1 Introduction

Assorted facts provoke our curiosity. A beverage surges in popularity. We wonder why. A coded message on a public sculpture resists concerted decryption efforts. We wonder what is so difficult about it. Data on galactic rotations indicate that some ‘dark matter’ has more mass than the entire previously known universe. We wonder what it is.

The wondering has no practical aim. It is a pure curiosity that often arises when we are baffled, amazed, or intrigued. The detached character of the attitude makes especially clear that it sets an epistemic goal for us. The goal is knowledge.

It might be that curiosity is satisfied only by knowledge because an attitude qualifies as curiosity only if it includes an aim to know. This would explain the role of knowledge in curiosity, but it would not explain our interest in knowledge. The fact would remain that our perplexity, amazement and the like routinely provoke curiosity, with its aim to know, and not some less cognitively demanding sort of interest. Wondering makes knowledge seem worth having.

We might feel satisfied by accepting an answer that we mistakenly think that we know, or by accepting a reasonable hypothesis about the answer. But a satisfied feeling is not what wonder and curiosity have us seeking. They have us seeking to find out the correct answer. Finding out is coming to know. Nothing less than knowing wholly fulfills the cognitive aspirations to which wondering and curiosity give rise.

What is it about knowledge that presents itself to us as being intellectually satisfying? The answer offered here, in a few preliminary words, is that in knowledge a purely intellectual interest is satisfied by truth-indicative reasons. Knowledge implies that we have such reasons and that they bear a relation to the known proposition that is additionally satisfying. A full answer to this question is a principal goal here. It will be argued that a traditional conception of knowledge confirms and explains the impression that knowledge satisfies our purely intellectual interest.

Knowledge is readily found appealing for another reason. Knowing seems to be a good way for a mind to relate to a fact. Knowing is more attractive, just because of what it is, than is having a true belief without knowledge. This is an appearance of greater intrinsic goodness.Footnote 1 It will be argued that the same traditional conception of knowledge also accounts quite well for this evaluation.

The prevalent alternative conceptions of knowledge do not explain the credibility of the view. We shall see illustrations of that failure. This is reason to regard the traditional conception as an especially helpful way to think about what knowledge is.

2 Intellectual fulfillment

A tradition in epistemology has it that knowledge is justified belief that is connected to the truth so as to avoid the Gettier problem. Here is a statement of that conception of knowledge, formulated in a way that will be useful for the explanations that we seek.

Traditional Account of Knowledge (TAK). S knows X if and only if

(1) X is true,

(2) S believes X,

(3) S has adequate epistemic justification for believing X, and

(4) the justification that S has for believing X is non-accidentally related in the proper way to the truth of X.

The first three conditions in TAK are formulated in a fairly standard way. They require a true belief that is appropriately justified. The fourth TAK condition is broadly familiar too. It blocks Gettier cases. The condition says roughly that the justification of a known belief must be rightly related to its truth. It should be further clarified before TAK is applied to the explanatory tasks at hand.

2.1 The fourth condition

The condition requires that knower’s justification for the known belief be properly linked to ‘the truth of’ the proposition, that is, linked to some factual state of affairs that suffices for the proposition’s truth. The linked fact might be a state of affairs that makes the proposition true. But it need not be that very fact. The fourth condition of TAK allows a properly linked fact to be one that is sufficient for the truth of the known proposition. This leeway has the following advantage.

The condition allows us to know a future truth when our current factual basis is unexpectedly preempted. Such knowledge does exist. We know that the planet Mercury will cease to exist. The Sun will eventually expand so as to incinerate a planet at Mercury’s distance. Our justification for this fact secures for us the knowledge that Mercury will cease to exist. We know that, even if what does terminate Mercury’s existence is the obliterating impact of some currently unknown asteroid, or some other unforeseen event. Our knowledge of the sufficiency of the fact about the Sun’s expansion lets us know that Mercury will be destroyed, no matter how that actually happens. The required connection to a sufficient fact in the fourth condition of TAK provides for this. The nature of the sufficiency does not matter for the sake of having knowledge—causal, logical, metaphysical, even material sufficiency will do. What is needed is properly linked justification in support of the fact. If it is a sufficient fact and not the true proposition itself, then knowledge also requires justification in support of its sufficiency for the truth of the known proposition. We have this in the case of Mercury’s destruction. We have properly non-accidental justification for the fact that the Sun will expand and for the fact that the Sun’s expansion would obliterate Mercury. This meets the fourth condition for knowing that Mercury will cease to exist, however that actually happens.

Another detail of TAK’s fourth condition plays a role in the explanations to be offered here. The condition requires the justification to be non-accidentally related in a proper way to a truth-sufficient fact. This requirement incorporates a lesson of the Gettier literature. Some accident occurs in every example of justified true belief that is not knowledge. The relevant type of accident is unclear. Accidents are involved in some cases that are instances of knowledge, such as when a knower luckily has just the right visual perspective or the knower accidentally comes across a felicitous explanatory hypothesis. Still, in each Gettier case there is some knowledge-preventing sort of accident in the relation of justification to sufficient fact. The belief ‘just happens to be right,’ even though it is justified. The fourth condition in TAK accommodates that lesson by prohibiting any such accident.

Finally concerning TAK’s fourth condition, its use of ‘in the proper way’ renders TAK implicitly circular. In the context of TAK, this phrase is short for ‘in the way that occurs in knowledge’. This circularity prevents TAK from formulating any sort of reductive theory of knowledge. It does not prevent the condition from having informative content. The TAK fourth condition states that in any case of knowledge, there is a non-accidental relation of a justification to a truth-sufficient fact. That is a substantial constraint. Stating the requirement without the circularity would involve replacing ‘in a proper way’ with an expression that excludes exactly the knowledge-blocking kinds of accidents. This would solve the Gettier problem. TAK does not attempt to do that. Rather, what its fourth condition requires to hold between a justification and a truth-sufficient fact is whatever turns out to be the non-accidental relation that occurs in knowledge. This is all we need for present purposes. It has enough content to enable us to use TAK to give the explanations that we seek.Footnote 2,Footnote 3

2.2 The interest in meeting the first three conditions

It is straightforward that when we are curious about something, we are interested in gaining true beliefs about it. TAK straightforwardly accommodates this interest by requiring true belief for knowledge.

Meeting the third condition of TAK requires having justification. The TAK terminology of ‘adequate epistemic justification’ is typical in statements of the traditional conditions. ‘Epistemic’ here means ‘pertaining to knowledge’ and ‘adequate’ here means ‘adequate for knowledge’. So ‘adequate epistemic justification’ also induces a circularity. The expression of the condition says that a certain type and strength of justification is needed for knowledge, and it identifies the justification by implicit reference to knowledge.

In this case, though, the circularity can be usefully avoided. The relevant notions of justification and adequacy can be explicated in a way that facilitates the explanations that we seek. First, ‘adequate’ justification for knowledge is enough justification to make confident belief an intellectually appropriate response to the proposition. In other words, having confidence in the proposition, in light of the justification, is rational for the sake of a pure interest in the truth of the matter. Having a confident belief is being sure that the proposition is true. Being sure does not imply being as intensely confident as possible. Being sure is having confidence without having any doubts.

Given that making this confidence rational in this way is what the justification does, the nature of what does the justifying also can be informatively specified. This rationalizing of confidence is a role for our evidence, provided that ‘our evidence’ is broadly understood. It is ‘ours’ in that it is mentally internal to us. It includes whatever we have in mind that contributes to informing us. It consists in those considerations that indicate to us where the truth lies. They indicate the truth to us in that their bearing on truth is available to our reflective attention. This sort of truth indication is accomplished by a wide range of mental phenomena, from abstract arguments about necessary facts to experiential states such as perceptual representations and apparent memories.

Evidence so construed provides the justification needed for knowledge. Our evidence justifies the proposition well enough to make our being sure an attitude that is intellectually rational for us to take toward the proposition. Being sure is a doxastic response that fits a purely intellectual interest in the truth of the matter.Footnote 4

With this elaboration, the contribution of TAK’s justification condition to intellectual satisfaction can be efficiently explained. The condition requires that our being sure of the known proposition is a rational response to our evidence in the matter. When we meet the condition, we have reason to believe a proposition from evidence that gives assurance of its truth. Having truth-assuring evidence manifestly contributes to being in an intellectually satisfying relation to the proposition. Reasons that support truth sufficiently to make being sure a rational attitude are reasons that give conclusive indications as to where the truth lies. The evidence excludes any good reason to doubt. Such indications plainly answer to an interest in the truth for its own sake, that is, a purely intellectual interest. That is what our evidence for a known proposition supplies.

Having a true belief that is backed by evidence that makes it rational to be sure is thus more intellectually satisfying than is having a mere true belief. This justification is part of how knowledge fulfills our cognitive interest in the facts.

This is a claim about fulfilling an intellectual aspiration. It is not a claim about feeling fulfilled. We do not always get a sense of satisfaction from learning an answer to a question that we have wondered about. Our concern for finding the answer can fade. For instance, we might wonder what fraction of the flowers in a garden are roses. When counting each flower becomes tedious, the question can lose its appeal. Yet if we doggedly persist until the count is completed and the fraction is known, we do satisfy the intellectual goal.

The rational fulfillment of an intellectual interest is quite analogous to the fulfillment of an intention to act by acting on the intention. The act need not yield any feeling of satisfaction. Quite similarly, the knowledge we gain from the satisfaction of our curiosity may give us no sense of satisfaction. The content of the knowledge might even spoil our getting any good feeling out of the intellectual satisfaction. The content might be a troubling fact. If we wonder whether we got a bargain when we made a spur-of-the-moment purchase and we learn that it was no bargain, then that knowledge satisfies our curiosity to find out. But we may well feel vexed rather than satisfied. Also, we might feel entirely satisfied by a fully justified but false answer that does not give us knowledge. Furthermore, in order for us to take some answer to fulfill our curiosity, we need not have much justification to regard the answer as true. If we are impatient with the pace of cosmological inquiry into the nature of dark matter, we might find some mildly plausible conjecture satisfying enough to lose interest. We are fallible about when we do know an answer and we are capable of feeling satisfied with an answer that we do not know.

The present claim about the satisfaction that is accomplished by meeting the TAK justification condition is narrowly rational. Our wondering poses an intellectual goal of finding out the truth. The justification helps knowledge to meet the goal. We usually do desire to know. We usually are gratified by the goal’s fulfillment. But the claim is that justification for the known proposition contributes to meeting the intellectual goal that is set by our curiosity, however we feel about that.

2.3 The interest in meeting the fourth condition

The fourth condition of TAK requires some proper sort of non-accidental relation between justification the person has for the belief and some truth-sufficient fact. Again, the statement of the condition in TAK offers no solution to the Gettier problem. Its use of ‘in a proper way’ holds a place for an expression of some condition that blocks exactly the knowledge-preventing accidents in the relation of justification to fact. TAK implies that some such relation solves the Gettier problem. That information about what the condition requires is sufficient to enable us to understand how the condition contributes to the intellectual fulfillment of knowledge.

In the Gettier examples the justification is the sort needed for knowledge. It gives a rationally satisfactory assurance of the truth of the belief. Introspectively, it seems as satisfactory as knowledge. But the justification is disconnected from the truth. Putting the point in terms of TAK’s fourth condition, no truth-sufficient fact has a proper relation to the justification. For instance, there is the classic Nogot sort of case. The protagonist Dilbert believes that someone who works in his office owns a Ford. Dilbert’s justification for the belief derives from evidence about Nogot. Nogot is a co-worker who drives Ford that he only appears to own. He has long been borrowing it from an obliging relative. Someone else in Dilbert’s office, Havit, does own a Ford. Havit’s Ford ownership makes Dilbert’s belief true, but it does not contribute to Dilbert’s justification for the belief. His justification traces only to Nogot. The deceptive source of Dilbert’s justification thus makes it an accident that he is right that someone in his office owns a Ford. Dilbert's belief is not knowledge.

Once we are apprised of the possibility of justification that is detached in some such way from the truth that it justifies, it becomes clear to us that rational support of this accidental sort does not fully satisfy our interest. The potential for this shortcoming is not manifest when we think about the typical cases of justified true belief. They are cases of knowledge with its properly connected justification. When we appreciate Gettier cases, though, we see that not all justified true beliefs give us the full connection to the truth that wonder and curiosity have us seek.

TAK’s anti-Gettier condition says that the justification is related in the way that is required for knowledge to some fact that is sufficient for the belief to be true. Again, the presence or absence of this link is not introspectively discernible by the subject. The link does not make knowledge seem more satisfying to the subject, or otherwise better. If we reflectively compare this sort of connection to its absence in Gettier cases, however, we find that having the fourth condition’s non-accidental link of justification to truth better fulfills our intellectual interest in the proposition. When this sort of state of affairs does obtain, what backs our confidence that our belief is right is rooted in something that is evidently sufficient for it to be right. Our minds are placed in close rational accord with the pertinent truth by some fact that we are justified in regarding as sufficient for that truth. Moreover, if we are in this relation to a sufficient fact, then our purely intellectual interest truth of that matter has been satisfied.Footnote 5

2.4 A summary of the satisfactions

Meeting the TAK conditions is intellectually fulfilling. When we meet them, we believe the truth of the matter. We do not merely happen to get it right. We are rationally assured of it. We do not receive this rational assurance of the truth from a source that merely happens to coincide with the fact of the matter. We have determined what the truth is by having evidence for it that derives from some fact that suffices for it. This is the TAK explanation of why knowledge is intellectually satisfying.

We can readily appreciate that our having an answer that meets these conditions fulfills the intellectual aim given by wonder, perplexity and curiosity. This ease is an asset of TAK. We have the ready impression that the aim is to know. Our easy recognition of this aim is best explained by locating the intellectual satisfaction of knowledge in plain view. This is true of knowledge as it is conceived by TAK.

3 Alternative accounts

Other conceptions of knowledge do not offer a comparably good explanation of how knowledge is more intellectually satisfactory than true belief. Their fundamental liability is that they do not imply anything that is comparably satisfying to our having a factually connected rational backing for the belief. Two alternative accounts will serve as illustrations.

3.1 Truth-tracking

We “track the truth” when our believing of a proposition is counterfactually dependent on the proposition’s being true. Here is the classic conception of knowledge as truth-tracking:

Counterfactual Account of Knowledge (CAK) S’s true belief in X is knowledge iff S would believe X, if X were true and S were to use S’s actual belief-forming method, and S would not believe X, if X were not true and S were to use S’s actual belief-forming method.Footnote 6

The truth-tracking conditions of CAK require that we believe by a method that inclines us toward a certain otherworldly accuracy. In possible worlds where the proposition is true that are very much like the actual world, the method has us believing it; in the possible world most like the actual world where the proposition is untrue, the method has us not believing it.

Truth-tracking does not account for the intellectual satisfaction that knowledge provides. Our having the required doxastic inclinations entails nothing intellectually satisfactory about our actual cognitive relation to the proposition. CAK requires only that our method would have us believing, and not believing, in synch with what the proposition’s truth-value would be. It does not matter how poorly that method informs us about the proposition.

Examples show how intellectually deficient a truth-tracking method can be. Even a thoroughly superstitious method can have the requisite counterfactual truth-sensitivity. Suppose that Creed is curious about the outcome of a forthcoming lottery and consults a fortuneteller. In order to secure Creed’s trust, the fortuneteller employs for the first prediction a reliable covert informer who rigs the lottery. Creed’s method of forming a belief about the lottery outcome is to take the fortuneteller’s word for it. That method tracks the truth of the prediction under the circumstances.Footnote 7 The first prediction is in accord with the truth of the matter, and it would have been in accord, since the prediction does derive, and would have derived, from the fortuneteller’s lottery-fixer. Yet Creed believes the prediction out of a superstitious faith in the fortuneteller. This basis for belief does not satisfy an intellectual interest in the question. Creed’s attitude would have been in considerably more reasonable if Creed had withheld judgment about the lottery outcome.

This is a case of truth-tracking without knowledge, and so it is a counterexample to any truth-tracking theory that allows Creed’s method to count as an appropriate sort of truth-tracking. But our present aim in considering truth-tracking theories is not to assess their success at classifying cases. The aim is to determine whether they explain how knowledge fulfills intellectual interest. The problem is that their required disposition to track the truth can be cognitively empty. What suffices for the tracking is anything that gives the right counterfactual features to a belief-forming method: it must match the proposition’s truth value in some way or other, no matter how irrationally. Truth-tracking need not do us any intellectual good.

A truth-tracking theory can do better than CAK at capturing the extension of knowledge. By the same means it can do better at requiring the fulfillment of intellectual interest. The theory can require that we track a truth specifically because we are inclined to believe in accordance with the epistemic justification that we have. This is a source of a truth- tracking disposition that does contribute to our intellectual fulfillment. This sort of truth-tracking depends on our having something that satisfies our intellect in a way that we have noted: it gives us rational assurance that the proposition is true.

If this sort of truth-tracking method is required, though, then the tracking itself does not help. The counterfactual implications of the tracking are superfluous. The justification does the work. We can see this by imagining away the tracking while leaving in place the justification.

Suppose that Cie becomes peripherally aware of a mark on the ceiling and wonders what it is. She looks at it carefully. Her belief that a green spiral is on the ceiling is quite well justified by her visual experience. She knows that a green spiral is there. Cie almost moved her head differently. Had she done so, a cosmic ray that was passing harmlessly through her brain would have had a psychological impact. The effect would have been a momentary inversion of Cie’s color experience. She would have perceived the spiral as red and believed accordingly.

In this case Cie’s method does not track the truth. She would have failed to believe the truth in the slightly different circumstances that almost obtained. She would have used the same visual method and, as a result of the inversion, she would not have believed that the spiral was green.Footnote 8 Nevertheless, Cie’s curiosity about the mark on the ceiling is satisfied by her actual perceptual knowledge that it is a green spiral. The narrowly avoided inversion from the cosmic ray does not prevent her from receiving that satisfaction from this knowledge. Cie’s failure to have a truth-tracking disposition concerning the proposition thus does not do harm her intellectual position. It is the justification that contributes to her intellectual fulfillment. No such contributuion is made by any counterfactual relation of her believing to its truth-value.Footnote 9

3.2 Williamson’s account

Timothy Williamson denies that knowledge has an analysis into constituent conceptual elements (Williamson 2000, 33) This denial is compatible with TAK. TAK does not offer an analysis. It offers necessary and sufficient conditions that give a conception of what knowledge is. TAK is compatible with other conceptions of knowledge. Williamson proposes one that is consistent with TAK. It is quite different. We shall see that Williamson’s account does not have TAK’s explanatory assets.

Williamson proposes the following as a distinguishing characteristic of knowledge:

Williamsonian Account of Knowledge (WAK). Knowledge is the most general propositional attitude that is a factive mental state. (Williamson 2000, 39)

Here is what Williamson means by this characterization. A mental state is a condition that a mind is in, like believing, being regretful, and feeling elated, rather than a process like discovering and proving. A factive attitude is an attitude that must have a truth as its propositional content, like apprehending and acknowledging. The most general one of these is the one that is entailed by all of them. Putting these ideas together, what WAK asserts is that knowledge is the attitudinal state that both entails the truth of its propositional content and is entailed by all such attitudes.

According to Williamson, this account of knowledge makes intelligible the importance that knowing has to us:

On this account the importance of knowing to us becomes as intelligible as the importance of truth. Factive mental states are important to us as states whose essence includes a matching between mind and world and knowledge is important to us as the most general factive stative attitude. (Williamson 2000, 40)

Williamson does not assert the existence of a connection between the importance of knowledge to us and its satisfying our intellectual interest in the truth. But whatever satisfies this interest matters to us by doing so. Consequently the importance of knowledge to us must at least include whatever makes it satisfy this interest. So it is reasonable to look to WAK for an explanation of why knowledge is more intellectually fulfilling than is mere true belief.

WAK does not yield this explanation. First, all true beliefs have the importance to us of ‘states whose essence includes a matching between mind and world’. So the importance to us of that matching cannot be something that makes knowledge more intellectually satisfying than any other true belief.

It might be thought to make a difference here that Williamson claims knowledge to be a state that is essentially a relation to truth. True belief is a state that is only contingently related to truth.

The metaphysical claim regarding the essence of knowledge is doubtful. The mental state natural kind here seems to be belief, not knowledge. It is plausible that some states that actually are instances of knowledge might have been false beliefs instead. It is plausible that some states that once were true beliefs became knowledge when they became appropriately justified. But the metaphysical issue can be set aside. However it turns out, there is the de dicto necessity that any true belief is a matching of mind to world. So any true belief does the matching that is claimed to be essential to the state of knowledge. On reflection it is clear that it is the actual matching that is important for our intellectual interest in how things are. A possibility of the same state not having matched the world makes no difference to this interest. So the contingent matching of a true belief is just as good for achieving this purpose as is knowing the truth.

The rest of the WAK characterization does not imply anything that better fulfills our interest than does true belief. Knowledge is an attitude, but so is belief. Knowledge is a mental state, but so is belief. Williamson argues that true belief is not a mental state (Williamson 2000, 27–30). Suppose so. Nevertheless, reflection indicates that it does not matter to satisfying the interest whether the mind-to-world match occurs within the mental state of knowing a fact, or between the mental state of belief and an external fact.

Finally, WAK tells us that every factive stative propositional attitude entails knowing the content of the attitude. If the account is correct, then this entailment is something significant about the other factive stative propositional attitudes. They all imply knowing their propositional contents. But we seek an explanation of what makes the knowledge itself intellectually fulfilling. This interest is satisfied by what knowledge itself provides. It does not affect this satisfaction that certain other mental states entail the knowledge. If they were all additionally helpful in fulfilling curiosity, apart from the knowledge that they entail, then implying one of them would be implying something additionally satisfying. But factive stative attitudes are generally not otherwise intellectually satisfying. This is most vividly illustrated by being amazed or being puzzled that something is so. Such attitudes provoke further curiosity about their contents rather than adding to the satisfaction of knowing the contents.

In sum, the WAK conception of knowledge does not explain how it offers anything better than true belief for satisfying curiosity.

3.3 Conclusion concerning fulfilling intellectual interest

TAK is better than these two other conceptions of knowledge at explaining how it provides intellectual satisfaction. Section 7 offers grounds to generalize from this comparison.

4 Value

In some of our highest-minded moments we seek knowledge just because of what it is. We regard it as intrinsically valuable. TAK can be seen to bear out this evaluation.

4.1 A test of value

There is no decisive way to justify attributing intrinsic goodness to knowledge. This is not surprising, since we have no decisive way to justify attributing intrinsic value to anything. We can identify a rational basis for judging, though. It is reasonable to think that we can test for intrinsic goodness by gaining a certain sort of evidence for its presence.

In some cases when we consider a possible state of affairs on their own, the state inspires in us an admiring or approving response to some sort of unity that it displays.Footnote 10 In other words, some possible states of affairs impress us as having an attractive sort of coherence; they have an appealingly unified interrelationship among constituents of certain sorts. For instance, something that we can find attractive about a sunset is the interrelationship among the colors, shapes, and lighting in the scene. For a different sort of instance, the organization of a complex and efficient assembly line can prompt our admiration and our esteem. It is reasonable to think that these positive affective responses to some sorts of unity are evidence that the states of affairs that are the objects of our affection have some intrinsic goodness in virtue of a sort of unified order that impresses us.Footnote 11 Furthermore, it is plausible that the more appealing to us is the unity displayed, the greater is the degree of intrinsic value that is thereby supported.Footnote 12

4.2 Applying the value test to the first three conditions

Any true belief gives some evidence of being intrinsically good by our test of displaying an attractive unity. When we think of someone who is accepting a factual proposition as a fact, we are disposed to view favorably this relation between attitude and reality. Accepting a proposition as true appealingly accords with its being true.

Knowledge is intrinsically better by our test. Belief in a true proposition might be disconnected in various ways from that truth. A closer integration between mind and world is more appealing. Knowledge is more attractively unified in this way.

TAK accounts for that unity. Any one belief is a small part of the perspective on reality of the believer. The entire perspective can make believing the proposition unreasonable, conjectural, or subject to serious doubt. These are cognitive incongruities. They detract from the overall coherence of belief in the proposition by such a mind. In contrast, when the TAK justification condition is met, the justifying aspect of the mind’s perspective—the evidence that the person has—gives assurance that the relevant part of the world is as it is believed to be.Footnote 13 The assurance is sufficient to remove reasonable doubt about the proposition. The evidence includes experiences or memories that fit together in a rational way with the belief being true. This gives more unity to the relation of the whole mind to the fact. Thus, by our test of the attractiveness of the unity, the justification condition asserted by TAK contributes to knowledge being intrinsically better than true belief.

4.3 Applying the value test to the fourth condition

Suppose that the justification for some belief only happens to coincide with the truth of the proposition. For instance, the justifying evidence for a truth might be a product of convincing confabulations. That evidence has nothing to do with what makes the proposition true. If so, then there is something merely fortuitous in the resulting relation of mind to believed fact. The fourth condition of TAK tells us that knowledge is not that way. It requires a relation of the person’s justification to something sufficient for the belief’s truth that is non-accidental in the way that is needed for knowledge. This constraint requires that the justifying basis for the belief is properly linked to some appropriate fact. The connection between basis and fact is not a rationally unexpected fluke; the relation is as the person’s evidence makes it seem to be. This further unity gives grounds for attributing greater intrinsic value to knowledge. The resulting state of affairs includes an attractively closer bond among belief, rational basis, and fact.

4.3.1 An objection to the evaluation

Jonathan Kvanvig has argued that the gory details of any effective Gettier-proofing fourth condition on knowledge undermine the value of meeting the condition.Footnote 14 The argument goes as follows. A Gettier-proofing condition must distinguish knowledge from the most similar examples that are not knowledge. Sometimes the difference between knowledge and a Gettiered justified true belief is a minor matter of causes or circumstances. In the Nogot case described above, for example, it is the difference between legally possessing a Ford and borrowing one. Such differences clearly have no intrinsic value. A solution to the Gettier problem will often count some such unimportantly differing circumstance as crucial to meeting its anti-Gettier condition. The solution will be a gerrymander that separates some cases by worthless differences. Thus, meeting the anti-Gettier condition will not ensure have any value of its own that might make knowledge intrinsically better than Gettiered justified true belief.

4.3.2 Reply

This reasoning depends on the value contributed by a Gettier-proofing condition being possessed by any differentiating element that distinguishes a state of affairs of knowing from a Gettiered justified true belief. If the value really had to be possessed by whatever local facts make the difference, then the worthless causal differences that sometimes do the differentiating would indeed prevent the condition from entailing any added value, as the argument alleges. But the difference-making element need not have any value. Instead, that element can make something else intrinsically good by helping to constitute a larger state of affairs that is intrinsically good. This is what happens when the fourth condition of TAK is met. Meeting that condition enhances the unity of the relation of a mind to its world.

We can use our Nogot example to illustrate this. Dilbert has a justified true belief that someone in his office owns a Ford. Nogot’s apparent ownership is a case of borrowing. Owning need not be better than borrowing in any way. Nevertheless, because Dilbert’s justification derives from the borrowing Nogot, the justification is detached from any genuine instance of Ford ownership by a co-worker. In all Gettier cases there is some such discrepancy. It worsens the believer’s cognitive situation. The nature of the disqualifying accidents has proven to be fiendishly difficult to specify. But they are clearly present. They dislocate the justification. They bring about some disunity in the mind-to-world relationship. Meeting TAK’s fourth condition disallows these disconnecting accidents. It requires the knower to be in a better relation to the proposition. The fact that some particular causal, intellectual, or environmental details are part of the value-enhancing work need not be apparent by considering them on their own. It may become apparent that they are improvements only when we observe their contribution to the mind-to-world relationship that is achieved by meeting the whole condition.Footnote 15 We can become aware of the greater appeal by considering the instance of knowledge that is the closest counterpart to the Nogot case. In that sort of situation, Dilbert has his justifying basis for his Ford ownership belief from evidence about someone otherwise like the actual Nogot but who does own a Ford. This state of affairs presents to us on reflection a more attractive unity in its mind-to-world connection. Dilbert’s perspective on the situation is in closer accord with that situation. Our test for intrinsic value gives us reason to think that this is better.

4.4 Conclusions about TAK and value

The third and fourth TAK conditions on knowledge make it intrinsically better than mere true belief. It has been easy to see the attraction of meeting the TAK conditions. This ready appeal is an asset of TAK. It is in accord with the ease with which people judge that knowledge is more desirable for its own sake.

This explanation does not imply that all knowledge is equally intrinsically valuable. The claim is only that the intrinsic goodness of knowing exists in any instance. Knowing some altogether trivial fact about the number of peas in a pod need not be obviously or significantly intrinsically valuable. On reflection, though, we can find something positive even in such uninteresting and practically worthless knowledge. Being cognizant of a fact is always an attractively unifying relation between a mind and a part of its world.

5 More value

The intrinsic value that we have been discussing does not exhaust the value of knowledge. Knowledge can have extrinsic value by contributing to variously valuable ends. Knowledge might have moral value too, but it need not. Positive moral value is value that makes a positive contribution to something of moral significance. No moral significance is immediately implied by knowledge having positive intrinsic value, that is, having goodness in virtue of its intrinsic nature. It might be that, as some consequentialists assert, promoting anything that is intrinsically good makes some positive contribution toward an act’s moral permissibility. But that is just one sort of consequentialist view of the bearing of intrinsic goodness on morality. According to other consequentialist theories, the morally relevant intrinsic goods are only those that consist in some positive affective state such as pleasure or contentment. Moving away from consequentialism, it might be that all morally relevant goodness consists in an exercise of good character. In various moral views the intrinsic value of knowledge makes no moral contribution.

Intrinsic goodness has no undisputed bearing on ethics. The account offered here of the intrinsic value of knowledge is morally neutral. It allows knowledge to play any moral role or none at all.

The explanation of intrinsic goodness relying on TAK is not the whole story about the intrinsic value of knowledge. The proposal is that the TAK conception of knowledge accounts for the intrinsic goodness that any knowledge appears to have, just because it is knowledge. Factors other than this appealing unity might give some knowledge additional intrinsic value. Knowing an important truth or a profound truth might be intrinsically better than knowing a truth of lesser significance. Knowing an explanation might be better than knowing an isolated fact. Knowing one’s own deepest reflective convictions might be an especially good relation of knower to known fact. In contrast, some knowledge of bad facts, some knowledge by bad knowers, and some bad combinations of knowledge and knower, might be intrinsically bad. Meeting the TAK conditions gives all knowledge some intrinsic goodness. But is finite and can be outweighed. It is highly plausible that someone's knowing that millions have been murdered is an intrinsically bad state of affairs. The evil in the known fact greatly overwhelms the good relation of mind to fact that makes it a case of knowledge.

6 An alternative explanation of the value

Leading nontraditional conceptions of knowledge offer no comparably good account of its intrinsic value. The most promising alternative is a virtue-theoretic approach. It gives a prominent place to valuable things. But we shall see that the value is only contingently present and it is not possessed by the knowledge itself.

Here is an illustrative virtue theory that has been defended by John Greco:

Virtue Account of Knowledge (VAK) Knowledge is belief that is true primarily because of the intellectual virtue that is manifested in the believing.Footnote 16

VAK requires that knowers have a virtuous intellectual disposition toward accuracy and that when they know, some such disposition is primarily responsible for why it is a truth that they believe. The VAK account thus implies good things about knowers and knowledge acquisition. Knowers possess a virtue. The event of acquiring knowledge involves exercise of a virtue. But what is intrinsically good about the knowledge is not something good about people or something good about how knowledge is acquired. The value is in the knowing relationship, however it is entered. We can see that knowledge has this value, apart from any virtue that the believing involves, by considering a case in which the intrinsic value of knowledge is present without the virtue.

Suppose that Walker wants to enjoy his country stroll. To avoid learning of any threat to his enjoyment, Walker has been exercising intellectual vice to close his mind. He is willfully inattentive to signs of the impending weather. Yet Walker irresistibly hears an unmistakable boom of thunder. He thereby knows that thunder sounded. Since the belief was not virtuously acquired, VAK does not count Walker as having this knowledge.Footnote 17 The example might constitiue an objection to VAK. But the present point is about explaining value. When we ignore the vice in Walker’s doxastic efforts and concentrate on his cognition of the fact that thunder sounded, we can find something that earns our reflective esteem. We can find the mind-to-world connection that we appreciate about knowledge. Walker’s belief that thunder sounded is an instance of a sort of unity that attracts greater reflective esteem than does a luckily true belief. By our test of value, this is evidence for its having more intrinsic goodness than the true belief. A virtue theory such as VAK does not account for that value. It yields no explanation of why Jones’s belief that thunder sounded is intrinsically better than any other true belief.

Once we take note of this it becomes apparent that even when exercising virtue is involved in knowing it is not the exercise of virtue that makes knowledge intrinsically better than true belief. The virtuous activity is something else that is good. The value of having and exercising the virtue are not in the right place to explain the value in the state of knowing. It is the reflective appeal of the knowing relationship itself that we seek to explain. Its appealing unity is our evidence for its intrinsic goodness. This evidence is not explained by adverting to good things about how the state is acquired or sustained.

VAK can be revised to give a conception of knowledge that does entail the value. The revision requires an exercise of a particular intellectual virtue: the virtue of believing on the basis of justifying evidence that derives properly from some evidently sufficient fact. If a virtue theory requires an exercise of this virtue, then the virtue theory implies that the TAK conditions are met. This gives the state of knowledge reflective appeal. But it also renders superfluous the distinctively virtue theoretic aspect of the approach. It does no additional good in accounting for the value of knowing to require that the evidence be employed in an exercise of intellectual virtue that acquires or sustains the knowledge.

Greco uses his virtue theory to offer a different explanation of the value of knowledge (Greco 2012). The other explanation identifies as the value of knowledge the achievement of attaining a true belief by virtuous believing.Footnote 18

The merit in intellectually virtuous achievement is also off target. Achievement is valuable. But it is not the value in knowledge. This becomes apparent by considering cases in which the value of the achievement varies while the value of the knowledge remains constant. Suppose that Keen is presented sequentially with two very nearly indiscernible photos of a complicated forest scene. In an astounding exercise of memory and visualization, Keen discerns, and thereby knows, that there is some small difference between the two–one leaf that is depicted in the photos is at a slightly differing angle. Keen’s cognitive achievement is superlative. But Keen’s knowing that a difference exists between the photos has no more intrinsic value than does his knowledge that some easily noticed difference exists between two other forest photos. Similarly, proving a theorem by an ingenious method is a greater cognitive achievement than is proving the theorem by a routine technique. The reasoning might show greater creativity or mathematical acumen. But the knowledge of the theorem has the same intrinsic value.

In such pairs of cases, one cognitive process is more impressive than another. The method of acquiring the knowledge is more discriminating, more intelligent, or more creative. The factual knowledge that results, however, is equally intrinsically good.

A different view could be taken of what the achievement is. The achievement in the belief acquisition could be held to consist in its arriving at a truth. If so, then the achievement of truth by all instances of knowledge could be counted as equally valuable. But then any true belief accomplishes the same achievement. We would have no explanation of why knowledge is intrinsically better. Thus, the intrinsic value of the knowledge is not the merit of the cognitive achievement.

7 Other accounts

Other nontraditional conceptions of knowledge exist. One of them might seem to be especially well equipped to explain its intrinsic value better than does TAK. This alternative requires direct acquaintance with the known fact.Footnote 19 Let’s consider how well that idea fares. There is a problem. After noting it, we shall see reason to think that no other conception can do as well as TAK.

Requiring direct acquaintance with the known fact looks promising as a way to entail greater intrinsic value. Acquaintance is some maximally close intellectual connection to part of one’s world. Being in this sort of relation might well constitute a greater unity that exceeds the value of meeting the TAK conditions for knowledge.

Suppose that the greater unity does have greater intrinsic value. There is a problem for any view requiring that much unity in order to have knowledge at all. It implies that the intellectual satisfaction of knowledge is less prevalent than it actually is. We often gain the satisfaction despite not being acquainted with the fact. Scientifically established truths provide an abundance of examples: some randomly chosen examples are the facts that some ravines are products of glacial erosion, that table salt is composed of sodium chloride, and that birds have reptilian ancestors. We are not directly acquainted with facts like those. The cognitive relation to them that science affords is mediated by observational data and theory. For most of us, teachers or texts further mediate the relation. Nevertheless when we study the relevant science our relation to such facts is quite intellectually satisfying.

Knowledge affords this complete intellectual satisfaction. TAK accounts for this. Scientific findings like these do meet the TAK conditions. The justification that we have for scientifically established propositions properly relates us to some truth-sufficient facts in nature. The relation obtains via our justifying bases for the theories. An acquaintance view is thus not as helpful as is TAK in explaining the range of our intellectual satisfactions.

Also, suppose that an acquaintance condition on knowledge implies knowledge to have greater unity than is implied by the TAK conditions, and therefore the acquaintance condition implies knowledge to have greater intrinsic value. This is not an explanatory advantage, unless we have reason to think that the value of the TAK conditions is insufficient to account for the degree of intrinsic value that knowledge really has. We do not have any such reason. In fact, we have reason to be modest about how much value knowledge entails. Knowledge of insignificant truths is not very impressive. In the most insignificant cases—knowledge of entirely unimportant quantities and the like—it is not always obvious that the knowing is intrinsically good. The entirely unexciting factual contents lack any appeal. This can distract us from apprehending the limited but definite value that is always afforded by the unity between mind and world. An account of knowledge that makes the value loom too large thus incurs a liability.

There is a further problem. It affects the explanatory prospects of any conception of knowledge that is substantially different from TAK. If an account does not have approximately the same extension as TAK, then it will have more trouble with counterexamples. If it has roughly the extension of TAK, then its conditions do this in some relatively unfamiliar way. This can be illustrated by WAK, Williamson’s conception of knowledge as the most general factive stative propositional attitude. However close is WAK’s extension to TAK’s, WAK is considerably more innovative. Its conceptual ingredients are less readily recognizable as characterizing knowledge. This makes trouble for citing anything like WAK to explain the prevalent and enduring interest in having knowledge. Aspects of knowledge that account for the interest must be sufficiently accessible for them to have been accessed by the many who have been drawn to knowledge for its own sake. The TAK conditions have an explanatory advantage here (assuming that a solution to the Gettier problem need not appeal to something conceptually exotic in specifying the knowledge-preventing accidents). The TAK conditions are relatively familiar and their attractions are relatively accessible.

8 Conclusion

Relying on any true belief has the same practical utility as does relying on knowledge of the proposition. The fact that truth makes all of the practical difference has been recognized since Plato’s Meno. In spite of this recognition, true belief without knowledge has never seemed as appealing.Footnote 20 Aiming to know the truth always figures centrally in serious inquiry. This perennial aim is understandable in light of the assets of knowledge that are on display in the TAK conception. According to its distinctive third and fourth conditions, knowledge requires having a confidence-supporting justification that is properly connected to an appropriate fact. We have seen two valuable things that this accomplishes. The rational relation to a fact makes knowledge intellectually satisfying. It also gives a more attractive unity to the states of affairs in which the TAK conditions obtain, thereby enriching their intrinsic value. This traditional conception thus explains two central ways in which knowledge matters to us.