For most of the history of philosophy, the definition of knowledge was settled. Going back to Plato's Theaetetus, knowledge was understood as justified true belief — abbreviated by philosophers as JTB. To know that something is true, three things had to be the case:
- You believed it (the belief condition)
- It was actually true (the truth condition)
- You had good reasons for believing it (the justification condition)
This definition seemed obvious. If any of the three conditions were missing, you couldn't really claim to know the thing in question. If you didn't believe it, you didn't know it. If it wasn't true, you didn't know it (you just thought you did). And if you had no good reason for your belief — if it just happened to be right by accident — that wasn't knowledge either.
For two thousand years, philosophers refined this definition without seriously challenging it. Then in 1963, a young philosophy professor at Wayne State University named Edmund Gettier published a three-page paper called "Is Justified True Belief Knowledge?" The answer he gave was: no.
The Counterexamples
Gettier's paper presented two short scenarios. They are simple enough to summarize.
Case 1: Smith and Jones are both applying for a job. Smith has strong evidence for the following proposition: "Jones will get the job, and Jones has ten coins in his pocket." (Smith was told by the company president that Jones would be hired, and he counted the coins himself.) From this, Smith infers a more general proposition: "The man who will get the job has ten coins in his pocket." Smith is justified in believing this. He has good evidence.
But unbeknownst to Smith, he himself (not Jones) is the one who will get the job. And unbeknownst to him, he himself happens to have ten coins in his pocket.
So the proposition "The man who will get the job has ten coins in his pocket" turns out to be true. Smith believes it. Smith is justified in believing it. All three JTB conditions are met. But does Smith know this proposition? Clearly not. He believes the right thing for the wrong reasons. He stumbled onto the truth by accident.
Case 2: Smith has strong evidence that Jones owns a Ford. He has seen Jones drive a Ford for years; Jones offered him a ride in a Ford that very morning. So Smith forms a justified belief: "Jones owns a Ford." From this, he infers a disjunctive proposition: "Either Jones owns a Ford or Brown is in Barcelona." (Smith picks Brown, an acquaintance, and Barcelona at random — he has no idea where Brown is.) Smith is justified in believing this disjunction, because he is justified in believing one of its disjuncts (the Jones one).
But it turns out Jones does not own a Ford. He sold it last week and is now driving a rental. And by sheer coincidence, Brown actually is in Barcelona — Smith had no way to know this.
So the disjunction "Jones owns a Ford or Brown is in Barcelona" is true (because of Brown). Smith believes it. Smith is justified in believing it. All three JTB conditions are met. But does Smith know the disjunction? Again, no. The truth of his belief has nothing to do with the reasons that justify it.
What Gettier Showed
In just two examples, Gettier demonstrated that justified true belief was not sufficient for knowledge. There could be cases where someone believed something true, had good reasons to believe it, and yet failed to know it — because the connection between the reasons and the truth was broken.
What was missing? In both cases, Smith got lucky. His justification did not actually track the fact that made his belief true. The Ford he thought Jones owned was a phantom; the truth came from elsewhere. The job he predicted Jones would get was not the job he actually predicted; the truth came from a different fact about a different person.
Philosophers call this epistemic luck — when a belief happens to be true in a way that has nothing to do with the reasons for holding it. Gettier showed that JTB does not rule out epistemic luck, and so JTB is not enough for knowledge.
The Decades of Response
The Gettier problem produced one of the most active research programs in twentieth-century epistemology. Philosophers tried to fix the JTB definition by adding a fourth condition that would rule out Gettier-style cases. Several main strategies emerged.
No-false-lemmas. Maybe knowledge requires that your justification not depend on any false beliefs along the way. Smith's belief that "the man who will get the job has ten coins in his pocket" depends on the false belief that Jones will get the job. If we add the condition that you cannot have inferred your belief from any false premises, the Gettier cases are excluded.
The trouble: it is easy to construct Gettier-style cases that don't involve any false premises. The strategy patches the original cases but not the underlying problem.
Causal theories. Maybe knowledge requires the right kind of causal connection between the fact and the belief. Alvin Goldman developed this in 1967. The trouble: causation is the wrong kind of thing for many beliefs (mathematical truths, future events, abstract facts). It works for some beliefs but not others.
Reliabilism. Maybe knowledge requires that your belief was produced by a reliable process — one that tends to produce true beliefs. This is a popular contemporary view. It handles most Gettier cases reasonably well. But it has its own difficulties.
Sensitivity and safety. Other philosophers argue that knowledge requires that your belief would have been false if the fact were false (sensitivity), or that you couldn't easily have been wrong (safety). These are subtle modal conditions that try to capture what is missing in Gettier cases.
Knowledge first. Timothy Williamson, in his influential book Knowledge and Its Limits, argued that the whole project of defining knowledge in terms of more basic conditions was misguided. Knowledge, he claimed, is not built up out of belief, justification, and truth. It is its own thing — a basic, irreducible mental state. The Gettier problem persists because we are looking for the wrong kind of analysis.
Why It Matters Beyond Philosophy
The Gettier problem might sound like an academic curiosity — and in one sense, it is. Most everyday uses of the word "knowledge" are not affected by Gettier cases. But the problem has wider implications.
It highlights how easy it is to be right for the wrong reasons. Smith was correct that the man who would get the job had ten coins in his pocket. He just had no real basis for being correct. His correctness was an accident of circumstance.
This pattern shows up in many domains — politics, science, even daily judgment. People can hold true beliefs based on bad reasoning, and the truth of the belief doesn't make the reasoning any better. They can also hold false beliefs based on excellent reasoning, and the falsity of the belief doesn't necessarily indict their methods.
The Gettier problem trains a kind of intellectual humility. The fact that you turned out to be right doesn't mean you knew. The fact that your reasons were good doesn't mean your conclusion is true. Knowledge requires both — and a connection between them that is not just luck.
Edmund Gettier wrote one famous paper in his career. Three pages, two examples, no jargon. It was enough to overturn a definition that had stood for two thousand years.
