Justified True Belief

Start with the question in its most ordinary form. What is the difference between knowing something and merely being right about it?

The answer philosophy settled on is that knowledge has three parts. You believe it. It is true. And you are justified in believing it.

Each condition is doing work, and the quickest way to see that is to remove them one at a time.

Remove truth and you have a justified false belief. This is extremely common and not a failure of anybody's reasoning: for centuries excellent people held excellent reasons for believing that the sun went round the earth. They were wrong, and they were not being stupid. We do not say they knew it.

Remove justification and you have a lucky guess. Somebody picks a number, they happen to be right, and nobody thinks they knew. Being right by accident is a fact about the world rather than an achievement of yours.

Remove belief and the case is stranger and more interesting. Consider somebody who has all the evidence, could recite it, and still cannot bring themselves to accept the conclusion. A person who has been told their diagnosis and has not taken it in. We do not say they know, and that tells you knowledge has something to do with what you actually hold rather than merely what you have been shown.

So: belief, truth, justification. Three conditions, each necessary, and together, for a very long time, they looked sufficient.

Now the observation that makes this chapter worth reading rather than memorising.

Only one of the three is up to you.

Truth is decided by the world and you have no vote. Belief mostly happens to you; you cannot decide to believe it is raining, you can only look out of the window. Justification is the only part you can work on, improve, demand of others, or get wrong through carelessness.

Which means that the whole practical content of the theory of knowledge sits in the third condition. How good do your reasons have to be? What makes a reason good? Can a reason be good and still lead you somewhere false, and if so, what exactly is it doing?

Every disagreement in this book turns out to be a disagreement about that condition. And it is worth knowing before we go any further that the three-part definition, which is about to be broken quite badly in the next chapter, did not stop being useful when it stopped being a definition.

It remains the fastest diagnostic anybody has. When an argument has gone wrong, ask which of the three is missing. Is the disagreement about what is actually the case, or about whether the reasons are any good, or about what somebody genuinely holds as opposed to what they are saying?

Those are three completely different conversations, and most unproductive arguments are two people having two of them at once.