A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Epistemic luck is the phenomenon by which a belief turns out true in a way that is too coincidental to count as knowledge. Gettier cases are one species of epistemic luck, but the concept is broader: a stopped clock that is right twice a day yields a true belief by luck; a clairvoyant who accidentally guesses correctly has a true belief by luck. Two modal conditions have been proposed to exclude epistemic luck: sensitivity (if p were false, S would not believe p) and safety (S could not easily have had a false belief formed in the same way). Safety is generally preferred because sensitivity yields counterintuitive results for known necessary truths.
Evaluate safety and sensitivity against a range of cases: lottery beliefs, knowledge of necessary truths, standard perceptual beliefs. Check whether each condition correctly classifies the cases as knowledge or not-knowledge.
From your study of justified true belief and the Gettier problems, you know that a belief can be justified and true yet still fail to count as knowledge. Gettier showed this with cases where the justification and the truth come apart — the belief is true, but not for the reason the justification provides. Epistemic luck is the broader phenomenon that Gettier cases exemplify: a belief turns out true in a way that is too coincidental, too disconnected from the believer's epistemic position, to constitute genuine knowledge.
Consider a stopped clock that happens to display 3:15 at the exact moment you glance at it — and it is, in fact, 3:15. You have a justified belief (you checked a clock, which is normally a reliable method), and the belief is true (it really is 3:15). Yet something has clearly gone wrong: the clock is broken, and you would have formed the same belief at any other time of day. Your being right is a matter of luck, not epistemic competence. The concept of epistemic luck captures this intuition and extends it beyond Gettier's original cases to any situation where truth and justification coincide accidentally.
Two modal conditions have been proposed to formalize what epistemic luck is missing. The sensitivity condition says: S's belief that p is sensitive if, were p false, S would not believe p. The stopped-clock belief fails sensitivity — if it were not 3:15, you would still believe it is 3:15, because the clock always says 3:15. The safety condition says: S's belief that p is safe if S could not easily have formed a false belief in the same way. The stopped-clock belief fails safety too — in the vast majority of nearby possible situations where you check the same clock, the belief is false. Safety is generally preferred over sensitivity because sensitivity generates counterintuitive results for necessary truths: "if 2+2 were not 4, would you still believe it?" is a meaningless counterfactual, since there is no possible world where 2+2 is not 4, yet we clearly have knowledge of arithmetic.
The deeper lesson is that knowledge requires more than accidentally getting things right. Justification ensures your belief is responsibly formed; truth ensures the world cooperates; but epistemic luck shows that these two conditions can be satisfied independently, with no real connection between them. The anti-luck conditions — safety and sensitivity — attempt to bridge this gap by requiring that the belief-forming method be modally robust: it must track the truth not just in the actual world but across nearby possible situations. This connects epistemic luck to reliabilism and virtue epistemology, both of which seek to explain what makes knowledge more than a lucky coincidence between belief and reality.
Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.