1,203 218 250KB
Pages 15 Page size 595 x 792 pts Year 2007
It's Not What You Know that Counts Mark Kaplan The Journal of Philosophy, Vol. 82, No. 7. (Jul., 1985), pp. 350-363. Stable URL: http://links.jstor.org/sici?sici=0022-362X%28198507%2982%3A7%3C350%3AINWYKT%3E2.0.CO%3B2-Y The Journal of Philosophy is currently published by Journal of Philosophy, Inc..
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/jphil.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact [email protected].
http://www.jstor.org Fri May 18 08:48:29 2007
350
THE J O U R N A L OF PHILOSOPHY
IT'S N O T WHAT YOU KNOW T H A T COUNTS*
I
N 1963, Edmund Gettier published a paper called "Is Justified True Belief ~ n o w l e d ~ e ? "in' which he offered two counterexamples to the claim that this question should be answered in the affirmative. He described two cases in which a person has, for some proposition P, validly inferred P from some proposition which she was justified in believing but which was, lamentably, false. His claim: these are two cases in which a person has justified true belief but fails to have knowledge. Anyone who is the least bit familiar with Anglo-American epistemology of the last two decades is aware of what the upshot of Gettier's paper was. A great many philosophers in a great many articles acknowledged that Gettier's putative counterexamples were genuine and undertook to amend the analysis of propositional knowledge which Gettier had shown to be defective. Indeed, if the sheer volume of philosophical work published on a given topic is a guide to the importance of that subject to philosophy, then it is fair to say that the proper emendation of the justified-true-belief analysis of ' S knows that P' became almost immediately one of the major problems of epistemology. Now, over twenty years after the publication of Gettier's paper, it seems to be worth asking wherein exactly its importance lies. For it is a remarkable fact that very little of the vast literature spawned by its publication seems to have been devoted to a serious attempt to locate the importance of Gettier's counterexamples. My purpose in this paper is to argue that this fact is not only remarkable-it is ominous. I intend to argue that the assumption lying behind the constructive responses to Gettier's paper-to the effect that Gettier's examples show that a historically important definition of knowledge is in need of revision-is mistaken. I intend to argue that neither is the definition of knowledge Gettier showed defective of historical importance nor is its revision of contemporary importance-that the moral to be drawn from Gettier's counterexamples is that what you know doesn't matter. *I a m indebted to Joan Weiner for rxtensive discussions, enlightening suggestions a n d frank criticism. I would also like to acknowledge the helpful advice offered me by my other colleagues at T h e University of Wisconsin/Milwaukee and (during a leave of absence) at T h e University of Pennsylvania, to w h o m I read versions of this paper. In particular, I would like to thank Margaret Atherton, Donna Catudal, J o h n Koethe, Alexander Nehamas, Robert Schwartz, a n d Robert M'achbroit. A shorter version of this paper was read at the Pacific Division meetings of the APA in March 1985. ' ~ n a l ~ s zxxrrr.6, s, n.s. 96 (June 1963): 121-123.
0022-362X/85/8207/0350501.40
01985 T h e Journal of Philosophy, Inc.
I T ' S N O T WHAT YOU K N O W THAT C O U N T S
35 1
Let us begin by taking a historical perspective. A philosophical problem often derives its life from the fact that it is a problem for some doctrine advanced by one or more great philosophers of the past. Thus if the analysis of knowledge Gettier attacked were historically important-if Gettier were posing a problem for a historically important conception of knowledge-the response to his counterexamples would be understandable. There is indeed indication that many philosophers have thought that the analysis of knowledge Gettier attacked is a historically important doctrine. Gettier himself suggests that "Plato seems to be considering some such definition [of ' S knows that P ' ] at Theaetetus 201 and perhaps accepting one at Meno 98" (121). And his respondents seem to agree. Roderick Chisholm calls the justifiedtrue-belief conception of knowledge "the classical c ~ n c e ~ t i o n . " ~ Brian Skyrms, again citing Plato, calls it "the traditional definition of knowledge."3 And in a published address, a large excerpt of which was selected by Michael Roth and Leon Galis to serve as a n introduction to their well-known collection of papers on the "Gett i e r - p r ~ b l e m , "A. ~ M. MacIver traces what he takes to be a clear continuum from Plato's threefold attempt to define knowledge in the Theaetetus through Russell's discussion of knowledge in the thirteenth chapter of T h e Problems of Philosophy to Ayer's definition in T h e Problem of Knowledge-one of the definitions Gettier singled out for attack. "The general line of argument," writes MacIver, "can be indicated by drawing examples from Plato and Russell indiscriminately" (7). But although the picture of Gettier's having revived interest in a venerable Platonic philosophical enterprise is attractive, it is not the least bit obvious from the Platonic texts themselves that this picture is accurate. It is certainly true that, in Theaetetus 201, Plato has Theaetetus propose that we should think of knowledge as true belief "with a n account." But consider what Plato has to say by way of explicating this notion of a n account. Theaetetus and Socrates jointly explore three proposals, Socrates suggesting each and, each time, securing the agreement of Theaetetus. The first proposal is that an account consists in "making one's thought plain by means of speech, with expressions and names" (206d), that an account is "a sort of image of thought in speech" ( 2 0 8 ~ )the ; second is that to give an account Theory of Knowledge, 2nd ed. (Englewood Cliffs, N.J.: Prentice-Hall, 1977), p . 102. "The Explication of ' X Knows that p'," this JOIIRNAL., L X I V , 12 (June 22, 1967): 373-389, p . 373. 4 ~ n o w i n g(New York: Random House, 1970).
352
T H E JOIiRNAL OF PHILOSOPHY
is just "to go through the thing, element by element, till one has gone through the whole" ( 2 0 8 ~ )the ; final proposal is that to give a n account is "to state some mark by which the thing one is asked for differs from everything else" ( 2 0 8 ~ )Each . proposal having been criticized, the project is abandoned. Whether Plato had given u p on the project or later revived it is a matter of conjecture. What is clear is that none of the three proposals is anything like what one would expect from a philosopher w h o was trying to say what it is to give a justification for believing a proposition is true or what it is to know that a proposition is true. T h e proposals appear rather to be the product of a philosopher who was trying to say what it is to know a nonpropositional object like ( t o use a n example from the Theaetetus itself, 207a) a wagon-what it takes to identify a wagon or what it takes to know what a wagon is. T h e nature of propositional knowledge seems not to be at issue at all. Let us consider, then, Meno 97-98, where Plato has Socrates say that what distinguishes knowledge from mere true opinion is that the former is "tethered" by "working out the reason." Once again the temptation to think that Plato is defining propositional knowledge of the sort Gettier has i n mind is hard to sustain once one notices how Socrates elaborates upon his claim. "That process, my dear Meno, is recollection as we agreed earlier. Once they [true opinions] are tied down, they become knowledge a n d are stable." As is well known, the doctrine of recollection is invoked earlier i n the Meno to explain how we can seek to identify (and inquire into the nature of) Platonic abstract objects such as virtue. T h e doctrine states that such inquiry is actually a n attempt at recollecting a n innate knowledge of these objects-knowledge present i n us at birth. It is knowledge of abstract objects such as virtue whose explication Plato is seeking, and it is only the innate stable knowledge of such objects that Plato seems willing to call "knowledge." It is, of course, possible that what the Platonic texts appear, o n the surface, to be saying is not what they are saying in fact. But if this is so-if Plato is really entertaining, let alone endorsing, some definition of 'S knows that P'-then there is a lot of work that needs to be done to show how a "deep" reading of Plato is supposed to bear this out.5 And until that is done, the claims made by 'And, to win over the many scholars in whose opinion n o such reading can be sustained. See, for example, John McDowell's translation of, a n d commentary o n , the Theaetetus (New York: Oxford, 1973), pp. 116, 230, 252; Nicholas White, Plato on Knowledge and Reality (Indianapolis: Hackett, 1976), p p . 176/7; Alexander Nehamas, "Episteme a n d Logos in Plato's Later T h o u g h t , " Archiu fur Geschichte der Philosophie, I.xl.1, 1 (1984). T h e last-mentioned work also contains numerous references to the work of other like-minded scholars. T h e translations from the
I T ' S N O T W H A T YOU K N O W T H A T C O U N T S
353
Gettier and others to the effect that their philosophical enterprise has a Platonic ancestry can only be taken with a grain of salt. T h e same can be said for the claim that Gettier has posed a problem for "the traditional definition of knowledge." Descartes is certainl y a prominent member of the epistemological tradition and just as certainly a philosopher who concerned himself with the question of what counts as propositional knowledge. But Gettier's counterexamples can pose no difficulty for Descartes's conception of propositional knowledge. T h e Gettier cases turn on the possibility of a person's having a justified, yet false, belief-a possibility Descartes would never have acknowledged. In holding that the expression 'knowledge' applies only to true and evident cognition, Descartes denied just what Gettier needs to be granted if his counterexamples are to have force: namely that fallible means of justification can yield knowledge.6 But if the appeal to the philosophical tradition of centuries past does not carry conviction, perhaps some appeal to twentiethcentury concerns will. Philosophical problems need not be old to be urgent. There are plenty of issues that have been of lively epistemological concern yet were not-and perhaps could not have been-of interest to a Plato (or even a Descartes). Consider the project, identified with the logical positivists, of trying to define cognitive significance. It was a project that was motivated by an antipathy toward metaphysics which would have astonished a Plato or a Descartes. And though it was a pursuit that might well have delighted ~ u m e , it ' was not upon Hume's wouldbe enthusiasm that the advocates of the project rested their case. T h e urgency of the project derived from its relation to an ethic of rational inquiry which the positivists were seeking to advance-an ethic according to which cognitively insignificant claims would be ruled out as potential explanations, claims bearing the same cognitive significance would be treated as equivalent, and theoretical --
~--
Theaetetus reproduced in this paper are due to McDowell ( o p . cit.), a n d those from the Meno are due to W. C. K. Guthrie [in E. Hamilton and H. Cairns eds, Plato: T h e Collected Dialogues (Princeton, N.J.: University Press, 1961)]. In this respect, of course, the cartesian skeptic is in entire agreement with Descartes-differing only in her lack of sanguineness about the prospects of our having knowledge. T h u s such a skeptic every bit as much as Descartes will want to say that Gettier has posed n o problem for the definition of knowledge as she understands the term. None of this is to say, however, that o u r conception of knowledge bears n o interesting relation to Descartes's. See below, 361/2 and fn 12. 'See the closing paragraph of An Enquiry concerning H u m a n Understanding. I t is cited by A. J. Ayer as "an excellent statement of the positivist's position" in his introduction to Ayer, ed., Logical Positivism (New York: Free Press, 1959), p. 10.
354
T H E JOLIRNAL OF PHILOSOPHY
terms would be introduced a n d taken seriously i n inquiry only when their cognitive significance was adequately stated. For the philosophers who took u p the task of defining cognitive significance, the classification of claims as cognitively significant (or cognitively insignificant) was central to the proper conduct of inquiry. Consider the project of providing a definition of lawlikeness-of that property of generalizations which enables them to receive evidential support from the observation of their instances. Again we have a concern that could not have found expression i n the corpus of a Plato or a Descartes (but that would quite probably have found favor with Hume). And again, more important, we have a project whose impetus came from a concern to define the proper conduct of inquiry. T h e participants in the project were concerned that accidental generalizations not be confused with lawlike ones-that we should not become confident of the truth of accidental generalizations on the strength of our having established that some of their instances hold; that we should be able to distinguish genuine laws from spurious ones. Appraising the lawlikeness of generalizations was important to these authors' conception of how inquiry ought be conducted. I d o not mean to imply that I endorse the two projects just mentioned or that I think they have met, or will meet, with success. Insofar as I mean to praise them, it is as projects that were motivated by a concern about the proper conduct of inquiry. Whether or not they succeeded in advancing or clarifying the state of the art of inquiry, this is precisely what these projects were meant to do. It seems to me that it was this very motivation that made them important projects. So it might be with the project to define 'S knows that P'. It might be that, like the projects just described, the attempt to solve the Gettier problem is legitimately aimed at advancing or clarifyi n g the proper conduct of inquiry. If so, it seems to me that the solution of that problem has a legitimate claim upon our attention. But is the attempt to solve the Gettier problem aimed at advancing or clarifying the proper conduct of inquiry? One can hardly fail to be struck by how little the authors of the work in question have had to say about the matter. It may, of course, be that it is so obvious how solving the Gettier problem will enhance our understanding or prosecution of inquiry that it simply goes without saying. But I am afraid that the real explanation-though it too has gone largely without saying-is quite different. T o see that this is so, let us begin by considering how knowledge
IT'S N O T W H A T YOU K N O W T H A T C O U N T S
355
q u a justified true belief might figure in the enterprise of inquiry. Imagine that you have been engaging in inquiry. Being a responsible inquirer, you have carefully weighed evidence and argument and have come to the conclusion that the weight of evidence clearly favors P and, so, you have concluded that P is true. Suppose you now ask yourself, "But do I know that P?" Notice that, on the justified-true-belief analysis of knowledge, there is nothing to find out, nothing to do. Having already satisfied yourself that P is true and that the evidence supports your contention that P is true, you have ipso facto already satisfied yourself that you have justified true belief. From where you sit, determining whether you believe P with justification and determining whether you know that P come to the same thing. But then, far from being integral to your pursuit of inquiry, distinguishing the propositions you know from those you don't know is, on the justified-true-belief analysis, a fifth wheel. 'Knowledge' turns out to be nothing more than an honorific you may bestow on those of your beliefs which you consider justified should using the term 'justified' alone seem tiresome. T h i s result is not, of course, a problem for those who would solve the Gettier problem. They are, after all, no defenders of the justified-true-belief analysis of 'S knows that P ' . Their problem is worse. Let us return to the case above, and suppose that we have on hand some analysis of 'S knows that P' which solves the Gettier problem by, in effect, disqualifying justified true belief inferred from an essential false premise. Suppose again that you have carried out inquiry, come to believe that P on good evidence, and that you now ask, with this new analysis of knowledge in mind, "But do I know that P?" Once again, there is nothing to find out. Insofar as you are satisfied that your belief in P is well founded, you will ipso facto be satisfied that you have not inferred P from a false premise-otherwise you would not think you had good reason for concluding that P. Just as on the justified-true-belief analysis, determining whether you believe P with good reason and determining whether you know that P come to the same thing. But then the project of trying to solve the Gettier problem stands guilty of two counts of insufficient motivation. First, on any analysis of ' S knows that P' that solves the Gettier problem, the enterprise of distinguishing the propositions you do know from the propositions you don't know will turn out to be every bit as unnecessary to the proper conduct of inquiry as it was on the justifiedtrue-belief analysis. Second, determining what you know will come to the same thing as determining what you are justified in believ-
356
THE JOLTRIVALOF PHILOSOPHY
ing, whether you adopt the justified-true-belief analysis of knowledge or solve the Gettier problem by finding and adopting some analysis that disqualifies justified true belief inferred from a n essential false premise. T h u s for the purpose of distinguishing propositions you know from those you don't know, it makes no difference which of the two analyses of ' S knows that P' you adopt. But then, to the degree to which the project of solving the Gettier problem is supposed to be motivated by some problem that the justified-true-belief analysis of knowledge poses for the understanding or execution of the proper conduct of inquiry, that project appears to rest upon a bogus motivation. There seems to be n o problem.8 None of this is to say that the three-way distinction, between your believing P with justification, your having justified true belief, and your having justified true belief that is (at the very least) not inferred from an essential false premise, is a distinction without a difference. From where you sit, you cannot distinguish among these three states of affairs. But from where I sit, things look different. I can be in a position to tell that your belief in P is justified yet false-even though you cannot. I can be in a position to tell that you have a justified true belief inferred from a false premise-even though you cannot. But my ability to make these discriminations can be significant here only if it is important to the proper conduct of inquiry that I exercise that ability. But is it important? Suppose that right now I can tell that your belief that P, though perfectly justified and arrived at by the most scrupulous application of our best methods of inquiry, is unluckily based upon a crucial false belief. Some of your evidence, though admirably gathered and evaluated, was nonetheless misleading-a sample turned out, ' ~ o t i c ethat this argument affects equally those analyses of 'S knows that P' which seek correctly to classify the Gettier cases without talking about justificationanalyses that appeal to the existence of causal relations between events i n the world and an agent's beliefs or to the truth of certain counterfactuals relating the truth of propositions and the credence an agent places i n these propositions. For, o n such analyses, it is presumably still true that any evidence you have that renders you unwilling to believe that you know that P will tpso jacto render you unwilling to think that your (continued) belief that P is justified. Notice as well that the argument is not, in essence, novel. Similar arguments to the effect that your justified true beliefs are indistinguishable for you from your merely justified beliefs (but not for me-see below) have been advanced by M1illiam Rozeboom in "Why I Know So Much More than You Do," Amerlcan Philosophical Quarterly, I V , 4 (October 1967): 281-290, and, more recently, by Isaac Levi in T h e Enterprise of Knowledge (Cambridge, Mass.: MI-rPress, 1980), sec. 1.9. Levi's assessment of the significance of Gettier's counterexamples for first-person (presenttense) knowledge appraisals is much the same as my own.
IT'S N O T WHAT YOU KNOW T H A T C O U N T S
357
against all odds, to be extremely unrepresentative of the population from which it was randomly selected. O n the basis of your evidence, you did what any reasonable investigator would have done, and should have done, in your place: you inferred Q and then, from Q, you inferred that P. P is true, but, unfortunately, Q is false. Here is a case in which, from where I sit, the difference between the justified-true-belief analysis of knowledge and, say, a justifiedtrue-belief-based-on-no-essential-false-premise analysis of knowledge is apparent: I can tell that, according to the former, you know that P and that, according to the latter, you d o not. T h u s if I were to adhere to the justified-true-belief analysis of knowledge, I would conclude that you know that P. If I were to adhere to the other analysis, I would conclude that you do not know that P. But what difference does it make which I adhere to or what I conclude? It is not as if your performance as an investigator were at issue. I have already granted that you have done as well as anyone could have done in your place-that your performance was exemplary. Any worry I might conceivably have about the propriety of your conduct as an inquirer has already been laid to rest. Thus which analysis I adopt-whether I decide that you know that P or not-can have nothing to d o with any concern I have with the propriety of your performance as an inquirer. But perhaps all this shows is that what is at issue in this case is not the propriety of your conduct as a n inquirer, but rather the propriety of my undertaking to criticize your route to your belief that P. Even though your inferring that P is above criticism in the case under discussion, the inference itself is not. Your argument for P rests upon the false assumption that Q. And though it would be inappropriate for me to fault you for having been convinced by that argument, it might well be appropriate for me to point out to you the flaw in your argument (and apprise you of my reason for believing P is true) should I have the opportunity. What might then be at stake in my deciding what analysis of 'S knows that P' to adopt-and in deciding whether you know that P or not-is whether it would be appropriate for me to criticize your argument for P. This, however, cannot be quite right. T h e appropriateness of my criticizing your reasoning (your route to P) and apprising you of my own will depend on a great variety of circumstances: how well I a m acquainted with you, how I think you will take my criticism, whether I think you are likely to spot your error on your own, how I assess the educational value for you (supposing you were a stu-
358
THE JOURNAL OF PHILOSOPHY
dent of mine) of your finding and correcting your errors without my intervention, a n d so on. Deciding what analysis of '5' knows that P' to adopt and (thus) whether to say that you know that P will not of itself determine whether the circumstances render it a p propriate for me to point o u t the flaws in your argument. But perhaps this is because what is at issue in the case under discussion is not the propriety of my criticizing you (which depends upon extraneous circumstances) but simply the vulnerability of your argument to criticism. Perhaps, i n deciding whether your justified belief i n P constitutes knowledge that P , I a m deciding whether your argument for P-however well crafted-ultimately holds u p under critical scrutiny (and hence whether my intervention i n your inquiry would be even prima facie warranted). T h i s is a n attractive suggestion. It is easy to see that the task of determining whether arguments stand u p under critical scrutiny is central to the proper conduct of inquiry. T h e trouble is that it is hard to see how either the execution or the understanding of that task could possibly hang on our solving the Gettier problem. If I want to decide whether your argument stands u p under critical scrutiny, it is clear what I should do: I should consult what evidence I can muster, reflect upon my intuitions about justification, and decide whether, from where I sit, the argument provides a good reason for believing its conclusion-whether the argument is one I can endorse. It is a familiar routine. After all, determining whether arguments stand u p under critical scrutiny stands at the very heart of the process of arriving at justified belief. T h e enterprise of trying to arrive at justified belief is nothing more than the enterprise of trying to arrive at a belief supported by reasons that will stand u p under critical scrutiny. In asking, of the argument you have produced in the case under discussion, "Does that argument stand u p under critical scrutiny?" I am asking a kind of question I have occasion to ask virtually every time I engage in inquiry. Of course, for all its familiarity, my routine for deciding whether I should endorse a given argument contains much that is obscure. Were it otherwise, one would expect we would have a far more comprehensive inductive logic than we have in fact. We thus have good reason to be grateful for any contribution a philosopher can make toward clarifying or improving our routines for appraising arguments. But it is hard to see how attention to Gettier's counterexamples-how a solution to the Gettier problem-could possibly make such a contribution. Consider the case under discussion-a case structurally identical with the cases presented in Gettier's original article. From where I sit, the question whether your argument for P stands u p under
I T ' S N O T W H A T YOLT KNOW T H A T C O U N T S
359
critical scrutiny could hardly be more straightforward. Incorporating, as it does, a crucial false premise, your argument is clearly one I cannot endorse. What then is the problem to which Gettier's counterexamples are supposed to be alerting me? Surely none of us needs to be taught that a n argument with a true conclusion does not carry conviction if that conclusion rests upon a false premise. But, viewed as a contribution to the enterprise of understanding or improving our routines for determining what arguments stand u p under critical scrutiny, that is all Gettier's counterexamples can possibly mean to be teaching us. T h e conclusion can only be that Gettier's counterexamples make n o contribution whatsoever to that enterprise: they pose n o problem that anyone interested in the evaluation of arguments needs to solve. None of this is to deny that it is possible for me to take another approach to the problem of determining when, in general, your arguments stand u p under critical scrutiny. It is certainly true that I might choose to view this problem as just the problem of determining what other conditions a justified true belief must satisfy i n order to qualify as knowledge. But what would be the point? My canon of justification is already invested in the role of guiding me in my evaluation of arguments. And, as we have seen, the Gettierproblem in n o way impeaches the performance of my canon of inquiry in that role. But if Gettier's counterexamples present n o difficulty for my evaluation of your performance as a n inquirer, nor for my fulfillment of my duties as a critic, nor for my assessment of the vulnerability of your argument to criticism, what difficulty do they present? T h e answer would seem to be that they present n o difficulty whatsoever-or, at least, n o difficulty of any import to the understanding or improvement of rational inquiry. Not only does the attempt to solve the Gettier problem fail to be a response to any problem that would have bothered Plato or Descartes; it now looks to be a project that cannot possibly address any problem w e have in understanding or prosecuting the enterprise of inquiry. For, as we have seen, Gettier's counterexamples pose n o problem for any of our attempts to carry out, or to assess others' attempts to carry out, good investigative procedures. Of course, even if the project of solving the Gettier problem cannot address any problem we have in understanding or prosecuting the enterprise of inquiry, it is still possible that a solution to the Gettier problem (i.e., a general account of when it is appropriate to attribute propositional knowledge to persons) might entail some substantive thesis concerning the way i n which reasoning does (or should) proceed. T h u s , it might be thought, it is still possible that
360
T H E JOURNAL OF PHILOSOPHY
we might learn something about how reasoning does (or should) proceed by trying to solve the Gettier problem. And so, it might be argued, it is entirely appropriate that philosophers concerned with the proper conduct of inquiry should include among their projects the project of trying to solve that problem.g But this argument proceeds too quickly. T h e mere fact that a solution to the Gettier problem might entail something substantive about the proper conduct of inquiry does not mean that we stand to learn anything about reasoning by solving the Gettier problem. We stand to learn something about reasoning by solving the Gettier problem only if we have antecedently decided that we need a general account of when it is appropriate to say of a person that she knows that P. Otherwise, the fact that the solution to the Gettier problem entails a thesis about the proper conduct of inquiry will not, and should not, lend that thesis any credibility at all. But in that case the possibility that the solution to the Gettier problem might entail such a thesis cannot itself provide a motivation for seeking to solve the Gettier problem. What is needed, in addition, is an independent ground for deciding that our theory of rational inquiry should incorporate a solution to the Gettier problem-a ground for thinking that Gettier counterexamples are of sufficient importance, indeed that knowledge attribution is of sufficient importance, to require an adequate theory of rational inquiry to offer a n account of knowledge attribution that successfully deals with Gettier's counterexamples. But it is just such a ground for which we have searched in vain. It is important to note that it is not the concept of knowledge, but rather our conception of knowledge, that is to blame. Were we but seventeenth-century philosophers, it might not be so difficult to convince ourselves that knowledge attribution plays an important role in the proper conduct of inquiry. After all, Descartes's preoccupation with specifying the conditions under which one has propositional knowledge suffered no lack of motivation. For Descartes, determining what propositions you know-as opposed to merely believe with justification in our sense of 'justification'-was central to the proper conduct of inquiry.'' Of course, what enabled 9 ~ h i view, s that the point of solving the Gettier problem is to learn something about reasoning, has been suggested to me a number of times in conversation and has been expressed and pursued in print by Gilbert Harman. See, for example, his Thought (Princeton, N.J.: University Press, 1973) and "Using Intuitions about Knowledge to Study Reasoning: A Reply to M'illiams," this J O I ' R S A I . , I.xx\', 8 (AUgust 1978): 433-438. ''See the Discourse on Method and the Rules for the Direction of the Mind, especially Rule 11.
IT'S NOT WHAT YOU KNOW THAT COUNTS
36 1
him to accord knowledge so central a role in inquiry was his view that one can distinguish what one knows from what one merely believes with justification in our sense. Knowledge, for Descartes, is a state of cognition different-and, from the knower's point of view, discernibly different-from what we call "justified belief." According to Descartes, a known proposition leaves unmistakable psychological evidence of its being known. Thus Descartes's call to (in effect) attend to what you know as opposed to what you merely believe with justification, was a contentful exhortation. We, o n the other hand, do not think we are endowed with the special cognitive equipment Descartes thought we had. Since we are saddled with a psychology that (rightly) does not admit special states of knowing that are, from the agent's point of view, discernibly different from states of justified belief, a contemporary call to attend to what you know-as opposed to what you merely believe with justification-would simply be confused. Given our conception of knowledge, all we can do by way of seeking knowledge is seek justified belief and hope that this justified belief will satisfy whatever other conditions a justified belief must satisfy in order to qualify as knowledge. This being so, it is not hard to see why the enterprise of specifying what those conditions are looks so purposeless. For if all we can do by way of seeking knowledge is seek justified belief, then, to secure agreement o n how rational inquiry is to be conducted, we need only secure agreement on the canons of justification-it does not matter whether we agree or not on what knowledge is. It is thus a feature peculiar to our conception of knowledgethat knowledge is indistinguishable from the agent's point of view from merely justified belief-which dooms the analysis of knowledge to irrevelance in helping us to understand and advance the proper conduct of inquiry. Yet the selfsame feature of our conception of knowledge is partly responsible for our feeling the force of Gettier's counterexamplesfor our feeling that Gettier has posed a problem that any adequate analysis of knowledge would have to solve. We noted earlier that Gettier's counterexamples pose no problem for Descartes's conception of knowledge because Descartes denied that any belief that was justified via fallible means of justification could count as knowledge. It was not, of course, that Descartes was a skeptic. Rather he was convinced that there are available to us demonstrably infallible methods of justification which could yield a substantive science. And he held that only beliefs justified via such methods-beliefs thus discernibly different from the agent's point of view from beliefs that are justified via fallible methods-count as knowledge.
362
THE J O U R N A L OF PHILOSOPHY
Convinced as we are that there are no such methods available to us, we must admit, on pain of denying that we know much of anything, that what distinguishes fallibly justified beliefs that constitute knowledge from fallibly justified beliefs that do not constitute knowledge is some feature of those beliefs which is undetectable by the agent. It is only because we (unlike Descartes) make this admission, that there can be for us (as there cannot be for Descartes) a question as to what feature of an agent's fallibly justified beliefs might, in a manner undetectable by that agent, distinguish those which constitute knowledge from those which do not. l t In other words, it is only because we make this admission that there can be for us the question to which the justified-true-belief analysis attempts (but unsuccessfully, as Gettier shows) to provide an answer.I2 T h e upshot is that a single feature peculiar to our conception of knowledge-the fact that an agent cannot distinguish what she merely believes with justification from what she knows-is responsible both for our viewing Gettier as having indeed found a problem for the analysis of knowledge and for the pointlessness of coming u p with such an analysis. Hence the claim, advertised earlier, that the moral to be drawn from Gettier's counterexamples is that what you know doesn't matter: insofar as we grant that Gettier has posed a problem that an adequate analysis of knowledge would need to solve, we must also grant that, for the purpose of understanding and advancing rational inquiry, we have n o reason to take the enterprise of providing such an analysis-or, indeed, the practice of knowledge attribution itself-the least bit seriously. A final comment. It may seem strange that I have been using the decisiveness of Gettier's counterexamples to the justified-true-belief analysis of propositional knowledge to argue that, on our conception of knowledge, what you know does not much count. It might seem tempting to reply that the very decisiveness of Gettier's counterexamples-the near unanimity with which philosophers have "Let alone a question of what unjustified beliefs might, by virtue of their truth and the manner in which they are acquired, also count as knowledge. I2Indeed the response to Gettier's counterexamples can be seen as a reflection of both o u r conceptual distance and o u r conceptual proximity to o u r seventeenthcentury philosophical ancestors. T h e distance is reflected in the fact that, presupposing as they d o that we can achieve nothing better than fallibly justified belief and that knowledge is psychologically indistinguishable from mere fallibly justified belief, Gettier's counterexamples pose n o problem for the seventeenth-century conception of knowledge. T h e proximity is reflected by the fact (evidenced by the near unanimity with which Gettier's counterexamples have been judged to be genuine) that, n o more than Descartes do we want to apply the honorific 'knowledge' to a belief inferred from a belief that is not itself known.
IT'S NOT WHAT YOCTKNOW THAT COUNTS
363
taken them to show that the justified-true-belief analysis is defective-bears evidence that our conception of knowledge is alive and well. But that would be to miss the point. Whether there are clear cases of propositional knowledge and, indeed, whether the prospects of arriving at a widely accepted solution to the Gettier problem are bright, is not the issue. Even if it were easy to arrive at a solution-even if there were widespread agreement as to what intuitions needed to be satisfied and what intuitions could safely be sacrificed for the purpose of arriving at this solution-there would still remain the question of what the solution was good for. T h e absence of any good answer to this question, I have been arguing, undermines the project of trying to solve the Gettier problem. Of course, if the project of trying to solve the Gettier problem were straightforward, the fact that it lacked motivation would not be so troubling, the wasted effort not so great. But, as we are all aware, it is not straightforward. We have already witnessed twenty years of extremely subtle and ingenious work which, far from producing promise of a widely accepted definition of 'S knows that P', has given rise instead to competing programs that d o not even try to accommodate the intuitions of their rivals. What is done is done. But the future need not remake the mistakes of the past. My message is that it is time to stop and face the unpleasant reality that we simply have n o use for a definition of propositional knowledge. T h e message is not all bad. After all, insight into the phenomenon of inquiry was there to be had for those who looked before 1963, and there is every reason to suspect that it awaits us now. MARK KAPLAN
T h e University of Wisconsin/Milwaukee