Why Should Inquiring Minds Want to Know

Why Should Inquiring Minds Want to Know?: Meno Problems and Epistemological Axiology

Meno: In that case, I wonder why knowledge should be so much more prized than right opinion, and indeed how there is any difference between them.

Socrates: Shall I tell you the reason for your surprise, or do you know it?

Meno: No, tell me.

Socrates: It is because you have not observed the statues of Daedalus. Perhaps you don’t have them in your country.

Meno: What makes you say that?

Socrates: They too, if no one ties them down, run away and escape. If tied, they stay where they are put.

Meno: What of it?

Socrates: If you have one of his works untethered, it is not worth much; it gives you the slip like a runaway slave. But a tethered specimen is very valuable, for they are magnificent creations. And that, I may say, has a bearing on the matter of true opinions. True opinions are a fine thing and do all sorts of good so long as they stay in their place, but they will not stay long. They run away from a man’s mind; so they are not worth much until you tether them by working out the reason. That process, my dear Meno, is recollection, as we agreed earlier. Once they are tied down, they become knowledge, and are stable. That is why knowledge is something more valuable than right opinion. What distinguishes one from the other is the tether.^{^[i]}

National Enquirer commercials tell us that some people want to know. I have no idea what such a desire has to do with reading tabloid journalism, but the avowal of wanting to know interests me. Maybe this desire is shared by all; at the very least, curiosity is universal. Curiosity may amount to a desire for knowledge, or perhaps it might be explained in other terms, such as a desire for understanding or for finding the truth. Perhaps none of these, even. Maybe the desire is only one of being able to make sense of one’s experience of the world. Or maybe the important matter is not the existence of any desire at all. Perhaps, that is, it is not desire as such that drives the search, but rather some need or interest or purpose.

The questions raised by these meandering thoughts all have to do with the internal, psychological constitution of typical human beings. Such psychological questions lead naturally to axiological ones, for we can wonder whether what we desire and value is really valuable or desirable. So, regardless of whether humans desire or seek knowledge, is knowledge valuable? Or is merely getting to the truth the appropriate goal? Why should one value knowledge if one already has understanding, or what would be the use of knowledge about a phenomenon once one had made sense of it?

Such axiological questions appear first in the Meno, where Socrates addresses the question of what makes knowledge better than mere true belief. The question raises the possibility that knowledge isn’t nearly so important as we might suppose. Here I will identify the discipline of epistemology with inquiry into the nature, extent, and importance of knowledge. Given this stipulation, the Socratic issue raises the specter that epistemology itself is not very important. Epistemology has long been a central subdiscipline of philosophy, but deserved centrality must be earned by argument, for there is no a priori certainty that the relationship between mind and world, the nature of cognition and the appropriate standards for evaluating it, should be addressed employing the concept of knowledge. Reflecting on philosophical inquiry into cognition might lead us to the conclusion that addressing epistemological questions is ancillary to the fundamental tasks of philosophy. The hypothesis that I want to consider and argue for is the hypothesis that epistemology is really not central to the fundamental philosophical tasks of understanding the relationships between mind, world, and language; that epistemology is born of confusion, that it arises when we legitimately desire or value some things which are confused with knowledge (things such as infallibility, incorribility, permanence, unrevisability, metaphysical certainty, or the capacity to stand in the face of any amount of further learning). The result is that we think knowledge important, thereby committing ourselves to the centrality of epistemology to philosophy, when knowledge is really not all that important nor is epistemology central to our philosophical interests.

Epistemological Axiology, Knowledge, and Truth

In the Meno, Socrates and Meno discuss the value placed on knowledge over true opinion, but there are other problems of epistemological axiology besides this one. One could have arisen for Socrates, for a presupposition of the problem he discusses is that the value of true belief should be obvious. It is easy to see, however, that there is a question to be answered here as well: what value does true belief have over false belief, over belief that is merely empirically adequate, or over beliefs that make sense of experience? Furthermore, once time has introduced Gettier, we have the further question of what makes knowledge better than justified true belief^{^[ii]} (notice that this question is interesting and important even if one thinks that justification is not necessary for knowledge, for even if knowledge is, say, reliably produced true belief, we might have to conclude that it offers us nothing of value beyond justified true belief). The history of epistemology reveals much discussion of the question of the differences between truth and empirical adequacy, knowledge and true belief, justified true belief and true belief, and knowledge and justified true belief. But there is little attention paid to Meno problems, problems of epistemological axiology. It is on those problems that I want to focus here.

Let us begin, then, by asking what would justify the pre-eminence of knowledge, or what is the same thing, the pre-eminence of epistemology, in philosophical inquiry regarding cognition, regarding the connection between mind and world. What do we want out of cognition that would make us focus so quickly on knowledge? A quick and easy answer might be that it is knowledge itself that we want, but this answer is hasty. First, the history of epistemology is not on the side of this proposal, for it shows other concerns. Starting with Plato and Socrates, the importance of knowledge over true belief is explained, not on the basis of the intrinsic value of the former over the latter, but in other terms. For Plato, knowledge is “tethered”, and true belief is not; for Descartes, knowledge is not open to doubt and cannot be undermined by further learning, whereas true belief is and can. The history of epistemology suggests that knowledge is valuable because it is partially constituted by other properties that are obviously valuable. The problem for these defenses of the importance of knowledge is that the properties cited, properties such as incorrigibility, infallibility, permanence, “tetheredness”, metaphysical certainty, and the like, though immensely valuable, simply are not among the constituents of knowledge. Such a history suggests that the focus on knowledge results from some sleight of hand whereby knowledge is confused with that which is truly valuable. Furthermore, some acquaintance with the literature spawned by Gettier should make us wonder why we should want knowledge, if knowledge is ungettiered justified true belief. What interest do we have in such? I want to get to the truth, and I want to be sure I have. I’m sure you’re roughly the same in this regard, and it is important to notice that no mention of the concept of knowledge, especially no mention of anything ungettiered, need be made to understand such desires.

You might think I haven’t tried very hard to find the importance of knowledge, so I’ll work a bit harder at defending my thesis. First, I will consider what it would be like for cognition to be ideal, to be everything we could ever want out of it. This ideal, we shall see, may imply knowledge, but is not identical to it. So, the theory of knowledge is a mere footnote to what ideal cognition would get us. Of course, cognition is not ideal, and a bit more realistic approach to it might reveal an important place for epistemology, so I will consider what happens when we make adjustments for various features of the human condition that make this ideal impossible for us to achieve. Backing off of the ideal will help us grasp what types of interaction between mind and world we find most valuable, and whether the concept of knowledge ever plays a central role in the kinds of interactions that are to be valued most. I will argue that knowledge plays only an ancillary role at best, and hence that epistemology deserves no central place in philosophy.

Here’s what I want: every time I consider a proposition, I want to be able to tell immediately whether or not it is true, and for every proposition whose truth value affects me in any way, I want to have considered that proposition and stored that information so that no cognitive mistake ever causes me not to get what I want. Well, not quite; I exaggerate a bit. There are some things I guess I prefer not to remain ignorant of: the intimate details of my parents’ sex life, for example. But the desire for ignorance here is a product of interests I have other than purely cognitive ones. If my only interests were purely cognitive ones, what I said earlier would be true. Of course, were I nothing more than a cognitive machine, I might want other things as well, such as omniscience, but this will do for a start.

Others might agree with the idea behind what I just wrote, but will balk at this way of putting the desire. They may, for example, prefer to avoid the concept of truth altogether. They might argue that we have no interest in truth but only in that which is empirically adequate, or that which makes sense out of the entirety of our experience. I am somewhat sympathetic to such claims. At times, I think all that matters is having an understanding of things that can stand up to further testing. As any good student of scientific methodology knows, true hypotheses purportedly do just that, and so we are tempted to value true beliefs in virtue of their ability to withstand testing. But such thinking is confused. True beliefs are no more immune to defeat than false ones. Experts can deceive us, pockets of misleading evidence are commonplace, and auxiliary hypotheses confound our testing procedures.

There are other ways of attempting to secure the importance of truth, however, that I will only mention but not pursue here. Davidsonian arguments concerning the role of truth in interpretation,^{^[iii]} and externalist theories of mental content^{^[iv]} with their implications regarding the rarity of false belief come to mind. Alternatively, we might balk on theological grounds at the idea that truth is unimportant; if so, maybe epistemologists should be in Religious Studies departments! Since I don’t want to focus here on these issues, I’ll grant provisionally the value of truth, even though serious questions remain about the defense of this importance.

Putting aside the question of truth, let’s return to the picture of cognitive ideality above. There I noted that I want to be able to tell immediately and directly, without any special effort, the truth value of any claim that has or will have any effect on my life. I want cognitive excellence of a certain sort with minimal effort. I don’t think I’m idiosyncratic in this regard; I think we all want efficiency in producing desired output. We want a maximally efficient cognitive machine.

Alas, wishes are not horses. One of the lessons of experience is that such machinery is simply unavailable. So if we can’t have maximally efficient cognitive machinery, what might we want short of that? One answer is provided by Descartes. He wants some way of guaranteeing that his beliefs are true, and he wants to be able to secure such a guarantee without having to leave the comfort of his warm stove. A little less lazy than I, but not much. I like it.

Notice, however, that neither what Descartes nor I want should be confused with knowledge. In both cases, knowledge would be a mere by-product of getting what we want. So these desires provide no foundation for the centrality of epistemology to the philosophical enterprise. Hence, if these desires reflect what is valuable from a cognitive point of view, epistemology could become central only by confusing knowledge with something else distinct from it.

Recall, however, that we are supposed to be trucking in reality here, and Descartes’ ideal is every bit as chimerical as my earlier ideal for the lazy. Our goal should not to provide an account of cognitive ideality simpliciter; it should be, rather, to elucidate a realistic ideal, one that is humanly possible to achieve. Descartes’ ideal simply cannot be achieved; there is nothing or almost nothing about which we can be metaphysically certain. So the epistemological heaven described by Descartes turns out to be mere fancy.

Cartesianism is attractive in part because, like Descartes, we all want some guarantee that we have the right beliefs. First, we abhor the doublemindedness constituted by instability in the cognitive realm. We do not want to be in a condition where we keep changing our minds from one moment to the next about what to believe, nor is it good for us to be continually in the Pyrrhonian state of suspension of belief. But mere stability of belief--fixation of belief--is not what we want either. We are all aware of the delusional capacities of human beings, and we don’t want to live a lie. What we want is fixation on the right beliefs, not mere fixation of belief. We want fixed, true belief. If Descartes’ ideal were humanly achievable, it would get us fixed, true belief. But it isn’t. So we must ask, how do we retreat from Descartes?

Truth and Justification

Let’s cut to the chase and see what can be made of the idea that an appropriate retreat from Descartes will lead us to the promised land of epistemology. Here’s a try. If you want to fix true belief, you should want to be responsive to indicators of truth, or what is the same thing, to evidence of truth. When you are responsive to such, the beliefs on which you fix will be justified, or warranted.

This proposal attempts to sneak in a concept of evidence or justification or warrant in terms of truth indication or likelihood of truth, but the attempt should be rejected at this point. The defense of the theory of justification claims that if you desire fixed, true belief, you should want to be sensitive to indicators of truth. If the proposal means to include among such indicators signs that are only suggestive of truth, it should be rejected at this point. If I want chocolate, I’ll want to go to a store that sells it. I won’t be inclined to take counsel from those who tell me which stores look like they sell chocolate, or are likely to sell chocolate, over counsel from those who tell me which stores in fact sell chocolate. Recall that the only value we have identified is fixed, true belief. So all we are entitled to claim at this point is that if a belief is true and fixed, it is valuable, and if it is not true and fixed, it is not valuable. So, if a concept of evidence is legitimate here, the only things that will count as evidence are signs that imply the truth of what they signal. That is, no role has been found as yet for the epistemological pith of evidenced, false belief. Yet, if we have found a place for a concept of evidence, surely there must be such a role; so we should resist at this point the suggestion that the traditional epistemological concern with justification or warrant has found a place in our discussion.

One might claim that we have failed to find a place for warrant because we have ignored the difference between the single case and the long run. Truth may be all that matters in the single case, it might be claimed, but there is also the matter of what will happen in the long run. The strategies for battle taught at West Point might fail in a single case, but they are taught because they will generate more victories in the long run. Just so, the procedures, mechanisms, and methods that produce mostly true beliefs should be valued because they will produce more true beliefs over the long run than procedures, mechanisms and methods that produce beliefs unlikely to be true.

This response is unsuccessful. First, it does not follow from anything said so far, or anything that might reasonably be added, that more of one’s beliefs will be true in the long run if one uses methods or procedures likely to get one to the truth. Second, and more important, the defense is irrelevant. For if there are methods, mechanisms, and procedures that make it likely that we will get to the truth by employing them, there are other methods, mechanisms, and procedures that are even better. Consider the mechanism operative when and only when I believe the truth; ponder the method employed when one believes p if and only if p is true; contemplate the procedure followed by accepting p just in case it is true. If we want focus on the long-term prospects of getting to the truth and if we want to adjust our counsel regarding strategies of belief-formation so as to maximize getting to the truth and avoiding error, why not prefer the very best strategies such as those just delineated? If we do, we get the earlier result that true belief is all that matters; justification drops entirely out of the picture.

One might argue that justification or warrant doesn’t arise only when one adopts optimal methods for getting to the truth; satisfactory methods, when employed properly, can also generate warrant. This response is of no use whatsoever in our context. For we do not yet have any reason for thinking that justification is important. What I want to know is whether traditional epistemology is worth doing, and to this point all we have found is that true belief is valuable. So the project is not to construct a theory of justification or warrant; we are asking instead why anyone should undertake such a project at all. The attempt above was to argue that the project is important on the basis of a distinction between features of a single case and features of a long-run pursuit of truth. That attempt failed, and any attempt to avoid that argument by distinguishing between optimal and adequate results of inquiry is a complete non sequiter.

There are, however, two other ways I can think of to try to find a place for justification or warrant in a theory that grants the importance of true belief. One standard maneuver is to begin talking about means and ends, with truth as the end and justification as the means to it.^{^[v]} The other maneuver pessimistically resigns on the task of clarifying justification in terms of the goal of truth, claiming instead that it is itself valuable independently of any relation it might have to the goal of having fixed true beliefs.^{^[vi]} I will begin with the more pessimistic approach.

The position that justification is valuable independently of the importance or value of truth ought to strike us as an utterly mysterious one. It is akin to developing statistical categories in baseball that have nothing to do with winning baseball games. We keep statistics on batting average, slugging percentage, numbers of home runs, stolen bases, earned run average, fielding percentage, etc., because each of these has something to do with success in the game, i.e., winning. But suppose we introduce a further category: what percentage of times you step on home plate as you begin running toward first base, and claim that the lower percentage, the better (left-handers have an obvious advantage in this category, which, this left-hander holds, is all for the good). Puzzled, you query why anyone should be interested in this statistic. What does it have to do with success in the game of baseball? I answer that there is no connection, it’s just a valuable characteristic to have independently of any role it might play in winning games. You’d walk away perplexed by such a claim, I submit. I further submit that the same reaction is appropriate when it is claimed that justification has a value completely independent of the value of truth. The point of cognition is to get to the truth (so we have assumed here, at any rate), and the things we cite when we want to defend the truth of what we believe are usually (what we take to be) justifications of the truth. If that isn’t what justification is, if it is not connected to the truth in any interesting way at all, I don’t see why we’d be any more interested in it than in what percentage of times batters hit home plate on their way to first base.

If justification does not have intrinsic value, then perhaps it has instrumental value. Instead of thinking of justification as having value independently of truth, perhaps we will find its value in its relationship to truth: truth is the goal of inquiry and justification the means to it. There is even a way of seeing the first position, that justification must have intrinsic value, as arising out of this conception of the relationship between justification and truth. One gets forced into the intrinsic value position by assuming a restrictive account of what can count as a connection to truth. Once one restricts the possible connections, the next step is to become pessimistic about the prospects of connecting justification and truth on the restricted possibilities envisaged. Finally, uncomfortable with idea that justification is unimportant, one decides that no connection at all is needed; justification is valuable in itself, and not on the basis of any connection to truth.

The undue restrictions on possible connections to truth come by ignoring other possibilities. In particular, an ambiguity gets ignored in the concept of a means to a goal. This concept is ambiguous between intentional means and effective means. In the arena of action, the first concept is instanced when a person performs a certain action with the intention of realizing a certain goal. If I am chosen to take a shot from halfcourt at a Chicago Bulls basketball game for one million dollars, I will perform certain actions as a means to the goal of making the shot. I will, for example, face the basket; I’ll even shoot the ball. But I will perform no action that constitutes an effective means toward winning the million dollars. For, to be an effective means, the action must make it objectively likely that the goal is realized, or at least more likely than it would have been otherwise. In many cases, however, there simply are no effective means available. In the example, nothing I could do will make it likely that I make a shot from halfcourt, and nothing I could do will even raise the likelihood of my making such a shot. After all, I’m not as young as I used to be; maybe I’m beyond the point of even throwing the ball that far. Nonetheless, one would still try, and in trying, adopt some intentional means to the goal of making the basket.

When justification is conceived of as appropriately connected to the truth only when it is an effective way to truth, it is easy to become pessimistic about the prospects of a satisfactory account of the connection between justification and truth. I don’t want to vouch for the correctness of any particular complaint here, but one of the most worrisome difficulties has come to be called the New Evil Demon Problem.^{^[vii]} If justification must be an effective means to the truth, then inhabitants of evil demon worlds have hardly any justified beliefs. But how could that be? After all, the same is not true of us, even though they might, after all, be us. Or so the argument runs, at any rate. Plagued by this difficulty, one might simply give up on the idea that justification and truth are connected, resorting to the mysterious idea that justification is intrinsically valuable independently of any connection to the truth.

Notice that if we do not limit our conception of means to effective means but also countenance intentional means, more can be said. Some goals cannot be achieved directly, requiring the adoption of some means in achieving them. In the case of winning a million dollars at a Bulls game, that goal cannot be achieved directly. The first means that I must adopt is that of making a basket from midcourt. So that becomes my secondary goal. But note that it too cannot be achieved directly, so I must adopt some means of achieving it. Notice that this process of developing means toward goals stops when I get to actions that I can control with relative ease. I can, pretty directly, turn and face the basket. If you’re a fan of basic actions in action theory, you might want to insist that I go further, perhaps all the way to tryings: in order to reach the goal of turning and facing the basket, I must try to do so. But I don’t want to become mired in action theory here. I only want to point out that if we countenance intentional means to a goal, we can develop an account of when means should be adopted and how the ways in which we ordinarily approach such goals count as means to achieving them. If we talk only of effective means, nothing in the example is coherent. I have to make a basket in order to win a million dollars, but nothing I do is effective in making the basket. The only basis of evaluation available is, thus, whether or not I make the basket; no discussion of appropriate means can serve as a basis for positive evaluation in spite of missing.

Care must be shown in extending this discussion which is at home most in the arena of action to the arena of belief, for we do not want to assume that beliefs are voluntary in the way actions are.^{^[viii]} Perhaps something like the following is what we are after. The goal of truth for belief is relatively remote; it is not a property we can tell directly and immediately whether a belief has. So we should try to have, or value, beliefs with some other property, one that we can tell directly and immediately whether a belief has. In order to count as an analogue of an intentional means to a goal in the arena of action, this property must be one that is appropriate (or the best we can do), by our own lights, for getting to the truth. Such a description generates the appropriate analogue in the arena of belief of a concept most at home in the arena of action, and it does so without requiring that beliefs are voluntary.

These points yield a lesson for those who wish to defend the importance of justification by appealing to the distinction between means and ends, or between intrinsic and instrumental value. The very first question such an approach must be able to answer is why the distinction between means and ends is introduced in the first place. The only adequate response must appeal in some way to the mediacy of the goal and the immediacy of the means, to our inability to simply achieve the goal directly and immediately in the case of action, to our inability to tell immediately and directly whether the goal has been achieved in the arena of belief. Yet, if only effective means count as means, this answer is unavailable. For no property a belief has is such that it is necessarily likely to attach to a true belief and also one that is easier to tell a belief has than the property of truth itself. Thus, the position that justification has a value derived from that of truth becomes indefensible when means are limited to effective means, for such a defense simply must be able to explain why the talk of means arises in the first place. Only by countenancing intentional means can such an explanation be developed, however.

So, if justification is instrumentally valuable in virtue of its relationship to the truth, that relationship must be conceived in terms of intentional means to the goal of truth rather than in terms of effective means. Intentional means theorists can say that truth is not a property of a belief that we can tell immediately and directly whether it is present, so we should adopt some means for getting to the truth. Moreover, in order to fulfill this motive for introducing the concept of a means to the truth, the property that results from following such means must be internalist in character--it must be such that one can tell by reflection or introspection alone that a belief has that property; it must be a discernible property of a belief.^{^[ix]}

Effective means theorists might not succumb yet, however. They may want to insist that even if justification must be introduced into the discussion in the guise of intentional means to the goal of truth, we won’t have justification unless the intentional means are also effective means. This ploy is not directly relevant to the question of the importance of justification, but I will make one brief comment about it. I would remind the reader that the most thorough investigation of such an approach--Larry BonJour’s^{^[x]}--ends in the deepest skepticism about the existence of justified beliefs, a skepticism strong enough to entail all other forms of skepticism about justification.

Effective means theorists still have one complaint. They may complain that we are showing a prejudice against approximations. They may say, that is, that even if the very best would be for us to always and only get to the truth, some failures are closer approximations to that ideal than others, and hence more valuable. They may insist that we are exhibiting a kind of childish response of saying that if we can’t have the very best, we don’t want anything at all. A more mature approach would be to recognize the unmatched value of always finding the truth, and yet grant the value, albeit lower, of always holding beliefs that are highly likely to be true.

An appropriate retort to this insult is to point out that maturity often culminates in senility. Something like this culmination occurs here, I think. You might buy a TV that works half the time instead of one that works only ten percent of the time; but that decision gives you no reason to prize your acquisition. The old joke is that close doesn’t count except for hand grenades, government work, and jazz. The point of the joke is not that approximations are sometimes valuable, but rather that in certain areas getting close to what would be the goal in other domains just is to have fulfilled the goal in other areas. That is, the goal is simply different, and easier to achieve, in some areas than in others. So the appeal to the value of approximations simply won’t work.

There is, however, one appeal to approximations that would work. If we first established that the goal could not be achieved, or could not be achieved directly, then we could turn to ways of getting to the goal that might only yield approximations of the ideal rather than fulfillment of it. Yet, if it is the best we can do, no more could reasonably be expected, and the results of doing the best we can should be valued. If, for example, the best TV’s made only work half the time, then in purchasing such a machine, you’ve purchased something valuable (as long as there is something valuable in watching TV). Such a response, however, returns us to the land of internalism. The best we can do, the best processes to instantiate, the best methods to follow, must be understood first in terms of the doxastic analogue of intentional means to a goal. The moral of the story is that the centrality of epistemology to the philosophical enterprise presupposes the internalistic character of warrant or justification; without it, all we get is the value of truth over error.

Let me make it clear that I am not endorsing the internalist proposal here, for I think that this approach faces difficulties of its own. For one thing, there must be a strong emphasis on the immediacy and directness of telling which beliefs are true, for a quite natural response to the claim that we can never tell, immediately and directly, which of our beliefs are true would be incredulity. “Of course I can tell, even immediately and directly, that some of my beliefs are true--for example, that I am sitting in a room, that the temperature is not over 1000 degrees celcius, and so on for a long list of other beliefs.” Call this the Moorean response.^{^[xi]} Internalists must reject the Moorean response, and to do so plausibly, they must clarify the notions of immediacy and directness in such a way that the Moorean response is blocked. Perhaps this can be done, and perhaps not. One very suspicious way of answering the Moorean response appeals to infallibility: we can’t immediately and directly tell that any of our beliefs are true, because we might be wrong about them. Given this response, justification would have to be an unmistakeable property of a belief. Such a reply to the Moorean response is suspicious precisely because of its appeal to infallibility. The history of epistemology is littered with the carnage of appeals to infallibility, and claims about the infallibility or unmistakeability of justification seem equally ill-fated.

Perhaps a better route would be to focus the issue of the metaphor of transparency. The idea would be that if you are justified in believing a claim, then if you reflect on the question of whether the belief in question is a good one to hold given the goal of getting to the truth and avoiding error, you’d be justified in believing that the belief is appropriate to the epistemic goal, i.e., it would be obvious to you that your belief is justified. Truth has no such universal transparency. A belief might be true and you nonetheless conclude, on reflection, that it is false, that it is not a good one to hold given the goal of getting to the truth and avoiding error, and that it is not justified. So, perhaps, the only epistemically important concept of justification is one for which transparency holds.^{^[xii]}

There are costs of such a theory, however. For most theories, e.g., foundationalist, coherentist, reliabilist, or evidentialist, do not honor this metatheoretical constraint. If not, then no matter what else one might say about such theories, they are irrelevant to the task of defending the importance of epistemology. They may constitute theories of one requirement for knowledge, but they do nothing to show that there is anything cognitively important other than true belief.

The fact that standard theories of justification typically do not yield the result that justification is transparent to reflection may lead us to question the transparency defense of justification. We may decide, that is, that transparency is too strong a requirement on a theory of justification, leaving the above defense of justification still susceptible to the Moorean response. We may then begin to wonder whether the Moorean response can be answered at all. If it cannot be answered, then there will be no defense of the importance of justification. But even if the Moorean response cannot be answered, there is a useful lesson here. For the Moorean response threatens to undermine the only available way of defending the importance of justification by appeal to the distinction between means and ends, and that defense is clearly internalist in character. The lesson of this result is that there is simply no place in a defense of the importance of epistemology by appeal to the distinction between means and ends for externalist theories of justification. Such theories make use of a tow by internalism across the waters from philosophical ignominy toward the land of useful theory, only to eschew internalism on sighting land, failing to take into account that they drown without the tow. I’m not saying the internalist will make it to shore; the two may drown together. but at least the internalist has this to be said on his behalf: he was, after all, making a significant effort, the only one with any hope of success.

Knowledge and the Gettier Problem

Though I have not defended the acceptability of the internalist defense of the importance of justification, I want to grant the importance of both truth and justification for the rest of this paper. For even if both of these are important, the importance of epistemology has not yet been shown, since knowledge is more than justified true belief. Let me repeat that I do not mean to suggest that there is no such thing as knowledge or that it would be improper to develop a theory of it; what I’m questioning is the importance of doing so. Why should we care at all what knowledge is, even if it matters to us what truth and justification are and which of our beliefs have these properties?

This question becomes more poignant when we reflect on the difference between knowledge and justified true belief. Knowledge, it is said, is not accidental; the connection between belief and truth is not accidental and the connection between justification and truth is not accidental either. Why should we be interested in such a concept? Suppose all of your beliefs are justified and true; what difference would it make if you didn’t have any knowledge, if all your justified true beliefs were only accidentally so? You would be an epistemological Mr. Magoo, accidentally getting things correct, right and left. But so what? Why should you care?

Of course, I don’t doubt that some do care. But maybe we care, when we do, because we desire something even stronger than knowledge, perhaps something like the metaphysical certainty Descartes describes. We’ve been down this path already, however, and it is time for reality therapy to have its effect. Once we adopt more realistic desires, it is not clear that it matters at all whether we ever instance the ordinary concept of knowledge.

You might think something in the literature on the nature of knowledge might provide ammunition for defending the importance of knowledge. A brief look at some representative approaches suggests otherwise. What we will find is that the capacity to address axiological issues in an interesting way is strongly correlated with failure at solving the Gettier problem, but I must caution that my argumentation here will be sketchy at best. As our continental brethren might put the point, argumentation will appear in what follows in “the privative mode.”

Some approaches to the Gettier problem focus on the role falsehoods play. According to such approaches, the absence of knowledge is a result of evidence that contains falsehoods, or presupposes falsehoods, or confirms falsehoods.^{^[xiii]} Such approaches suggest that knowledge might be important because it insulates us from error, beyond the object of belief itself. That viewpoint, however, is simply mistaken. First, among the lessons of the preface and fallibility paradoxes are that we nearly always have evidence in support of some false propositions; I also think that these same paradoxes can be used to show that the justifications for those aware of their own fallibility nearly always presuppose falsehoods. Second, the possibility of statistical knowledge undermines many of these approaches. A statistical sample can deviate in statistically significant ways from what it confirms about a population, and still be used to gather knowledge about that population. For example, a sample might give us knowledge that most swans are white even though the actual percentage of white swans is significantly different than our sample confirms. We can agree that insulation from error is important, but the lesson here is that one can’t defend the importance of knowledge by citing perhaps unfortunate features of the human condition that knowledge has no power to displace.

Another approach to the Gettier problem imputes to knowers a truth-tracking power.^{^[xiv]} When you lack knowledge, on this approach, either you would still believe the claim if it were false or there are close counterfactual circumstances in which it is still true in which you would not believe it. Yet, these counterfactuals alone are of little interest to us. A benevolent demon might just devote himself to making the appropriate counterfactuals true so that I count as a truth-tracker; another, malevolent demon might make those same counterfactuals false regardless of the quality of my evidence. The truth or falsity of these counterfactuals alone is of no special interest to us, so if their truth is what makes the difference between justifed true belief and knowledge, then knowledge is of no special importance either.

There is a reply here, that it is not the truth of the counterfactuals that interests us, but the explanation of their truth. We want the counterfactuals to be true in virtue of capacities or abilities we have at finding the truth, so even though the mere truth of the counterfactuals may not be important or valuable, their truth in virtue of our powers is. If the counterfactuals are true because of the activity of some demon, they don’t interest us much; but if they are true in virtue of our cognitive powers, they are interest us, and properly so.

The question we must ask, however, is what all this has to do with knowledge. It is not hard to see that knowledge is not to be identified with justified true belief, even where the true belief occurs in virtue of our cognitive powers in such a way as to sustain the truth of the counterfactuals above. The closest such suggestion one can find in the literature is Nozick’s truth-tracking theory, but this proposal is nothing more than Nozick’s intuitive theory which he immediately amends in the face of obvious counterexamples.^{^[xv]} These counterexamples imply that production of justified true belief by suitably impressive cognitive powers is neither necessary nor sufficient for knowledge. So even if we could defend the importance of impressive cognitive architecture as above, such a defense simply does not yield the conclusion that knowledge is important.

There is another problematic feature of this proposal. This reply ignores the role a suitable, cooperative environment plays in making the counterfactuals in question true. When the counterfactuals are true, they are never true solely in virtue of our cognitive powers. The environment itself must also be suitable for the operation of those powers. If we had infallible powers, this claim would not be true, but we do not have such powers, and hence the cooperation or suitability of the environment is required for the truth of the counterfactuals in question.

What is attractive about the idea that we want to have a hold on the truth in virtue of our own abilities, I submit, is that we would like to be in control of the cognitive affairs in question so that our hold on the truth is not fortuitous or accidental. But once we see that it is never solely in virtue of our cognitive powers that we find the truth, we must grant there is always a bit of fortuitousness present when we find the truth. Without a suitable environment, none of our powers would be sufficient. So if it is fortuitousness that we hope to eliminate, we are hoping for something that cannot be had short of possessing infallible powers of discernment. I grant that such powers are surely desirable and valuable, but any attempt to salvage the importance of epistemology that appeals to the value of infallible powers of discernment will surely fail. For knowledge simply has little to do with infallibility.

Even if fortuity cannot be eliminated completely, one might still wonder why it wouldn’t be valuable to eliminate some elements or kinds of it even if others can’t be eliminated. So, since to know something is to eliminate a certain kind of fortuity regarding one’s true belief, it is valuable to have knowledge even if there are others kinds of fortuity that are not eliminated by knowledge.

One problem with this response is that it is simply not obvious that there is a distinctive kind of fortuity that is ruled out by knowledge, except by stipulating that the kind in question is just that kind eliminated by the whatever closes the gap between justified true belief and knowledge. It is often pointed out that the reason that justified true beliefs can fail to be knowledge is that they can be accidentally true, but that gives us no reason to think that knowledge can be defined in terms of non-accidentally true, justified belief, nor in terms of justified true belief that possesses some particular kind of non-accidentality. No one taking themselves to provide an analysis of knowledge also takes themselves to be analyzing the notion of accidentality or some kind of it when giving an account of knowledge. Instead, what is being done is to give an account of knowledge on which the condition which plugs the gap between justified true belief and knowledge falls under the genus of non-accidentality. This characterization leaves entirely open whether there are is any kind of fortuity or accidentality that is ruled out by knowledge, except where that kind is stipulatively picked out in terms of the species of non-accidentality that plugs the gap between justified true belief and knowledge. If that is how we are conceiving of kinds of accidentality, however, I doubt there is any way to sustain the above claim that ruling out this kind of accidentality is valuable unless any other stipulated species of the genus in question is similarly valuable to eliminate. And this latter claim is highly dubitable. Consider the fortuity I experience of having many more beliefs about a particular locale within the state of Texas than other locales. Eliminate that fortuity and I won’t have more detailed information about any locale over any other, and that implies that I would lose a certain depth of understanding that is valuable. So, what is needed and lacking in the above proposal is a defense of the claim that the particular kind of fortuity that knowledge rules out is valuable.

Worse yet, as we’ve imagined the situation, we are considering individuals who are unusually blessed by luck. Not only does it fail to be valuable to eliminate all kinds of fortuity, there is some fortuity that is immensely valuable. One kind is the fortuity of having no knowledge but yet having only justified true beliefs and a broad range of them. This kind of fortuity I’d be glad to have. Even more, I’d prefer this kind of luck to the absence of luck of those who know a lot. And for those who have the luck of knowing everything the believe, it is simply not obvious why the way in which they fail to be lucky is to be preferred over the luck possessed by those who only have justified true beliefs. There is simply no uniform answer to the question about when fortuity or non-fortuity is valuable, and all I can see available for a defense of the importance of knowledge here would be a claim to the effect that the absence of any kind of fortuity about true belief is always a good thing. It isn’t, and hence the importance of knowledge cannot be defended in this way.

Another approach to the nature of knowledge claims that if you don’t have knowledge then either your cognitive equipment is not functioning as it was designed to function or it is not operating in an environment suitable to that design. So suppose I regularly have justified true belief but no knowledge; then either I’m an engineering nightmare (perhaps, better put, an engineer’s nightmare) or I’m not in Kansas anymore. Upset, I’m not; I’ll just blissfully smile and bless the gods for my justified true beliefs while some epistemologists sing their lament. I’d join in the refrain if there were an argument that my condition will lead, or likely will lead, to failures of justified true belief in the future, but no such argument is available.

Another approach to the Gettier problem is the defeasibility approach,^{^[xvi]} and we might wonder whether there is some special value attaching to justified beliefs that are not defeated. I agree there is, but knowledge is not undefeated justified true belief. One of the lessons of that literature is that not all defeaters undermine knowledge; some defeaters are misleading ones and some are not^{^[xvii]} (perhaps because some defeaters are ultimately overridden by other information and others are not). If a defeater is misleading, it does not undermine the claim to knowledge; if it is not misleading, it does. So even if we value undefeated justification, that gives us no reason to value knowledge over justified true belief, for knowledge does not eliminate defeaters. And if there are defeaters present, I don’t see how the distinction between defeaters that undermine knowledge and those that do not is of any significant importance to human interests, needs, or purposes. One might answer like this: when you are subject only to the misleading defeaters, you could learn everything epistemically relevant to your belief and still be justified in believing what you presently believe. That is true, but it is also true when you are subject to non-misleading defeaters, for there aren’t any justified truths that one with all the evidence would fail to be justified in believing.

Knowledge, Assertion, and Understanding

There are, however, two other ways to try to answer the gettierized version of the Meno problem that strike me as a bit more interesting than the attempts considered so far. The first attempt claims that we should want to avoid accidentally true beliefs because they show that we lack an accurate picture of the interrelationship between things, that we lack an adequate understanding of the explanatory connections in nature. The second attempts to link the conditions for knowledge with speech act theory. I turn first to the lack of understanding approach.

The lack of understanding approach insists that we should want to understand what makes what true, and having only accidentally justified true beliefs bars such understanding. This approach is interesting to me because it suggests that there are certain intellectual virtues that we will not possess if our beliefs are only accidentally true. Understanding is one such virtue; perhaps wisdom is another. This approach thus suggests that the virtues themselves are important, and that mere justified true belief is no definitive mark of them. I agree on both counts, but unfortunately this emphasis on the virtues will not rescue epistemology, for such an approach to defending the importance of that discipline is at least too strong and perhaps too weak as well. It is too strong because knowledge doesn’t require explanatory understanding; instead, it is at most implied by this kind of understanding. One doesn’t need to have any explanatory understanding of a phenomenon in order to learn (say, by testimony) of its occurrence. Thus, if explanatory understanding is what is crucially important beyond justified true belief, we’d have a reason for replacing epistemology with the project of constructing an adequate account of explanatory understanding. The theory of knowledge might compose an addendum to such a project, but the point would remain that we should spend less of our time doing epistemology and more time thinking about understanding. This viewpoint is one I am quite sympathetic with, but I will not defend it further here.

The answer may also be too weak, for I am not certain that explanatory understanding implies knowledge. It may be that a suitable range of true beliefs, or justified true beliefs, will be enough for explanatory understanding. If I can correctly answer all questions directly from information I possess about, say, the rise and fall of Comanche dominance of the southern plains of North America in the late seventeenth century through 1875, then, I’m inclined to think, I have some understanding of that phenomenon. Do I have knowledge? Maybe yes; maybe no. On behalf of knowledge ascription, there is ordinary language: if you heard me give all the answers, you would certainly say that I know a lot about the subject. Ordinary language, however, is far from decisive. All it shows for sure is that you’d have good evidence for concluding that I have knowledge. Furthermore, on the usual theories of knowledge, I could give all those answers from information possessed and still lack knowledge. I can answer in this way and lack justification, and I can answer in this way and still have only accidentally justified true beliefs.

Of course, such vagaries may lead you to wonder whether being able to answer questions on the basis of information possessed is enough for understanding. I grant the need for a good argument here, and I’m not sure there is one. But at least this much can be said. Whether a person understands seems to be principally a matter of that person seeing the relationships between various (true) propositions. The contrary suggestion would have to be, then, that though someone might correctly see the relationships in question, they don’t understand because, perhaps, it is only accidental that they see these connections. That strikes me as strained; I’m inclined to conclude that, if it is only accidental that they see these connections, then it is only accidental that they have the understanding that they have. I grant that there is more by way of confession than argument here, but the response that strikes me as the most natural one is to posit the possibility of understanding by accident rather than insist that understanding implies knowledge.

Regardless of how that issue is resolved, the above appeal to loss of understanding when our justified true beliefs are only accidentally so doesn’t justify a focus on knowledge, so let us turn to the speech act attempt to see if it fares any better. The theory of speech acts posits conditions under which we are licensed to make various assertions--call these conditions assertibility conditions. For your action of asserting p to be legitimate, you must satisfy the assertibility conditions for p. The speech act rescue of epistemology posits that assertibility conditions include all the conditions for knowledge. First, you are required to believe p in order to legitimately assert it. We often reprimand each other’s assertions by pointing out, “You don’t really believe that,” to which we might hear the reply, “Yes, you’re right, I shouldn’t have said that; I’m only feeling sorry for myself.” Second, you must have good reasons for what you say, and what you say must be true. My father sometimes says that earthquakes are much more common in the twentieth century than in prior centuries. I question how he can say this since he has no evidence for it. His only reason for asserting it is that it fits his eschatological views quite well. (I’m going to assume without argument here that fitting his eschatological views is not reason enough for him to believe what he does. He agrees in this case; his sheepishness at admitting he has no better reason confirms it.) It turns out that he’s right, and there is available scientific confirmation of what he thinks. But he didn’t have the evidence, and so shouldn’t have made the claim (stubborn cuss that he is, an inheritable trait by the way, he won’t acknowledge the point, but that is another matter...). On the matter of truth, it is easy to elicit a retraction by showing that what a person said is false. I say, “Foley is not here,” and you point him out, and I take it back. Finally, failure to satisfy the Gettier condition for knowledge engenders the same type of retraction. Consider the Nogot-Havit case, in which Nogot convinces you that he has a Ferrari, from which you infer that someone in the room owns a Ferrari. If you assert that someone in the room owns a Ferrari, we can get you to take it back by showing you that Nogot in fact does not own a Ferrari. You are required to withdraw your assertion as much as you are when the other conditions for knowledge fail.

Such considerations are supposed to convince us that epistemology is a central subdiscipline of philosophy because the conditions for knowledge are assertibility conditions on utterances.^{^[xviii]} I do not think the defense is successful, however. First, I’ll engage in a bit of border skirmishing, to the effect that the connections between legitimate assertion and knowledge are weaker than the above suggests. Afterwards, I’ll get to the main reason the proposal fails.

Let’s begin by being precise about the proposal. First, it claims a logical relationship between the assertibility of p and knowing p: p’s being assertible by S entails S’s knowing p, where p’s being assertible is understood to mean that it is permissible for S to assert p. This claim is false. Consider someone who takes Pascal’s advice of going to Mass and hoping for the best; in line with such advice, a person may sincerely avow that God exists even though that person does not (yet) believe it. Furthermore, no one in such a condition need be moved to retract the assertion upon complaint that the assertion is not backed by belief. Or, again, consider someone, moved by William James’ arguments, ^{^[xix]} comes to believe that God exists and asserts it, all the while knowing that there is insufficient evidence to confirm this claim. Such a person need not retract the claim when the absence of evidence is noted. For a third case, imagine Churchland or Stich asserting, “I believe nothing that I assert,” something their writings imply.^{^[xx]} On the speech act proposal, no such sentence is assertible, but it is easy to see that neither would be guilty of any impropriety in his assertion, even if it should turn out that they are philosophically mistaken in claiming there are no beliefs. This latter case is a special instance of skeptical and Pyrrhonian assertion. (As an aside, such possibilities of appropriate assertion show that Moore’s paradox, the paradox that arises from a person’s asserting that a certain proposition is true but that he or she doesn’t believe it, is not always paradoxical; it is paradoxical only given certain assumptions about how standards for belief and standards for assertion are correlated.) Skeptics are quite comfortable asserting that knowledge is not possible, and Pyrrhonian skeptics not only assert that knowledge is not possible but that it is best not to hold any beliefs at all (though the interpretation of their belief/appearance distinction is quite subtle). In each of these cases, assertion is legitimate even though the question of whether it is backed by knowledge is left open.

So it would appear that the most that can be claimed is that knowing p is ordinarily a requirement for legitimately asserting p. What, precisely, does such a claim mean, however? It is, of course, senseless to claim that one thing ordinarily entails another. Perhaps the notion of a requirement should be understood differently. Instead of thinking of requirement in terms of entailment, perhaps we should think of it as a defeasible relation, so that x can require y even though x&z does not require y. If we understand the notion of requirement in this way, we might still be able to claim that the assertibility of p requires knowing p even though in the unusual cases described above additional factors come into play so that the assertibility of p in those circumstances does not require knowledge.

Even such retrenchment cannot save the proposal, however. Recall the way in which the proposal began. We look for ways in which we can require a person to take back an assertion, and it turns out that the conditions for knowledge are among those ways. There are, however, two quite different things a person might be doing in taking back an assertion. The person might be taking back only what is said, or she might be taking back the saying of it. So, for example, if Joe says, “I don’t know why I keep trying to be friendly, nobody likes me at all,” and Mary says, “Joe you don’t really believe that; you’re just upset,” Joe might apologize for saying what he knows is false. Such cases support the view that belief is ordinarily a requirement on assertion because the retraction involves a taking back of the saying itself. Moreover, when people assert things without any good reason whatsoever, we reprove such utterances, realizing that even though what is said might be true, those individuals have no business saying so. So once again, it is the saying itself that is at fault. But when new information is presented that undermines an assertion, only a retraction of what is said is in order. In the Nogot-Havit case,^{^[xxi]} pointing out that Nogot does not own a Ferrari should lead you to retract your claim that someone in the office owns one. It would be quite bizarre to hear you apologizing for having made the original claim; it is what was said that was mistaken, not the saying of it. Again, if I assert what is false, and you show me that it is false, I’ll retract my statement. But I wouldn’t say that my uttering of it was out of order.

This distinction thus suggests that there is a better account of assertibility than the speech act proposal aimed at rescuing epistemology. Instead of conditions of knowledge being assertibility conditions, only belief and justification are among such conditions. So here’s the story we should tell about the connection between cognition and assertion. The appeal to knowledge in account for assertibility is superfluous, for all the explanatory work can be done by the concept of justified belief itself. When the saying itself is inappropriate, that is so in the ordinary case where standards of assertion and standards of belief converge because justified belief was not present when the assertion was made. When what was said must be retracted, that is so because justification for belief is no longer present. That is all the explanation that is needed, and no appeal to concept of knowledge is involved in it.

There is, however, still a smidgen of a difficulty. For when you find out that you’ve been gettiered, there is some residual embarrassment or regret for your assertions. If you assert that someone in this room owns a Ferrari on the basis of believing that Nogot owns one, and then are told that Nogot has deceived you, you might not only retract your statement, but experience some embarrassment or regret for the assertion. Are you thereby showing the inappropriateness of assertion in the absence of knowledge? No. You are showing that you dislike being duped, and one can be duped because of truths that undermine knowledge and because of truths that do not undermine knowledge. You’d experience the same embarrassment or regret if the defeater mentioned were a misleading one. The classic case involving misleading defeaters is the case of Tom stealing a book from the library.^{^[xxii]} You see Tom steal the book, and have a justified true belief that he stole it. But his mother, an inveterate liar, tells the police that it was his twin brother, Tim, who stole it. The police know the story is concocted; they know she is an inveterate liar and will say anything to protect Tom. They also know Tom has no twin. But you don’t know all this, and if you were told what the mother said, you’d be every bit as inclined to take back the assertion that Tom stole the book and to experience some embarrassment or regret for having confidently claimed it. But the defeater is a misleading one; without having been told it, it does not undermine your knowledge, because it is so obviously farcical. Yet, even misleading defeaters, when discovered, undermine knowledge; that is what makes them defeaters in the first place. The lesson, then, is that your embarrassment, shame, or regret is not an indicator that knowledge is a prerequisite of appropriate assertion. Instead, it is only a sign that none of us are comfortable with the existence of information we are not aware that would undermine the justification of our beliefs. To have justifications that are immune from defeat is, of course, a valuable characteristic of a belief (provided justification itself is), for such justification has a kind of permanence to it--it is tethered, to use Socrates’ metaphor, and will not fly away. To conclude, however, that one should never say anything for which one lacks immunity from defeat is, to paraphrase William James, to show a preoccupation with not being duped.^{^[xxiii]}

Conclusion