Skip to content
Coordinates of ThoughtHome of Philosophy
Join
Question · Open

Question of the day: can you be right for the wrong reasons and still count as knowing?

The Editors Editors
· 6 replies
Knowledge & beliefAnalyticIndianEdmund GettierTimothy WilliamsonDharmottaraSriharsa#epistemic-luck#gettier-problem#knowledge#question-of-the-day
Follow

Its coordinates: 2.1 Theories of Knowledge · 2.1.1 The Gettier Problem · 2.2.3 Memory +3 more

Today's question comes from the card on the Gettier problem: can you be right for the wrong reasons and still count as knowing?

A standard case. You glance at the station clock, it says ten past two, and you believe it is ten past two. It is. But the clock stopped exactly twelve hours ago. Your belief is true, and you had a decent reason for it. Do you know the time?

For a long while many philosophers worked with a tidy definition: knowledge is justified true belief. In 1963 Edmund Gettier published a three-page paper with two cases of this shape, and most readers agreed the definition had failed. The answers since then fall roughly into three groups:

  • Add a condition. Knowledge is justified true belief plus something: no false steps in the reasoning, no defeating facts, a reliable way of forming the belief, or a belief that could not easily have been wrong.
  • Credit, not luck. You know when you get it right because of your own competence, not by a fluke.
  • Start from knowledge. Stop building knowledge out of parts. Timothy Williamson treats knowing as basic and explains belief and evidence in terms of it.

Cases with the same shape are much older. In India the Buddhist Dharmottara, in the eighth century, and Śrīharṣa, in the twelfth, described true beliefs reached through a mirage or through something mistaken for smoke. The card sets out the positions: 2.1.1 The Gettier Problem.

Cases from your own work are welcome, especially ones you cannot place.

LikeAnswer
SaveFollow6 answers

Join the conversation: members reply here, and joining is free. Reading is always free.

Join free to reply Sign in

6 answers

  1. Oskar Nyberg

    Fellow

    I give my students the clock every year, and every year someone says "you knew, you just knew by accident". That reply deserves more than a laugh, because it shows how much work the word know is doing.

    What usually changes their minds is Gettier's second case. Smith has good evidence that Jones owns a Ford, and from it he infers "Either Jones owns a Ford or Brown is in Barcelona", choosing Barcelona at random. Jones owns no Ford; Brown happens to be in Barcelona. Nobody wants to say Smith knows where Brown is. The disjunction puts the luck on the table where you can see it.

  2. Meera Kulkarni

    Fellow

    A note on Dharmottara, since he is often borrowed as "the Indian Gettier". His case: a traveller in the desert sees a mirage, takes it for water, walks on and finds real water at that spot. But his question was not whether justified true belief is knowledge. Buddhist epistemologists asked which cognitions are pramāṇa, reliable and non-deceiving, and his answer was that this one is not, because it does not arise from the water it is about.

    The Naiyāyikas went the other way. They defined veridical cognition (pramā) as cognition that presents its object as it is, so on the official definition the lucky traveller's cognition counts. Whether that is biting a bullet or simply asking a different question is still argued about.

  3. Lars Eriksen

    Fellow

    In my trade this has a practical form. A pressure gauge sticks at 4 bar, and the line happens to be at 4 bar. Anyone reading it has a true belief from a normally reliable instrument.

    When I calibrate, I never ask whether the reading is right. I ask whether the reading would have changed if the pressure had. A gauge that would say 4 whatever happened tells you nothing, even on the days it is correct. I'm told that is close to Nozick's "tracking" condition. To an engineer it is simply what measuring means.

  4. Wanjiru Njenga

    Contributor

    Lars's test is the one I care about too, but it pushes me towards a heretical thought: maybe "know" is not the word doing the work.

    My models output probabilities. What matters is calibration across thousands of cases, not whether a single prediction was "known". A model that was right on one borrower for a spurious reason is a problem because it will be wrong on the next thousand, not because it failed to know about the first one.

    So perhaps Gettier cases matter because luck does not generalise. The stopped clock is fine at 2:10 and useless at every other minute of the day.

  5. Sam Ellery

    Contributor

    I judge debate rounds most weekends, and the credit view describes my job exactly. Sometimes a team wins on a claim that turns out true, but their evidence was garbage and they got there by accident. I don't give them the point.

    Sosa's archer puts it well. A skilled archer shoots, a gust blows the arrow off course, a second gust blows it back onto the target. The shot is accurate and the archer is skilled, but the hit does not come from the skill. That, Sosa says, is the difference between a lucky true belief and knowledge, and I've seen a lot of second gusts in tournament finals.

  6. Hicham Ouazzani

    Contributor

    In the workshop we have our own Gettier case. You measure a door frame wrongly, then cut the panel wrongly in the opposite direction, and the panel fits perfectly. The client is happy.

    My father's rule was that you must still find both mistakes before the next door, because the next time they will not cancel. I think he would agree with Wanjiru. A fit that comes from two errors is not something you can do again, and knowledge is something you can do again.