For any person, there are some things they know, and some things theydon’t. What exactly is the difference? What does it take to knowsomething? It’s not enough just to believe it—wedon’t know the things we’re wrong about. Knowledge seemsto be more like a way of getting at the truth. The analysis ofknowledge concerns the attempt to articulate in what exactly this kindof “getting at the truth” consists.

More particularly, the project of analysing knowledge is to stateconditions that are individually necessary and jointly sufficient forpropositional knowledge, thoroughly answering the question:“What does it take to know something?” By“propositional knowledge”, we mean knowingthatsomething is the case, i.e., knowledge of a proposition. Forexample, if Susan knows that Alyssa is a musician, she has knowledgeof the proposition that Alyssa is a musician. Propositional knowledgeshould be distinguished from knowledge by “acquaintance”,for example when we say that Susan knowsAlyssa, or that sheknowsVancouver. (Knowing Susan may involve knowing somepropositions about Susan, but it is only the latter kind of knowledgethat is our focus here.)

“Knows that” is the canonical English expression forindicating propositional knowledge. We also often attributepropositional knowledge with so-called “knows-wh”locutions: Sooyung might know whether the shop is open, for instance,or Juan might know who won the election. These cases also involvepropositional knowledge: one knows these things by knowing theproposition that is the answer to the embedded question. Sooyung canknow whether the shop is open by knowing that the shop is not open.There is some controversy over the particular status of knowledge-how(i.e., “Ahmed knows how to ride a bicycle”)—seeStanley 2011 and his opponents discussed therein.

The propositional knowledge that is the analysandum of the analysis ofknowledge literature is paradigmatically expressed in English bysentences of the form “S knows thatP”,where “S” refers to the knowing subject, and“P” expresses the proposition that is known. Aproposed analysis consists of a statement of the following form:“S knows thatP if and only ifJ,” whereJ indicatesthatS satisfiesthe analysans—paradigmatically, a list of conditions thatare individually necessary and jointly sufficient forS tohave knowledge thatP.

It is not enough merely to pick out the actual extension of knowledge.Even if, in actual fact, all cases ofS knowing thatP are cases in whichS satisfiesJ, and all cases of the latterare cases of the former,J might fail as an analysis ofknowledge. For example, it might be that there are possible cases ofknowledge withoutJ, or vice versa. A proper analysis ofknowledge should at least be a necessary truth. Consequently,hypothetical thought experiments provide appropriate test cases forvarious analyses, as we shall see below.

Even a necessary biconditional linking knowledge to some stateJ would probably not be sufficient for an analysis ofknowledge, although just what more is required is a matter of somecontroversy. According to one version of the project, to analyzeknowledge is literally to identify the components that make upknowledge—compare a chemist who analyzes a sample to learn itschemical composition. On this interpretation of the project ofanalyzing knowledge, the defender of a successful analysis ofknowledge will be committed to something like the metaphysical claimthatwhat it is forS to know thatPis for some list of conditions involvingS and theproposition thatP to obtain. Other theorists think ofthe analysis of knowledge as distinctivelyconceptual—to analyse knowledge is to limn thestructure of theconcept of knowledge. On one version of thisapproach, the conceptknowledge is literally composed of morebasic concepts, linked together by something like Boolean operators.Some contributors to the literature write in terms ofdefinitions of knowledge. Consequently, an analysis issubject not only to extensional accuracy, but to facts about thecognitive representation of knowledge and other epistemic notions. Inpractice, many epistemologists engaging in the project of analyzingknowledge leave these metaphilosophical interpretive questionsunresolved; attempted analyses, and counterexamples thereto, are oftenproposed without its being made explicit whether the claims areintended as metaphysical or conceptual ones.^[1] In many cases, this lack of specificity may be legitimate, since allparties tend to agree that an analysis of knowledge oughtatleast to be extensionally correct in all metaphysically possibleworlds. As we shall see, many theories have been defended and,especially, refuted on those terms.

The attempt to analyze knowledge has received a considerable amount ofattention from epistemologists, particularly in the late20^th century, but no analysis has been widely accepted.Some contemporary epistemologists reject the assumption that knowledgeis susceptible to analysis.

• 1. Knowledge as Justified True Belief
  • 1.1 The Truth Condition
  • 1.2 The Belief Condition
  • 1.3 The “Justification” Condition
    • 1.3.1 The “Justification” Label
    • 1.3.2 Internalism and Externalism
    • 1.3.3 Kinds of Justification
• 2. Lightweight Knowledge
• 3. Counterexamples to the Tripartite Analysis
• 4. No False Lemmas
• 5. Modal Conditions
  • 5.1 Sensitivity
  • 5.2 Safety
  • 5.3 Relevant Alternatives
• 6. Doing without Justification?
  • 6.1 Reliabilist Theories of Knowledge
  • 6.2 Causal Theories of Knowledge
• 7. Is Knowledge Analyzable?
• 8. Epistemic Luck
• 9. Methodological Options
• 10. Virtue-Theoretic Approaches
  • 10.1 The “AAA” Evaluations
  • 10.2 Fake Barn Cases
• 11. Knowledge First
• 12. Pragmatic Encroachment
• 13. Contextualism
• Bibliography
• Academic Tools
• Other Internet Resources
• Related Entries

1. Knowledge as Justified True Belief

Much of the 20^th century literature on the analysis ofknowledge began with a critique of a three-part“traditional” analysis of knowledge. Although thisanalysis does derive from ideas articulated throughout the history ofepistemology, it’s a bit of a convenient fiction to say thatthis so-called “traditional” analysis was ever widelyaccepted; it was primarily articulated in the 20^th centuryby its attackers. See Kaplan 1985, Dutant 2015, Antognazza 2015, andLe Morvan 2017 on the mismatch between the canonical story and theactual intellectual history.

There are three components to the traditional analysis of knowledge.(For this reason, it is sometimes referred to as a“tripartite” analysis of knowledge.) According to thisanalysis, justified true belief is necessary and sufficient forknowledge.

The tripartite analysis of knowledge is often abbreviated as the“JTB” analysis, for “justified truebelief”.

Let’s consider the three conditions in turn.

1.1 The Truth Condition

Most epistemologists agree that what is false cannot be known. Forexample, eighty-four is not a prime number. Consequently, it couldnever be the case that anyone knows that eighty-four is a primenumber. One can only know things that are true. For this reason, thetraditional theory of knowledge includes, as one of the components forthe analysans forS knows that P, the condition that itis true thatP. (Sometimes, the explicit invocation oftruth is left out, and the relevant condition is listed simplyasP. See Nozick (1981: 680, n. 6).)

Sometimes, when people are very confident of something that turns outto be wrong, it can be natural to describe them as“knowing” that false proposition. Suppose, for example,that late in a sports match, the home team is so far behind that manyspectators are very confident that they will lose. If they turn aroundin the final seconds to achieve a miraculous comeback and win thegame, some people might be inclined to say that, prior to thecomeback, many spectators “knew” that the home team wouldlose. Allan Hazlett (2010) argues on the basis of data like this that“knows” is not a factive verb.^[2] Hazlett’s diagnosis is deeply controversial; mostepistemologists will treat such knowledge ascriptions as a kind ofexaggeration, and not as literally true. The analysis of knowledge isabout identifying the circumstances under which someone really doesknow a given proposition—not necessarily about the circumstancesunder which it might be conversationally acceptable to say that theyknow.

Relatedly, the idea that a proposition is true does not require thatanyone can know or prove that it is true. Not all truths areestablished truths. If you flip a coin and never check how itlanded, it may be true that it landed heads, even if nobody has anyway to tell. Truth is ametaphysical, as opposed toepistemological, notion: truth is a matter of how thingsare, not how they can beshown to be. So when we saythat only true things can be known, we’re not (yet) sayinganything about how anyone canaccess the truth. Aswe’ll see, the other conditions have important roles to playhere. Knowledge is a kind of relationship with the truth—to knowsomething is to have a certain kind of access to a fact.

1.2 The Belief Condition

The belief condition is only slightly more controversial than thetruth condition. The general idea behind the belief condition is thatyou can only know what you believe. Failing to believe somethingprecludes knowing it. “Belief” in the context of the JTBtheory meansfull belief, oroutright belief. In aweak sense, one might “believe” something by virtue ofbeing pretty confident that it’s probably true—in thisweak sense, someone who considered the home team the favourite to winthe game, even while recognizing a nontrivial possibility of theirlosing, might be said to have believed that the home team would win.Outright belief is stronger.^[3] To believe outright thatP, it isn’t enough to have apretty high confidence inP; it is something closer to acommitment or a being sure, or having settled the question. In thepaper that explicitly laid out the “justified true belief”analysis, Edmund Gettier (1963) actually listed three slightlydifferent tripartite analyses, which differed in how this conditionwas articulated: one was given in terms of “believes”, onein terms of “accepts”, and one in terms of “is surethat”. Epistemologists have tended to follow Gettier in treatingthese as effectively equivalent. For further discussion, see McGlynn2014: 25–9 and Ichikawa 2017-a: §7.1.

Although initially it might seem obvious that knowing thatPrequires believing thatP, a few philosophers have arguedthat knowledge without belief is indeed possible. Suppose Walter comeshome after work to find out that his house has burned down. He says:“I don’t believe it.” Critics of the beliefcondition might argue that Walter knows that his house has burned down(he sees that it has), but, as his words indicate, he does not believeit. The standard response is that Walter’s avowal of disbeliefis not literally true; what Walter wishes to convey by saying,“I don’t believe it”, is not that he really does notbelieve that his house has burned down, but rather that he finds ithard to come to terms with what he sees. If he genuinely didn’tbelieve it, some of his subsequent actions, such as phoning hisinsurance company, would be rather mysterious.

Colin Radford (1966) gave a more substantive critique of the beliefcondition. Suppose Albert is quizzed on English history. One of thequestions is: “When did Queen Elizabeth die?” Albertdoesn’t think he knows, but answers the question correctly.Moreover, he gives correct answers to many other questions to which hedidn’t think he knew the answer. Let us focus on Albert’sanswer to the question about Elizabeth:

• (E)Elizabeth died in1603.

Radford makes the following two claims about this example:

1. Albert does not believe (E).
2. Albert knows (E).

Radford’s intuitions about cases like these do not seem to beidiosyncratic; Blake Myers-Schutz & Eric Schwitzgebel (2013) findevidence suggesting that many ordinary speakers tend to react in theway Radford suggests. In support of (a), Radford emphasizes thatAlbert thinks he doesn’t know the answer to the question. Hedoesn’t trust his answer because he takes it to be a mere guess.In support of (b), Radford argues that Albert’s answer is not atall just a lucky guess. The fact that he answers most of the questionscorrectly indicates that he has actually learned, and never forgotten,such historical facts.

Since he takes (a) and (b) to be true, Radford holds that belief isnot necessary for knowledge. But both of these intuitions arecontroversial. Some epistemologists might deny (a), arguing thatAlbert does have atacit belief that (E), even thoughit’s not one that he thinks amounts to knowledge. See e.g.,David Rose & Jonathan Schaffer (2013). Alternatively, one mightdeny (b), arguing that Albert’s correct answer is not anexpression of knowledge, perhaps because it is insufficientlyconfident. See Nagel 2010 and Ichikawa 2024 on ways in whichuncertainty can undermine knowledge by undermining belief.

Maria Antognazza (2020) offers quite a different kind of challenge tothe belief condition on knowledge: she rejects the orthodoxphilosophical understanding of belief, according to which it is a kindof doxastic commitment that does not depend on whether or not itamounts to an apprehension of reality, holding instead that belief, byits nature, is a cognitive representation supported only indirectly byexperience. Belief, according to Antognazza, is fundamentally unlikeknowledge, in that by definition, one can only believe that which isnot directly apprehended. This isn’t merely a denial that beliefis a necessary condition for knowledge; on her view, believingP is actuallyincompatible with knowingP.This is not a mainstream or widely held approach to belief incontemporary epistemology, but Antognazza argues that it is supportedby considerations in the philosophy of mind, and representative ofimportant historical epistemological traditions.

1.3 The “Justification” Condition

Why is a third condition necessary? Why not say that knowledge is truebelief? The standard answer rationale has to do wtih cases where abelief might be true even though it is formed improperly. Suppose thatWilliam flips a coin and confidently believes—on noparticular basis—that it will land tails. If, coincidentally,the coin does indeed land tails, then William’s belief was true;but such a lucky guess, intuitively, can’t count as knowledge.For William to know, his belief must be grounded in adequateevidential support. Contemporary epistemologists often use“justification” as the name for the kind of support thatis needed here. Recognition of the need for something like thiscondition is often traced to Plato’sTheaetetus, whereSocrates points out that “true opinion” is in generalinsufficient for knowledge. For example, if a lawyer employs sophistryto induce a jury into a belief that happens to be true, this belief isinsufficiently well-grounded to constitute knowledge.

1.3.1 The “Justification” Label

Although “justification” is the standard contemporary namefor the condition in question, this label is a relatively recent(20^th century) development, and the name can carry somepotentially misleading connotations, especially for those not alreadyversed in the literature.

In an ordinary sense, something might count as “justified”only after someone has successfully undertaken the project ofjustifying it. Understood in this way, the idea that a truebelief must be justified in order to count as knowledge would meanthat in order forS to know thatP,S musthave already given a positive defense of their belief thatP,onedemonstrating that their belief is appropriate.Epistemologists who endorse a justification condition on knowledge donot typically mean their condition to be understood in such a strongway. Rather, they typically intend to suggest only that aS’s belief thatP is knowledge only if theirbelief isin fact appropriately supported byS’s evidence. This might be so, even ifS hasnot engaged in a project ofactively justifying their belief.Consider, for example, an ordinary person’s belief that fiveplus five equals ten. Most people have never attempted to justify thisbelief, and probably would be at a loss as to how to go aboutjustifying it. But for most people, that belief would qualify as aninstance of knowledge.^[4] For this reason, it is important, in understanding the role of thejustification condition in the analysis of knowledge, not to assumethat justification is a dialectical activity in which the believermust have engaged. Some epistemologists prefer to use different labelsfor this condition for this reason, such as the idea that a belief isrational orwarranted orevidentially well-supported.^[5]

Jonathan Ichikawa (2024: ch. 3) gives an additional critique of the“justification” label for this condition, arguing thatthis word carries a negative presupposition to the effect that beliefshould, by default, be considered suspect. Ichikawa argues that, forthis reason, the idea that beliefs need to be “justified”furthers unwarranted skeptical biases.

1.3.2 Internalism and Externalism

There is considerable disagreement among epistemologists concerningwhat the relevant sort of justification here is.Internalistsabout justification think that whether a belief is justified dependswholly on states, that are, in some sense,internal to thesubject. According to one common such sense of “internal”,only those features of a subject’s experience which are directlyor introspectively available count as“internal”—call this “accessinternalism”. According to another, only intrinsic states of thesubject are “internal”—call this “stateinternalism”. See theInternalist vs. ExternalistConceptions of Epistemic Justification entry (Pappas 2023) formore detail.

Earl Conee and Richard Feldman (2001) present an example of aninternalist view. They have it thatS’s belief thatP is justified if and only if believing thatP isthe attitude towardsP that best fitsS’sevidence, where the latter is understood to depend only onS’s internal mental states. Conee and Feldman calltheir view “evidentialism”, and characterize this as thethesis that justification is wholly a matter of the subject’sevidence. Given their (not unsubstantial) assumption that whatevidence a subject has is an internal matter, evidentialism implies internalism.^[6]Externalists about justification think that factors externalto the subject can be relevant for justification. For example, processreliabilists think that justified beliefs are those which are formedby a cognitive process which tends to produce a high proportion oftrue beliefs relative to false ones.^[7] We shall return to the question of how reliabilist approaches bear onthe analysis of knowledge in§6.1.

1.3.3 Kinds of Justification

It is worth noting that one might distinguish between two importantlydifferent notions of justification, standardly referred to as“propositional justification” and “doxasticjustification” (sometimes “ex ante”justification and “ex post” justification, respectively).^[8] Unlike that between internalist and externalist approaches tojustification, the distinction between propositional and doxasticjustification does not represent a conflict to be resolved; it is adistinction between two distinct properties that are sometimes called“justification”. Propositional justification concernswhether a subject has sufficient reason to believe a given proposition;^[9] doxastic justification concerns whether a given belief is heldappropriately. One common way of relating the two is to suggest thatpropositional justification is the more fundamental, and that doxasticjustification is a matter of a subject having a belief that isappropriately responsive to or based on their propositionaljustification.

The precise relation between propositional and doxastic justificationis subject to controversy, but it is uncontroversial that the twonotions can come apart. Suppose that Ingrid ignores a great deal ofexcellent evidence indicating that a given neighborhood is dangerous,but superstitiously comes to believe that the neighborhood isdangerous when she sees a black cat crossing the street. Since formingbeliefs on the basis of superstition is not an epistemicallyappropriate way of forming beliefs, Ingrid’s belief is notdoxastically justified; nevertheless, shedoes have goodreason to believe as she does, so she does have propositionaljustification for the proposition that the neighborhood isdangerous.

Since knowledge is a particularly successful kind of belief, doxasticjustification is a stronger candidate for being closely related toknowledge; the JTB theory is typically thought to invoke doxasticjustification (but cf. Lowy 1978).

2. Lightweight Knowledge

Some epistemologists have suggested that there may be multiple sensesof the term “knowledge”, and that not all of them requireall three elements of the tripartite theory of knowledge. Some haveargued that there is, in addition to the sense of“knowledge” gestured at above, another,weaksense of “knowledge” that requires only true belief (seee.g., Hawthorne 2002 and Goldman & Olsson 2009; the lattercontains additional relevant references). This view is sometimesmotivated by the thought that, when we consider whether someone knowsthatP, or wonder who of a group of people know thatP, often, we are not at all interested in whether therelevant subjects have beliefs that are justified; we just want toknow whether they have the true belief. For example, as John Hawthorne(2002: 253–54) points out, one might ask how many students knowthat Vienna is the capital of Austria; the correct answer, one mightthink, just is the number of students who offer “Vienna”as the answer to the corresponding question, irrespective of whethertheir beliefs are justified. Similarly, if you are planning a surpriseparty for Eugene and ask whether he knows about it, “yes”may be an appropriate answer merely on the grounds that Eugenebelieves that you are planning a party.

One could allow that there is a lightweight sense of knowledge thatrequires only true belief; another option is to decline to accept theintuitive sentences as true at face value. A theorist might, forinstance, deny that sentences like “Eugene knows that you areplanning a party” or “Eighteen students know that Viennais the capital of Austria” are literally true in the envisagedsituations, explaining away their apparent felicity as loose talk orhyperbole.

Even among those epistemologists who think that there is a lightweightsense of “knows” that does not require justification, mosttypically admit that there is also a stronger sense which does, andthat it is this stronger state that is the main target ofepistemological theorizing about knowledge. The remainder of thisarticle will set aside the lightweight sense, if indeed there is one,and focus on the stronger sense of “knows”.

3. Counterexamples to the Tripartite Analysis

It is nearly universally accepted amongst contemporary epistemologiststhat there are counterexamples to the analysis of knowledge asjustified true belief. Although most agree that each element of thetripartite theory isnecessary for knowledge, it is widelyaccepted that they are not themselvessufficient. In otherwords, there seem to be cases of justified true belief that still fallshort of knowledge. Such examples have actually been discussed sinceantiquity. Here, for example, is an example from the Indianphilosopher Dharmottara, c. 770 CE.

Imagine that we are seeking water on a hot day. We suddenly see water,or so we think. In fact, we are not seeing water but a mirage, butwhen we reach the spot, we are lucky and find water right there undera rock. Can we say that we had genuine knowledge of water? The answerseems to be negative, for we were just lucky. (qtd. in Dreyfus 1997:292)

Śrīharṣa gave similar cases in the 12^thcentury CE—see Das 2021a: §2.1. It’s typicallythought to be intuitive that in Dharmottara’s case, one does notknow that there is water at the location in question. But it alsoseems that all three conditions of the tripartite theory of knowledgeare met: it istrue that there is water in that spot, onebelieves that there is water in that spot, and one’sbelief isjustified or reasonable, since it is based onone’s sensory experience. So, it seems, cases like this showthat the tripartite analysis of knowledge has counterexamples.

The 14^th-century Italian philosopher Peter of Mantuapresented this similar case:

Let it be assumed that Plato is next to you and you know him to berunning, but you mistakenly believe that he is Socrates, so that youfirmly believe that Socrates is running. However, let it be so thatSocrates is in fact running in Rome; however, you do not know this.(from Peter of Mantua’sDe scire et dubitare, given inBoh 1985: 95)

Bertrand Russell gave a similar case in the early 20^thcentury:

If you look at a clock which you believe to be going, but which infact has stopped, and you happen to look at it at a moment when it isright, you will acquire a true belief as to the time of day, but youcannot be correctly said to have knowledge. (Russell 1923:113)^[10]

Although such cases have been discussed throughout the history ofEastern and Western epistemology, they came into far more theoreticalsalience in the second half of the 20th century, prompted especiallyby Gettier’s influential 1963 paper, “Is Justified TrueBelief Knowledge?”. Gettier presented two cases in which a truebelief is inferred from a justified false belief, including thisone:

Let us suppose that Smith has strong evidence for the followingproposition: (f) Jones owns a Ford. … Smith has another friend,Brown, of whose whereabouts he is totally ignorant. Smith selectsthree place-names quite at random, and constructs the following threepropositions:

(g) Either Jones owns a Ford, or Brown is in Boston;
(h) Either Jones owns a Ford, or Brown is in Barcelona;
(i) Either Jones owns a Ford, or Brown is in Brest-Litovsk.

Each of these propositions is entailed by (f). … Smith istherefore completely justified in believing each of these threepropositions. Smith, of course, has no idea where Brown is. Butimagine now that two further conditions hold. First, Jones doesnot own a Ford, but is at present driving a rented car. Andsecondly, by the sheerest coincidence, and entirely unknown to Smith,the place mentioned in proposition (h) happens really to be the placewhere Brown is. (Gettier 1963: 123)

Gettier observed that, intuitively, Smith does not know thedisjunction (h), despite having a justified true belief to thateffect. Gettier’s paper was enormously influential. Despitetheir older history, counterexamples to the tripartite theory havesince often come to be known as “Gettier cases”. Sincethey appear to refute the JTB analysis, many epistemologists,especially in the late 20th century, undertook the project ofattempting to amend that analysis to provide an adequate account. Thedifficulty in doing so with success is what is commonly referred to asthe “Gettier problem”.

As indicated in§1.3, one role of the justification condition was to rule out lucky guessesas cases of knowledge. A lesson of the Gettier problem is that itappears that even true beliefs that are justified can nevertheless beepistemically lucky in a way inconsistent with knowledge. Much of thefocus in the analysis of knowledge literature was on identifying anadditional condition, beyond justification, truth, and belief, thatmight explain why so-called “Gettier cases” aren’tknowledge. The hope was that articulating such a fourth conditionX would allow for a “JTB +X” analysisof knowledge that did not succumb to counterexamples.

Let us consider an instance of this attempt to articulate a“degettiering” condition.

4. No False Lemmas

One natural idea was to add the following fourth condition to thetraditional analysis:

4. S’s belief thatP is not inferred from any falsehood.^[11]

The subjects of Gettier’s cases straightforwardly violated thiscondition. When Smith believes thatP (where“P” designates thedisjunctioneither Jones owns a Ford or Brown is inBarcelona), he does so because he inferred it from the justifiedbut false propositionthat Jones owns a Ford. Theother counterexamples above might also be thought to involve inferencefrom falsehoods. In Russell’s case, for example, one ends upwith the true belief that it is, say, 1:00, but this is based in parton the false belief that one is looking at a clock that is functioningproperly. So a condition like this one might explain why these casesaren’t knowledge, pointing to the “missingingredient” left out by the tripartite theory.

However, this “no false lemmas” proposal is not widelyaccepted. There are apparent examples of Gettier cases that needinvolve no inference; therefore, there are possible cases of justifiedtrue belief without knowledge, even though this proposed fourthcondition is met. Suppose, for example, that James, who is relaxing ona bench in a park, observes an apparent dog in a nearby field. So hebelieves:

4. There is a dog in the field.

Suppose further that the putative dog is actually a robot dog soperfect that it could not be distinguished from an actual dog byvision alone. James does not know that such robot dogs exist; aJapanese toy manufacturer has only recently developed them, and whatJames sees is a prototype that is used for testing the public’sresponse. Given these assumptions, (d) is of course false. But supposefurther that just a few feet away from the robot dog, there is a realdog, concealed from James’s view. Given this further assumption,James’s belief in (d) is true. And since this belief is based onordinary perceptual processes, most epistemologists will agree that itis justified. But as in the counterexamples to the tripartite theory,James’s belief appears to be true only as a matter of luck, in away inconsistent with knowledge. So once again, what we have before usis a justified true belief that isn’t knowledge.^[12] Moreover, arguably, this belief is directly justified by a visualexperience; it is not inferred from any falsehood. If so, then the JTBaccount, even if supplemented with (4), gives us the wrong result thatJames knows (d).

Another case illustrating that clause (4) won’t do the job isthe well-known Barn County case, first given by Carl Ginet, andpublished by Alvin Goldman (1976). Suppose there is a county with thefollowing peculiar feature. The landscape next to the road leadingthrough that county is peppered with barn facades: structures thatfrom the road look exactly like barns. Observation from any otherviewpoint would immediately reveal these structures to be fakes:devices erected for the purpose of fooling unsuspecting motorists intobelieving in the presence of barns. Suppose Henry is driving along theroad that leads through Barn County. Naturally, he will on numerousoccasions form false beliefs in the presence of barns. Since Henry hasno reason to suspect that he is the victim of organized deception,these beliefs are justified. Now suppose further that, on one of thoseoccasions when he believes there is a barn over there, he happens tobe looking at the one and only real barn in the county. This time, hisbelief is justified and true. But since its truth is the result ofluck, it is exceedingly plausible to judge that Henry’s beliefis not an instance of knowledge. Yet condition (4) is met in thiscase. His belief does not appear to be the result of any inferencefrom a falsehood. Once again, we see that (4) does not succeed as ageneral solution to the Gettier problem.

5. Modal Conditions

5.1 Sensitivity

Another candidate fourth condition on knowledge issensitivity. Sensitivity, to a first approximation, is thiscounterfactual relation:

S’s belief thatP is sensitive if and only if,if were false thatP,S would not believe thatP.^[13]

A sensitivity condition on knowledge was defended by Robert Nozick(1981). Given one influential semantics for counterfactualconditionals (Lewis 1973), the sensitivity condition is equivalent tothe requirement that, in the nearest possible worlds in whichnot-P, the subject does not believe thatP.

One motivation for including a sensitivity condition in an analysis ofknowledge is that there seems to be an intuitive sense in whichknowledge requires not merely being correct, buttracking thetruth in other possible circumstances. This approach seems to be aplausible diagnosis of what goes wrong in at least some of thecounterexamples to the traditional theory. For example, inDharmottara’s desert water case, your belief that there is waterin a certain location appears to be insensitive to the fact of thewater. For if there were no water there, you would have held the samebelief on the same grounds—viz., the mirage.

However, it is doubtful that a sensitivity condition can account forall the counterexamples. It does so only in cases in which, had theproposition in question been false, it would have been believedanyway. As Saul Kripke (2011: 167–68) has pointed out, there arecounterexamples to the traditional analysis that are not like this.Consider for instance the Barn County case mentioned above. Henrylooks at a particular location where there happens to be a barn andbelieves there to be a barn there. The sensitivity condition rules outthis belief as knowledge only if, if there hadn’t been a barnthere, Henry would still have believed there was. But thiscounterfactual may be false, depending on the details of the case. Forinstance, it is false if the particular location Henry is examining isnot one that would have been suitable for the erecting of a barnfacade. Relatedly, as Kripke has also indicated (2011: 186), if wesuppose that barn facades are always green, but genuine barns arealways red, Henry’s belief that he sees ared barn willbe sensitive, even though his belief that he sees abarn willnot. (We assume Henry is unaware that colour signifies anythingrelevant.) Since intuitively, the former belief looks to fall short ofknowledge in just the same way as the latter, a sensitivity conditionwill only handle some of the intuitive counterexamples to thetraditional analysis.

Most epistemologists today reject sensitivity requirements onknowledge. The chief motivation against a sensitivity condition wasarticulated by Keith DeRose (1995: 27–28): given plausibleassumptions, a sensitivity condition on knowledge has unacceptableimplications. Assume that ordinary subjects know at least many of thethings we typically take them to know. For example, George, who cansee and use his hands perfectly well, knows that he has hands. This byitself is perfectly consistent with a sensitivity condition onknowledge, since if George didnot have hands—ifthey’d been recently chopped off, for instance—he wouldnot believe that he had hands.

But now imagine a skeptical scenario in which George does not havehands. Suppose that George is the victim of a Cartesian demon,deceiving him into believing that he has hands. If George were in sucha scenario, he would falsely believe himself not to be in such ascenario. So given the sensitivity condition, George cannot know thathe is not in such a scenario.

Although these two verdicts—the knowledge-attributing one aboutordinary knowledge, and the knowledge-denying one about the skepticalscenario—are arguably each intuitive, it is intuitivelyproblematic to hold them together. Their conjunction is, inDeRose’s term,abominable. A sentence like“George knows that he has hands, but he doesn’t know thathe’s not the handless victim of a Cartesian demon” isoften called an “abominable conjunction”; a sensitivitycondition on knowledge, combined with the nonskeptical claim thatthere is ordinary knowledge, seems to imply such abominable conjunctions.^[14]

Most contemporary epistemologists have taken considerations like theseto be sufficient reason to reject sensitivity conditions.^[15] But some epistemologists have continued to defend sensitivity conditions.^[16] See Becker & Black 2012 for some contemporary discussions.

5.2 Safety

Although few epistemologists today endorse a sensitivity condition onknowledge, the idea that knowledge requires a subject to stand in aparticular modal relation to the proposition known remains a popularone. In his 1999 paper “How to Defeat Opposition toMoore”, Ernest Sosa proposed that asafety conditionought to take the role that sensitivity was intended to play. Sosacharacterized safety as the counterfactual contrapositive ofsensitivity.

Sensitivity:
If it were false thatP,S would not believethatP.

Safety:
IfS were to believe thatP, it would not be false thatP.^[17]

Although contraposition is valid for the material conditional, Sosasuggests that it is invalid for counterfactuals, which is whysensitivity and safety are not equivalent. An example of a safe beliefthat is not sensitive, according to Sosa, is the belief that a distantskeptical scenario does not obtain. If we stipulate that George,discussed above, has never been at risk of being the victim of aCartesian demon—because, say, Cartesian demons do not exist inGeorge’s world—then George’s belief that he is notsuch a victim is a safe one, even though we saw in the previoussection that it could not be sensitive. Notice that although westipulated that George is not at risk of deceit by Cartesian demons,we didnot stipulate that George himself had any particularaccess to this fact. Unless he does, safety, like sensitivity, will beanexternalist condition on knowledge in the“access” sense. It is also externalist in the“state” sense, since the truth of the relevantcounterfactuals will depend on features outside the subject.

Characterizing safety in these counterfactual terms depends onsubstantive assumptions about the semantics of counterfactual conditionals.^[18] If we were to accept, for instance, David Lewis’s or RobertStalnaker’s treatment of counterfactuals, according to which thetruth value of a counterfactual conditional with a true antecedent isjust the truth value of its consequent, all true beliefs would countas safe according to the counterfactual analysis of safety.^[19] Sosa intends the relevant counterfactuals to be making a strongerclaim, requiring roughly that inall nearby worlds in whichS believes thatP, it is not falsethatP.

Rather than resting on a contentious treatment of counterfactuals,then, it may be most perspicuous to understand the safety conditionmore directly in these modal terms, as Sosa himself often does:

Evaluating analyses of knowledge that incorporate safety conditionscan be difficult, given the vagueness of the stated“nearby” condition. The status of potentialcounterexamples will not always be straightforward to apply. Forexample, Juan Comesaña (2005) presents a case he takes torefute the requirement that knowledge be safe. InComesaña’s example, the host of a Halloween party enlistsJudy to direct guests to the party. Judy’s instructions are togive everyone the same directions, which are in fact accurate, butthat if she sees Michael, the party will be moved to another location.(The host does not want Michael to find the party.) Suppose Michaelnever shows up. If a given guest does not, but very nearly does,decide to wear a very realistic Michael costume to the party, then hisbelief, based in Judy’s testimony, about the whereabouts of theparty will be true, but could, Comesaña says, easily have beenfalse. (Had he merely made a slightly different choice about hiscostume, he would have been deceived.) Comesaña describes thecase as a counterexample to a safety condition on knowledge. Likewise,some epistemologists (e.g. Neta and Rohrbaugh 2004:399) arguethat safety conditions fare poorly with respect to some versions ofGinet’s barn facade cases. However, it is open to a safetytheorist to argue that the relevant skeptical scenarios, thoughpossible andin some sense nearby, is not near enough in therelevant respect to falsify the safety condition. Such a theoristwould, if she wanted the safety condition to deliver clear verdicts,face the task of articulating just what the relevant notion ofsimilarity amounts to (see also Bogardus 2014).

Not all further clarifications of a safety condition will be suitablefor the use of the latter in an analysis of knowledge. In particular,if the respect of similarity that is relevant for safety is itselfexplicated in terms of knowledge, then an analysis of knowledge whichmade reference to safety would be in this respect circular. This, forinstance, is how Timothy Williamson characterizes safety. He writes,in response to a challenge by Goldman:

In many cases, someone with no idea of what knowledge is would beunable to determine whether safety obtained. Although they could usethe principle that safety entails truth to exclude some cases, thoseare not the interesting ones. Thus Goldman will be disappointed whenhe asks what the safety account predicts about various examples inwhich conflicting considerations pull in different directions. One mayhave to decide whether safety obtains by first deciding whetherknowledge obtains, rather than vice versa. (Williamson 2009: 305)

Because safety is understood only in terms of knowledge, safety sounderstood cannot serve in an analysis of knowledge. (Nor is itWilliamson’s intent that it should do so; as we will see below,Williamson rejects the project of analyzing knowledge.)

5.3 Relevant Alternatives

A third approach to modal conditions on knowledge worthy of mention isthe requirement that for a subject to know thatP, they mustrule out all “relevant alternatives”—that is,they must be able to rule out any relevant possibility inwhichP is false. Significant early proponents of thisview include Stine 1976, Goldman 1976, and Dretske 1981; more recentdefenders include Lewis 1996, Schaffer 2004, Ichikawa 2017-a,Blome-Tillmann 2009-b, and Gardiner 2020.

The idea behind this approach to knowledge is that for a subject toknow thatP, they must be able to “rule out”competing hypotheses to the possibility thatP—but thatonly some subset of all possibilities in whichP isfalse are “relevant” for knowledge attributions. Considerfor example, the differences between the several models that have beenproduced of Apple’s iPhone. To be able to know by sight that aparticular phone is the 16 Plus model, it is natural to suppose thatone must be able to tell the difference between the iPhone 16 Plus andthe iPhone 15 Pro; the possibility that the phone in question is adifferent model is a relevant alternative. But perhaps there are otherpossibilities in which the belief that there is an iPhone 16 Plus isfalse that do not need to be ruled out—perhaps, for instance,the possibility that the phone is not an iPhone but a replica made asa movie prop needn’t be considered. Likewise for the possibilitythat there is no object there at all, the phone-like appearances beingthe product of a Cartesian demon’s machinations. Notice that inthese cases and many of the others that motivate therelevant-alternatives approach to knowledge, there is an intuitivesense in which the relevant alternatives tend to be moresimilar to actuality than irrelevant ones. For this reason,the relevant alternatives theory and safety-theoretic approaches aresimilar, both in verdict and in spirit. As in the case of a safetytheorist, the relevant alternatives theorist faces a challenge inattempting to articulate what determines which possibilities arerelevant in a given situation.

6. Doing without Justification?

As we saw in§1.3, one motivation for including a justification condition in an analysisof knowledge was to prevent lucky guesses from counting as knowledge.However, the Gettier problem shows that including a justificationcondition does not rule out all epistemically problematic instances ofluck. Consequently, some epistemologists have suggested that positinga justification condition on knowledge was a false move; perhaps it issome other condition that ought to be included along with truth andbelief as components of knowledge. This kind of strategy was advancedby a number of authors from the late 1960s to the early 1980s,although there has been less discussion of it since.^[20] Kornblith 2008 and Ichikawa 2024: ch. 3 provide notableexceptions.

This section will consider attempts to amend the tripartite analysis,not by adding a fourth condition, but by replacing the justificationcondition with a different one.

6.1 Reliabilist Theories of Knowledge

One candidate property for such a state isreliability. Partof what is problematic about lucky guesses is precisely that they areso lucky: such guesses are formed in a way such that it is unlikelythat they should turn out true. According to a certain form ofknowledge reliabilism, it is unreliability, not lack of justification,which prevents such beliefs from amounting to knowledge. Reliabilisttheories of knowledge incorporate this idea into a reliabilitycondition on knowledge.^[21] Here is an example of such a view:

Simple K-Reliabilism replaces the justification clause in thetraditional tripartite theory with a reliability clause. As we haveseen, reliabilists about justification think that justification for abelief consists in a genesis in a reliable cognitive process. Giventhis view, Simple K-Reliabilism and the JTB theory are equivalent.However, the present proposal is silent on justification. Goldman 1979is the seminal defense of reliabilism about justification; reliabilismis extended to knowledge in Goldman 1986. See Goldman & Beddor2021 for a survey of reliabilism in general. Dretske (1985: 177)suggests that invoking reliability instead of justification in anapproach to knowledge better allows for knowledge in non-humananimals. (See also Dretske 1989: 95)

6.2 Causal Theories of Knowledge

Another move in a similar spirit to K-reliabilism replaces thejustification clause in the JTB theory with a condition requiring acausal connection between the belief and the fact believed;^[22] this is the approach of Goldman (1967, 1976).^[23] Goldman’s own causal theory is a sophisticated one; we will notengage with its details here. See Goldman’s papers. Instead,consider a simplified causal theory of knowledge, which illustratesthe main motivation behind causal theories.

Do approaches like Simple K-Reliabilism or the Simple Causal Theoryfare better than the tripartite theory with respect to Gettier cases?Although some proponents have suggested they do—see e.g.,Dretske 1985: 179; Plantinga 1993: 48—many of the standardcounterexamples to the JTB theory appear to refute these views aswell. Consider again the case of the barn facades. Henry sees a realbarn, and that’s why he believes there is a barn nearby. Thisbelief is formed by perceptual processes, which are by-and-largereliable: only rarely do they lead him into false beliefs. So it lookslike the case meets the conditions of Simple K-Reliabilism just asmuch as it does those of the JTB theory. It is also a counterexampleto the causal theory, since the real barn Henry perceives is causallyresponsible for his belief. There is reason to doubt, therefore, thatshifting from justification to a condition like reliability willescape the Gettier problem.^[24] Gettier cases seem to pose as much of a problem for K-reliabilism andcausal theories as for the JTB account. Neither theory alone seems tosucceed in stating sufficient conditions for knowledge.^[25]

7. Is Knowledge Analyzable?

Gettier’s paper launched a flurry of philosophical activity byepistemologists attempting to revise the JTB theory, usually by addingone or more conditions, to close the gap between knowledge andjustified true belief. We have seen already how several of theseattempts failed. When intuitive counterexamples were proposed to eachtheory, epistemologists often responded by amending their theories,complicating the existing conditions or adding new ones. Much of thisdialectic is chronicled thoroughly by Shope 1983, to which theinterested reader is directed.

After some decades of such iterations, some epistemologists began todoubt that progress was being made. In her 1994 paper “TheInescapability of Gettier Problems”, Linda Zagzebski suggestedthat no analysis sufficiently similar to the JTB analysis could everavoid the problems highlighted by Gettier’s cases. Moreprecisely, Zagzebski argued, any analysans of the form JTB+X,whereX is a condition or list of conditions logicallyindependent from justification, truth, and belief, would besusceptible to Gettier-style counterexamples. She offered what was ineffect a recipe for constructing Gettier cases:

• (1)Start with anexample of a case where a subject has a justified false belief thatalso meets conditionX.
• (2)Modify the case sothat the belief is true merely by luck.

Zagzebski suggests that the resultant case will always represent anintuitive lack of knowledge. So any non-redundant addition to the JTBtheory will leave the Gettier problem unsolved.^[26] We may illustrate the application of the recipe using one ofZagzebski’s own examples, refuting an attempt by Alvin Plantinga(1996) to solve the Gettier problem. Plantinga added a fourthcondition, according to which, for a belief to be knowledge thesubject’s belief-producing faculties must be working properly inan appropriate environment.

In step one of Zagzebski’s procedure, we imagine a case in whicha subject’s faculties are working properly in an appropriateenvironment, but the ensuing belief, though justified, is false.Zagzebski invites us to imagine that Mary has good eyesight—goodenough for her cognitive faculties typically to yield knowledge thather husband is sitting in the living room. Such faculties, even whenworking properly in suitable environments, however, are notinfallible—if they were, the condition would not be independentfrom truth—so we can imagine a case in which they go wrong.Perhaps this is an unusual instance in which Mary’shusband’s brother, who looks a lot like the husband, is in theliving room, and Mary concludes, on the basis of the proper functionof her visual capacity, that her husband is in the living room. Thisbelief, since false, is certainly not knowledge.

In step two, we imagine Mary’s misidentification of the occupantof the living room as before, but add to the case that the husband is,by luck, also in the living room. Now Mary’s belief is true, butintuitively, it is no more an instance of knowledge than the falsebelief in the first step was. Indeed, we have constructed a case quitelike the standard counterexamples to the tripartite theory.

Since the recipe is a general one, it appears to be applicable to anycondition one might add to the JTB theory, so long as it does notitself entail truth. The argument generalizes against all“non-redundant” JTB+X analyses.

One potential response to Zagzebski’s argument, and the failureof the Gettier project more generally, would be to conclude thatknowledge is unanalyzable. Although it would represent a significantdeparture from much analytic epistemology of the late 20^thcentury, it is not clear that this is ultimately a particularlyradical suggestion. Few concepts of interest have proved susceptibleto traditional analysis (Fodor 1998). One prominent approach toknowledge in this vein is discussed in§11 below.

Another possible line is the one mentioned in§2: to strengthen the justification condition to rule out Gettier casesas justified. In order for this strategy to prevent Zagzebski’srecipe from working, one would need to posit a justification conditionthat precludes the possibility of step one above—the onlyobvious way to do this is for justification to entail truth. If itdoes, then it will of course be impossible to start with a case thathas justified false belief. This kind of approach is not mainstream,but it does have its defenders—see, e.g., Sturgeon 1993 andMerricks 1995. Julien Dutant (2015) argues that such an approachbetter reflects the history of epistemology than contemporaryfallibilist approaches to justification do. Jonathan Sutton (2007) andClayton Littlejohn (2012) defend factive approaches to justificationon other grounds.

A third avenue of response would be to consider potential analyses ofknowledge that are not of the nonredundant form JTB+X.Indeed, we have already seen some such attempts, albeit unsuccessfulones. For instance, the causal theory of knowledge includes a clauserequiring that the belief thatP be caused by the fact thatP. This condition entails both belief and truth, and so isnot susceptible to Zagzebski’s recipe. (As we’ve seen, itfalls to Gettier-style cases on other grounds.) One family ofstrategies along these lines would build into an analysis of knowledgea prohibition on epistemic luck directly; let us consider this sort ofmove in more detail.

8. Epistemic Luck

If the problem illustrated by Gettier cases is that JTB and JTB+analyses are compatible with a degree of epistemic luck that isinconsistent with knowledge, a natural idea is to amend one’sanalysis of knowledge by including an explicit “anti-luck”condition. Zagzebski herself outlines this option (1994: 72). PeterUnger (1968) gives an early analysis of this kind. For example:

The first thing to note about this analysis is that it is“redundant” in the sense described in the previoussection; the fourth condition entails the first two.^[27] So its surface form notwithstanding, it actually represents asignificant departure from the JTB+ analyses. Rather than composingknowledge from various independent components, this analysis demandsinstead that the epistemic states are related to one another insubstantive ways.

The anti-luck condition, like the safety condition of the previoussection, is vague as stated. For one thing, whether a belief is trueby luck comes in degrees—just how much luck does it take to beinconsistent with knowledge? Furthermore, it seems, independently ofquestions about degrees of luck, we must distinguish between differentkinds of luck. Not all epistemic luck is incompatible withhaving knowledge. Suppose someone enters a raffle and wins anencyclopedia, then reads various of its entries, correcting many oftheir previous misapprehensions. There is a straightforward sense inwhich the resultant beliefs are true only by luck—for oursubject was very lucky to have won that raffle—but this is notthe sort of luck, intuitively, that interferes with the possession of knowledge.^[28]

Furthermore, there is a sense in which our ordinary perceptual beliefsare true by luck, since it is possible for us to have been the victimof a Cartesian demon and so we are, in some sense, lucky not to be.But unless we are to capitulate to radical skepticism, it seems thatthis sort of luck, too, ought to be considered compatible with knowledge.^[29]

Duncan Pritchard (2016) suggests that it might be better to expresssome of these ideas in terms of risk, rather than luck: one’sbelief isn’t knowledge if there was too much risk of one havingformed a false belief. For discussion, see Navarro (2023) orHirvelä (2025). These subtleties have received renewed attentionin recent decades, but they were also a focus in antiquity, especiallyin the Indian Nyāya tradition. See e.g., Das (2021b) onGaṅgeśa Upādhyāya’s 14th-century treatmentof epistemic luck.

As was the case for the safety condition discussed in§5.3, one challenge for proponents of an anti-luck or anti-risk conditionon knowledge is to specify the relevant notions—of luck, or ofrisk—in terms clear enough to make adequate predictions withintuitively plausible extensions. And if it is to serve in theanalysis of knowledge, it is also important to explain the relevantnotion of luck or risk without invoking knowledge. The challenge is,in effect, to state an analysis that is at once plausible,informative, and noncircular.

9. Methodological Options

As our discussion so far makes clear, one standard way of evaluatingattempted analyses of knowledge has given a central role to testing itagainst intuitions about cases. In the late 20^th century,the perceived lack of progress towards an acceptableanalysis—including the considerations attributed to Zagzebski in§7 above—led some epistemologists to pursue other methodologicalstrategies. (No doubt, a wider philosophical trend away from“conceptual analysis” more broadly also contributed tothis change.) Some of the more recent attempts to analyse knowledgehave been motivated in part by broader considerations about the roleof knowledge, or of discourse about knowledge.

One important view of this sort is that defended by Edward Craig(1990). Craig’s entry-point into the analysis of knowledge wasnot intuitions about cases, but rather a focus on the role that theconcept of knowledge plays for humans. In particular, Craig suggestedthat the point of using the category of knowledge was for people toflag reliable informants—to help people know whom to trust inmatters epistemic. Craig defends an account of knowledge that isdesigned to fill this role, even though it is susceptible to intuitivecounterexamples. The plausibility of such accounts, with a lessintuitive extension but with a different kind of theoreticaljustification, is a matter of controversy.

Another view worth mentioning in this context is that of HilaryKornblith (2002), which has it that knowledge is a natural kind, to beanalysed the same way other scientific kinds are. Intuition has a roleto play in identifying paradigms, but generalizing from there is anempirical, scientific matter, and intuitive counterexamples are to beexpected.

The “knowledge first” stance is also connected to thesemethodological issues. See§11 below.

10. Virtue-Theoretic Approaches

The virtue-theoretic approach to knowledge is in some respects similarto the safety and anti-luck approaches. Indeed, Sosa, one of the mostprominent authors of the virtue-theoretic approach emphasized safetyas part of his characterization of virtue. The virtue approach treatsknowledge as a particularly successful or valuable form of belief, andexplicates what it is to be knowledge in such terms. Like theanti-luck theory, a virtue-theoretic theory leaves behind the JTB+project of identifying knowledge with a truth-functional combinationof independent epistemic properties; knowledge, according to thisapproach, requires a certain non-logical relationship between beliefand truth.

10.1 The “AAA” Evaluations

Sosa has often (e.g., Sosa 2007: ch. 2) made use of an analogy of askilled archer shooting at a target; we may find it instructive aswell. Here are two ways in which an archer’s shot might beevaluated:

1. Was the shot successful? Did it hit its target?
2. Did the shot’s executionmanifest thearcher’s skill? Was it produced in a way that makes it likely tosucceed?

The kind of success at issue in (1) Sosa callsaccuracy.The kind of skill discussed in (2) Sosa callsadroitness. A shot is adroit if it is produced skillfully.Adroit shots needn’t be accurate, as not all skilled shotssucceed. And accurate shots needn’t be adroit, as some unskilledshots are lucky.

In addition to accuracy and adroitness, Sosa suggests that there isanother respect in which a shot may be evaluated, relating the two.This, Sosa callsaptness.

3. Did the shot’ssuccess manifest the archer’sskill?

A shot is apt if it is accuratebecause adroit. Aptnessentails, but requires more than, the conjunction of accuracy andadroitness, for a shot might be both successful and skillful withoutbeing apt. For example, if a skillful shot is diverted by anunexpected gust of wind, then redirected towards the target by asecond lucky gust, its ultimate accuracy does not manifest the skill,but rather reflects the lucky coincidence of the wind.

Sosa suggests that this “AAA” model of evaluation isapplicable quite generally for the evaluation of any action or objectwith a characteristic aim. In particular, it is applicable to beliefwith respect to its aim at truth:

1. A belief isaccurate if and only if it is true.
2. A belief isadroit if and only if it is produced skillfully.^[30]
3. A belief isapt if and only if it is true in a waymanifesting, or attributable to, the believer’s skill.

Sosa identifies knowledge with apt belief, so understood.^[31] Knowledge entails both truth (accuracy) and justification(adroitness), on this view, but they are not merely independentcomponents out of which knowledge is truth-functionally composed. Itrequires that the skill explain the success. This is in some respectssimilar to the anti-luck condition we have examined above, in that itlegislates that the relation between justification and truth be nomere coincidence. However, insofar as Sosa’s “AAA”model is generally applicable in a way going beyond epistemology,there are perhaps better prospects for understanding the relevantnotion of aptness in a way independent of understanding knowledgeitself than we found for the notion of epistemic luck.

10.2 Fake Barn Cases

Understanding knowledge as apt belief accommodates Gettier’straditional counterexamples to the JTB theory ratherstraightforwardly. When Smith believes that either Jones owns a Fordor Brown is in Barcelona, the accuracy of his belief is notattributable to his inferential skills (which the case does not callinto question). Rather, unlucky circumstances (the misleading evidenceabout Jones’s car) have interfered with his skillful cognitiveperformance, just as the first diverting gust of wind interfered withthe archer’s shot. Compensating for the unlucky interference, alucky circumstance (Brown’s coincidental presence in Barcelona)renders the belief true after all, similar to the way in which thesecond gust of wind returns the archer’s arrow back onto theproper path towards the target.

Fake barn cases, by contrast, may be less easily accommodated bySosa’s AAA approach. When Henry looks at the only real barn in acountryside full of barn facades, he uses a generally reliableperceptual faculty for recognizing barns, and he goes right in thisinstance. Suppose we say the accuracy of Henry’s beliefmanifests his competence as a perceiver. If so, we would have to judgethat his belief is apt and therefore qualifies as an instance ofknowledge. That would be a problematic outcome because the intuitionthe case is meant to elicit is that Henry doesnot haveknowledge. There are three ways in which an advocate of the AAAapproach might respond to this difficulty.

First, AAA advocates might argue that, although Henry has a generalcompetence to recognize barns, he is deprived of this ability in hiscurrent environment, precisely because he is in fake barn county.According to a second, subtly different strategy, Henry retains abarn-recognition competence, his current location notwithstanding,but, due to the ubiquity of fake barns, his competence does notmanifest itself in his belief, since its truth is attributable more toluck than to his skill in recognizing barns.^[32] Third, Sosa’s own response to the problem is to bite thebullet. Judging Henry’s belief to be apt, Sosa accepts theoutcome that Henry knows there is a barn before him. He attempts toexplain away the counterintuitiveness of this result by emphasizingthe lack of a further epistemically valuable state, which he calls“reflective knowledge” (see Sosa 2007: 31–32).

11. Knowledge First

Not every concept is analyzable into more fundamental terms. This isclear both upon reflection on examples—what analysis could beoffered ofhydrogen,animal, orJohn F.Kennedy?—and on grounds of infinite regress. Why should wethink thatknowledge has an analysis? Proponents of theso-called “knowledge first” project, championed especiallyby Timothy Williamson’s 2000 bookKnowledge and ItsLimits, typically suppose that the project of analyzing knowledgewas a mistake. Williamson does not think that knowledge is anuninteresting state, or that the notion of knowledge is somehowfundamentally confused. On the contrary, Williamson thinks thatknowledge is among the most fundamental psychological andepistemological states there are. As such, it is a mistake to analyzeknowledge in terms of other, more fundamental epistemic notions,because knowledge itself is, in at least many cases, more fundamental.As Williamson puts it, we should put “knowledge first”.Knowledge might figure into some analyses, but it will do so in theanalysans, not in the analysandum.^[33]

There is no very straightforward argument for this conclusion; itscase consists largely in the attempted demonstration of thetheoretical success of the knowledge first stance. Weighing thesebenefits against those of more traditional approaches to knowledge isbeyond the scope of this article.^[34]

Although Williamson denies that knowledge is susceptible to analysisin the sense at issue in this article, he does think that there areinteresting and informative ways to characterize knowledge. Forexample, Williamson accepts these claims:

• Knowledge is the most general factive mental state.
• S knows thatP if and only ifS’stotal evidence includes the proposition thatP.

Williamson is also careful to emphasize that the rejection of theproject of analyzing knowledge in no way suggests that there are notinteresting and informative necessary or sufficient conditions onknowledge. The traditional ideas that knowledge entails truth, belief,and justification are all consistent with the knowledge first project.And Williamson (2000: 126) is explicit endorsing a safety requirementon knowledge—just not one that serves as part of ananalysis.

One point worth recognizing, then, is that one need not engage in theambitious project of attempting to analyze knowledge in order to havecontact with a number of interesting questions about which factors areand are not relevant for whether a subject has knowledge. In the nextsection, we consider an important contemporary debate about whetherpragmatic factors are relevant for knowledge.

12. Pragmatic Encroachment

Traditional approaches to knowledge have it that knowledge has to dowith factors like truth and justification. Whether knowledge requiressafety, sensitivity, reliability, or independence from certain kindsof luck has proven controversial. But something that all of thesepotential conditions on knowledge seem to have in common is that theyhave some sort of intimate connection with the truth of the relevantbelief. Although it is admittedly difficult to make the relevantconnection precise, there is an intuitive sense in which every factorwe’ve examined as a candidate for being relevant to knowledgehas something to do with truth of the would-be knowledgeablebeliefs.

In recent years, some epistemologists have argued that focus on suchtruth-relevant factors leaves something important out of our pictureof knowledge. In particular, they have argued that distinctivelypragmatic factors are relevant to whether a subject hasknowledge. Call this thesis “pragmatic encroachment”:^[35]

The constitution claim here is important; it is trivial thatdifferences in pragmatic circumstances cancause differencesin knowledge. For example, if the question of whether marijuana use islegal in Connecticut is more important to Sandra than it is to Daniel,Sandra is more likely to seek out evidence, and come to knowledge,than Daniel is. This uninteresting claim is not what is at issue.Pragmatic encroachment theorists think that the practical importanceitself can make for a change in knowledge, without reliance on suchdownstream effects as a difference in evidence-gathering activity.Sandra and Daniel might in some sense be inthe same epistemicposition, where the only difference is that the question is moreimportant to Sandra. This difference, according to pragmaticencroachment, might make it the case that Daniel knows, but Sandradoes not.^[36]

Pragmatic encroachment can be motivated by intuitions about cases.Jason Stanley’s 2005 bookKnowledge and PracticalInterests argues that it is the best explanation for pairs ofcases like the following, where the contrasted cases are evidentiallyalike, but differ pragmatically:

Stanley argues that the moral of cases like these is that in general,the more important the question of whetherP, the harder itis to know thatP. Other, more broadly theoretical, argumentsfor pragmatic encroachment have been offered as well. Jeremy Fantl& Matthew McGrath (2009) argue that encroachment follows fromfallibilism and plausible principles linking knowledge and action,while Brian Weatherson (2012, 2024) argues that the bestinterpretation of decision theory requires encroachment.

Pragmatic encroachment is not an analysis of knowledge; it is merelythe claim that pragmatic factors are relevant for determining whethera subject’s belief constitutes knowledge. Some, but not all,pragmatic encroachment theorists will endorse a necessarybiconditional that might be interpreted as an analysis of knowledge.For example, a pragmatic encroachment theorist might claim that:

S knows thatP if and only if no epistemic weaknesswith respect to whetherP preventsS fromproperly using the proposition thatP as a reasonfor action.

This connection between knowledge and action is similar to onesendorsed by Fantl & McGrath (2009), but it is stronger thananything for which they argue.

Pragmatic encroachment on knowledge is deeply controversial. PatrickRysiew (2001), Jessica Brown (2006), and Mikkel Gerken (2017) haveargued that traditional views about the nature of knowledge aresufficient to account for the data mentioned above. MichaelBlome-Tillmann (2009a) argues that it has unacceptablycounterintuitive results, like the truth of such claims asS knowsthat P,but if it were more important, she wouldn’tknow, orS knew thatP until the question becameimportant. Stanley (2005) offers strategies for accepting suchconsequences. Other, more theoretical arguments against encroachmenthave also been advanced. For example, Ichikawa, Benjamin Jarvis, &Katherine Rubin (2012) argue that pragmatic encroachment is at oddswith important tenets of belief-desire psychology.

13. Contextualism

One final topic standing in need of treatment is contextualism aboutknowledge attributions, according to which the word“knows” and its cognates are context-sensitive. Therelationship between contextualism and the analysis of knowledge isnot at all straightforward. Arguably, they have different subjectmatters (the former a word, and the latter an epistemic relation ormental state). Nevertheless, the methodology of theorizing aboutknowledge may be helpfully informed by semantic considerations aboutthe language in which such theorizing takes place. And ifcontextualism is correct, then a theorist of knowledge must attendcarefully to the potential for ambiguity.

It is uncontroversial that many English words are context-sensitive.The most obvious cases are indexicals, such as “I”,“you”, “here”, and “now” (DavidKaplan (1977) gives the standard view of indexicals).

The word “you” refers to a different person, depending onthe conversational context in which it is uttered; in particular, itdepends on the person one is addressing. Other context-sensitive termsare gradable adjectives like “tall”—how tallsomething must be to count as “tall” depends on theconversational context—and quantifiers like“everyone”—which people count as part of“everyone” depends on the conversational context.Contextualists about “knows” think that this verb belongson the list of context-sensitive terms. A consequence of contextualismis that sentences containing “knows” may express distinctpropositions, depending on the conversational contexts in whichthey’re uttered. This feature allows contextualists to offer aneffective, though not uncontroversial, response to skepticism. For amore thorough overview of contextualism and its bearing on skepticism,see Ichikawa 2017-b or Rysiew 2023.

Contextualists have modeled this context-sensitivity in various ways.DeRose (2009) has suggested that there is a context-invariant notionof “strength of epistemic position”, and that how strong aposition one must be in in order to satisfy “knows” variesfrom context to context; this is, in effect, to understand thesemantics of knowledge attributions much as we understand that ofgradable adjectives. (How much height one must have to satisfy“tall” also varies from context to context.) Stewart Cohen(1988) adopts a contextualist treatment of “relevantalternatives” theory, according to which, in skeptical contexts,but not ordinary ones, skeptical possibilities are relevant. Thisaspect is retained in the view of Lewis 1996, which characterizes acontextualist approach that is more similar to quantifiers and modals.Blome-Tillmann 2009-b and Ichikawa 2017-a defend and develop theLewisian view in different ways.

Contextualism and pragmatic encroachment represent differentstrategies for addressing some of the same “shifty”patterns of intuitive data. (In fact, contextualism was generallydeveloped first; pragmatic encroachment theorists were motivated inpart by the attempt to explain some of the patterns contextualistswere interested in without contextualism’s semanticcommitments.) Although this represents a sense in which they tend tobe rival approaches, contextualism and pragmatic encroachment are byno means inconsistent. One could think that “knows”requires the satisfaction of different standards in differentcontexts, andalso think that the subject’s practicalsituation is relevant for whether a given standard is satisfied.

Like pragmatic encroachment, contextualism is deeply controversial.Critics have argued that it posits an implausible kind of semanticerror in ordinary speakers who do not recognize the putativecontext-sensitivity—see Schiffer 1996 and Greenough &Kindermann 2017—and that it is at odds with plausibletheoretical principles involving knowledge—see Hawthorne 2003,Williamson 2005, and Worsnip 2017. In addition, some of the argumentsthat are used to undercut the data motivating pragmatic encroachmentare also taken to undermine the case for contextualism; see againRysiew 2001 and Brown 2006.

Bibliography

• Almeder, Robert, 1999,Harmless Naturalism: The Limits ofScience and the Nature of Philosophy, Chicago and La Salle: OpenCourt.
• Alston, William P., 1991,Perceiving God: The Epistemology ofReligious Experience, Ithaca: Cornell University Press.
• Antognazza, Maria Rosa, 2015, “The Benefit to Philosophy ofthe Study of its History”,British Journal for the Historyof Philosophy, 23(1):161–184.doi:10.1080/09608788.2014.974020
• –––, 2020, “The Distinction in Kindbetween Knowledge and Belief”,Proceedings of theAristotelian Society, 120(3): 2020:277–308.doi:10.1093/arisoc/aoaa013
• Armstrong, D.M., 1973,Belief, Truth, and Knowledge,Cambridge: Cambridge University Press.
• Becker, Kelly, and Tim Black (eds.), 2012,TheSensitivity Principle in Epistemology. New York: CambridgeUniversity Press.
• Bird, Alexander, 2007, “Justified Judging”,Philosophy and Phenomenological Research, 74(1):81–110. doi:10.1111/j.1933-1592.2007.00004.x
• Blome-Tillmann, Michael, 2009a, “Contextualism,Subject-Sensitive Invariantism, and the Interaction of‘Knowledge’-Ascriptions with Modal and TemporalOperators”,Philosophy and Phenomenological Research,79(2): 315–331. doi:10.1111/j.1933-1592.2009.00280.x
• –––, 2009-b, “Knowledge andPresuppositions”,Mind, 118(470): 241–294.doi:10.1093/mind/fzp032
• Boh, Ivan, 1985, “Belief, Justification and Knowledge: SomeLate Medieval Epistemic Concerns”,Journal of the RockyMountain Medieval and Renaissance Association, 6:87–103.
• Bogardus, Tomas, 2014, “Knowledge under Threat”,Philosophy and Phenomenological Research, 88(2):289–313. doi:10.1111/j.1933-1592.2011.00564.x
• Brown, J., 2006, “Contextualism and Warranted AssertabilityManoeuvres”,Philosophical Studies, 130(3):407–435. doi:10.1007/s11098-004-5747-3
• Chisholm, Roderick, 1989,Theory of Knowledge,3^rd edition, Englewood Cliffs: Prentice Hall.
• Clark, Michael, 1963, “Knowledge and Grounds: A Comment onMr. Gettier’s Paper”,Analysis, 24(2):46–48. doi:10.2307/3327068
• Clarke, Roger, 2013, “Belief Is Credence One (inContext)”,Philosophers’ Imprint, 13:1–18.
• Cohen, Stewart, 1988 “How to Be a Fallibilist”,Philosophical Perspectives, Epistemology, 2: 91–123.doi:10.2307/2214070
• Comesaña, Juan, 2005, “Unsafe Knowledge”,Synthese, 146(3): 395–404.doi:10.1007/s11229-004-6213-7
• Conee, Earl, and Richard Feldman, 2004,Evidentialism:Essays in Epistemology, Oxford: Oxford University Press.doi:10.1093/0199253722.001.0001
• –––, 2001, “Internalism Defended”,American Philosophical Quarterly, 38(1): 1–18.Reprinted in Conee and Feldman 2004, pp. 53–82.doi:10.1093/0199253722.003.0004
• Craig, Edward, 1990,Knowledge and the State of Nature: AnEssay in Conceptual Synthesis, Oxford: Oxford University Press.doi:10.1093/0198238797.001.0001
• Das, Nilanjan, 2021a, “Śrīharṣa”,TheStanford Encyclopedia of Philosophy (Winter 2021 Edition),Edward N. Zalta (ed.), URL =<https://plato.stanford.edu/archives/win2021/entries/sriharsa/>.
• –––, 2021b,“Gaṅgeśa”,Journal of IndianPhilosophy, 49(2): 153–202.doi: 10.1007/s10781-021-09461-6
• DeRose, Keith, 1995, “Solving the Skeptical Problem”,The Philosophical Review, 104(1): 1–52.doi:10.2307/2186011
• –––, 2000, “Ought We to Follow OurEvidence?”Philosophy and Phenomenological Research,60(3): 697–706. doi:10.2307/2653824
• –––, 2009,The Case for Contextualism,New York: Oxford University Press.doi:10.1093/acprof:oso/9780199564460.001.0001
• Dougherty, Trent (ed.), 2011,Evidentialism and ItsDiscontents, New York: Oxford University Press.doi:10.1093/acprof:oso/9780199563500.001.0001
• Dretske, Fred, 1981,Knowledge and the Flow ofInformation, Cambridge, MA: MIT Press.
• –––, 1985, “Précis ofKnowledgeand the Flow of Information”, in Hilary Kornblith (ed.),Naturalizing Epistemology, Cambridge, MA: MIT Press,pp. 169–187.
• –––, 1989, “The Need to Know”, inMarjorie Clay and Keith Lehrer (eds.),Knowledge andSkepticism, Boulder: Westview Press, pp. 89–100.
• –––, 2005, “The Case againstClosure”, in Matthias Steup and Ernest Sosa(eds).Contemporary Debates in Epistemology,1^st edition, Malden, MA: Blackwell,pp. 13–25.
• Dreyfus, George B.J., 1997,Recognizing Reality:Dharmakirti’s Philosophy and Its Tibetan Interpretations,Albany, NY: SUNY Press.
• Dutant, Julien, 2015, “The Legend of the Justified TrueBelief Analysis”,Philosophical Perspectives, 29(1):95–145. doi:10.1111/phpe.12061
• Fantl, Jeremy, and Matthew McGrath, 2009,Knowledge in anUncertain World, New York: Oxford University Press.doi:10.1093/acprof:oso/9780199550623.001.0001
• Feldman, Richard, and Earl Conee, 1985,“Evidentialism”,Philosophical Studies, 48(1):15–34. doi:10.1007/BF00372404
• Fodor, Jerry, 1998,Concepts: Where Cognitive Science WentWrong, New York: Oxford University Press.doi:10.1093/0198236360.001.0001
• Gardiner, Georgi, 2020, “Relevance and Risk: How theRelevant Alternatives Framework Models the Epistemology ofRisk,”Synthese, 199 (1–2): 481–511.doi:10.1007/s11229-020-02668-2
• Gerken, Mikkel, 2017,On Folk Epistemology, Oxford:Oxford University Press.
• Gettier, Edmund L., 1963, “Is Justified True BeliefKnowledge?”,Analysis, 23(6): 121–123.doi:10.2307/3326922
• Gibbons, John, 2013,The Norm of Belief, OxfordUniversity Press. doi:10.1093/acprof:oso/9780199673391.001.0001
• Goldman, Alvin I., 1967, “A Causal Theory ofKnowing”,Journal of Philosophy, 64(12):357–372. doi:10.2307/2024268
• –––, 1976, “Discrimination and PerceptualKnowledge”,Journal of Philosophy, 73(20):771–791. doi:10.2307/2025679
• –––, 1979, “What Is JustifiedBelief?” inJustification and Knowledge, George S.Pappas (ed.), Dordrecht: D. Reidel, pp. 1–25.
• –––, 1986,Epistemology and Cognition,Cambridge, MA: Harvard University Press.
• –––, 1999, “InternalismExposed”,Journal of Philosophy, 96(6):271–293. doi:10.2307/2564679
• –––, 2009, “Internalism, Externalism, andthe Architecture of Justification”,Journal ofPhilosophy, 106(6): 309–338.doi:10.5840/jphil2009106611
• Goldman, Alvin I., and Bob Beddor, 2021, “ReliabilistEpistemology”,The Stanford Encyclopedia ofPhilosophy (Summer 2021 Edition), Edward N. Zalta (ed.), URL= <https://plato.stanford.edu/archives/sum2021/entries/reliabilism/>.
• Goldman, Alvin I., and Erik J. Olsson, 2009, “Reliabilismand the Value of Knowledge”, in Adrian Haddock, Alan Millar, andDuncan Pritchard (eds.),Epistemic Value, New York: OxfordUniversity Press, pp. 19–41.doi:10.1093/acprof:oso/9780199231188.001.0001
• Greco, John, 2009, “Knowledge and Success fromAbility”,Philosophical Studies, 142(1): 17–26.doi:10.1007/s11098-008-9307-0
• Greenough, Patrick, and Dirk Kindermann, 2017, “The SemanticError Problem for Epistemic Contextualism”, in Ichikawa 2017-b,pp. 305–320.
• Greenough, Patrick, and Duncan Pritchard (eds.), 2009,Williamson on Knowledge, Oxford: Oxford University Press.doi:10.1093/acprof:oso/9780199287512.001.0001
• Hazlett, Allan, 2010, “The Myth of Factive Verbs”,Philosophy and Phenomenological Research, 80(3):497–522. doi:10.1111/j.1933-1592.2010.00338.x
• Hawthorne, John, 2002, “Deeply Contingent A PrioriKnowledge”,Philosophy and Phenomenological Research,65(2): 247–269. doi:10.1111/j.1933-1592.2002.tb00201.x
• –––, 2003,Knowledge and Lotteries, NewYork: Oxford University Press. doi:10.1093/0199269556.001.0001
• Hirvelä, Jaakko, 2025, “Metaphysics of Risk andLuck,”Noûs, 59(2) :335–348.doi:10.111/nous.12516
• Ichikawa, Jonathan, 2011, “Quantifiers, Knowledge, andCounterfactuals”,Philosophy and PhenomenologicalResearch, 82(2): 287–313.doi:10.1111/j.1933-1592.2010.00427.x
• –––, 2017-a,Contextualising Knowledge:Epistemology and Semantics, Oxford: Oxford University Press.
• ––– (ed.), 2017-b,Routledge Handbook ofEpistemic Contextualism, New York: Taylor & Francis.
• –––, 2024,Epistemic Courage, OxfordUniversity Press.
• Ichikawa, Jonathan, and Benjamin Jarvis, 2013,The Rulesof Thought, Oxford: Oxford University Press.
• Ichikawa, Jonathan, Benjamin Jarvis, and Katherine Rubin,2012, “Pragmatic Encroachment and Belief-DesirePsychology”,Analytic Philosophy, 53(4): 327–342.doi:10.1111/j.2153-960X.2012.00564.x
• Ichikawa, Jonathan, and Carrie Jenkins, 2017, “On PuttingKnowledge ‘First’, in J. Adam Carter, Emma C. Gordon,and Benjamin Jarvis (eds.),Knowledge First: Approaches inEpistemology and Mind, Oxford: Oxford University Press, pp.113–131. doi: 10.1093/oso/9780198716310.003.0006
• Jackson, Elizabeth G., 2020, “The Relationship BetweenBelief and Credence”,Philosophy Compass, 15(6):1–13. doi:10.1111/phc3.12668
• Kaplan, David, 1977, “Demonstratives”, paper presentedat a symposium on Demonstratives at the March 1977 meetings of thePacific Division of the American Philosophical Association. Printed inJoseph Almog, John Perry & Howard Wettstein (eds.), 1989,Themes From Kaplan, Oxford: Oxford University Press,pp. 481–563.
• Kaplan, Mark, 1985, “It’s Not What You Know ThatCounts”,Journal of Philosophy, 82(7):350–63. doi:10.2307/2026524
• Kornblith, Hilary, 2002,Knowledge and Its Place inNature, Oxford: Oxford University Press.doi:10.1093/0199246319.001.0001
• –––, 2008, “Knowledge Needs NoJustification”, in Quentin Smith (ed.),Epistemology: NewEssays, Oxford: Oxford University Press, pp. 5–23.doi:10.1093/acprof:oso/9780199264933.003.0002
• Kripke, Saul A., 2011, “Nozick on Knowledge”, inPhilosophical Troubles: Collected Papers, Volume 1, Oxford:Oxford University Press, pp. 162–224.doi:10.1093/acprof:oso/9780199730155.003.0007
• Le Morvan, Pierre, 2017, “Knowledge before Gettier”,British Journal for the History of Philosophy, 25(6):1216–1238. doi: 10.1080/09608788.2017.1320968
• Lewis, David K., 1973,Counterfactuals, Cambridge, MA:Harvard University Press.
• –––, 1996, “Elusive Knowledge”,Australasian Journal of Philosophy, 74(4): 549–567.doi:10.1080/00048409612347521
• Littlejohn, Clayton, 2012,Justification and theTruth-Connection, Cambridge: Cambridge University Press.
• Lowy, Catherine, 1978, “Gettier’s Notion ofJustification”,Mind, 87(345): 105–108.doi:10.1093/mind/LXXXVII.1.105
• Lyons, Jack C., 2009,Perception and Basic Beliefs: Zombies,Modules and the Problem of the External World, New York: OxfordUniversity Press. doi:10.1093/acprof:oso/9780195373578.001.0001
• McGlynn, Aidan, 2014,Knowledge First?, Basingstoke:Palgrave. doi:10.1057/9781137026460
• McKinnon, Rachel, 2013, “Getting Luck Properly underControl”,Metaphilosophy, 44(4): 496–511.doi:10.1111/meta.12044
• Merricks, Trenton, 1995, “Warrant Entails Truth”,Philosophy and Phenomenological Research, 55(4):841–855. doi:10.2307/2108335
• Miracchi, Lisa, 2015, “Competence to Know”,Philosophical Studies, 172(1): 29–56.doi:10.1007/s11098-014-0325-9
• Myers-Schulz, Blake, and Eric Schwitzgebel, 2013,“Knowing ThatP without Believing ThatP”,Noûs, 47(2): 371–384.doi:10.1111/nous.12022
• Nagel, Jennifer, 2010, “Epistemic Anxiety and AdaptiveInvariantism”,Philosophical Perspectives, 24(1):407–435. doi:10.1111/j.1520-8583.2010.00198.x
• –––, 2013, “Knowledge as a MentalState”,Oxford Studies in Epistemology, 4:275–310. doi:10.1093/acprof:oso/9780199672707.003.0010
• Navarro, Jesús, 2023, “Epistemic Luck and EpistemicRisk”,Erkenntnis, 88(3): 929–950. doi:10.1007/s10670-021-00387-9
• Neta, Ram & Rohrbaugh, Guy, 2004, “Luminosityand the Safety of Knowledge,”Pacific PhilosophicalQuarterly, 85: 396–406.doi: 10.1111/j.1468-0114.2004.00207.x
• Nozick, Robert, 1981,Philosophical Explanations,Cambridge, MA: Harvard University Press.
• Owens, David, 2000,Reason without Freedom: The Problem ofEpistemic Normativity, London: Routledge.
• Pappas, George, “Internalist vs. Externalist Conceptions ofEpistemic Justification”,The Stanford Encyclopedia ofPhilosophy (Spring 2023 Edition), Edward N. Zalta & UriNodelman (eds.), URL =<https://plato.stanford.edu/archives/spr2023/entries/justep-intext/>.
• Phillips, Stephen H., and N.S. Ramanuja Tatacharya (trans.), 2004,Epistemology of Perception: Gaṅgeśa’s“Tattvacintāmaṇi”, New York: AmericanInstitute of Buddhist Studies.
• Plantinga, Alvin, 1993,Warrant and Proper Function,Oxford: Oxford University Press. doi:10.1093/0195078640.001.0001
• –––, 1996, “Respondeo”, in JonathanL. Kvanvig (ed.),Warrant in Contemporary Epistemology. Essays inHonor of Plantinga’s Theory of Knowledge, Lanham, MD:Rowman and Littlefield, pp. 307–378.
• Pollock, John J., 1986,Contemporary Theories ofKnowledge, Totowa, NJ: Rowman and Littlefield.
• Pritchard, Duncan, 2005,Epistemic Luck, New York: OxfordUniversity Press. doi:10.1093/019928038X.001.0001
• –––, 2016, “Epistemic Risk”,Journal of Philosophy, 113(11): 550–571.doi:10.5840/jphil20161131137
• Pritchard, Duncan, Alan Millar, and Adrian Haddock, 2010,The Nature and Value of Knowledge: Three Investigations, NewYork: Oxford University Press.doi:10.1093/acprof:oso/9780199586264.001.0001
• Radford, Colin, 1966, “Knowledge—By Examples”,Analysis, 27(1): 1–11. doi:10.2307/3326979
• Rose, David, and Jonathan Schaffer, 2013, “KnowledgeEntails Dispositional Belief”,Philosophical Studies,166(Supplement 1): 19–50. doi:10.1007/s11098-012-0052-z
• Roush, Sherrilyn, 2005,Tracking Truth: Knowledge, Evidence,and Science, Oxford: Oxford University Press.
• –––, 2012, “Sensitivity andClosure,” in Kelly Becker and Tim Black (eds.),TheSensitivity Principle in Epistemology, New York: CambridgeUniversity Press, pp. 242–268
• Russell, Bertrand, 1923,Human Knowledge: Its Scope andLimits, London: George Allen and Unwin Ltd.
• Rysiew, Patrick, 2001, “The Context-Sensitivity of KnowledgeAttributions”,Noûs, 35(4): 477–514.doi:10.1111/0029-4624.00349
• –––, 2023, “EpistemicContextualism”,The Stanford Encyclopedia of Philosophy(Winter 2023 Edition), Edward N. Zalta and Uri Nodelman (eds.), URL =<https://plato.stanford.edu/archives/win2023/entries/contextualism-epistemology/>.
• Schaffer, Jonathan, 2004, “From Contextualism toContrastivism”,Philosophical Studies, 119(1–2):73–104. doi:10.1023/B:PHIL.0000029351.56460.8c
• Schiffer, Stephen, 1996, “Contextualist Solutions toScepticism”,Proceedings of the Aristotelian Society,96(1): 317–333. doi:10.1093/aristotelian/96.1.317
• Schulz, Moritz, 2021, “Strong Knowledge, WeakBelief?”,Synthese, 199(3–4): 8741–8753.doi:10.1007/s11229-021-03180-x
• Shope, Robert K., 1983,The Analysis of Knowing: A Decade ofResearch, Princeton: Princeton University Press.
• Siscoe, Robert Weston, 2021, “Belief, Rational andJustified”,Mind, 130(517): 59–83.doi:10.1093/mind/fzaa021
• Sosa, Ernest, 1999, “How to Defeat Opposition toMoore”,Philosophical Perspectives, 13: 141–153.doi:10.1111/0029-4624.33.s13.7
• –––, 2007,A Virtue Epistemology: Apt Beliefand Reflective Knowledge (Volume I), New York: Oxford UniversityPress. doi:10.1093/acprof:oso/9780199297023.001.0001
• Stanley, Jason, 2005,Knowledge and Practical Interests,New York: Oxford University Press.doi:10.1093/0199288038.001.0001
• –––, 2011,Know How, New York: OxfordUniversity Press. doi:10.1093/acprof:oso/9780199695362.001.0001
• Steglich-Petersen, Asbjørn, 2010, “Luck as anEpistemic Notion”,Synthese, 176 (3): 361–377.doi:10.1007/s11229-009-9569-x
• Stine, G.C., 1976, “Skepticism, Relevant Alternatives, andDeductive Closure”,Philosophical Studies, 29(4):249–261. doi:10.1007/BF00411885
• Sturgeon, Scott, 1993, “The Gettier Problem”,Analysis, 53(3): 156–164. doi:10.2307/3328464
• Sutton, Jonathan, 2007,Without Justification, Cambridge,MA: The MIT Press.
• Unger, Peter, 1968, “An Analysis of FactualKnowledge”,Journal of Philosophy, 65(6):157–170. doi:10.2307/2024203
• Weatherson, Brian, 2012, “Knowledge, Bets, andInterests”, in Jessica Brown and Mikkel Gerken(eds.),Knowledge Ascriptions, New York: OxfordUniversity Press, pp. 75–103.doi:10.1093/acprof:oso/9780199693702.003.0004
• –––, 2024,Knowledge: A Human InterestStory, Open Book Publishers.
• Williamson, Timothy, 2000,Knowledge and Its Limits,Oxford: Oxford University Press. doi:10.1093/019925656X.001.0001
• –––, 2005, “Knowledge, Context, and theAgent’s Point of View”, in Gerhard Preyer and GeorgPeter (eds.),Contextualism in Philosophy: Knowledge, Meaning, andTruth, Oxford: Oxford University Press,pp. 91–114.
• –––, 2009, “Reply to Alvin Goldman”,in Greenough & Pritchard 2009, pp. 305–312.doi:10.1093/acprof:oso/9780199287512.003.0017
• Worsnip, Alex, 2017, “Contextualism and KnowledgeNorms”, in Ichikawa 2017-b, pp. 177-189.
• Wright, Crispin, 2004, “Warrant for Nothing (And Foundationsfor Free)?”Aristotelian Society Supplementary Volume,78(1): 167–212. doi:10.1111/j.0309-7013.2004.00121.x
• Zagzebski, Linda, 1994, “The Inescapability of GettierProblems”,The Philosophical Quarterly, 44(174):65–73. doi:10.2307/2220147

Academic Tools

	How to cite this entry.
	Preview the PDF version of this entry at theFriends of the SEP Society.
	Look up topics and thinkers related to this entry at the Internet Philosophy Ontology Project (InPhO).
	Enhanced bibliography for this entryatPhilPapers, with links to its database.

Other Internet Resources

• The Epistemology Research Guide

Related Entries

contextualism, epistemic |epistemic closure |epistemology: naturalism in |epistemology: social |epistemology: virtue |justification, epistemic: coherentist theories of |justification, epistemic: foundationalist theories of |justification, epistemic: internalist vs. externalist conceptions of |skepticism: and content externalism

Acknowledgments

As of the 2026 update, Jonathan Ichikawa has taken sole responsibilityfor updating and maintaining this entry. He especially thanks KurtSylvan for substantive suggestions and references, and AshleyLindsley-Kim for editorial help and research assistance. The 2012 and2017 versions were prepared collaboratively by Jonathan Ichikawa andMatthias Steup. For the 2012 version, the authors were grateful toKurt Sylvan for extremely detailed and constructive comments onmultiple drafts of this entry. Thanks also to an anonymous referee foradditional helpful suggestions. For the 2017 revision, the authorsgive thanks to Clayton Littlejohn, Jennifer Nagel, and Scott Sturgeonfor helpful and constructive feedback and suggestions. Thanks to BenBayer, Kenneth Ehrenberg, and Mark Young for drawing our attention toerrors in the previous version.