What is the difference between knowing something and merely believing it? Suppose two people both say the bus comes at 8:15. One read the timetable; the other guessed. If the bus does come at 8:15, both were right, but only the first knew. Philosophers have long tried to say exactly what makes the difference.
The classic analysis
In Plato's dialogue Theaetetus, Socrates considers the suggestion that knowledge is true belief accompanied by an account, a reason or explanation. This developed into what is now called the justified true belief analysis. On this view, a person knows a proposition when three conditions are met:
- Belief: they believe it. You cannot know something you do not believe.
- Truth: it is true. You cannot know something false; you can only think you know it.
- Justification: they have good reason or evidence for it. A lucky guess is not knowledge, because it lacks justification.
Each condition seems necessary, and many philosophers treated the three together as sufficient: meeting all three was thought to be enough for knowledge.
Gettier's challenge
In 1963 the American philosopher Edmund Gettier published a three-page paper, “Is Justified True Belief Knowledge?”, that changed the field. He described cases in which someone has a justified true belief that clearly is not knowledge.
In one case, Smith and Jones have applied for the same job. Smith has strong evidence that Jones will get it (the company president told him so) and that Jones has ten coins in his pocket (Smith counted them). Smith concludes: “The man who will get the job has ten coins in his pocket.” In fact Smith himself gets the job, and, unknown to him, he also happens to have ten coins in his own pocket. His belief is true and justified, but true only by luck. It seems wrong to say he knew it.
Bertrand Russell had described a similar case earlier: someone glances at a clock that has stopped, at the one moment when it happens to show the right time. Their belief is justified (clocks are usually reliable) and true, but they do not know the time.
What goes wrong
In these cases, the connection between the person's reasons and the truth of the belief is broken. They reach the truth by luck, not because their reasons track how things are. Philosophers call this epistemic luck, and the challenge is to explain how knowledge excludes it.
Proposed repairs
- No false lemmas: knowledge must not be based on any false belief. Smith reasoned from the false belief that Jones would get the job. But other cases involve no false step in reasoning, yet still seem not to be knowledge. In one, a person looks at a real barn in a county filled with fake barn facades they could not tell apart.
- Defeasibility: knowledge is justified true belief that would not be undermined by any further truth one does not know.
- Reliability or causal connection: knowledge must be produced by a process that reliably leads to truth, or the fact must cause the belief in the right way.
- Knowledge first: some philosophers, notably Timothy Williamson, argue that knowledge cannot be broken into simpler parts and should be treated as a basic concept.
No repair has won general agreement, and new counterexamples keep appearing. Some suspect the search for a perfect definition is misguided, while granting that it has taught us a great deal.
Why it matters
Gettier cases are not just puzzles. They show that being right is not the same as having good reasons, and that good reasons do not guarantee being right. In everyday life, the habit of asking “am I right for the right reasons?” guards against overconfidence after lucky successes.
Worked example: the lucky forecast
In January, Dana tells her friends that a certain company's share price will rise this year. Her reason: she read a report that the company was about to be bought by a larger firm. She has reasonable grounds, since the report came from a usually reliable financial newspaper. By December the share price has indeed risen sharply, and her friends are impressed.
But the takeover never happened; the report was wrong. The price rose for an unrelated reason: the company's new product became an unexpected hit. Dana's belief was true and justified, yet its truth had nothing to do with her reasons. It has the shape of a Gettier case, and most people would say she did not really know the price would rise. She was right, but not for the right reasons.
This matters in practice. If Dana and her friends treat her success as proof of skill, they may trust her next prediction more than they should. People who succeed by luck often become overconfident, because outcomes are easier to see than reasons. Investors, managers and gamblers all face this trap.
A good habit is to check reasons, not just results. After any success, ask: did things turn out well because of what I believed, or despite it? If your reasons turned out false, the success tells you little about your judgment. Over many cases a track record can separate skill from luck, but a single right answer cannot.