Paradoxes of Language and Truth

Some of the sturdiest-looking ideas we have are the words 'true' and 'false' and the notion of a collection of things. They feel too basic to cause trouble. Yet when language is allowed to talk about itself, these humble notions can be turned against us to produce a flat contradiction from premises no one wants to deny. These are antinomies in the sense of the previous lesson, and they are not idle riddles: one of them forced mathematicians to rebuild the foundations of their subject.

Understanding them pays off well beyond logic. Self-reference lurks wherever a rule applies to itself, a statement comments on statements, or a system tries to describe its own limits — in law, in computing, in claims that begin 'everything I say is...'. Seeing exactly where self-reference is safe and where it detonates is a lasting reasoning skill, and a reminder that even our most trusted concepts have edges.

The Liar: self-reference meets truth

Take the sentence 'This sentence is false.' Ask the obvious question: is it true? If it is true, then what it says holds — so it is false. If it is false, then it fails to be false, which is to say it is true. Either assumption yields its own opposite. The sentence cannot be consistently assigned either value. Notice the two ingredients working together: the sentence refers to itself, and it uses the word false, a truth predicate applied to that very self. Neither ingredient is exotic, yet their combination breaks the tidy rule that every statement is exactly one of true or false.

Why the Liar resists a quick fix

The natural first move is to ban the sentence as meaningless. But the paradox is slippery. 'The sentence in this box is false,' written alone in a box, seems perfectly meaningful, and the same trouble returns. A pair of sentences can do it too: card one reads 'the sentence on card two is true,' card two reads 'the sentence on card one is false' — neither refers to itself, yet together they loop into contradiction. Serious responses, from ranking languages into levels to allowing truth-value gaps, all give up some comfortable assumption about how truth attaches to sentences. There is no consensus solution, which is exactly why the Liar remains studied rather than solved.

Russell's paradox: the Liar strikes mathematics

The same self-referential shape appears with collections instead of sentences. Most sets are not members of themselves — the set of all cats is not a cat. So consider the set of all sets that are not members of themselves. Is that set a member of itself? If it is, then by its own definition it must not be; if it is not, then it exactly qualifies, so it must be. The contradiction is the Liar's twin, with 'is a member of' in place of 'is false.' What makes this graver is that it struck a theory intended to be perfectly rigorous.

Work the Liar slowly, because the discipline matters. Let L be the sentence 'L is false.' Suppose L is true. A true sentence says something that holds; L says that L is false; so L is false. That contradicts our supposition. Now suppose instead L is false. L says precisely that L is false, so what L says is the case; a sentence whose content is the case is true; so L is true. Again a contradiction. We assumed only that L is either true or false — the standard rule — and both branches collapsed. That is the antinomy: not a feeling of confusion but a derived contradiction from ordinary premises, signalling that one of those premises about truth and self-reference must be revised.

Self-reference by itself is not the villain. 'This sentence is written in English' refers to itself and is simply true. 'This sentence contains five words' refers to itself and is simply false, with no paradox. Even 'This sentence is true' — the Truth-Teller — creates no contradiction; it can consistently be called true or called false, it just gives no way to decide, which is a different defect. The lesson is diagnostic: what ignites the Liar is self-reference plus a negative use of the truth predicate that forces the sentence to contradict itself. Locating that precise combination is more useful than banning self-reference wholesale.

The Liar is ancient. It is standardly credited to Eubulides of Miletus, a Greek philosopher of the fourth century BCE, who is reported to have posed it in the form 'A man says that he is lying; is what he says true or false?' It is often linked to the older line attributed to Epimenides, a Cretan, that 'Cretans are always liars' — though, strictly, that remark can simply be false (some Cretan tells the truth) without contradiction, so it is a looser relative rather than the sharp paradox. The clean self-referential version is Eubulides'.

More than two millennia later the same shape reappeared with devastating effect. In 1901 the British philosopher Bertrand Russell discovered the paradox of the set of all sets that are not members of themselves. In a now-famous letter of June 1902 he communicated it to the German logician Gottlob Frege, whose life's work aimed to ground arithmetic in logic and whose second volume of the 'Grundgesetze der Arithmetik' was then at press. Frege's reply was gracious and stricken; he acknowledged that the paradox shook the foundation he had built. The episode helped drive decades of repair work — Russell and Whitehead's theory of types and, later, axiomatic set theories such as Zermelo-Fraenkel — designed to keep the self-membership loop from arising.

• 'The Liar is just a silly word game.' It derives a genuine contradiction from ordinary assumptions about truth, which is why logicians take it seriously.
• 'Any self-referential sentence is paradoxical.' Many are plainly true or plainly false; the Liar needs self-reference and a self-undermining use of 'false.'
• 'Russell's paradox is unrelated to the Liar.' They share one structure, with set membership playing the role that truth plays in the Liar.
• 'Epimenides stated the Liar paradox.' His 'Cretans are always liars' can be merely false without contradiction; the strict paradox is credited to Eubulides.

Write three self-referential sentences: one clearly true, one clearly false, and one that behaves like the Liar. For the paradoxical one, run both assumptions (true, then false) and record how each leads to its opposite. Then state, in a single sentence, the extra ingredient beyond self-reference that made the third sentence explode.

Think Like a Maester: When a statement is allowed to pronounce on its own truth, check whether it can quietly turn into its own denial — that loop, not self-reference alone, is where contradiction hides.

1. Walk through why 'This sentence is false' cannot be consistently called true or false.
2. What two ingredients must combine to produce the Liar paradox?
3. Why is 'This sentence is in English' not paradoxical, even though it refers to itself?
4. State Russell's paradox and explain how it mirrors the Liar.
5. Why is Epimenides' 'Cretans are always liars' a looser relative of the Liar rather than the strict paradox itself?

When language is allowed to talk about its own truth, ordinary notions can produce extraordinary trouble. The Liar sentence, 'This sentence is false,' cannot be consistently assigned either truth value, because self-reference joined to a negative truth predicate makes the sentence contradict itself. Self-reference alone is harmless, as plainly true or false self-referential sentences show; the paradox needs that extra, self-undermining ingredient. The same structure, with set membership in place of truth, gives Russell's 1901 paradox of the set of all sets that are not members of themselves — the discovery that Russell sent to Frege in 1902 and that helped force a rebuilding of logic's foundations. Traceable from Eubulides in the fourth century BCE to modern set theory, these puzzles are antinomies in the fullest sense: contradictions that demand we revise something we had trusted.

• What Is Truth?
• Paradoxes of Infinity