The definition of knowledge



Philosophers have a tendency to treat questions that revolve around the meaning of a particular concept as if they are more substantive than they in fact are. They get fascinated by some concept, such as knowledge, truth, beauty, or goodness, and then spend the rest of their lives—and, collectively, centuries or millennia—analyzing it. Some analysis, of course, is perfectly legitimate and can be interesting; but I think philosophers often go overboard in their never-ending obsession with such things as the nature of knowledge or truth or goodness or whatever. After all, these are just concepts, words that have arisen in the course of history, not objective realities such as science investigates. It seems disproportionate to obsess for years over a word like “knowledge”; why not obsess over the subtleties in meaning of any other concept, like house or book or city or river? Nearly all concepts are peculiar and have paradoxical features. Like, the concept of a river is odd: when is a body of water a river? Why is this body a creek and not a river? (Someone might say it’s a mere creek, another might disagree.) If you drain all the water out and then put in new water, is it the same river? What if you change the course of the river, even drastically? Is it the same river? Every concept, it seems, is puzzling in some respect or other. So why focus on knowledge, truth, and goodness? They’re “constructed” like any other word, just as puzzling as every other word. You can, perhaps, tease out some conceptual “essence,” but often the word will remain irremediably vague or, in some cases, contradictory, since a concept is an empirical object of natural language (and natural language is hardly a model of clarity or precision).


In short, I think much of philosophy is misguided. This is, in effect, what I argued in the following paper (in 2006), from a class I took many years ago in epistemology. [See also this post.] If you tease out the conceptual dimensions of the word ‘knowledge,’ you’ll find that knowledge, in the strictest sense, is impossible, and philosophers should stop trying to define it and explaining how it’s possible, what it means, what standards determine when something counts as knowledge, etc. All the obsession over the meaning of an empirical concept is almost unseemly; it’s like, leave the defining to the lexicographers, and go do something more productive with your time. Besides, there’s something a little neurotic, I think, about wanting absolute clarity on when it is you can justifiably attribute knowledge to yourself and when you can’t. This need for certainty—and intense aversion to philosophical skepticism—is a little off-putting.


Readers of my writings (what few readers there are) know I have a rather low opinion of intellectuals as a social category. Whether economists, philosophers, political scientists, "humanists," or whoever, they’re usually wrong and misguided.

***


Upon surveying the contemporary epistemological scene, one is struck by how many debates involve disagreement about the definition of ‘knowledge’. The internalism-externalism debate, for example, is made possible by conceptual ambiguity. Pure externalists claim that one need not have reasons (in the standard sense) for a belief in order for it to count as knowledge; what matters is that there be a certain kind of causal connection between the belief and the world.[1] Internalists, on the other hand, believe that the notion of knowledge implies that the believer has access to good reasons for his belief. Another example is the debate between contextualists and invariantists. Contextualists actually incorporate the ambiguity of ‘knowledge’ into their position: they claim that the standards of knowledge-attributions vary with the attributer’s context. This means that the truth-conditions of such attributions change, which means, effectively, that their meaning changes as well. Invariantists, of course, deny this; knowledge has the same standards in all contexts. More generally, philosophers tend to bandy about the word ‘knowledge’ as if its meaning is unproblematic; a typical example is this sentence of John Hawthorne’s: “Someone might know that his partner will never leave him and this may entail that she will never leave him for Mr. X”.[2] What does ‘know’ mean here? No explanation is offered. In this paper I will attempt to remedy the situation by giving an original analysis of knowledge, the main conclusion of which will be that if the word ‘knowledge’, as used by ordinary people, is taken to its logical conclusion, it is meaningless, and thus unachievable. I will then apply my ideas to a few standard epistemological problems.

The definition of knowledge


The best method of defining a word is to observe how it is commonly used. ‘Knowledge’ is used in fairly specific ways, which should make the task of defining it straightforward. In the sense in which we are interested in it, only propositions that are believed (and perhaps facts as well) can be objects of knowledge. The propositions must, moreover, be true: I cannot know that p if p is false. The third condition is that the believer not have any doubt about the truth of the proposition in question. This is illustrated by the plausibility of the following dialogue:


“I know that that’s a tree over there.”

“Really? You don’t have any doubt whatsoever? You’re certain of it?”

“Well, I’m not completely certain. I suppose it could be a tall bush.”

“Then you don’t know it’s a tree.”


If I doubt the truth of a proposition, obviously I do not know it. Therefore, my feeling certain is the third condition for attributing knowledge to me.[3] The fourth is internal justification. The following dialogue illustrates this. [I deleted the dialogue because of its excessive length.] [….]


In short, the fourth condition of knowledge is that the knower have the ability to (truly and justifiably) attribute knowledge to himself, which requires (1) that he be aware of all the information relevant[4] to his belief, and (2) (by implication) that he base his belief on the right reasons. In other words, for knowledge to be justifiably attributed to him, he must be able to know that he knows. In this context, “be able to know” means that in the moment in which he knows that p, he has access to all the evidence, all the information, all the reasons relevant to his knowledge that p, such that if someone asked him “Do you know that p?”, he would be able to adduce everything in p’s favor and everything against it and explain how it is that he knows that p.


The reader may object, “What are your reasons for introducing this severe fourth condition of knowledge (viz., that the knower be able to justifiably attribute knowledge to himself)? You still have not argued for it.” Actually, I have. I noted in the dialogue above that people do not commonly attribute knowledge to infants and animals (except in a loose sense); the reason can only be that infants and animals cannot know that they have knowledge. If they could (i.e., if they were self-conscious), then surely there would be no reason not to attribute knowledge to them. Now, in order for a being to know it has knowledge that p, it must have access to all the facts relevant to its knowledge that p: it must know, for example, what were the direct causes of its belief that p, what justification there is for believing that p, and so on. –More generally, however, my argument for the fourth condition is simply that people use the word ‘knowledge’ in such a way that this fourth condition is implied. For instance, imagine this dialogue between Dick Cheney and Donald Rumsfeld:


Dick: “John knows that we manipulated the intelligence we used to justify the war in Iraq.”


Donald: “Actually, he doesn’t know it yet. He thinks we did, but he can’t prove it.”


Although Dick and Donald know that John’s belief is true, they refrain from attributing knowledge to him because they know he does not attribute knowledge to himself. The implication of this example is that, in order for someone to have knowledge, he must be able to say, justifiably, that he knows he has knowledge.


The reader may have noticed that the fourth condition is suspiciously similar to the third: a knower’s ability to ascribe knowledge to himself seems equivalent to his certainty that his belief is true. There is a subtle difference, however. As mentioned above, certainty is a purely subjective state, while justifiable self-ascription of knowledge is both subjective and objective. It is a unity of the two: it is subjective awareness of the objective validity of one’s judgment that one is in a state of knowledge. The fourth condition implies the third, but the third does not imply the fourth.


I will anticipate here the later discussion of internalism and externalism, because otherwise the reader might object that I have ignored externalism in my definition of knowledge. In truth, I have merely combined the two opposing positions: I have incorporated externalism into an internalist definition. “Which form of externalism?” asks the reader. “Not any form in particular,” I reply. The knower must simply know all “external” facts relevant to his knowledge that p: for example, that it is not accidental that he is right about its being the case that p.[5] It would be superfluous, however, to list all the conditions commonly invoked by externalists in order to describe what it is that the knower knows; the point is that he has perfect knowledge of the portion of reality that bears on his belief that p. Similarly, anyone who justifiably attributes knowledge to him must have the same complete knowledge that he has. This criterion of “omniscience” (with regard to a particular aspect of reality) is not arbitrary or absurd; it is implied by the meaning of the verb ‘to know’.

The meaninglessness of knowledge

Having outlined the definition of knowledge, I will now expound its implication—viz., that knowledge in the strictest sense is impossible for humans to achieve. The reason for this may not be evident: it is not that omniscience (in the limited form under consideration) is impossible—for, after all, this is not necessarily the case—but rather that it is in the nature of homo sapiens not to be able to know if/when omniscience has been achieved. This is due to the peculiar nature of consciousness. Humans, unlike other animals, are self-conscious: they are aware of themselves. In being self-aware, they are also necessarily aware of the externality of the world that confronts them. They are aware of it as an “other” thing, a thing to which they do not have the same direct, intuitive access that they have to their own consciousness. That is, they do not have (intuitive) access to the “interior” of this object, as they do to their (self-)consciousness. Because the world (i.e., reality[6]) is necessarily seen by self-consciousness as external to it, truth is seen as external—for truth is but a propositional expression of (a given aspect of) the world. This means that the negation of a given proposition is always conceivable. Even propositions about (self-)consciousness can conceivably be negated, because, although consciousness has intuitive access to the “interior” of itself (as it does not to that of the external world), it is also its own other. It observes itself, it is opposed to itself; therefore, it is “opposed to” (or outside of) any truth about itself, and thus any proposition. Now, to say that the negation of a proposition is always conceivable is to say that the person who asserts the proposition may be wrong, which is effectively to say that he may not be aware of all the facts relevant to the truth or falsity of the proposition. This is, moreover, intuitively compelling. We can never know whether we have grasped all the facts relevant to a given proposition, because, as already stated, facts are external to us; thus, we would have to step outside ourselves to determine (beyond doubt) whether we had understood a given aspect of reality. We would have to attain a God-like perspective, from which negation would be inconceivable.[7] Because this is impossible—because, that is, negation is always conceivable—knowledge is impossible.


There is another argument for that conclusion, though it amounts to the same thing as the foregoing. Certainty, as we have seen, is a necessary condition for knowledge. In other words, S does not know that p unless S has no doubts that p. What does it mean to doubt that p? It means to be aware of the possibility that p is false. Now, we have seen that in the case of every proposition p, it is possible for S to be aware that p may be false, simply by imagining not-p. Whether S is in fact aware of this has no bearing on the question of whether S has knowledge, because the relevant condition for knowledge is that S can justifiably attribute knowledge to himself, not that S does attribute knowledge to himself. All that matters, then, is that S can conceive that not-p: this is enough to show that S cannot know that p (because imagining that not-p is equivalent to doubting that p, and doubt is incompatible with knowledge). In fact, complete certainty is, arguably, inconceivable, and hence meaningless. What would it be like not to be aware of the possibility that any proposition p can be false? All that is necessary for this awareness is that one understand the meaning of not-p; but what would it be like not to understand the meaning of not-p? That seems unimaginable, meaningless; thus, complete certainty is meaningless, which means that knowledge itself is. –In short, it will always be possible to doubt that p (and an absence of doubt is meaningless); therefore, it will never be possible to know that p—and, moreover, knowledge that p is meaningless.


Maybe my definition of doubt seems too loose. It permits such statements as “I see a computer in front of me, but because I can imagine the truth of the proposition ‘I do not see a computer in front of me’, in a sense I doubt there is a computer in front of me—even though I believe there is.” It seems strange to believe a proposition but doubt it at the same time. Nevertheless, there is nothing paradoxical about this. I believe that the universe was created in a “big bang” but admit I may be wrong—which means, effectively, that my belief is subject to doubt. Similarly, it is possible to believe that I am writing on a computer and yet to doubt it, albeit in some nearly “empty” way. In any case, the fact that philosophers debate these sorts of questions—that they ask themselves, “Do I know there is a computer here?”, “Do I know I am not a brain in a vat?”—proves they have some kind of doubt—and, as we have seen, all kinds are incompatible with knowledge.


The question remains, however, why these epistemological problems—namely, the debates over skepticism, over internalism and externalism, over contextualism, over the proper definition of knowledge—have proven as recalcitrant as they have. Why has so little progress been made by epistemologists in the last forty years? The answer is that they have been approaching the problems from the wrong angle. They have been searching for a (single) account of knowledge that can both preserve our knowledge of ordinary truths and explain the power of the skeptic’s intuitions. They have been trying to formulate a single definition of knowledge—and justification—that would provide a way around skepticism and Gettier-style problems while permitting people to retain their belief in ordinary, commonsensical knowledge. This project was doomed from the start. No definition can serve both purposes, for ‘knowledge’ is not used in the same way in the skeptic’s context (or, in general, the epistemologist’s context) as it is in the ordinary context. One has only to read Wittgenstein’s On Certainty to see that he would have predicted the theoretic floundering of the last few decades, for the simple reason that epistemologists have ignored his main insight: philosophical knowledge is meaningless. Unfortunately, he did not express his insight with precision. Although the concept of knowledge is, in a sense, meaningless, it is not completely meaningless. It is not gibberish. Its meaninglessness is analogous to that of my doubt that I am not a brain in a vat: this doubt is not fully conceptualizable, not imaginable in the way that my doubt about the origin of the universe is. Rather, it is merely an awareness of the abstract possibility that in some metaphysical world I am somehow a brain in a vat. Similarly, I cannot imagine what it would be like to have knowledge in the strict sense in which I defined it above; I can only abstractly conceive the possibility of not having doubts, of having direct access to metaphysical truth. Epistemologists have gone astray in thinking that literal knowledge (as opposed to “loose”, “metaphorical” knowledge; see below) is possible; their attempt to justify this conviction has led them into quagmires of casuistry.


Wittgenstein was wrong, however, to think that the origin of all this philosophical confusion was that philosophers had taken the word ‘knowledge’ out of its ordinary context and given it a new meaning (if he did in fact believe this). In reality, they had simply had the audacity to take the word seriously: they had tried to give it a consistent meaning that adhered to the ways in which ordinary people used it. This task was impossible, because ordinarily ‘knowledge’ is used in a loose sense—one might almost say a metaphorical sense. If you asked a normal intelligent person if he agreed with the definition of knowledge as “a true belief the truth of which one is certain, on the basis of his internal justification for the belief”, he would say, “Yes”, because that is how the word is used. But if you observed how he used the word in his daily interactions, you would see he understood both “certainty” and “internal justification” loosely. For instance, suppose you said, while indicating an object in the distance, “You know that’s a tree over there, right?” He might say, “Of course I know it! It’s tall, it has leaves, and it’s brown.” Upon approaching it, though, he sees it is a papier-mâché model of a tree. “I guess I didn’t know it was a tree after all,” he says. His judgment that he had known it was a tree was false, because, first, his belief was false, and second, his internal justification was not sufficient. This kind of outcome is conceivable in every scenario, which means that people never have knowledge (for they are never in a position to ascribe knowledge to themselves). It is conceivable even with regard to the statement that I know I have hands, because the division between me and the objective world (or objective truth) ensures that my internal justification is not sufficient. (I may, for example, be deluded by an evil demon, à la Descartes.) Even the statement that I know I exist is subject to doubt, because I can imagine the possibility that in some metaphysical realm I do not “really” exist in the way I think I do. In other words, that I can imagine the truth of not-p (where p is the proposition in question) is sufficient to rule out certainty.


The different kinds of knowledge


It would be impracticable, however, to stop using the word ‘knowledge’ in social interactions. Thus, we are forced to posit two kinds of knowledge: that which is impossible and that which is presupposed in daily life. We may call the first kind “reflective knowledge”, or “absolute knowledge”, and the second “conventional knowledge”. The second is not knowledge per se, but, loosely speaking, it can be called such. To try to analyze in detail the standards according to which conventional knowledge-attributions are made would be difficult, because they vary between contexts. Broadly speaking, though, S has “conventional knowledge” that p if S is fairly certain that p and S has some sort of compelling internal justification—and, in some contexts, external justification (see below)—for believing that p. Examples of such knowledge would include my beliefs that the earth is approximately spherical, that I live in St. Louis, and that I will spend Christmas in Rhode Island. I “know” each of these statements. Notice that beliefs need not be true in order to count as (conventional) knowledge. Only in retrospect is the truth-criterion relevant: if I learn that the earth is not spherical, I will think, “Obviously my belief that the earth was spherical did not count as knowledge”; still, at the time, it did. Otherwise we are never justified in attributing knowledge to ourselves (since we are never in a position to decide, beyond doubt, that a given belief is true). The reader may find it problematic that the same belief can count as knowledge at one time but not at another—because people are inclined to think a belief is either true or false, and therefore that it either is knowledge or isn’t—but this ambiguity in my account accurately reflects the ambiguity in the common use of the word ‘knowledge’. According to the precise definition of the word, the reader’s intuition is indeed correct; but according to the word’s conventional use, there is no “objective truth” about whether a given belief is “actually” known. Rather, it is known if someone believes it is known. Conventional knowledge exists only by convention, only in and through the ways in which the word ‘knowledge’ is conventionally used; therefore, it is as logically imprecise as convention. The essentialist definition of knowledge I gave above is the logically (and intuitively) valid, but unrealistic, extrapolation from convention; the non-essentialist definition is the logically (and intuitively) problematic, but realistic, extrapolation. It will collapse in logical paradoxes if examined too closely, but to examine it closely would be, effectively, to misunderstand it.


I should note that on the extreme end of the spectrum of forms of conventional knowledge—which spectrum ranges from the most lenient to the most stringent criteria for knowledge-attributions—is knowledge that satisfies Gettier requirements. “Goldman’s barn case” is a kind of Gettier example:

S is driving in the country and stops in front of a barn. Unbeknownst to S, S is looking at one of the few genuine barns in an area spawned with barn facsimiles. The facsimiles are so realistic that if S had stopped in front of any of them, S would have been tricked into thinking he was looking at a real barn. But by luck S has stopped in front of one of the few real barns left in the area. The standard intuition is that S does not know he is looking at a real barn, because he could easily have had the same belief while looking at a facsimile.[8]

If the (potential) knowledge-attributer is aware of all the information, unlike S, he will not attribute knowledge to S because the truth of S’s belief is accidental. Only if S has all the relevant information will knowledge be attributed to him. This kind of knowledge falls under the concept “conventional knowledge” (despite the stringency of its conditions as compared with those of knowledge-attributions in more mundane contexts) because, like all conventional knowledge, it is merely an approximation of absolute knowledge, such that neither S nor the attributer can ever be sure he has all the information. Even if S is aware of the Gettier conditions that the knowledge-attributer is aware of (and thus, according to the attributer, has knowledge), they both may be unaware of other (relevant) conditions that might conceivably exist, and which therefore vitiate S’s knowledge. Hence, even knowledge that (from the attributer’s perspective) satisfies “Gettier clauses” is fallible, like all conventional knowledge.


The distinctions I have made suggest solutions to several traditional epistemological problems—for example, the controversy surrounding Academic Skepticism (to use the terminology of Sextus Empiricus[9]). Academic Skeptics claim we cannot have knowledge of “epistemically interesting” types of propositions, that is, propositions that are “generally thought to be known given what we ordinarily take knowledge to be”.[10] According to the arguments I have proposed, skepticism cannot be refuted. People cannot have reflective (or absolute) knowledge. Perhaps the reader will retort that I am contradicting myself: “you say we cannot have knowledge, but clearly you think that that proposition itself can be known!” While I do think the proposition is true, I do not think it can be known for certain. So I am not contradicting myself. “So you accept skepticism? Does it not disturb you that all your beliefs may be false?—that you have no knowledge?” No, it doesn’t disturb me in the least. I must admit that I have never been as terrified of skepticism as most philosophers are. That I do not have absolute knowledge vis-à-vis any of my beliefs seems to me like a purely scholastic problem. It does not mean, for example, that my beliefs are probably false. Skeptical possibilities (such as the “evil demon” and the “brain in a vat” hypotheses) are so ridiculous that I feel entitled to ignore them, even while knowing[11] that they remain possibilities. Moreover, they are often patently deluded. For instance, to suggest that I may be a brain in a vat is effectively to suggest a self-contradiction. The I that I am is right here; the word ‘I’, as applied to myself, has no meaning outside of my immediately existing, immediately self-present self here in this world. It is absurd to say I am two beings at once—I am this self and a brain in a vat somewhere in the ‘background’. Nevertheless, because I acknowledge the possibility that in some metaphysical world hidden from me self-contradictions may somehow be true, I cannot say that I know I am not a brain in a vat. –In short, skepticism is irrefutable, but it should be ignored.


The internalism vs. externalism debate is another pseudo-problem. As we have seen, the two positions are united in the true definition of knowledge, such that the debate between them becomes moot. The internal justification that is a prerequisite for knowledge incorporates external justification: i.e., in order to know that p, one must have perfect knowledge of everything that directly relates to one’s belief that p. Regarding conventional knowledge, we have seen that S’s possession of internal justification is almost always a prerequisite for attributing knowledge to S; in many cases (such as Gettier contexts), the attributer requires also that S be externally justified. If asked, most people would probably say that external justification (such as that the truth of the belief in question is not accidental) is always necessary; however, in practice people often ignore this criterion when attributing knowledge to themselves and others. Because the nature of conventional knowledge varies between contexts (since people do not always use ‘knowledge’ in the same way, with the same criteria determining their attributions)—in other words, because there is no “truth of the matter” about whether a given person has knowledge; rather, he has knowledge simply if someone attributes it to him—we are forced to say that external justification is sometimes necessary and sometimes not. It depends on the context and on the knowledge-attributer.


My position may seem similar to contextualism; however, the parallels are superficial. A contextualist thinks that if the attributer is in a low-standards context, the person to whom knowledge has been attributed does indeed know that p; if, on the other hand, the attributer is in a high-standards context, the person does not necessarily know that p (even if the ‘p’ in question is the same proposition that the low-standards person knew). According to my position, neither person has knowledge in the strict sense, but in the low-standards context, the person does have conventional knowledge. The reason, however, is that we have done away with the truth criterion (except when the attributer has it in mind); if we had not, even in the low-standards context the person would not necessarily have knowledge. This is the flaw in contextualism: it forgets that truth is unaffected by a lowering of standards. That I lower the standards for my knowledge-attribution does not make it any more likely that the person to whom I am attributing knowledge has grasped truth. We may illustrate the point by taking an example from John Hawthorne’s paper “The Case for Closure”[12]: a newspaper prints the statement that Manchester United beat Coventry City. According to contextualism, a reader who does not consider the possibility that the statement is a misprint knows that Manchester United beat Coventry City, while a reader who does consider such a possibility does not know that Manchester United beat Coventry City (because he is unsure whether it is a misprint). As it turns out, the statement was a misprint; thus, the reader in the low-standards context could not have “known” that the statement was true, if knowledge implies the statement is in fact true (as a contextualist thinks it does). Nevertheless, a contextualist is committed to saying that he did know it; therefore, contextualism is false. The position I have outlined is more tenable, in that it denies that (in every context) p need be true in order for one to know that p. This does seem counterintuitive, but it adheres to conventional practice.


If my account is correct, where does it leave epistemologists? For one thing, it means that many of them have wasted the last forty years. Not all have, though. The foundationalism-coherentism debate is left intact, as is the empiricism-rationalism debate. These issues are unrelated to the definition of knowledge; they bear only on (an individual’s or a community’s) collective understanding. Understanding should be distinguished from knowledge (in both of its senses): it is a “system” or a “schema”—though not necessarily a well-organized, consistent one—that both is an interpretation and serves as a framework for interpretations of experience. It can be loosely defined as the collective body of “knowledge” of a given entity (whether individual or communal); however, it is not necessarily true. The areas of epistemology that are involved in analyzing the structure of understanding, but not in asking whether we know a given proposition, are left unharmed by my arguments.


“Still,” the reader may ask, “are you saying we don’t know that, e.g., 5 + 2 = 7, or that trees generally have leaves, or that bachelors are unmarried males? If so, what do you mean?” Basically, the thrust of my arguments is that to ask such questions is misguided. To say we don’t know that p, when p is something like “5 + 2 = 7”, is largely meaningless. Nevertheless, the existence of these sorts of questions (and their stubborn resistance to being answered) shows that, in a sense, we do not, after all, absolutely know, beyond any conceivable doubt, that 5 + 2 = 7.


It may help, in this context, if I make use of one last classification, one last category of knowledge. (I probably should have introduced it earlier, but—better late than never.) Aside from knowledge in the strictest sense (“absolute knowledge”, or “reflective knowledge”) and knowledge in the loosest sense (“conventional knowledge”), there is knowledge in the sense of ‘My doubt that p is empty, meaningless, in that I cannot imagine what it would be like for p to be false’. We may call this “unconditional knowledge”, since there are no conditions under which we can concretely imagine it is false. It can be subdivided into two classes: statements whose justification is a priori, and statements whose justification is a posteriori. Such propositions as “Grass turns green in the spring”, “All cubes have twelve edges”, and “I have hands” are unconditionally, and conventionally, known, though not absolutely known. (Unconditional knowledge may be seen as a subclass of conventional knowledge, since it too is but an approximation of absolute knowledge.) The criteria for unconditional knowledge-attributions are that both the attributer and the person to whom knowledge is attributed have internal justification and certainty so overwhelming that they cannot imagine “what it would be like” for the proposition in question to be false. If it were, their entire “understanding” would be thrown into confusion: their cognitive/conceptual framework for interpreting experience would collapse, and they would lose faith in all their former beliefs. Even a proposition like “President Bush went to New Orleans after Hurricane Katrina hit the city” is unconditionally known (by me), because if it were false I would lose faith in the methods I use to interpret experience, the methods of collecting evidence and forming beliefs, and thus I would come to doubt nearly everything I had formerly known. Examples of propositions that are conventionally but not unconditionally known are “I will spend Christmas in Rhode Island”, “I will go to class on Thursday”, “The common belief that there are nine planets in the solar system is correct”, and “Geologists are right that the earth is about four billion years old”.

It may be asked, “Are statements that satisfy Gettier-clauses unconditionally known, or merely conventionally known?” This question is difficult to answer; however, it is also insignificant. It treats classification as an end in itself. If one’s goal is to attain insight into the nature of knowledge, there is no point in trying to decide whether statements that pass Gettier-standards satisfy the criteria on which unconditional knowledge-attributions are based. What matters is that these statements are not absolutely known, nor are they as fallible as the loosest kinds of conventional knowledge. The notions of conventional and unconditional knowledge are not—and need not be—defined with complete precision, in an attempt to make them somehow “correspond” to an objective reality; they are merely useful tools for conceptualizing knowledge-attributions.


Incidentally, it would not be futile for epistemologists to analyze—as they are apt to—how it is that people come to have unconditional knowledge (or even conventional knowledge) of certain beliefs; however, I think that this task might be more appropriate for psychologists than philosophers. Or maybe they could work together. I do not think, though, that the dialogue (between philosophers, psychologists, or whomever) should be framed in the dogmatic way it has been in recent decades; rather, it should be informed by an open-minded, “relativistic” attitude, grounded in recognition that knowledge-attributions do not, and should not, adhere to a single set of criteria (be they “internalist”, “externalist” or whatever). But regardless of the ways in which epistemology will proceed in the coming years, its practitioners have, for the past few decades, been misled by a lack of preciseness and a conflation of categories. In order to make progress they should start defining their terms more carefully, by observing how ordinary people use them.


This paper has left various questions unanswered; its purpose, however, was only to sketch the framework of a way out of the confusion that defines the contemporary epistemological scene.


[1] See Richard Feldman, “Justification is Internal,” in Contemporary Debates in Epistemology, eds. Mattias Steup and Ernest Sosa (Malden, MA: Blackwell Publishing Ltd., 2005), 272, 273. [2] John Hawthorne, “The Case for Closure,” ibid., 40. [3] It is unclear why most epistemologists have not recognized this. For the concept of certainty has no place in their definitions of knowledge. [4] Whether “relevant” should be understood narrowly or broadly need not concern us here. The point is that the knower must be aware of all the information that might conceivably bear on the belief in question (i.e., that he would need to know in order to make an informed and correct decision for or against the belief). [5] See Robert K. Shope, “Chapter 1: Conditions and Analyses of Knowing,” in The Oxford Handbook of Epistemology, ed. Paul K. Moser (New York: Oxford University Press, 2002), 35, http://www.questia.com. [6] Or, as Wittgenstein defines it, “all that is the case”. [7] If God existed, he would not be able to think false thoughts. He would see reality—he would be reality—in its totality and nothing else; thus, negation would not exist for him. [8] Berit Brogaard, “Can Virtue Reliabilism Explain the Value of Knowledge?” [9] See Peter Elien, “Chapter 11: Skepticism,” in The Oxford Handbook of Epistemology, ed. Paul K. Moser (New York: Oxford University Press, 2002), 337, http://www.questia.com. [10] Ibid. [11] In this context (as in most others), I use ‘knowledge’ in the sense of ‘conventional knowledge’, not ‘reflective knowledge’. [12] See Contemporary Debates in Epistemology, eds. Mattias Steup and Ernest Sosa (Malden, MA: Blackwell Publishing Ltd., 2005), 39.

Wright's Writing

© 2014-2020 by Chris Wright