What is epistemology? A beginner’s guide to the theory of knowledge

When people in my reading group ask what epistemology is, I give them a short answer and then a lucky guess. The short answer: epistemology is the part of philosophy that studies knowledge. It asks what knowledge is, where it comes from, how our beliefs are supported, and how far our knowing can reach. The name is built from the Greek epistēmē, knowledge, and logos, account or study, though the English word itself is a nineteenth-century coinage. The questions are much older, and they were asked in more than one place.

The lucky guess comes next, because it shows why the subject exists at all.

Question one: what is knowledge?

Suppose you glance at the clock on the station wall. It says twenty past two, so you believe it is twenty past two, and you are right. Unknown to you, the clock stopped yesterday at exactly twenty past two. Do you know the time? Most people say no. You had a true belief, and a reasonable one, since station clocks usually work. But you were right by luck.

That small case (versions of it are usually traced to Bertrand Russell) contains the central problem. Knowledge is not just believing something true, because a wild guess can be true. Plato saw this in the Theaetetus, a dialogue in which Socrates and a gifted young mathematician test three definitions: knowledge is perception, knowledge is true judgement, knowledge is true judgement with an account. Each fails, and the dialogue ends without an answer. Beginners are often annoyed by that. I think it is one of the most honest endings in philosophy.

For a long time the third definition, tidied up as "justified true belief", was treated as roughly right. In 1963 Edmund Gettier published a paper of about three pages describing cases in which someone has a justified, true belief that is still not knowledge, because the justification and the truth are connected only by accident. The stopped clock has exactly that shape. Since then epistemologists have tried to name the missing ingredient. Reliabilists, following Alvin Goldman, say a belief must come from a process that reliably produces truths. Safety theorists say the belief must be one that could not easily have been false. Timothy Williamson and others have turned the project round: perhaps knowledge cannot be analysed into parts at all, and should be treated as basic.

Question two: where does knowledge come from?

The second question is about sources. In the European tradition this became an argument between empiricists, who held that all our ideas come from experience (for John Locke the mind begins empty), and rationalists, who held that reason can know some substantive truths on its own. Kant tried to settle the quarrel by arguing that experience supplies the content and the mind supplies the structure, and that neither works alone.

Classical Indian philosophy asked a related but different question, and it is worth learning early because it changes how you see the first. Indian epistemologists spoke of pramāṇas, the means or instruments of valid cognition, and argued about how many there are and which can be reduced to others. The Cārvāka materialists accepted only perception. Buddhist logicians such as Dignāga accepted perception and inference, and treated testimony as a kind of inference. The Nyāya school, whose founding sūtras are attributed to Gautama (Akṣapāda), counted four: perception, inference, comparison and testimony, each irreducible. The Mīmāṃsakas counted five or six.

A common beginner's mistake is to translate pramāṇa as "knowledge". It names the instrument, not the result. The European debate often asks which faculty does the real work; the Indian debate counts the instruments and argues about whether one can be derived from another. Those are different questions, and both are good ones.

Take testimony. You believe that Antarctica exists, though you have never been. Is your belief really an inference ("reliable sources say so, and reliable sources are usually right")? Or is being told something by a competent speaker a source of knowledge in its own right? A child believes her grandmother when told the river is too deep to cross. She has no theory of her grandmother's reliability, yet it seems odd to say she does not know the river is deep. Dignāga's answer and the Nyāya answer still map the two sides of this argument today.

Question three: how are beliefs supported?

Even justified beliefs rest on other beliefs. What do those rest on? An ancient puzzle, associated with the Pyrrhonist Agrippa, says the chain of reasons must either go on for ever, loop back on itself, or stop somewhere arbitrary. Foundationalists accept the third option but deny that it is arbitrary: some beliefs, such as beliefs about how things appear to you, need no further support. Coherentists say beliefs are justified by how well they hang together, like strands of a web. A further dispute, between internalists and externalists, asks whether you must be able to tell from the inside what makes your belief justified, or whether it is enough that it was formed in the right way.

Question four: how far can knowledge reach?

Finally there is scepticism: the worry that we know much less than we think, perhaps nothing at all. Every tradition I read has a version. Greek Pyrrhonists suspended judgement; al-Ghazali described a real crisis of doubt in his autobiography; Descartes used doubt as a method to reach certainty; Nāgārjuna refuted others' positions while declining to hold one himself. Most contemporary epistemologists are fallibilists: they think knowledge is real but never comes with a guarantee.

Why the comparison matters

Here is the argument I make to newcomers. The habit of asking "what is knowledge?" and the habit of asking "what produced this cognition, and was it a sound instrument?" pull in different directions, and a beginner needs both. The first teaches you to notice luck. The second teaches you to ask, of any belief, where it came from.

Indian writers also discussed luckily true cognitions long before 1963. The eighth-century Buddhist Dharmottara described a traveller who sees a mirage, believes there is water ahead, and happens to be right because water lies hidden nearby. Some specialists doubt that such cases are strictly parallel to Gettier's, since on Indian accounts a genuine pramāṇa guarantees the truth of what it delivers. That debate is worth having. The point is not who got there first. It is that two traditions, asking differently, kept running into the same troubling fact: being right is not the same as knowing.

That is also why epistemology is not a dry technical subject. The person who forwards a rumour that happens to be true, the jury weighing a confident witness, the doctor trusting a test result: all of them face Gettier's problem and Dignāga's question in practice.

If you want to begin, read the Theaetetus in a modern translation, slowly, and stop to argue with it. Bertrand Russell's The Problems of Philosophy (1912) is still a clear first book on knowledge and its sources. Then find a good introduction to the pramāṇa debate, and notice the moments when its questions do not fit the ones you started with. Those moments are where most of the learning happens.