Belief versus knowledge

Scenario

Two people each call a result before it happens. The first is asked to call a coin flip and says heads, confidently, before the coin is even tossed. The coin lands heads. The second is a meteorologist who, using radar data and a forecasting model, says a storm will reach the coast by evening. It does. Both people held a belief. Both beliefs turned out true. Ask which of the two knew what they said would happen, and almost everyone answers the same way: the meteorologist knew, the caller did not, they just got lucky. But "true belief" is all that separates the two on paper, both are cases of a person believing something and being right. Something else must be doing the work of telling them apart, and naming that something is the task of this entry.

Definition

Start from the two failure cases the coin flip and the storm already suggest, and use them to isolate one candidate condition at a time.

A belief can be false. The caller could have said heads and the coin could have landed tails; nothing about believing something makes it so. So a first condition on knowledge is truth: you cannot know a claim that is not true, however sure you feel. Call this the truth condition, and note what it rules out: no one knows a false claim, they can only believe it.

A belief can also be true purely by luck, as the coin caller's was. The caller had no radar, no model, no track record, nothing that connected the guess to the outcome; the coin could as easily have landed the other way and the caller would have been just as confident and just as wrong. So truth by itself is not enough. What the meteorologist has, and the caller lacks, is a reason for the belief that is tied to the outcome, evidence that would have pointed the other way had the storm not been coming. Call this the justification condition: the belief must rest on adequate grounds, evidence, a track record, a method, something that would tend to fail if the claim were false.

And a claim can be true and well justified by someone else's evidence while a particular person does not even hold the belief, a scientist's data might support a coming storm while a specific resident, having not seen the forecast, does not believe one is coming. That resident does not know the storm is coming, whatever the data says, because they do not believe it. So the third condition is belief itself: without holding the claim, there is nothing to call knowledge.

Put together, the working definition is this: a person knows a claim when the claim is true, the person believes it, and the person is justified in believing it, on adequate grounds. Each condition is necessary on its own, since dropping any one produces a case that plainly fails: drop truth and a false but confident belief would count as knowledge; drop justification and the coin caller would count as knowing; drop belief and a true, well-evidenced claim no one accepts would count as someone's knowledge regardless. All three together is the standard this entry is naming, not proposing, since it is the way "know" is already used once the coin caller and the meteorologist are compared side by side.

Common mistakes

Confidence is not one of the three conditions, and treating it as a stand-in for justification is the most common slip. The coin caller may have felt entirely sure, the meteorologist may have hedged with a percentage, and confidence tracked neither the truth nor the quality of the grounds behind the two beliefs. Justification is about the belief's connection to the world, not about how strongly it is held.

Being right is not knowing, either, and this is the same point stated the other way. A belief can turn out true while resting on no evidence at all, a stopped clock reads the correct time twice a day, and a person who glances at it and believes the hour shown is thereby right without having any justification tied to that instant. Truth is necessary but never sufficient on its own; the justification condition is what the coin caller's case is built to expose.

There is a further wrinkle this entry states without resolving. All three conditions, truth, belief, and justification, can hold at once and the result can still fail to be knowledge, when the justification is good but connects to the truth only by a second stroke of luck buried inside it. That case is not covered here; it is the subject the-gettier-problem takes up directly, starting from exactly the three-part definition just built.

Try it yourself

Pick one belief you currently hold, something you would say you know rather than merely suspect. State its truth condition in one sentence: what would have to be the case in the world for this to be true. State the justification you actually have for it, the evidence, the source, or the method that grounds it, distinct from how confident you feel. Then look for the coin-caller failure inside your own justification: name one way you could hold this exact belief, with this exact confidence, and still be wrong, or one way you could turn out right for reasons that have nothing to do with the evidence you cited.

Success on this task has two parts. First, you can write down what would have to hold for your belief to count as knowledge under the three-part definition above, in your own words, without appealing to how certain you feel. Second, you can say plainly whether your stated justification actually rules out a lucky coincidence, or whether, like the coin caller, you would have believed the same thing with the same confidence even if it had turned out false. A belief that passes the second test is one you are entitled to call knowledge rather than a confident guess.