What is epistemology? How we know what we know

Epistemology is the branch of philosophy that studies knowledge and justified belief: what it means to know something rather than merely believe it, where knowledge comes from, and whether we can know anything at all. The word joins the Greek episteme, knowledge, with logos, study. Its questions sound abstract, but they are the ones behind every argument about evidence, from a courtroom to a news feed.

In short

• Epistemology asks three big questions: what knowledge is, where it comes from, and how far it reaches.
• The classic answer to the first, that knowledge is justified true belief, fails on cases where someone is right by luck, and no fix has won agreement.
• On sources, rationalists stress reason and empiricists stress experience; skeptics ask whether either can rule out radical error.

What questions does epistemology ask?

Its questions fall into three groups, according to the Stanford Encyclopedia of Philosophy's overview.

• What is knowledge? What separates knowing that it's raining from guessing it, or believing it on bad grounds?
• Where does knowledge come from? Perception, memory, reasoning, the testimony of other people: which of these give knowledge, and how?
• How far does it reach? Can we know the future, other minds, the external world, or right and wrong? Or is some of that out of reach?

A related question runs through all three: what makes a belief justified, that is, reasonable to hold given what you have to go on? You can be justified and wrong (a careful forecast that fails) or right and unjustified (a lucky guess). Epistemology studies the gap between those. It is one of the main branches of philosophy, next to metaphysics, ethics and logic.

Is knowledge just justified true belief?

For a long time, many philosophers thought so. On this view, which goes back to Plato's dialogue Theaetetus, you know something when three conditions hold: you believe it, it is true, and you have good reason to believe it. Each condition does work. You can't know what you don't believe; you can't know something false, only think you do; and a lucky guess isn't knowledge.

In 1963 Edmund Gettier published a short paper with two cases in which all three conditions held and yet the person plainly didn't know. One of them, about a job and ten coins, is walked through step by step in the note on what philosophy is. An older case makes the same point, and it is easier to remember. Bertrand Russell described it in 1948, as the Stanford Encyclopedia's entry on the analysis of knowledge recounts.

Worked example: the stopped clock

Alice walks into a room and glances at the clock on the wall. It says 2:00. The clock has always been reliable, so she believes it is 2:00. In fact the clock stopped exactly twelve hours earlier, and by chance it really is 2:00. Test her belief against each condition.

1. Belief. Alice believes it is 2:00. Condition met.
2. Truth. It is 2:00. Condition met.
3. Justification. She read a clock that has always worked, which is how anyone tells the time. Condition met.
4. Verdict. She has a justified true belief. Does she know it is 2:00? Almost everyone says no. If she had walked in a minute later, she would have believed the same thing and been wrong.

What the case shows is that being right, for good reasons, can still be luck. The reasons Alice had weren't connected to the truth in the right way. Philosophers call this kind of case a Gettier case, and the search for what's missing is called the Gettier problem.

How have philosophers tried to fix the definition?

Each proposed fix adds something to justified true belief, or replaces justification with something else. Here is how four of the best known handle Alice.

Proposal	The extra condition	What it says about Alice
No false lemmas	Your belief mustn't rest on any false step	She relied on the false belief that the clock was working, so she doesn't know
Reliabilism (Alvin Goldman)	Your belief must come from a process that reliably produces truths	Reading a stopped clock is not a reliable process, so she doesn't know
Safety (Ernest Sosa and others)	You couldn't easily have been wrong	A minute later she would have been wrong, so she doesn't know
Knowledge first (Timothy Williamson)	Knowledge can't be broken into simpler parts; it is the basic idea	The case shows why every analysis fails, not which part is missing

Each proposal handles the clock, and each meets its own counterexamples. In the fake barn case, given by Carl Ginet and published by Goldman in 1976, Henry drives through a county full of barn façades and by chance looks at the one real barn. He relied on no false step, so the first fix says he knows, though most people say he doesn't. Decades on, there is no agreed definition, and some philosophers take that as Williamson does: knowledge may be one of the ideas we use to explain others.

Where does knowledge come from: reason or experience?

This is the oldest dispute in the field. Rationalists, such as René Descartes, Baruch Spinoza and Gottfried Leibniz in the 17th century, held that some important knowledge comes from reason alone, independently of the senses. Empiricists, such as John Locke and David Hume, held that experience is the only source of our ideas and of knowledge about the world, according to the Stanford Encyclopedia's entry.

Position	Its claim	An everyday example
Rationalism	Some knowledge comes from reason alone	You know that if Ana is taller than Ben and Ben is taller than Cy, Ana is taller than Cy, without measuring anyone
Empiricism	Knowledge of the world comes from experience	You know the milk is sour because you tasted it
Skepticism	We know much less than we think, perhaps nothing about the outside world	You can't rule out that you're dreaming right now
Foundationalism	Justification rests on basic beliefs that need no further support	"I seem to see something red" needs no argument behind it
Coherentism	Beliefs are justified by fitting together, like a web	A detective's theory is believed because every clue fits it

Hume pressed empiricism hard. All our reasoning about the future, he argued, assumes that nature will go on as it has, and that assumption can't be proved by reason or by experience without arguing in a circle. This is the problem of induction, and it bears directly on inductive reasoning and on how science learns from evidence. Immanuel Kant later tried to combine the two camps, arguing that experience gives the content of knowledge while the mind supplies its structure.

Can we know anything at all?

The skeptic says perhaps not. Descartes, in 1641, imagined an evil demon able to deceive him about everything his senses report. A modern version is the brain in a vat: a brain kept alive and fed signals by a computer would have exactly your experiences. The skeptic's argument, as the Stanford Encyclopedia sets it out, runs: if you know you have hands, you know you aren't a handless brain in a vat. You can't know you aren't one. So you don't know you have hands.

Replies differ. G. E. Moore, in a 1939 lecture, held up his hands and said he was more certain he had them than he could ever be of the skeptic's premises, so he would reject a premise instead. Contextualists say "know" sets a higher bar in a philosophy seminar than in daily life, so the skeptic is right in one setting and wrong in the other. Some, following Descartes himself, look for beliefs that can't be doubted; his was "I think, therefore I am". The doubt itself is old: Plato's allegory of the cave imagines prisoners who take shadows for the whole of reality.

Why does epistemology matter outside philosophy?

Its questions come back whenever people decide whom and what to trust. How much should you believe a source you can't check? Most of what anyone knows comes from testimony, other people's word, and epistemologists study when that transmits knowledge and when it doesn't. When does a run of confirming examples justify a general rule? How should a belief change when new evidence arrives? Those are the questions that critical thinking puts into practice, and that formal tools such as Bayes' theorem try to make exact.

Check yourself

• Without looking back: walk through the stopped clock case and say which condition of justified true belief it shows isn't enough.
• What is the difference between what a rationalist and an empiricist would say about how you know 7 + 5 = 12?
• State the skeptic's brain-in-a-vat argument in three steps, and give one reply to it.

Sources

• Epistemology — Stanford Encyclopedia of Philosophy
• The analysis of knowledge — Stanford Encyclopedia of Philosophy
• Rationalism vs. empiricism — Stanford Encyclopedia of Philosophy
• Skepticism — Stanford Encyclopedia of Philosophy
• Epistemology — Internet Encyclopedia of Philosophy