5.2 The Traditional Analysis of Knowledge

What this lecture covers

This lecture works through the traditional analysis of propositional knowledge as justified true belief (JTB), testing each of its three conditions in turn with examples and puzzle cases. It ends by introducing the regress problem facing justification and the contrast between internalist and externalist responses to it.

You'll come away able to state the JTB analysis, explain why "knowledge entails truth" only in one of two possible readings, evaluate whether belief is really necessary for knowledge using cases like blindsight, and distinguish coherentism, foundationalism and externalism as responses to the regress of justification.

Key ideas

• The JTB analysis: to know that P, a subject must have a true belief that P, and be justified in believing it—A.J. Ayer's version substitutes "sure, and has the right to be sure" for belief and justification.
• Scope ambiguity in "knowledge entails truth": "if S knows P, P must be true" is plausible, but "if S knows P, P is necessarily true" is false, since one can know contingent facts like one's own existence.
• Roughly true vs precisely true: apparent counterexamples like "I know France is hexagonal" dissolve once you distinguish the loose and strict senses of "hexagonal."
• Truth is not the same as universal belief: something can be true even if nobody believes it (continental drift before it was discovered), and false even if everyone believes it (geocentrism).
• Testing the belief condition: cases like unwittingly correct quiz guesses and blindsight raise the question of whether knowledge always requires conscious belief, with "unconscious belief" as one possible patch.
• The infinite regress of "knowing that you know": requiring that knowing P entails knowing that you know P leads to an infinite, impossible chain of higher-order knowledge claims, so this requirement must be rejected.
• The regress of justification: justifying a belief by appeal to other beliefs threatens an infinite regress, addressed by either coherentism (beliefs justify each other as a coherent web) or foundationalism (certain basic beliefs are self-justifying).
• Internalism vs externalism: internalism requires all justifying factors to be accessible to the subject's awareness; externalism allows justification to rest on factors like reliable causal connections that the subject cannot introspect, illustrated by a dog's reliable ability to detect a nearby cat by smell.

Walkthrough

The three conditions of knowledge (0:11)

Millican states the traditional JTB analysis: P must be true, S must believe P, and S must be justified in believing it. He notes Ayer's alternative phrasing—being "sure" and having "the right to be sure"—as an improvement, since weak, merely probable belief (such as a 60% confidence about tomorrow's weather) doesn't seem to count as knowledge.

Does knowledge require truth? (1:11)

Millican carefully distinguishes two readings of "if S knows P, P must be true": one claims P is necessarily true (false, since we can know contingent facts like our own existence), the other claims that necessarily, if S knows P, P is true (correct). He resolves apparent counterexamples—like claiming to "know France is hexagonal"—by distinguishing loose and strict senses of the predicate, and separately stresses that truth is not the same as widespread belief, citing continental drift and geocentrism.

Does knowledge require belief? (6:16)

Using the example of a quiz contestant who guesses capital cities correctly despite denying he knows them, and the neurological phenomenon of blindsight (where a person can reliably point to objects they claim not to see), Millican probes whether belief is always necessary for knowledge. He considers "unconscious belief," supported by consistent guessing patterns, as one way to preserve the belief requirement, while noting blindsight makes this harder to sustain since there's no long-term memory pattern involved.

Against requiring self-reflective knowledge (9:19)

Millican dismantles the idea that knowing P requires knowing that you know P, showing this generates an infinite and humanly impossible regress of higher-order knowledge claims, and concludes that knowledge cannot require full conscious self-reflection all the way through.

Justification and the regress problem (11:23)

Turning to the justification condition, Millican explains why true belief alone isn't knowledge—a lucky guess or an unreliable source that happens to be right doesn't count. He then sets out the regress problem: justifying a belief by other beliefs invites the question of how those are justified, threatening an infinite chain. Two traditional responses are presented: coherentism, where interlocking beliefs justify each other, and foundationalism, where some beliefs (Descartes's included) are self-evidently secure without further justification.

Externalism and the case of the dog (14:24)

Millican introduces externalism as a modern alternative that drops the requirement that justifying factors be consciously accessible. He illustrates this with a dog that knows a cat is nearby by smell, without any capacity to intellectually justify that knowledge; what matters instead is a reliable causal link between the smell, the cat's presence, and the dog's detection. He suggests extending this same idea to human knowledge, since we've already seen that knowledge cannot require knowing how you know.

Before you watch

• Watch the previous lecture introducing the topic of knowledge and the methods of conceptual analysis, since this lecture directly applies those methods to the JTB analysis.
• It helps to recall the discussion of scepticism and Descartes's foundationalism from earlier lectures, referenced here when discussing responses to the regress of justification.

Check your understanding

1. What are the three conditions in the traditional justified-true-belief analysis of knowledge?
2. Why is "if S knows P, then P is necessarily true" false, even though "necessarily, if S knows P, then P is true" is plausible?
3. How do the quiz-guessing and blindsight cases challenge the idea that knowledge always requires conscious belief?
4. Why does requiring "if you know P, you must know that you know P" lead to an infinite regress?
5. What is the key difference between coherentism and foundationalism as solutions to the regress of justification, and how does externalism offer a different kind of solution?