is insufficient if the believer avoids inquiry or dismisses counterevidence. The principle therefore couples a support threshold with responsibility for how the evidential record was formed.

How would you explain it like I'm…

Check Before You Believe

Clifford's Principle says you should only believe something when you have good reasons for it. And you should not cover your ears when someone shows you a reason you might be wrong. Believing without good reasons counts as doing something wrong, even about small things.

The Duty to Have Evidence

Clifford's Principle is an idea from philosophy about believing things responsibly. It says that believing something without enough good evidence is actually a moral wrong, not just a mistake. It matters whether you had the right to believe it, based on the evidence you had and the evidence you looked for, not only whether it turned out true. It also says you must not ignore evidence that goes against you, and collecting only reasons you like is not enough. The rule applies to everyone and to every belief, not just important ones.

Moral Duty of Evidence

Clifford's Principle turns evidential responsibility into a moral duty: it is wrong to believe something on insufficient evidence. It judges a belief not just by whether it is true but by whether the believer was entitled to it, given the evidence they had and the evidence they sought. In its strict form it applies to everyone and every belief, not only to public or high-stakes claims. It also includes a duty not to ignore relevant evidence: piling up reasons that support what you already think does not count if you avoided inquiry or brushed off counterevidence. So the principle has two parts, a threshold of support and responsibility for how you gathered your evidence.

Clifford's Principle is an ethics-of-belief thesis that makes evidential responsibility a moral demand: believing on insufficient evidence is wrong. Its assessment targets entitlement rather than truth, whether the agent was justified in believing given the evidence available and the evidence they sought. In its stringent formulation it is universal in scope, applying to any agent and any belief rather than reserving scrutiny for public or high-stakes matters. A companion duty not to ignore relevant evidence gives it a procedural dimension: accumulating congenial reasons does not suffice if the believer evades inquiry or dismisses counterevidence. The principle thus couples a support threshold with responsibility for how one's evidential record was formed, distinguishing it from purely truth-oriented or purely prudential norms of belief.

Scope of Application¶

Ethics of belief. Evaluates assent, doubt, inquiry, and revision as morally significant practices.
Public reasoning. Clarifies why unsupported conviction may be criticized even before harmful conduct occurs.
Testimony and trust. Examines responsibilities when agents rely on others rather than direct evidence.
Disagreement. Separates honest evidential differences from selective attention or facile dismissal.

Clarity¶

Application requires naming the proposition, the agent's evidence, what further inquiry was reasonably available, and why the support was insufficient. The principle concerns responsible assent, not the psychological impossibility of controlling every passing impression. Inclusion test: Identify an actual belief, the evidence available to its agent, a claim that the support is insufficient, and moral criticism of accepting or retaining it. Exclusion test: Exclude ordinary requests for proof, courtroom burdens, purely descriptive accounts of cognition, and advice to seek information that carries no moral judgment. Nearest boundary: Evidentialism can say justification depends on evidence; Clifford's principle adds the stronger thesis that believing on insufficient evidence is morally wrong. Exit condition: The case exits the principle when evidence no longer governs permissibility or when the evaluation becomes solely prudential, legal, or predictive.

Manages Complexity¶

The principle compresses evidence gathering, attention, assent, and moral responsibility into one demanding test. Its severity is analytically useful, but disagreements about sufficiency, voluntary control, and practical urgency must remain visible.

Abstract Reasoning¶

State the proposition and degree of assent.
Inventory favorable evidence, counterevidence, and accessible inquiry.
Assess sufficiency relative to the claim rather than hindsight about truth.
Check whether relevant evidence was ignored or dismissed.
Make the moral inference separately from epistemic and practical criticism.

Knowledge Transfer¶

The pattern transfers where a domain treats information handling as an accountable condition of commitment. It does not automatically convert every uncertain forecast or incomplete database into a moral violation.

Neighborhood in Abstraction Space¶

Clifford's Principle sits in a crowded region of the domain-specific corpus (29^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Epistemic Norms & Informal Fallacies (16 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08