The Tripartite View of Knowledge
The tripartite view is not a convincing definition of knowledge — its conditions are individually necessary but not jointly sufficient, and the failure of NFL and reliabilism to patch the Gettier problem reveals that truth condition analysis is structurally doomed. Virtue epistemology offers a more convincing account by making truth flow from intellectual virtue rather than treating it as an independent condition
- The tripartite view of knowledge claims that S knows P if and only if: S believes P, S has justification for P, and P is true — these are proposed as individually necessary and jointly sufficient conditions - A strength of the theory is how intuitively its criteria account for what Zagzebski called knowledge's 'cognitive contact with reality' — truth ensures the belief is not false, belief provides the cognitive element, and justification explains how the belief relates to its truth in a non-accidental way. A lucky guess, even if true, does not feel like knowledge, and justification seems to rule this out - I will consider three criticisms: that the conditions are not individually necessary, that they are not jointly sufficient (Gettier), and that truth condition analysis itself is the wrong approach to defining knowledge. The first is weak. The second motivates two attempted patches — no false lemmas and reliabilism — both of which ultimately fail. This failure motivates the third and most crucial criticism: that truth condition analysis is structurally doomed. I will argue that virtue epistemology offers a more convincing account of knowledge than any truth-conditional theory
Section 1: The Conditions Are Not Individually Necessary
- This objection is weak. On belief — AI outputs are not knowledge because knowledge requires a cognitive state and AI has no mind. The self-deception case is interesting but we might say the person has unconscious belief, which still satisfies the condition. On truth — we would not say people in the past knew the sun orbited the earth; we would say they believed it and were wrong. Knowledge differs from mere belief precisely because only the former is necessarily true. On justification — the quiz case is an exception, and definitions should not be built on exceptions - The justification condition is the weakest of the three and will come under increasing pressure as the essay progresses — this foreshadows the deeper problems ahead
: The individual necessity objection
- The tripartite view claims each of the three conditions is individually necessary — remove any one and knowledge disappears - This can be challenged for each condition: - Against belief: AI can output 'Paris is the capital of France' without having a mind or beliefs — is this not knowledge without belief? Or consider self-deception: someone who deep down knows they are in love but consciously denies it seems to have knowledge without explicit belief - Against truth: people in the past held the justified belief that the sun orbits the earth, and everybody considered this knowledge — did they not know this, even at the time? - Against justification: in a general knowledge quiz, we say someone knows the capital of a country even if they cannot justify their answer — the true belief alone seems sufficient
Section 2: The Conditions Are Not Jointly Sufficient: Gettier, NFL and Reliabilism
- The Gettier cases have genuine force. The intuition that Smith does not know is near-universal. Attempts to criticise the specific cases — for instance arguing Smith is not genuinely justified in Case 2 since he constructed a disjunction without evidence for the second disjunct — are interesting but do not defuse the general problem. Even if the specific cases have flaws, the underlying point stands: justified true belief can be accidentally true in a way that excludes knowledge - NFL handles the original Gettier cases cleanly but the barn facade case is a genuine counterexample it cannot answer. The suggestion that Henry inferred his belief from a general false belief that he was in a county of real barns is not convincing — we can stipulate that Henry was absentmindedly perceiving each barn without making any such inference. NFL therefore cannot cover all logically possible Gettier-style cases - Reliabilism is weaker than NFL within truth condition analysis because it cannot even handle the original Gettier cases — the most clear-cut instances of non-knowledge. NFL at least gets those right. Goldman's sensitivity amendment helps with barn facade but the increasing complexity of the amendments suggests something is fundamentally wrong with the approach rather than just the specific conditions chosen
Gettier: The Gettier argument
- Gettier argues that justification, truth and belief are not jointly sufficient for knowledge — it is possible to have all three without having knowledge - Case 1: Smith has strong evidence that Jones will get a job, having been told so by the company president. He has also counted ten coins in Jones's pocket. He infers: 'the man who will get the job has ten coins in his pocket.' In fact Smith gets the job, and happens to have ten coins in his own pocket. Smith has a justified true belief — but clearly does not know, since his belief is true in virtue of facts he was entirely unaware of. His getting it right was pure luck - Gettier's basic point: if knowledge were justified true belief, you could gain knowledge by luck — but knowledge is not a matter of luck
: Response 1 — No False Lemmas
- A false lemma is a false belief from which a conclusion is inferred - In the Gettier case, Smith infers his conclusion from the false belief that Jones will get the job — this is a false lemma - The no false lemmas theory adds a fourth condition: S knows P only if S did not infer P from any false belief - This handles the original Gettier case: Smith's belief is disqualified as knowledge because it was inferred from a false lemma - However the barn facade case shows NFL is insufficient. Henry is driving through a county full of convincing fake barns. On the one occasion he looks at the single real barn, he forms the true belief 'there's a barn' — and this belief was not inferred from any false lemma, it came directly from perception. Yet Henry clearly does not know — it was pure luck that he happened to look at the real one rather than a fake
: Response 2 — Reliabilism
- Reliabilism replaces the justification condition with a causal one: S knows P if and only if S's true belief that P was produced by a reliable cognitive process — one that generates a high percentage of true beliefs, such as memory or perception - The motivation is that the justification condition keeps creating problems because a person's reason for a belief and the actual cause of the belief's truth can come apart — and that gap is where luck enters. By requiring the truth to be caused by a reliable process, reliabilism aims to close that gap - However reliabilism fails even on the original Gettier case. Smith's belief that Jones will get the job came from testimony — a generally reliable process. His belief that Jones has ten coins in his pocket came from perception — also reliable. Reliabilism would therefore classify Smith's belief as knowledge. But it clearly is not - Goldman attempts to rescue reliabilism by adding a sensitivity condition — S must be able to distinguish between relevant alternatives. Henry cannot distinguish the real barn from the fakes, so his belief is disqualified. But this amendment faces further counterexamples and begins to look ad hoc — a patchwork of additions rather than a principled theory
Section 3: Truth Condition Analysis Is Structurally Doomed: Zagzebski and Virtue Epistemology
- Virtue epistemology is not without problems. It is somewhat vague — what exactly counts as an intellectual virtue? And the barn facade case requires careful handling — Henry seems to be exercising perfectly attentive perception. But Zagzebski can respond that genuine intellectual virtue includes sensitivity to one's epistemic environment: a truly virtuous inquirer in an unfamiliar environment would have noticed something was off and been more cautious. These are problems that can be worked on within the framework rather than structural problems that doom the whole approach - The key strength of virtue epistemology over all truth-conditional theories is that it does not patch the Gettier problem — it dissolves it. By making truth flow from virtuous inquiry rather than treating it as an independent condition, the accidental connection Gettier exploits can never arise
Zagzebski: Zagzebski's structural argument
- Zagzebski argues that the failure of NFL and reliabilism is not accidental — it is the inevitable result of how truth condition analysis works - The problem is that truth condition analysis conjoins independent conditions. Because the conditions are independent of each other, you can always construct a case where they are all satisfied but the connection between them is accidental — and that accidental connection is exactly what Gettier exploits - This is not a problem with any particular set of conditions — it is a structural problem with the method itself. NFL and reliabilism failed not because they chose the wrong conditions but because they followed the wrong approach. This explains why 60 years of attempts to patch the tripartite view have produced no consensus — philosophers have been trying to solve a problem that cannot be solved from within truth condition analysis
: Virtue epistemology as the alternative
- Zagzebski's positive proposal is virtue epistemology: rather than asking what conditions must be satisfied for a belief to count as knowledge, we should ask what kind of person a knower is - Knowledge is true belief that flows from the exercise of intellectual virtue — careful reasoning, intellectual honesty, open-mindedness, perceptual attentiveness - The crucial move is that truth is no longer an independent condition bolted alongside belief and justification. Instead the truth flows from the virtuous act of inquiry — it is produced by it, not merely correlated with it. This breaks the Gettier structure: in Gettier cases the truth does not flow from anything virtuous Smith did — his getting it right was purely accidental, entirely disconnected from his inquiry - This also explains why Gettier cases feel wrong to us — not just that they do. Smith's failure is not that he could have been wrong, but that his being right had nothing to do with him as a thinker - An important feature of virtue epistemology is that intellectual virtue includes fallibilism — the genuinely virtuous inquirer holds beliefs with appropriate tentativeness, remaining open to revision. This means when a belief later turns out to be wrong, virtue epistemology can say the inquirer had warranted belief rather than knowledge — not because knowledge requires certainty as infallibilism claims, but because genuine intellectual virtue involves knowing the limits of one's own epistemic position. This is a more sophisticated and defensible position than infallibilism
: Comparison with Craig
- Craig also rejects truth condition analysis but for different reasons. He argues knowledge is a social concept whose function is to identify reliable informants — when we say someone knows something we flag them as a trustworthy source. Knowledge is defined by its pragmatic role in a community, not by intrinsic conditions - Craig diagnoses the disease brilliantly — he explains why truth condition analysis has always failed. But he does not tell us what knowledge is, only what we do with the word. That is a step sideways rather than forwards - Furthermore Craig's view struggles to explain why communities can be simply wrong — entire communities once endorsed flat-earth beliefs. Craig has no resources to say those communities were mistaken rather than that flat-earth belief genuinely was knowledge at the time. Virtue epistemology has no such problem — a virtuous inquirer tracks truth, not community consensus
- The conditions of the tripartite view are individually necessary: the objections in Point 1 do not produce convincing counterexamples, and the truth and belief conditions in particular seem indispensable - However the conditions are not jointly sufficient: Gettier's cases demonstrate this with near-universal intuitive force, and two serious attempts to patch the theory — NFL and reliabilism — both ultimately fail, with reliabilism failing even on the original Gettier cases - More fundamentally, Zagzebski shows this failure was inevitable: any theory conjoining independent conditions will always be Gettier-vulnerable. This is a structural problem with truth condition analysis itself - Craig offers a pragmatic alternative but only tells us what knowledge is for, not what it is — and cannot explain why communities can be simply wrong - Virtue epistemology is the most convincing account: by making truth flow from intellectual virtue rather than treating it as an independent condition, it dissolves the Gettier problem and explains why Gettier cases feel wrong in a way no truth-conditional theory can. Its inclusion of fallibilism as an intellectual virtue also handles the problem of beliefs that later turn out to be wrong without retreating to infallibilism - The tripartite view is therefore not a convincing definition of knowledge