International Journal for the Study of Skepticism 1 (2011) 317

brill.nl/skep

Disagreement, Skepticism, and the Dialectical Conception of Justication

Markus Lammenranta
University of Helsinki markus.lammenranta@helsinki.

Abstract It is a common intuition that at least in some cases disagreement has skeptical consequences: the participants are not justied in persisting in their beliefs. I will argue that the currently popular non-dialectical and individualistic accounts of justication, such as evidentialism and reliabilism, cannot explain this intuition and defend the dialectical conception of justication that can explain it. I will also argue that this sort of justication is a necessary condition of knowledge by relying on Craigs genealogy of the concept of knowledge. I will then respond to the accusation that the dialectical conception leads to radical skepticism. My response is partly concessive. It does lead to skepticism in areas where controversy prevails, such as philosophy, politics and religion, but this sort of skepticism is quite intuitive. Finally, I deal with the objection that my defense of skepticism about philosophy is self-refuting. Keywords disagreement; dialectic; genealogy; justication; skepticism

Ancient skeptics argued that people should suspend belief if they disagree and cannot rationally resolve their disagreements. In doing so, they seemed to understand justication dialectically and socially (Lammenranta 2008, forthcoming). The received view of justication in contemporary analytical epistemology, in contrast, is non-dialectical and individualistic. That is why it is no surprise that epistemologists have not paid much attention to disagreement. The reason why it became a hot topic quite recently seems to be that we do have the intuition that disagreement has at least in some cases skeptical consequences while non-dialectical and individualistic accounts of justication have diculties in explaining this or so I claim. In order to defend my claim, I will discuss a typical case in which we have the intuition that disagreement prevents justication and knowledge and argue that the received individualistic views cannot explain the intuition. This speaks for the dialectical and social view. I will also defend my diagnosis of this
Koninklijke Brill NV, Leiden, 2011 DOI 10.1163/221057011X554124

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

case and the dialectical conception of justication by relying on Edward Craigs genealogy of the concept of knowledge. I will argue that Craigs hypothesis about the point of the concept supports the view that dialectical justication is a necessary condition of knowledge. I will then respond to the objection that the dialectical conception leads to radical skepticism. I will try to show that it leads at most to urbane skepticism,1 a form of skepticism that is restricted to controversial issues in philosophy, science and religion. This is not a problem because this sort of skepticism is quite intuitive and plausible. Finally, I deal with the objection that my defense of skepticism about philosophy is self-refuting.

1. Epistemic Peerage Initially, it may seem that the standard non-dialectical accounts give no epistemic signicance to disagreement. This is because they are individualistic, and disagreement is a social phenomenon. They make the conditions of justication concern the individual subject, restricting the justifying factors to the subjects mental states or the causal sources of those states. For example, evidentialism takes justication to be a function of the subjects experiences and beliefs, and reliabilism takes it to be a function of the causal origin of those beliefs.2 So it may seem that what other people believe whether they disagree or not is irrelevant to justication. The matter is not so simple. It is true that disagreement as such has no epistemic signicance according to these individualistic accounts. What is relevant are the subjects beliefs about disagreement. Both evidentialism and reliabilism allow that beliefs about disagreement may aect the justication of other beliefs: they can defeat the justication of these other beliefs. Let us take an example of a typical case, Adam Elgas horse race, about which we do have the intuition that disagreement prevents the participants from having justication for their beliefs:
We are to judge the same contest, a race between horse A and horse B. Initially, I think that you are as good as me at judging such races, and you think in the same

1 On the distinction between rustic Pyrrhonism and urbane Pyrrhonism, see Barnes (1997, 612). 2 It might be pointed out that the etiology of belief can extend to other people and their beliefs. However, reliabilists typically identify the sources of belief individualistically with the psychological processes (Goldman), intellectual faculties (Sosa) or introspectively accessible grounds (Alston) of the subject whose belief is epistemically evaluated.

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

way about me, but then we realize that we disagree: I believe that horse A won the race, and you believe that horse B won. (Elga 2007, 486)

The intuition is that neither of us is justied in persisting in our beliefs: we should both give them up and look for further evidence. The individualistic accounts try to explain the intuition by appealing to defeaters. According to this story, I rst have a justied belief that horse A won, but when I learn that you disagree, I obtain a defeater for this belief: my belief that A won is no longer justied. How is this supposed to work? First of all, it is clear that my belief that you disagree does not alone have defeating power. I must also believe that you are as good as me at judging such things. I must believe that you are my epistemic peer. Two necessary conditions for epistemic peerage are typically given3:
Evidential equality. Two persons are evidentially equal relative to the question whether p if and only if they are equally familiar with the evidence relevant to the question whether p. Cognitive equality. Two persons are cognitively equal relative to the question whether p if and only if they are equally competent or reliable in assessing the evidence relevant to the question whether p.

So what is supposed to defeat my justication for believing that p is my belief that you are my epistemic peer and you believe that not-p. If I believe this, I am no longer justied in believing that p. This is how individualism attempts to explain the intuition that disagreement prevents justication in some cases.

2. Undercutting Defeaters What are defeaters? A defeater is a belief or some other mental state that makes some other belief lose its justication. Assume that I am justied in believing that p on the basis of evidence e. Then I form a new belief d. My belief d defeats my justication for believing that p if and only if e and d do not justify me in believing that p. John Pollock (1986, 389) distinguishes between two kinds of defeaters: (1) Rebutting defeaters for my belief are my reasons to believe that my belief is false. (2) Undercutting defeaters are my reasons to believe that my evidence

3 See, for example, Kelly (2005, 1745), Christensen (2007, 1889) and Lackey (2010 and forthcoming).

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

does not support or indicate the truth of my belief. I assume, following Pollock (1989, 150), that a rebutting defeater d cannot have less justication than the defeated belief p. In other words, the evidence for d cannot be weaker than the evidence for p. Otherwise, d would itself be defeated by p. If the evidence for p and the evidence for d are equally strong, p and d defeat each other, and neither is justied. Pollock calls this collective defeat. Which kind of defeater is relevant in the case of peer disagreement? Let us take rst the option that believed disagreement provides an undercutting defeater. Assume that e is good evidence for p, and that I believe that p on the basis of e. When I now learn that you, whom I take to be my epistemic peer, believe that not-p on the basis of e, I get evidence that e is not good evidence for p. So I seem to have an undercutting defeater for my belief that p. However, this is not clear, because, as Thomas Kelly (2005, 190) points out, I also have equally strong evidence that e is good evidence for p. Our being epistemic peers and equally competent in evaluating the common evidence, your believing that not-p on the basis of e is evidence that e is not good evidence for p, and my believing that p on the basis of e is evidence that e is good evidence for p. My new total evidence includes thus the following: (1) (2) (3) (4) (5) evidence e, I believe that p on the basis of e, you believe that not-p on the basis of e, we are both equally reliable in evaluating evidence e, and e cannot be good evidence for both p and not-p (the uniqueness thesis).

The conjunction of 3, 4 and 5 constitutes my evidence for believing that

(UD) e is not good evidence for p.

It is thus a potential undercutting defeater for my belief that p. The problem is that I have also a rebutting defeater for this undercutting defeater because 2 and 4 give me evidence for believing that
(RDUD) e is good evidence for p.

Because UD and RDUD are equally justied for me and because I know that they cannot both be true (5), there occurs what Pollock calls collective defeat. UD and RDUD defeat each other, and I am not justied in believing either. So I no longer have a defeater for my justication for believing that p. Kelly concludes that peer disagreement does not have epistemic signicance. It does not make our beliefs unjustied, which is counterintuitive in the horse race case and other similar cases.

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

One may object to Kellys conclusion and insist that I still have a defeater of some kind for my belief, because my conscious suspension of judgment about the question whether e is good evidence for p is also an undercutting defeater for my belief that p. So undercutting defeaters need not be beliefs. Other attitudes, like suspension of judgment, can also work as a defeater. Richard Feldman (2006, 2323) and Michael Bergmann (2005, 426) defend this view: If I consider whether my evidence supports p and I have to suspend judgment about the matter, I am not justied in believing that p. This may be intuitive.4 However, two problems arise when understanding defeaters in this way. First, it is not clear that UD and RDUD are equally justied for me. When all my evidence is taken into account, it seems more probable to me that it is you who have made a mistake in evaluating the evidence. If this is true, there is no collective defeat: only UD is defeated, in which case my belief that p remains undefeated and justied. I will discuss this sort of evidence in more detail in connection with rebutting defeaters and concentrate now on the second problem. The whole idea that disagreement provides undercutting defeaters for our beliefs presupposes that you and I literally share the evidence, that we possess exactly the same evidence. Assuming that evidence covers private perceptual experiences and memory experiences, it is clear that we cannot literally share our evidence. I do not have your experiences, neither have you mine. So if evidential equality means that we share or possess the same evidence, there are no evidential equals and no epistemic peers. And individualism fails to explain our intuitions concerning disagreement. Familiarity with the evidence could be understood more loosely. It is enough that we tell each other about our evidence. Then we both attain testimonial evidence about each others evidence. Feldman (2006, 233) says that evidence about evidence is evidence. He seems to mean that my evidence about your evidence for not-p is also evidence for not-p. This may be so, but it is important to keep in mind that this is testimonial evidence. I cannot attain perceptual evidence in this way. For example, in the case of the horse race, we do not share our evidence. My evidence for my belief that horse A nishes ahead of horse B consists of my perceptual experience: it appears to me that A nishes ahead of B. Your

4 Also Pollock and Cruz (1999, 2001) discuss a case in which a collectively defeated conclusion retains its defeating power.

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

evidence consists of your perceptual experience: it appears to you that B nishes ahead of A. After disclosing our evidence to each other, we get evidence about each others evidence. I learn that your evidence supports your belief, and you learn that my evidence supports mine. This gives neither of us an undercutting defeater: Because we dont have the same evidence, it may very well be that my evidence supports my belief while your evidence supports yours. One of us simply has misleading evidence.

3. Rebutting Defeaters How do we then explain the intuition that I am not justied in persisting in my belief in the horse race case and other similar cases? The only way seems to be to appeal to rebutting defeaters. The idea is that when I learn that you believe that horse B won on the basis of your evidence, I get evidence that horse B won, even though I do not share your evidence. So I do not have evidence only for the proposition that horse A won, but also for the proposition that horse B won. The evidence for the latter proposition is a rebutting defeater for my belief that horse A won.5 If this suggestion is to work, my evidence for horse B must be at least equally strong as my evidence for horse A. We may assume that your perceptual evidence for your belief is equally strong as my perceptual evidence for mine, but this does not mean that I possess equally strong evidence for both propositions, because I do not share your perceptual evidence. I have at most testimonial evidence about it. It seems clear that this testimonial evidence for the proposition that horse B is the winner cannot be as strong as my direct perceptual evidence for the proposition that A won. So the former cannot be a rebutting defeater for the latter. It is rather the other way around. If one thinks that my perceptual evidence is not strong enough to do the work alone, I have also other evidence that supports my belief over yours. I have simply many more reasons to doubt the proposition that B won than to doubt the proposition that A won. First, you may be lying or joking or teasing me when you claim that B won. So it may not really appear to you that horse B won. On the other hand, I know very well that I am not the one who is lying or joking if one of us is. So I have no similar reasons to doubt my own belief that A won. Second, assuming that you are sincere, there are still many

This is the way Kelly (2010, 1502) tries to explain the skeptical intuition.

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

reasons to suspect that it is you who made a mistake. For example, you may have gotten something in your eye, and your evidence was therefore unreliable. Perhaps you were drunk, or perhaps your eyesight is faulty, and so on. The point is that there is a large number of possible mistakes that I cannot rule out in your case but that I can rule out in mine. Surely I know that there was nothing in my eye and that I was not drunk. So it is epistemically more probable that you are the mistaken party instead of me. Jennifer Lackey (2010, 2778; forthcoming) calls the evidence that I have about my own experiences, beliefs, intentions and reliability, but that I lack about yours, personal evidence. She argues that if my belief enjoys a very high degree of justied condence and I have personal evidence supporting it, I am justied in persisting in this belief in the face of peer disagreement. This gives a wrong result in the horse-race case assuming that I have a high degree of justication for my belief. However, one could insist that my belief in such a case can only enjoy a low degree of justied condence. Perhaps the race was close, and it was not easy to say which horse won. Now Lackey says that I am no longer justied in my belief, or at least I should substantially lower my condence in it. I still have similar personal evidence supporting my belief, but she does not think that in the case of low justied condence my personal evidence breaks the symmetry. She gives no grounds for the omission of personal evidence in such cases, and it is, indeed, dicult to nd such grounds from reliabilism or evidentialism. The omission is thus completely ad hoc if individualism is true. So both my perceptual evidence and my personal evidence support my belief over yours. Therefore my belief that you disagree does not give me a defeater. Of course, one could point out, as Feldman (2005, 116) does, that our situation may still be symmetric. You can rely on similar considerations that support your belief. So, from an impartial point of view, evidence on both sides is equally strong. However, it is hard to see how this is relevant if we assume an individualist account of justication. If justication depends on my individual point of view on my beliefs and experiences it does not matter how things appear to an impartial observer. From my point of view, there is no symmetry: my evidence supports my belief rather than yours. And according to individualism, it is this point of view that decides the justication of my beliefs. The problem with individualism regarding cases like the horse race is that my perceptual evidence and my personal evidence give stronger support to my belief that p than my testimonial evidence about your evidence gives to not-p. This is why the latter evidence cannot defeat my justication for believing that p. If it were required that my evidence for p be independent of the

10 M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

disagreement, as David Christensen (2007, 198) and Adam Elga (2007, 492) suggest, I could not count my perceptual experiences as a part of my evidence, but I still would have my personal evidence that supports my belief. So such an independence requirement may not be enough to secure a balance between my positive and negative evidence for p. Furthermore, the requirement is completely unmotivated and ad hoc if justication is understood individualistically: Why should I disregard a part of my evidence as a response to disagreement if individualism is true? The only motivation that Christensen (2007, 198) gives for the independence requirement is that my appealing to my original evidence would beg the question against you. The term question-begging is sometimes used for arguments that are formally circular. However, nothing is formally circular in this case. So Christensen can only mean that my appealing to my perceptual evidence begs the question in a dialectical sense. I beg the question in this sense when I defend my belief by reasons that you would not nd acceptable. This motivation for the independence requirement is, however, not available for Christensen assuming he is an individualist. According to individualism, justication does not require such non-question-begging evidence: my justication does not depend on what you nd acceptable.

4. The Dialectical Conception of Justication I dont deny that the independence requirement and symmetry considerations are intuitive. It is only that individualism cannot explain them. They can be explained only if we accept the dialectical conception of justication that does require that my evidence be also acceptable to you. Assuming that this conception is true, neither my perceptual evidence nor my personal evidence is able to justify my belief, because you who disagree with me have reasons to doubt that evidence. I should have evidence that is independent of the dispute, evidence that you could accept. If our situation is symmetric in the way Feldman supposes, we have both internal evidence for our beliefs, but because my evidence is also a reason to doubt your evidence and your evidence is a reason to doubt mine, neither of us has good evidence according to the dialectical conception, and we should both give up our beliefs. So only the dialectical conception of justication respects the independence and symmetry considerations. Thus it seems that we need the dialectical conception of justication to explain our intuitions about certain cases of disagreement. According to it, justication is roughly a matter of defensibility not just to oneself, as some

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317 11

coherentists may have it, but defensibility to others. Defensibility to oneself cannot handle the cases intuitively, because it requires just coherence the absence of defeaters and in cases such as the horse race, we can both have coherent beliefs. What we lack is the capacity to defend our beliefs for each other, and this explains why our beliefs are not justied. So only the dialectical conception gets the right result. Thomas Kelly (2010, 1712) dismisses the dialectical conception of justication too quickly: he appeals to Timothy Williamson (2004; 2007, 23841) who notes that the view he calls dialectical conception of evidence would hand an easy victory to a skeptic who does not accept anything as evidence. Obviously, it is impossible to rationally persuade such a skeptic. If justication required this, it would be impossible. Kelly and Williamson seem to assume that dialectical justication requires that we be able to defend ourselves against all comers even the global skeptic. Perhaps, this is also what the Pyrrhonists presupposed when they argued that we should suspend all belief. However, the assumption is unreasonably strong. We can see how it can be avoided after considering Craigs genealogical account of our concept of knowledge.

5. Craigs genealogy We get further support for the dialectical conception of justication and our diagnosis of the horse-race case by applying the genealogical method of Edward Craig (1990, 117). Craig asks us to imagine a primitive community that does not yet have a concept of knowledge and to consider how that community could benet from having that concept. When we have a hypothesis about the purpose or the role of the concept of knowledge, we can then try to gure out what kind of concept would best serve the purpose or t the role. In this way, we get the concept of proto-knowledge. Then we can try to understand how our current concept of knowledge could have evolved from it. (See also Pritchard 2009, 801.) Craigs hypothesis is that the concept is needed for picking out dependable informants. The person who knows makes a good informant. We can now ask what properties we would want our informants to have. It is clear that we want them to have true beliefs about the questions we are interested in, but, as Craig notes, we also want them to have a property by which we can detect them, and this property must be reliably connected to truth. As some reliabilists have noted, this idea supports a reliabilist account of knowledge: It seems clear that we choose informants by virtue of their reliability. Someone having

12 M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

a reliable vision and standing on a hill is a good informant about what is happening in the valley.6 I think this is right. Knowledge does require a true and reliably formed belief, but it requires more. Often we are not in a position to assess the reliability of potential informants. In such cases, it would be very useful if we could ask them how they know what they claim to know and if they could defend their beliefs for us. People who can defend themselves and respond to our challenges make better informants.7 So Craigs hypothesis about the point of knowledge attribution supports the view that knowledge requires dialectical justication in addition to reliability.8 The hypothesis also supports my diagnosis of the horse race: It is clear that we would not take each other to be good informants about the winner of the race. Neither would a third person who did not himself see the race and who was looking for a trustworthy informant: Even though one of us may very well have a true and reliably formed belief about the matter, she is not in a position to say which one of us has such a belief. Neither can she choose one of us on the basis of the internal evidence that we each have for our beliefs, because this evidence is equally strong on both sides. So neither reliabilism nor evidentialism explains why we are both poor informants for each other and for such a third person.9 Only the dialectical conception can do this: it is because we cannot defend our beliefs for each other or the third party in a way that is dialectically eective. To sum up, I have assumed that there are some cases, such as Elgas horse race, in which disagreement has skeptical consequences and have argued that individualistic accounts, such as evidentialism and reliabilism, cannot explain this. The attempt to appeal to defeaters fails, and the attempts to appeal to independence requirements, symmetry considerations and questionbeggingness are unmotivated and ad hoc. These sorts of considerations are relevant only if the dialectical conception of justication is true. It is only this conception of justication that entails that we should evaluate disagreements

6 See, for example, Sosa (1991, 275) and Pritchard (2009, 805). The example is Pritchards. 7 Fricker (2008, 41) notes that the capacity to give reasons is an important indicator property of a good informant, not discussed by Craig. 8 Let me point out that this view does not make knowledge impossible for children and animals, because mere reliability is in many cases enough to make a good informant. In these cases, their beliefs enjoy the status of default dialectical justication. See below. 9 Both reliabilism and evidentialism entail that one of us may very well satisfy all the conditions of knowledge and thus be a good informant though intuitively neither of us is.

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317 13

from a neutral point of view and that has thus the power to explain the skeptical consequences. Finally, I need to respond to Kellys and Williamsons accusation that the dialectical conception leads to more radical skepticism.

6. Objectication We have so far been working with the concept of proto-knowledge. It has been enough to explain our intuitions in simple cases of disagreement, such as the horse race. In order to get a more detailed account of dialectical justication and to respond to Kellys and Williamsons claim about its more radical skeptical consequences we need to consider how our current concept could have evolved from the proto-concept. Craig (1990, 8297) calls this process the objectication of the concept. Craig focuses on the third-person applications of the concept, but we also apply the concept to ourselves. What is the point of doing that? It is not plausible to suppose that the purpose of my attributing knowledge to myself is to pick out myself as an informant to myself: If I already have the information, I am not in need of an informant.10 A more plausible answer is that, when I attribute knowledge to myself, I thereby volunteer myself as an informant to somebody else. So while the point of third-person applications of the concept is to pick out a good informant, the point of rst-person applications is to volunteer oneself as such an informant. I already argued that a good informant needs to have dialectical justication for her belief. Kelly and Williamson argued that it is impossible to meet this requirement because we cannot defend ourselves for the skeptic who does not accept any premises. However, if Craigs hypothesis is correct, we do not need to convince the skeptic. We just need to convince those who are looking for information about some question. Assume that I need information about some topic and I am evaluating you as a possible informant. Of course, I am interested in whether you have reasons that would convince me about the truth of your belief. I dont care whether they could convince the skeptic. The same is true about self-attributions of justied belief. If I volunteer myself as an informant to you, I take

10 Hookway (1990, 2078) discusses a case in which I can use myself as an informant to myself. However, this is a special case and does not explain most self-attributions of the concept.

14 M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

myself to have reasons that would convince you. If we cared about convincing the skeptic, sharing of information would become impossible. This suggests that attributions of justication are context-sensitive. They have a hidden indexical term. When I say that you are justied in your belief, I say in eect you have reasons that would convince me. When I, on the other hand, say that I am justied in my belief, I say that I have reasons that would convince you. Of course, it follows that my self-attribution of justication would be false if you were a skeptic. However, this skeptical consequence is restricted to the skeptics context. In most other contexts, our attributions of justication would be true. However, this suggestion does not yet give us our current concepts of justication and knowledge. It makes attributions of justication and knowledge too context-sensitive. The content of my self-attribution would vary with whom I am talking to. There are also good reasons from an informationsharing point of view why this is not so. One problem is that the suggestion would make it very dicult to volunteer oneself as an informant. It would require that we be able to keep track of what each individual person would accept as good reasons. We do not usually have such information. Furthermore, we also recommend people as informants to somebody else. This would require that this other person would accept the same reason as we do, and we must be aware of this fact. This is also information that we rarely have. It is clear that the practice of giving and asking for reasons enhances the sharing of information. When this practice has continued for some time, people learn what beliefs are accepted as reasons for other beliefs and what beliefs are accepted without needing further reasons. This further facilitates the sharing of information. Now people have some conception of what kind of reasons informants are expected to possess. This makes it easier to decide when to volunteer oneself as an informant and when to recommend somebody else as an informant to others. So the relevant context is not the individual subject who is looking for an informant. It is composed of the social group, the members of which are sharing information with each other. They have common beliefs about which sources of belief are reliable and under which conditions these sources most likely produce true beliefs. Beliefs that are taken to be based on those sources under those conditions are accepted without needing further support. We may say that they enjoy the status of default justication in the context. They need be defended only if there are specic reasons to doubt their truth or reliability, in which case the default status is lost (see Williams 2001, 14850).

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317 15

So the dialectical conception of justication does not lead to global skepticism. Our need to share information requires that many beliefs have the status of default justication. Beliefs that have this status need not be defended. In order to constitute knowledge they need just to be true and reliably formed and to be taken to be such in the relevant social context.

7. A Skeptical Epilogue It may still be argued that there are many contexts in addition to the context of the global skeptic where the dialectical conception leads to skepticism. There are many areas where controversy prevails, such as politics, religion and philosophy itself. Assuming that disagreements in these areas are genuine and rationally irresolvable, it follows from the dialectical conception that we lack knowledge and justied beliefs about such matters. It may be hard for many people, including philosophers, to accept this consequence, but is it so counterintuitive? Let us focus on the case of philosophy, which has a special importance to my discussion. I nd it quite intuitive that we lack knowledge about many philosophical questions. How often do we in fact attribute knowledge to some party in a philosophical dispute? It seems that it would not be taken to be appropriate. Craigs method conrms this: We would not take disputing philosophers to be good informants about philosophical truths. Neither would we, as philosophers, volunteer ourselves as informants to each other or to laymen. We say instead that people should consider the reasons for and against dierent positions themselves and make up their own minds. It may be further claimed that my account of justication is self-refuting: it follows from it that I am not justied in believing it. This is so if I have not managed to prove my case to other philosophers. Of course, I have tried to defend my view by relying on intuitions and other reasons that they could nd acceptable, but I have no illusions about being successful. This sort of thing rarely happens in philosophy. So let us assume that I am not successful and that many of you do not nd my reasons acceptable. Should I thus conclude that I am not justied in persisting in my view? Self-refutation is a problem only if I believe that my account is true. If I dont believe it, I can very well concede that I am not justied in believing it. Reecting on my attitude to my own account, I must say that I do not fully believe it. This is as it should be given the persistent disagreements in epistemology and philosophy more generally. I am not convinced that I alone am right and all the others defending competing views are wrong. On the other

16 M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317

hand, it seems to me that the considerations I raise at least favor my account. That is why I accept it and try to defend it. It seems that in controversial issues, such as philosophy, the proper attitude is acceptance rather than belief. As several philosophers11 have noticed, there is an important distinction between belief and acceptance. They emphasize two central dierences between these attitudes: (1) Acceptance, unlike belief, is under our voluntary control. (2) Acceptance does not entail belief. We can thus accept something we do not fully believe and use it as a premise in theoretical or practical reasoning. I would like to add a third one: (3) Knowledge requires belief rather than acceptance. If I were not myself convinced about a matter, I would not volunteer myself as an informant about it. Acceptance is not enough. However, what is the point of doing philosophy if it does not give us knowledge? Bertrand Russell (1967, 91), another philosophical skeptic, insists that philosophy is still valuable, because even though it cannot tell us how things really are, it can tell us how they could be. So even if I may not have succeeded in showing that the dialectical view is true, I may have managed to show that it at least oers a coherent view of how things could be epistemically. This is something that I may be justied in believing.12 The dialectical conception of justication does seem to have skeptical consequences concerning philosophy itself.13 This may be why epistemologists have been reluctant to accept it. They want naturally to defend their own profession. However, skepticism about philosophy is far from being counterintuitive. Philosophy is so full of controversy that it would sound very strange if somebody claimed to know the right answers to philosophical questions. The dialectical conception explains this strangeness, because, according to it, the claim would be false.14

See especially Cohen (1992) and Alston (1996). I follow more closely Alstons account of the distinction. 12 Van Fraassen (1980, 12) defends a similar view about scientic theories, which he calls constructive empirism. According to it, accepting a scientic theory involves as belief only that it is empirically adequate, that it saves the phenomena. It does not involve the belief that the theory is true. 13 The same is true of religion and politics where I also nd skepticism intuitive. I have here focused on philosophy because of the accusation of self-refutation. 14 I would like to thank Robert Audi, Raul Hakli and Diego Machuca as well as the audience of the conference on Responsible Belief in the Face of Disagreement at VU University Amsterdam in 2009 for their helpful comments.

11

M. Lammenranta / International Journal for the Study of Skepticism 1 (2011) 317 17

References
Alston, W. P. (1996). Belief, Acceptance, and Religious Faith, 327 in Jordan, J. and HowardSnyder, D. (eds.), Faith, Freedom, and Rationality: Philosophy of Religion Today. Lanham: Rowman & Littleeld. Barnes, J. (1997). The Beliefs of a Pyrrhonist, 5891 in Burnyeat, M. and Frede, M. (eds.), The Original Sceptics: A Contoversy. Indianapolis: Hackett. Bergmann, M. (2005). Defeaters and Higher-Level Requirements, The Philosophical Quarterly, 55: 41936. Christensen, D. (2007). Epistemology of Disagreement: the Good News, The Philosophical Review 116: 187217. Cohen, L. J. (1992). An Essay on Belief and Acceptance. Oxford: Clarendon Press. Craig, E. (1990). Knowledge and the State of Nature. Oxford: Clarendon Press. Feldman, R. (2006). Epistemological Puzzles about Disagreement, 21636 in Hetherington, S. (ed.), Epistemology Futures. Oxford: Clarendon Press. Fricker, M. (2008). Scepticism and the Genealogy of Knowledge: Situating Epistemology In Time, Philosophical Papers 37: 2750. Hookway, C. (1990). Scepticism. London: Routledge. Kelly, T. (2005). The Epistemic Signicance of Disagreement, Oxford Studies in Epistemology 1: 16796. . (2010). Peer Disagreement and Higher-Order Evidence, 11174 in Feldman, R. and Wareld, T. (eds.), Disagreement. Oxford: Oxford University Press. Lammenranta, M. (2008). The Pyrrhonian Problematic, 933 in Greco, J. (ed.), The Oxford Handbook of Skepticism. Oxford: Oxford University Press. . (Forthcoming). Skepticism and Disagreement, in D. Machuca (ed.) Pyrrhonism in Ancient, Modern, and Contemporary Philosophy. Dordrecht: Springer. Lackey, J. (2010). What Should We Do When We Disagree? Oxford Studies in Epistemology 3: 27493. . (Forthcoming). A Justicationist View of Disagreements Epistemic Signicance, in A. Haddock, A. Millar, and D. Pritchard (eds.), Social Epistemology. Oxford: Oxford University Press. Pollock, J. L. (1986). Contemporary Theories of Knowledge. London: Hutchinson. . (1989). How to Build a Person: A Prolegomenon. Cambridge: The MIT Press. Pollock, J. L. and Cruz, J. (1999). Contemporary Theories of Knowledge, 2nd edition. Lanham: Rowman & Littleeld. Pritchard, D. (2009). Knowledge. Basingstoke: Palgrave MacMillan. Russell, B. (1967). The Problems of Philosophy. Oxford: Oxford University Press. Sosa, E. (1991). Intellectual Virtue in Perspective, 27093 in his Knowledge in Perspective: Selected Essays in Epistemology. Cambridge: Cambridge University Press. Van Fraassen, B. C. (1980). The Scientic Image. Oxford: Clarendon Press. Williams, M. (2001). Problems of Knowledge. Oxford: Oxford University Press. Williamson, T. (2004). Philosophical Intuitions and Skepticism about Judgment, Dialectica 58: 10953. . (2007). The Philosophy of Philosophy. Oxford: Blackwell.