You know something. Or do you?
Picture this: You’re on a jury. The evidence is overwhelming. The defendant’s alibi crumbles. The prosecutor’s timeline is airtight. You vote guilty with total confidence. Then, days later, a hidden camera emerges—showing the actual crime. Everything you believed was true, but for the wrong reasons. You were right by accident.
That’s the vertigo that Edmund Gettier induced with a paper so short it could fit on a napkin. In 1963, he published three pages that demolished the definition of knowledge that had stood since Plato: justified true belief.
Here’s the classic formula: If you believe something, you have good reasons for it, and it’s actually true—then you know it. Simple, elegant, taught for centuries. Gettier showed it’s a house of cards.
He constructed cases where all three conditions are met, yet something is clearly missing. In one, you have strong evidence that Smith will get the job—your boss told you, the interview went perfectly. So you believe ‘Smith will get the job.’ You also believe ‘Smith has ten coins in his pocket’ (you saw him count them). Using logic, you combine the two: ‘Either Smith will get the job and has ten coins, OR Jones will get the job and has ten coins.’ That’s a justified true belief—because as it turns out, Jones gets the job, and unbeknownst to you, Jones also has ten coins. Your belief is true, justified, but not knowledge. You just got lucky.
The real saboteur isn’t luck—it’s a sneaky little rule of logic called OR-introduction.
Most people miss this. The power of OR is the hidden engine of Gettier’s counterexamples. Disjunction introduction lets you take any true statement and add an ‘or’ with any other statement—even a false one. The result is still true, but the truth might come from the wrong branch. You can be perfectly rational, perfectly justified, and still end up with a belief that is true only because of a backdoor.
This isn’t just a philosophical puzzle. It’s a live wire in AI, law, science, and everyday argumentation. When a machine learning model spits out the right answer but for the wrong reasons, do we say it ‘knows’? When a jury convicts based on evidence that only accidentally points to the truth, is that justice? When you’re absolutely certain you’re right—but you can’t trace the path from evidence to conclusion without a hidden ‘or’—are you really knowing?
A perfectly justified belief can be true by accident — and that’s not knowledge.
Gettier’s three pages didn’t just start a war in epistemology. They forced us to confront a uncomfortable truth: certainty is a feeling, not a guarantee. The most rigorous reasoning can be undermined by a logical formality. The most confident conviction can be a lucky guess dressed in justification.
So the next time you catch yourself saying ‘I know it’—pause. Ask yourself: Did I get here through a solid chain of evidence, or did I just take the OR exit? Because the difference between being right and knowing is the difference between a lucky guess and a reliable connection to truth. And that difference is everything.
Certainty is a feeling, not a guarantee. The Gettier problem reminds us that being right is not the same as knowing.
FAQ
Q: Isn't the Gettier problem just a philosophical curiosity with no real-world impact?
A: No. It exposes a fundamental flaw in how we evaluate knowledge in AI, law, science, and everyday reasoning. If a system is justified but only accidentally correct, we can't trust it.
Q: What's the practical implication for me?
A: It means you should question the difference between being right and being reliably right. In hiring, investing, or decision-making, ask: Are my conclusions justified in a way that excludes luck?
Q: What's the contrarian take?
A: Some philosophers argue that Gettier problems are solved by adding a fourth condition, but that's a band-aid. The real lesson is that knowledge is inherently probabilistic—we should embrace uncertainty rather than chase impossible certainty.