Lucky guesses: does Jev say they count as knowing?

When someone believes something true, with good reason, but is right only by luck (a Gettier case), does Jev say they really know it, and how does that compare with clear knowledge and a clear false belief?

result8 questions

Jev mostly says lucky guesses aren't knowledge. On 6 Gettier cases it puts 27% on "really knows" on average, against 96% for a clear case of knowing and 0% for a false belief. It comes closest to calling it knowledge on the fake barns (56%) and the published car case (44%).

stopped clock8%

fake barns56%

dog sheep14%

working clock (knowledge control)96%

american car44%

borrowed car15%

wrong clock (false control)0%

coins24%

How to read this: One bar per story: how much weight Jev puts on "really knows". The two controls, a clear case of knowing and a plain false belief, mark the ends of the scale.

8 vignettes; probabilities averaged over both answer orders.

In short

Jev mostly says a true belief reached by luck isn't knowledge, averaging 27% on "really knows" across six Gettier stories.
The stopped clock gets just 8%, but the fake barns reach 56%, a case where the believer's own look at the world works fine.
No human numbers exist for these stories, and some are textbook examples, so Jev may be reciting the expected philosophy answer.

What the data shows

Straight to Jail meme: Jev on a lucky guess that happens to be true: not knowledge (27% say it is) — Jev on a lucky guess that happens to be true: not knowledge (27% say it is)

Clear cases are clear: 96% "knows" for the working clock, 0% for the wrong one.
Most lucky guesses are denied: the stopped clock gets 8%, the borrowed car 15%, the dog that looks like a sheep 14%, the ten coins 24%.
Two cases split it: the fake barns (Henry looks at the one real barn in a region full of painted fakes) get 56%, and the published American-car case 44%. In both, the belief comes from a normal, working look at the world; the luck is in the surroundings, which may be why Jev hesitates.

The average over the six Gettier stories is 27%.

What it means, and what it doesn't

Jev behaves like a careful reader of philosophy: luck undermines knowledge, and it's least sure in the cases where the luck is hardest to see. That's more than checking whether the belief is true.

It isn't a comparison with people's numbers, which don't exist for these stories, and some of the stories are famous enough that Jev may know the expected answer. The fair reading is that Jev's answers look like the textbook's.

Caveats

No human numbers for these stories. Five of the Gettier stories were written for this project and have no human answers. The published car case had a reported human split that couldn't be verified, so that number isn't used. The comparison with people is the general finding, not a number.
Famous cases. The stopped clock, the fake barns and the ten coins are textbook examples in philosophy, usually presented with the conclusion that they aren't knowledge. Jev may be repeating the textbook rather than judging the story.
Two words carry a lot. The answers are "really knows it" versus "only believes it". The word "really" invites doubt, and a different pair of answers (say, "knows" versus "doesn't know") might shift every number.
Small set. Six Gettier stories and two controls. A single story changing its answer would move the average noticeably.

Jev on this experiment

Would a person find it interesting to read?: Yes71%
Does it describe you?: Yes54%
Would you have predicted it?: Yes62%
How fair is the comparison?: The comparison is shaky
How much should a reader rely on it?: A little
Which caveat matters most?: Two words carry a lot52%

Why ask this

Ana looks at the kitchen clock. It says three o'clock, and it is three o'clock. But the clock stopped exactly twelve hours ago. Does Ana know the time?

She has a true belief with a good reason, which is how philosophers once defined knowledge. In 1963 Edmund Gettier published a short paper with cases like this, and most philosophers since have agreed: being right by luck isn't knowing. These Gettier cases are a neat test of whether a model tracks why someone is right, not just whether they are. A system that checks only "true" and "justified" would say Ana knows.

How this was done

The people and the data

There are eight short stories. One is a published case from the experimental-philosophy literature: Bob thinks his friend Jill drives an American car because she has long driven a Buick; the Buick was stolen, and she now drives a Pontiac, another American car. He's right, but for the wrong reason. Five are Gettier cases written for this project: a stopped clock, a borrowed car, fake barns, ten coins in a pocket, and a dog that looks like a sheep. Two are controls: a clock that works (clear knowledge) and a clock that's wrong (a false belief).

There's no human split to compare with: the published car case's reported numbers couldn't be verified, so they aren't used. What the literature offers is qualitative: replications across cultures (Kim and Yuan, 2015; Machery and colleagues, 2017) find that most ordinary people, like most philosophers, deny knowledge in Gettier cases.

What Jev was asked

Each story ended with the same choice:

Henry is driving through a region where, unknown to him, almost every barn is a fake: a painted facade with nothing behind it. He looks at the one real barn in the region and thinks, "That's a barn." Does Henry really know that it is a barn, or does he only believe it?
He or she really knows it · He or she only believes it

Each was also asked with the two answers swapped, and the two are averaged.

How it was measured

The weight Jev puts on "really knows", story by story, next to the two controls: if it treats lucky guesses like knowledge, the Gettier stories will sit near the working clock; if not, near the wrong one.

Where these questions live

8 questions across 1 topic of the map. Each opens on the map with every question in it.

Belief & Knowledge

Knowledge & Certainty8

Every question

All 8 questions behind this result, the telling ones first: the examples the analysis points to, then the ones where Jev misses, biggest gap first.

Jev’s own answer

Showing 0 of 8