In three pages in 1963, Edmund Gettier broke the classical definition of knowledge. The paper showed that a belief can satisfy all three JTB conditions — justified, true, belief — and still not be knowledge. Sixty years of epistemology has not produced consensus on the fix. For PKM, the problem is a reminder that even impeccably-sourced true claims can be lucky accidents.
The Original Cases
Gettier's two cases share a structure: a justified belief whose justification goes through a false premise, but which happens to be true for unrelated reasons.
Case 1 — The Job and the Coins. Smith is justified in believing "Jones will get the job" (the boss said so). Smith observes Jones has ten coins in his pocket. Smith infers "the person who gets the job has ten coins in their pocket." In fact, Smith gets the job, and Smith happens to have ten coins in his pocket (which he didn't know). The inferred belief is true and justified, but it is true for a completely different reason than the justification suggests.
Case 2 — Ford and Barcelona. Smith has strong evidence that Jones owns a Ford. Smith infers "either Jones owns a Ford or Brown is in Barcelona" (a disjunction is true if either disjunct is true). In fact, Jones does not own a Ford, but Brown is by coincidence in Barcelona. The disjunction is true, the belief is justified, but the truth comes from the wrong side.
Both are cases of justified true belief. Neither feels like knowledge.
The General Pattern
The Gettier pattern is: a belief is justified (has adequate evidential support), true (accidentally, for reasons the justification does not track), and held (the knower believes it) — yet the connection between justification and truth is a lucky coincidence, not a reliable tracking relationship. Knowledge, whatever it is, seems to require that the truth and the justification be connected through something non-accidental.
Attempted Fixes
Philosophers have proposed several supplements to the JTB triad.
- No-false-premises condition: knowledge requires that no essential step in the justification be false. This handles Gettier's original cases but fails for modified cases that don't involve explicit false premises.
- Reliabilism: knowledge requires that the belief be produced by a reliable process. A process that occasionally produces true beliefs by coincidence is not reliable.
- Tracking theories (Nozick): a knower's belief must track the truth — they would not believe it if it were false, and would believe it if it were true.
- Safety conditions: in nearby possible worlds, the belief remains true. Lucky guesses fail this because in slightly different circumstances they would be wrong.
- Virtue epistemology: knowledge is true belief produced by the exercise of intellectual virtues (careful reasoning, accurate perception, etc.).
None has achieved consensus. The problem may not have a tidy analytic solution.
Why This Matters for PKM
Gettier cases happen in real knowledge work. A vault is full of claims justified by sources that happen to be true for reasons the source doesn't capture. A statistic from a blog post that the blogger miscited — but the statistic happens to be correct from the original study. A claim supported by a chain of reasoning with a subtle error — but the conclusion is right anyway. A biographical fact about a practitioner pulled from one profile that turns out to be wrong about the year but right about the event.
Each is a Gettier case in miniature: the vault holds a justified true belief, but the justification does not track the truth. Treating this as knowledge — especially when linking it to other notes and building on it — risks compounding hidden structural fragility even as surface accuracy holds.
Practical Hedges
A PKM-aware response does not try to eliminate Gettier cases (that is not possible) but to reduce their downstream damage.
- Track source chains, not just sources: if a blog cites a study, record both. When something later contradicts the claim, you can walk back the chain rather than re-deriving it.
- Prefer primary sources for load-bearing claims: the further from the primary source, the more opportunity for Gettier drift.
- Build redundant justification for high-stakes claims: two independent justifications for the same claim are harder to both be Gettier-lucky in the same direction.
- Date claims and flag re-verification: some Gettier cases become visible only when new information arrives; scheduled review catches them.
The Deeper Lesson
The Gettier problem is a permanent reminder that certainty is not available. Even the most carefully sourced claim may be accidentally right. This is not a reason to abandon justification — it is a reason to hold even justified beliefs with appropriate tentativeness, which is exactly the fallibilist stance the best PKM practice already endorses.
Key Points
- Gettier showed JTB can be satisfied without knowledge — a belief can be justified, true, and held, yet only accidentally correct
- The pattern: justification goes through a false premise or disconnected path, the conclusion happens to be true anyway
- Proposed fixes (no-false-premises, reliabilism, tracking, safety, virtue epistemology) all help but none achieves full consensus
- PKM regularly encounters Gettier-like situations — cited claims that happen to be right despite upstream errors
- The response is defensive: track source chains, prefer primary sources, build redundant justification, schedule re-verification
- The deeper lesson is fallibilism: even impeccable-looking justification may be lucky
Open Questions
- Can source-tracking automation surface Gettier risk by flagging claims whose justification is thin or indirect?
- Does the Gettier problem have specifically worse forms in LLM-mediated PKM (where model output may be right for reasons the model cannot verify)?
References
- Gettier, E. (1963). "Is Justified True Belief Knowledge?" Analysis
- Stanford Encyclopedia of Philosophy — "The Analysis of Knowledge"
- Nozick, R. (1981). Philosophical Explanations — tracking theory
- Sosa, E. — virtue epistemology framework