Self-refuting claims

Why does the claim that every truth is merely relative end up denying itself?

Someone says that all truth is relative, and someone else says "is that true absolutely?" — and the room treats this as a small victory, a debating trick that ends the conversation without anyone having learned anything.

I want to ask whether it is one. If the move is legitimate it is remarkably powerful: it needs no evidence and no agreement from the other side, since it uses only the claim itself. And if it is that powerful, we should be suspicious about what it establishes, and about what it leaves standing.

Start by making the move precise, because "it contradicts itself" does several jobs at once.

Take the claim unrestricted: every truth is true only relative to some framework. Since it is offered as a truth, ask whether it falls inside its own scope. There are two answers and both cost something.

Say it is inside. Then it too is true only relative to a framework — true for the relativist, while the person whose framework says truth is absolute is, by the relativist's own account, saying something true for them. The claim survives, but can no longer be asserted at anyone. It has become a report of what the speaker's framework contains, with no purchase on a listener who does not share it, and the relativist has lost what made the position worth stating.

Say instead it is outside — that this one claim is absolutely true. Then a truth exists which is not relative, and the claim is false by its own content. Here "denies itself" is literal: being true entails being false. That is a proof of falsity, not a taunt.

Now separate two ways a claim can turn on itself, because they differ in force.

One is what the claim says. "Every truth is merely relative", "there are no truths", "no sentence has a determinate meaning" — apply the content to itself and it yields its own falsity. The claim cannot be true, full stop.

The other is what the act of saying it does. "I am not now speaking." "I cannot write a word of English." Nothing in the content is contradictory — both could be true of someone, and a third party could report them truly. What fails is the performance: the claim is unassertible by this speaker in this act and remains available to everyone else. J.L. Mackie set out a taxonomy along these lines in the 1960s, distinguishing pragmatic self-refutation from the absolute variety (recalled). Refuting a speaker is not refuting a proposition.

Neither is a paradox. "This sentence is false" admits no assignment of true or false at all; a self-refuting claim is merely false, and demonstrably so.

Two cases show the pattern working. In the Theaetetus, Socrates turns Protagoras' doctrine that man is the measure of all things back on itself: if what each person thinks is true for them, and most think the doctrine false, then by his own principle their judgement stands. This is the peritrope, the turning about. Whether it succeeds is disputed, since Protagoras can index every truth to a person and accept that his doctrine is false for them — the first horn, taken deliberately.

The second is modern. The verification principle held that a statement is meaningful only if verifiable in experience or true by definition — and is itself neither, so by its own criterion it is meaningless. A.J. Ayer's reply, in later editions of Language, Truth and Logic, changed the claim's status rather than its content: the principle is a proposal about how to use the word "meaningful", and a proposal cannot be false, only useless.

That reply is the key to the subject, because every escape has one shape: move the claim out of its own scope, by restricting it ("all truths other than this one are relative") or reclassifying it as a rule of use. Formally this always works, so the question is what it costs. The exemption is usually ad hoc, motivated by nothing but survival. The restricted claim often no longer says what the position wanted said — the relativist's thought was that there is no standpoint outside all frameworks, and has just claimed one. And exemptions invite a regress: if this claim may be excepted, why not others?

A third response refuses to escape: Wittgenstein ends the Tractatus by declaring his own propositions nonsense, a ladder to be climbed and thrown away (6.54) — an acceptance of self-defeat, coupled with the claim that a proposition can show what it cannot say. Whether that is profound or an evasion is still argued.

So what does self-refutation establish? That a claim cannot be true in unrestricted form — not that the thought behind it is false. Global relativism dies to the peritrope; particular relativisms, about etiquette or taste or some contested class of moral judgements, are untouched, because they were never unrestricted. The argument constrains formulation, and mistaking it for a refutation of a position is the commonest way it is abused.

That is where it sits among its neighbours here. The objection appears in speaking of the unspeakable, where "no positive statement about God is true" looks like a positive statement about God, and Maimonides answers by reclassifying it as a claim about our language — exactly the escape above, with exactly its bill. That piece asks whether the escape works there; this one asks what kind of weapon the objection is.

Two refinements connect the pieces.

Self-refutation is a test, not a subject matter. Nothing about relativism makes it uniquely vulnerable; the vulnerability comes from a claim being universal and about truth, which puts it inside its own quantifier. Any sweeping enough thesis is exposed the same way — sceptics, verificationists and eliminativists about meaning have all faced a version.

And the strength of the test is inseparable from its narrowness. It requires no concession from an opponent because it uses only what they said — and for that reason it reaches only what they said. Reformulate, and it must be run again against the new claim, which is the loop that makes these debates long rather than decisive.

Every term the collection defines is gathered in the glossary.