Is Truth the Primary Epistemic Goal? (Epistemische Studien)

Markus Patrick Hess

Is Truth the Primary Epistemic Goal?

ontos verla

Bibliographic information published by Die Deutsche Bibliothek

Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliographie; detailed bibliographic data is available in the Internet at http://dnb.ddb.de

@J transaction

6

North and South America by Transaction Books Rutgers University Piscataway, NJ 08854-8042 [email protected]

United Kingdom, Ire, Iceland, Turkey, Malta, Portugal by Gazelle Books Services Limited White Cross Mills Hightown LANCASTER, LAI 4XS [email protected]

Livraison pour la France et la Belgique: Librairie Philosophique 1. Vrin 6, place de la Sorbonne; F-75005 PARIS Tel. +33 (0)143540347; Fax +33 (0)143 544818 www.vrin.fr

©20 I0 ontos verlag P.O. Box 1541, D-63133 Heusenstamm www.ontosverlag.com ISBN 978-3-86838-062-0 2010 No part of this book may be reproduced, stored in retrieval systems or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwIse without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use of the purchaser of the work Printed on acid-free paper ISO-Norm 970-6 FSC-certified (Forest Stewardship Council) This hardcover binding meets the International Library standard

Printed in Germany by buch bucher dd a9

1. Introduction "Is truth the primary epistemic goal?" This dissertation examines the three kinds of answers that have been offered to this question, and their l systematic import. First, there are those who affirm it is; second, there are those who deny it is; and thirdly, there are those who argue that truth is . simply not an epistetTIic goal at all. 2 The third reply shouldn't be confused with denying that the truth is a primary epistemic goal, since the second argument is not an argument to chase truth from the epistemic domain, but is simply about finding a less fundamental place for truth in the valuestructure that organizes epistemic goals. Going further into this dialectic, the naysayers can be subdivided into two groupings. In the first group are those who deny the claim that there is a single fundamental epistemic goal at all. For them, there is always a plurality of epistemic goals present. Kvanvig gives the following list of epistemic goals: "epistemic goals include knowledge, understanding, wisdom, rationality, justification, sense-making and empirically adequate theories in addition to getting the truth and avoiding error. ,,3 On the other side there are those who assume that there is a single fundamental 1 At this point I thank my supervisor Wolfgang Spohn for formulating the working theme of my dissertation as an open question. Different working titles, like "Truth in epistemology" or "The centrality of truth in epistemology", would have been possible. As I see it, the benefit of thinking about an open question is that you start every day anew to think about the question and its implications. 2 That is because the term "goal" might be seen as reserved exclusively for practical action. In Chapter 2 I evaluate several arguments that raise doubt about truth being an epistemic goal. 3 By giving this list, Kvanvig does not present an argument against epistemic value n10nism. He needs to defend the further assumption that all goals on the lists are equally good. Moreover, within this inquiry we shall recheck the alternative epistemic goals on Kvanvig's list in order to see whether all goals on the list are actually independent from the truth goal. If not, the dependent goals are merely means for attaining the truth goal. At least we can assume that know ledge and truth are both epistemic goals, because the knowledge goal cannot be reduced to the truth goal (see Kvanvig 2005, p. 287).

7

epistemic goal, which is not truth but an alternative epistemic goal. We will call this second subgroup "alternative monism". So alternative monism is also in opposition to epistemic value pluralism concerning the order of epistemic goals and epistemic values. Instead of continuing with the labels of "yea-sayers"and "naysayers" it is time to give the accurate labels for the groups introduced. The former are clearly committed to a monism, because for them truth is the primary epistemic goal,4 so I will call them "epistemic value monists".5 The naysayers are pluralists, because they state that we have more than one fundamental epistemic value; thus, I call them "epistemic value pluralists". The distinction between monism and pluralism should not be applied to the concept of truth itself. 6

1. A NORMATIVE QUESTION It is in1portant to emphasize that the scope of my work is not empirical. I do not intend to answer the question I have raised by referring to matters of fact. Therefore the prima facie objection that many people are not factually interested in the pursuit of truth is ruled out. Looking for a factive answer is interesting and enlightening, but I think that we will obtain more philosophical insights from understanding our question as a normative question. Marian David nicely expresses the idea that the truth goal as an ought-to-do aim. He writes: The idea that something is the primary epistemic goal has more the force of an "ought" than the force of an "is." It says that, to the extent that we do have epistemic goals, we ought to have them because we have the one singled out as primary. Obviously, it is rather risky to make claims about what we actually 4 When I use the term "monism" in an unqualified way I refer to epistemic value monism and not to alternative monism. 5 Goldman (1999) has introduced the term "veritism" for the yea-sayers. 6 One might ask whether truth consists of one thing or many things, especially in relation to the functions that truth plays in many different discourses (e.g. in science, in moral theory, even in art and literature).

8

want and about why we want what we want, without asking or studying many of "us." Even if all of us do have the epistemic goals discussed earlier (or at least closely related ones), and even if these goals are in fact connected in the manner mentioned above, some of us might not realize that their epistemic goals are so connected. It would then be wrong to say that we have other goals because we have the goal that is in fact the primary one; all that can be said is that we ought to. (David, 2005, p. 302-303; emphasis in original)

David clearly expresses the case for saying that aiming at the truth contains . a norlnative claim. Furthermore he argues that it is easier to determine an answer to the normative question than to the empirical question. The latter couldn't be completely answered unless we polled everybody, an obviously impossible task. Additionally, even if we could provide samples that helped us to create a descriptive picture, it wouldn't get us any closer to answering the question of whether the truth as an epistemic goal is a justified normative requiren1ent, because we evaluate the normative claims with respect to whether they are reasonable or not, and not whether they are empirically testable. A normative claim is not refuted once we have detected that many persons factually ignore it. Quite the contrary: We might say that the requirement is even in need of enforcement, once we realize how many people do not follow it. Second, episten1ic evaluation itself sets normative standards. Evaluative concepts can be characterized by their double role of being norn1ative and descriptive at the same tin1e. 7 Note that our observations of ignorance, for instance, are already led by our normative intuitions. We cannot describe ignorance without taking into account the criteria of justified belief. If we describe how a believer is actually ignorant we have already prescribed how the believer should actually perform his belief act, because the concept of ignorance is already a norn1ative concept which provides us with normative standards of evaluation. It is clear that we have learned a great deal about the human beings from the study of their mechanisms of acquiring and managing information. This empirical research has and will continue to have a great 7 Spohn (2002) has introduced this term and has discussed it using the example of a theory of rationality.

9

impact on epistemology. I want to emphasize, however, that epistemology is a normative discipline and therefore epistemologists are allowed to reconsider normative question. Epistemology is normative because central epistenlic concepts, like epistemic justification or rational belief, are normative. We are dealing with formal evaluative concepts, here, and not context-specific epistemologies, such as one might examine in the history of the sciences, for instance.

2. TRUTH AS THE PRIMARY EPISTEMIC GOAL Let us proceed to the statement of the problem that has aroused the most controversy, which is whether the truth can or cannot be the fundamental epistenlic goal. To repeat: it is in generally accepted that truth is an epistemic goal. However, the question remains whether it is the fundamental one. Whether it is or not depends, partly, on how we construe fundamental. So let us begin our approach to the problem here. There are several ways in which we speak of something as fundamental. We could mean that it requires no proof. We could mean that we build our arguments upon it. However, the meaning that holds the greatest interest for us defines "fundanlental" in terms of what takes precedent in the order of derivation. Consider two items, A and B: if A is fundamental to B, then it is possible to derive B from A but not vice versa. Furthermore, if A and B are related and A is allegedly fundamental to B, then we can say with respect to value attribution that B has only instrumental value. This presents us with a clear hierarchy of values: B's value depends on its relation to A. How does truth as a goal fit into this discussion? Assume that truth is represented by A, and coherence by B. Coherence will then be a derivative value, in as much as it gains its salience only with relation to the search for truth. The truth goal, on the other hand, is fundamental. You could not derive the truth goal fronl coherence. 8 Relatedly, the value of coherence is 81 do not want to nlaintain that coherence is an epistemic goal, because in a very

10

instrulnental because coherence is valued for being conducive to truth. Strictly speaking, the instrumental value is defined as a means of attainment. Any n1eans helpful for attaining the central goal becomes derivatively valuable in so far as the central goal is valuable. We can further conclude the following: if the truth goal had no value, then the Ineans to obtain the truth would not have value either. To epistemic value pluralists, the claim that truth is the fundamental value seems unjustified on the level of our above definitions. That is to say, it isn't the case that any other epistemic value has only derivative value. Their central intuition is that we can be in a situation where we do not attain truth but something else that is also epistemically valuable. In contrast, epistemic value monists are committed to the view that one can only attain something of epistemic value by attaining truth. This is the central debate of my dissertation in a nutshell. So far, we have clarified what type of question we are confronted with. We have seen what the prin1acy claim of epistemic value monism lies in. To complete our gloss on the terms of the controversy, we need to confront another question: Does our debate presuppose a certain concept of truth? It seems intuitively reasonable to expect that the choice between the many concepts or theories of truth would affect the discussion. Yet our argument is that no commitment to a particular theory of truth is involved in accepting the primacy of the truth goal. 9 The reason is: Epistemology in strict sense coherence is only a means to attain truth. In the literature coherence is sometimes treated as though it were an epistemic goal itself, but I'd advocate restraint here. Coherence, while it might be epistemically valuable, is not an epistemic goal. Firstly, this would unnecessarily inflate the list of epistemic goals. Secondly, we would confuse the difference between a means and a goal. A means is not per se a goal. For instance, my belt is helpful to keep my trousers up. In a situation where I have lost my belt somewhere in the house, finding the belt can be the goal of my inquiry. But my search for the belt is only a means to my real goal, of getting dressed, in the same way the buttons on my shirt and my shoelaces are. I think this example is helpful for understanding what epistemologists actually do when they reflect on what coherence is and in what sense coherence is truthconducive. 9 There are several works at the intersection between epistemology and truth theories. See for instance Sosa (1993, 200 1b) and Williams (1986). But none of these works

11

general is not determined by any rigid conception of truth, except for the everyday usage of the concept of "facts". 10 Another issue which quite obviously must be discussed is the matter of doxastic control. Normally we associate goals, especially personals goals, with deliberation and personal planning. How does the term "goal" fit in the epistemic value domain? The best way to reflect on this issue is to take up the question of the role the will plays in our belief formation, since this is what is assun1ed in deliberation or personal planning. Furthermore, in a more theoretical vein, it seems important to decide whether epistemic goals in general, or the truth goal in particular, imply a commitment to doxastic voluntarism. Doxastic voluntarism assumes that our will has control over our beliefs. Another assumption of doxastic voluntarism is that the formation of beliefs is similar to or even falls under the category of action. 11 The raising of my arm is controlled by my will, and the raising of my arm is an action. Doxastic voluntarists claim that the formation of a belief works similarly to the raising of an arm. As I take it, there is no commitment to doxastic voluntarism. It is not implied by the acceptance of truth as an epistemic goal; the truth goal provides at best an indirect control of our belief system. Epistemic goals and especially the truth goal provide an indirect goal because I can discard false beliefs. At this point one might object that already indirect control is a form of control and therefore the acceptance of the truth goal presupposes doxastic voluntarism. I do not want to go into greater detail; however, I think that voluntarism is primarily interested in direct control. So far I have been attempting to clear the field, explaining all the implications of our question and, as importantly, what it doesn't imply: come to the conclusion that a particular theory of truth has to be accepted by epistemologists. Pritchard (2004) has given a nice work in comparing deflationism in epistemology with the deflationism of truth. 10 Epistemology in general is not affected by the problems and paradoxes that theorists of truth investigate, because the latter do not attempt to complete the analysis of truth. See Lehrer (2000, pp. 26-32) for further discussion. Goldman (1999, Ch. 2) argues for a mixed account between a correspondence theory of truth and a deflationary theory of truth. 11 See Bennett (1990), who thinks that believing is voluntary.

12

either a position on the theory of the truth, or a position in the psychology of belief. In the following section I explain how I situate my project within epistemology. 3. VALUE-DRIVEN EPISTEMOLOGY I situate my project within value-driven epistemology. Value-driven epistemology is focused on which epistemic goals (and values) are central for epistemology. Traditional epistemology proceeds by presupposing certain epistelnic values and epistemic goals, and then going forward to make claims about true beliefs, proof, etc., without checking to see whether its presuppositions are justified. Value-driven epistemology has a narrow and a broad scope. The narrow scope of the project is concerned with a particular problem, which is called the Meno Problem in the literature after Plato's dialogue. There, the intriguing question is posed about what makes value lies in knowledge in comparison to mere true belief. In the Meno, Socrates seems to show that, according to the criteria of practical success, knowledge and mere true belief are equally successful. The possibility that knowledge has more value in terms of practical success is excluded from the start. 12 The value question refers to the distinctively epistemic value of knowledge. Hereafter, I will call the Meno problem the "value problem of knowledge". At this point we do not need to go into more details of a discussion that has accrued an impressive amount of literature. However, epistemic value pluralists take it that the upshot of the value problem is that epistemic value monists do not have the resources to explain why knowledge should possess this distinct epistenlic value.

12 This assumption is taken for granted here, even though in the epistemological literature some authors defend the idea that the additional value of knowledge has its roots in knowledge's guaranteeing more practical success than mere true belief, because knowledge has more stability than mere true belief. Olsson (2007), for instance, argues that we can identify the additional value of knowledge in terms of practical success.

13

Besides the narrow problem of the specific value of knowledge, there is the broader scope to value-driven epistemology. In that it investigates which values are fundamental in epistemology. Here we have to differentiate between epistemic values and practical values. What exactly is an epistemic value? What is an epistemic goal? What is an epistemic norm? Goals and norms go together under the assumption that norms for a goal x are conducive to attaining x. So, norms are understood instrumentally in the sense that they are helpful in attaining a certain goal. Concerning truth as an epistemic goal, we suppose that there exist certain norms that are helpful for attaining truth. For each episten1ic goal we will have one or more corresponding epistemic norms. A difficulty for the epistemic value pluralist is to spell out all the possible norms that are conducive to attaining his plurality of epistemic goals without Inisidentifying any of the used norms. The above merely characterizes our conceptual grasp of how the terms "epistemic goal" and "epistemic norm" are used; it doesn't give us a fullyspelled-out definition. Note that the crucial distinction here is between the domains in which epistemic and practical values inhere: purely epistemic values primarily serve epistemological matters. We can approach the question of episten1ic values from the point of view of preferences concerning the value of knowledge and the value of truth. Just reconsider the preference question: Would we prefer having true beliefs acquired in a responsible way over having mere true beliefs or not? Deciding what my preferences are expresses my epistemic values. So, if we prefer true belief acquired in a responsible way over just having true belief, then we'd take this as the expression of a value: Some allegedly prefer responsible acquired belief over n1erely true belief. These preference tests provide a starting point from which we can proceed to learn more about epistemic values. 13 Epistemic values are not decided by our preferences in the domain of practical values; something can be epistemically valuable (say, the question of the truth of string theory in physics) without having a concrete 13 This, the reader may have noted, brings us back to the narrow scope; our point, however, is to show that the broad scope also allows us to ask similar preference questions.

14

practical value. The distinct value of knowledge is an epistemic value: it is detached from practical value. This is not the same as saying that truth does not have any practical value at all. Quite the contrary: It is legitimate both within the domain of the practical and of the epistemological to ask what practical value knowing certain truths has. This question refers to the relation between the epistemic value domain and another value domain outside of it. What is the contribution of value-driven epistemology to traditional epistemology itself? The broad scope of value-driven epistemology, is concerned with the value structure of epistemology. It concentrates on which values are fundamental in epistemology and how all the central values are interrelated. Such a study will offer quite fruitful results. A good starting point for understanding the value-structure is to examine the debate between epistemic value pluralism and epistemic value monism, for in that debate, all the central questions are posed. 14 The narrow scope provides a further criterion for accepting a theory of knowledge. Jonathan Kvanvig (2003) has smartly suggested that the question about the value of knowledge should function as a check on theories of knowledge, so that, if a given theory of knowledge is incapable of explaining the added value of knowledge, it should then be discarded. But the narrow scope might introduce other, attention-worthy epistemic goals. Kvanvig concludes his book by arguing for the thesis that there is no theory of knowledge that can explain the added value of knowledge, which would seem, given his criterion, to justify a heavy dose of therapeutic nihilisln. 15 But then he supposes that there are other phenomena which are worthy of study. On this score, Kvanvig supposes that understanding is an important but neglected epistemic goal. This example brings out one of the prospects opened by value-driven epistemology - to wit, the discovery of fundamental, but so far neglected, epistemic goals. As a consequence of Kvanvig's suggestion to the value problem of knowledge, it seems that epistemology has to be expanded. In short, epistemic value pluralists may be attempting to broaden the focus of 14 Riggs (2008, p. 310) agrees that the debate between epistemic value pluralism and epistemic value monism is of key importance for value-driven epistemology. 15 See Kvanvig (2003).

15

our epistemological interest. In the following, I will outline some of the proposals that have been introduced in the literature. Brian Ellis (1988) for instance has argued that the primacy of the truth goal in mainstream epistemology has blocked our recognition of the plurality of epistemic goals: having created a false value-structure, epistemologists are unable to accord inlportance to epistemic goals that lie outside of the primacy of the truth goal. Wayne Riggs (2007) more specifically claims that epistemology needs expansion concerning the family of concepts of "epistemic credit", "responsibility", "luck" and "attribution" .16 Riggs's requirement presupposes that we have to understand knowledge as a cognitive accomplishment. The believer is doing something actively right when she acquires knowledge, and for that she should be credited. John Greco works on a similar theory. He expresses this line of thought as follows: [...] when we attribute knowledge to someone we mean to give the person credit for getting things right. Put another way, we imply that the person is responsible for getting things right. The key idea here is not that knowledge requires responsibility in one's conduct, although that might also be the case, but that knowledge requires responsibility for true belief. Again, to say that someone knows is to say that his believing the truth can be credited to him. It is to say that the person got things right owing to his own abilities, efforts, and actions, rather than owing to dumb luck, or blind chance, or something else. (Greco 2003, p. Ill; emphasis in original)

Responsibility as a category of epistemic evaluation is not new to

16 The expansion of epistemology is in part politically motivated. I mean 'political' not in the sense that epistemic value theorists see themselves as a social movement, as did, for instance, the Logical Empiricists in the early stage of the movement. Value-driven epistemology reflects about the actual goals of epistemology itself. I am indebted to Guy Axtell for making me aware of the distinction between the political side of value-driven epistemology and its argumentative side. See Axtell (2008a) for more details. I personally have sympathies with the political side of epistemic value pluralism, but I would like to keep the discussion on the argumentative level. In general, value-driven episten1010gy does not strive to revise epistemology. Riggs, for instance, states that value-driven epistemology is not about adding new inquiries to epistemology, but, rather, is directed at questioning the legitimacy of ongoing projects (see Riggs 2007, p. 308).

16

epistemology.17 There are several suggestions in the literature about correctly embedding responsibility in epistemology. On this score Greco and Riggs argue for a second track of epistemic evaluation: it is not only the acquisition of truth that is important, but also the involvement of the cognitive abilities of the believer in attaining truth. IS Jonathan Kvanvig (2005) has expressed a more radical view concerning the expansion of epistemology. The starting point for Kvanvig's discussion is metatheoretical. He argues that traditional epistemology needs to widen its scope by allowing for new phenomena. For him, the lack of this broadness has resulted in the overlooking of many cognitive phenomena by traditional epistemology. Kvanvig redefines epistemology's actual aim as being "to investigate successful cognition". By changing epistemology's interests from knowledge to the broader concept of successful cognition, Kvanvig hopes to open the door to many epistemically valuable phenomena that knowledge-centric epistemology has ignored, among them: thinking, reasoning, inquiring, revision of one's own belief system, understanding, theoretical wisdom, rational presupposition, responsible inquiry. The crucial shift in assumptions, Kvanvig thinks, comes when we see that cognitive success can't be adequately expressed in terms of truth alone. Kvanvig writes: In my view, the fundamental function of cognition and the fundamental intentional attitudes involved in cognitive activity can be characterized without appeal to the concept of truth. (Kvanvig 2005, p. 293)

Kvanvig separates two levels of analysis. The first level is concerned with epistemologists themselves, what they take to be the central components of 17 One standard objection to reliabilism is that it cannot incorporate into its argument a good account of the believer's responsibility for obtaining good results. See the classical paper by Zagzebski (2000). See also Montmarquet (1987) and Riggs (1997) for earlier versions of this view. Or see Kornblith (1983) actiontheoretic form of epistemic justification. 18 I use the more neutral term "involvement" because responsibility entails more than the involvement of an ability. Responsibility also references the notion of a duty or an obligation. Greco appears to use the term "responsibility" ambiguously, because we can either use it in the strong sense of a duty, or in the softer sense of the believer's capacity to respond to certain stimuli. But these two usages of "responsibility" are not completely compatible one with the other.

17

any epistemological theory; while the second level is defined by the objects of epistemological theorizing, namely, the believers themselves. Kvanvig is teasing out a point about ordinary epistemology: just because truth is central to the construction of epistemology for epistemologists, this doesn't mean that truth holds a similar importance for ordinary persons. Kvanvig's notion does seem borne out by ordinary life. Nonetheless, Kvanvig's observation could sin1ply mean that epistemological theorizing is independent from what preference believers have in everyday life. 19 In order to complete this survey of the dialectics between epistemic value pluralism and epistemic value monism, we need to clarify what assun1ptions the two positions have in COlnmon. We can summarize the common ground in three theses: (CO1) Truth is an important epistemic goal. (C02) Truth is valuable. (C03) There are several epistemic goals.

Epistemic value pluralists accept the importance of the truth goal, just not the claim to truth's primacy. Another way of defining this position is to say that epistemic value pluralists 'relativize' truth to other important goals. Episten1ic value pluralists suppose that all epistemic goals are equally important. How does episten1ic value pluralisn1 distinguish itself from epistemic relativism? Epistemic value pluralism shares a certain ground with relativism as well as with epistemic monism. Like relativism, epistemic value pluralism doubts that there exists a single standard of correctness under which we can sort all our epistemic claims. On the other hand, unlike classical relativism, epistemic value pluralists do value truth, and they do endorse the existence of facts which are independent of us. Relatedly, epistemic relativists do endorse a relativism about facts,

19 Because of this independence Kvanvig's point is moot. This description nicely illustrates the motivation that some epistemic value pluralists share. They fear that with truth as the primary epistemic goal we will result in an "absolutism", as Elgin has called it, which will falsely constrain the further development of epistemology (see for instance Elgin 1996).

18

whereas epistemic value pluralists do not. 20 If we accept the truth as an important epistemic goal, then we must believe that truth is valuable for us. Therefore it is no surprise that for both camps, truth is valuable. It remains to be seen how that is construed, and whether the truth has merely a practical value or an intrinsic one?1 The third thesis may appear peculiar at first sight, if we understand epistemic value monism as a strict monism. Then we might expect it to programmatically disallow any but one epistemic goal. But this is not correct: epistemic value monism easily accepts the existence of a plurality of epistemic values and goals, but organizes them into a structure in which truth must always be fundamental. Once we have specified a list of the central epistemic goals, epistemic value monists are inclined to clarify how those epistemic goals are ordered. This brings us to the core of the debate between epistemic value monism and epistelnic value pluralism, which is not about the possibility that there are many epistemic values and epistemic goals, but about their ordering. Each camp argues for a certain order among epistemic goals. 4. STRUCTURE OF THE BOOK The general structure of the dissertation is the following: Chapter 2 and 3 prepare the ground for our question. Chapters 4 to 7 evaluate the central objections given by epistemic value pluralists. In Chapter 2 I review the arguments given by Richard Rorty and Donald Davidson. They argue that truth cannot be an epistemic goal because (a) goals refer to practical action and (b) we cannot say when we have actually attained truth. I will show that Rorty and Davidson's argument that truth is not an epistemic goal are not conclusive. Their 20 See Weinberg (2007) and Pritchard (2009b) for a description of such a nloderate relativisnl. 21 The distinction between the instrumental and the intrinsic value of truth does not affect the dialectical positioning between epistemic value monism and epistemic value pluralism. To be precise, there is a tendency in the camp of epistemic value pluralism to treat the value of truth as an intrinsic value. See for instance Zagzebski (2003) as a paradigmatic representative of epistemic value pluralism.

objections are worthwhile for differentiating epistemic from practical goals. In the second part of the chapter I show where the idea of the truth goal has its origin, and after that I characterize the truth goal as an appealing epistemic ideal that is independent from personal acceptance of the epistemic goal. This chapter begins our investigation because it is necessary to show that we are justified in speaking of truth as an epistemic goal. In Chapter 3 I further analyze in what the value of truth is. With the acceptance of the tnlth goal it is self-evident that we must attribute some value to truth. It is important, then, to understand what kind of value truth actually has. Here we face the difficult question of whether truth is only valuable for practical reasons or whether it has non-instrumental value. In the first part of the chapter I argue that we must value truth at least on practical grounds. In the second part, I clarify the motivation for thinking that truth also has non-instrumental value. Finally I survey all of the arguments that are supposed to support this motivation, and find them all defective. This chapter prepares the ground for understanding properly how epistemic and non-epistemic values are related. 22 Chapter 4 argues for a schema of defensible requirements that derive fron1 the truth goal. The traditional view is that there are two such requirements, believing all truths and avoiding false beliefs; but I find these to require covering too much ground relative to the truths that we need for our epistemic project. I argue in this chapter that in place of these two requirements we can derive more narrowly tailored defensible requirements from the truth goal, which would be to believe significant truths. This chapter opens the real debate between epistemic value monism and epistemic value pluralism. Riggs's argument concludes that we need an overarching goal that regulates between the two extreme requirements of believing all truths and avoiding error. If Riggs is right, I will argue, then it would follow that truth cannot be the primary goal because the requirements following from it makes it necessary to introduce another overarching goal.

22 Questions about the moral value of truth (which have their own interest) are excluded from our discussion.

20

In Chapter 5 we enter the main ground of the discussion. In what sense is truth the pritnary epistemic goal? Epistemic value tnonists, as I take it, defend a prima facie claim concerning the primacy of the truth goal. Epistemic value pluralists have attempted to challenge this prin1acy claim by giving three counter-examples. I argue that all three counterexamples can be appropriately responded to without altering the the monist's prima facie claim and that therefore the primacy claim stays intact. In Chapter 6 I discuss what I call "alternative monisn1". Alternative monists suppose there is a more fundamental epistemic goal than truth, which we construe, for the purposes of the chapter, as the goal of being a rational or responsible believer. I argue against this account by stating that being a rational believer cannot be the alternative fundamental goal because it does not provide the guidance provided by the truth goal. In Chapter 7 I discuss the value problem of knowledge. Epistemic value pluralists see this problem as a challenge to the primacy of the truth goal because a second source of epistemic value must be assumed in order to solve the value problem of knowledge. I argue that epistemic value monism can explain what the added value of knowledge consists in without accepting the dual evaluation thesis. In the conclusion, which is Chapter 8, I give a short description of the main results of my work. In Chapter 9, the appendix, I apply the framework of epistemic value monism to the problem of epistelnic relativism and conclude that epistemic value monism is a suitable fran1ework to overcome that problem. The key idea is that epistemic value monists are not committed to a certain set of epistemic norms. They are only committed to a certain requirement of the correctness of epistemic norms.

2. Can Truth Be an Epistemic Goal? Richard Rorty and Donald Davidson have objected to the notion that truth is an epistemic goal because truth is not recognizable as a practical goal. By evaluating their arguments we can gain a better understanding of the idea of what it means for truth to be an epistemic goal. I examine those suggestions critically, and then present an outline of what it would mean to model goals not in tenns of the role they play in practical life, but as regulative ideal. Kornblith (2001) has helpfully outlined his own notion of how goals can be seen as ideals. I end with a brief discussion of the setnantics of the 'ideal' and its relevance to epistemic values. 1. RORTY ON TRUTH Rorty defends a radical form of deflationism about truth. 23 For him, truth is not a metaphysically heavyweight notion and he therefore holds that "there may be little to be said about truth" (Rorty 1995, p. 281). Rorty does not support the thesis that truth does not exist, nor does he maintain that truth is not important to us. In fact, Rorty thinks that the word "true" has important usages. Therefore, it would be unfair to Rorty's point to think of him as a truth nihilist. 24 Rorty, on the contrary, holds that there are three iInportant usages of the word "true" (Rorty, 1986): a) endorsing or performative usage, b) cautionary usage, c) disquotational usage.

Not only is he a deflationist about truth, but he also contends that the importance of nlany other distinctions and conceptions on the philosophical landscape has been exaggerated (see Blackburn 1998, p. 168). Pascal Engel argues that Rorty is a radical detlationist, but not, as many other critics have stated, a relativist (see Engel 2002, p. 137). Rorty himself thinks that many critics have understood hinl as a truth nihilist because he gave up the notion of correspondence (see Rorty 1998, p. 2). 24 Nietzsche can be seen as a truth nihilist (see Blackburn 2005, Ch. 4). 23

23

The endorsing or performative usage can be found when one person states something and another person says, "this is true", in the sense that he endorses the first person's statement. The meaning of the utterance "this is true" can be translated by the phrase "I believe this, too" or, more formally, "I believe (too) that p." Putting its meaning that way, it becomes clear that the first usage is just a signal of agreement. The cautionary usage indicates that there is a difference between justified belief and true belief. We can see this in the following phrase: "Your belief about X is justified, but it is not true." This applies to cases where we seem to have had good evidence for some belief, but it turned out that it was false. Suppose, for instance, that my brother is normally a reliable informant, but sometimes he is not. For instance, he does not want me to know how he spends Friday nights in the week of final exams. Last Friday he told me that he was studying late at night in the library, but in reality he went drinking with his friends. So, on the one hand, I am justified in believing that my brother was studying, because normally he accurately informs me; but, on the other hand, this time he didn't. The cautionary usage of the word "true" helps us to differentiate between true belief, on the one hand, and false belief for which we appeared to have some justification, on the other. The disquotational use of the word "true" indicates the indirect way of saying what the sentence said to be true states. The sentence "'Caesar was murdered' is true" states in an indirect way that Caesar was murdered. Thus, according to Rorty, apart from these three usages of the word "true", there is nothing to be said about the nature of truth. Rorty employs an understanding of the term "goal" that restricts its meaning to its ordinary usage. Rorty assumes that talk of a "goal" presupposes that there exists a point at which we can arrive in the end. This model, he claims, shows that putting together goals and truth will not do: It [truth] is not what common sense would call a goal. For it is not even something to which we might get closer, much less something we might realize we had finally reached. To try to make truth approachable and reachable is to do what Davidson deplores, to humanize truth. (Rorty 1995, p. 298; emphasis in the original)

In the introduction to his book Truth and Progress Rorty additionally writes:

24

If pragmatists cannot offer a theory of truth, what can they do? They can point out, I argue in the first essay of this volume, that truth is not a goal of inquiry. If "truth" is the name of such a goal then, indeed, there is no truth. For the absoluteness of truth nlakes it unserviceable as such a goal. A goal is something you can know that you are getting closer to, or farther away from. But there is no way to know our distance from truth, not even whether we are closer to it than our ancestors were. (Rorty 1998, p. 3-4)

Rorty says several puzzling things. What is central for our discussion is his understanding of what a goal is. This is the semantic launching pad for his attack on the notion that the truth could be a goal. For hinl a goal is defined by the conditions of its attainment, which are spelled out in terms of social action - e.g., reaching a certain point in space and time. This is an ordinary language determination of the meaning of 'goal', which is instantiated in everyday situations. Suppose my wife and I want to visit friends in Berlin. Suppose further we drive with our car and my wife ask Ine during the journey, "How far is Berlin?", then I am able to answer, understanding that Berlin is a 'fixed' goal towards which we are traveling in the car. I can detennine how far away Berlin is, and observe the distance diminishing as we approach it. This is, for Rorty, a paradigm case of goals and approaching goals. A goal is always an end point combined with clear instructions about how one can get to that end point. This definition of goal does not go together with what he calls the "absoluteness of truth. "But why should we believe that Rorty's understanding of a goal exhausts the semantic possibilities of it? Isn't it arbitrary to hold such a strict characterization of a goal (particularly if Rorty rules out the idea of theoretical goals). 25 I think Rorty commingles two different aspects of a goal. First, a goal can be identified or determined. Second, after determining the goal, we can attempt to attain it. Let's look at how this plays out in ordinary talk. Suppose that I want to become a millionaire. I can define or determine the goal of becoming a millionaire as having one million Euros on my bank account. This goal is clearly identifiable with regard to the amount of 25 For the distinction between theoretical and practical goals, see Grundmann (2002, pp. 122-123).

25

money on my bank account. Determining whether I have reached the goal of being a millionaire is fairly easy; but this leaves me with the problem of attaining it. That, it turns out, is hard, and is no way determined by simply setting the goal. There are innumerable paths to the attainment of the goal. As this example clearly shows, we make a distinction between the identifiability and the attainability of a goal even in everyday life. Rorty is not clear about this distinction when he uses the word "goal". So what is the actual problem for Rorty? Is it the determinability of the truth goal? Or is it its attainability? It seems to me that Rorty is happy to agree that goals can be determined in the epistemic sphere - it is the attainability of truth that poses the crucial problem for Rorty. Rorty states that we never can be sure of having attained truth, which points to Rorty's rigid sense of what goal talk is about. This impression is supported by another argument that Rorty presents against the idea that truth is an epistemic goal. 2. PRAGMATISM AND TRUTH Rorty's argument is rooted in a line of thought that goes through the tradition of American Pragmatism. 26 Rarty endorses pragmatism as an umbrella notion for all the things he favors. He writes:

26 Richard Rorty does not give a fair treatment to James's and Peirce's theories of truth. See Fuhrmann (2006) for a good discussion of James's account of truth, and see Misak (1991) for Peirce's theory of truth. American Pragmatists were often criticized for reducing theoretical goals to practical goals. See, for instance, the debate between John Dewey and Bertrand Russell in Dewey/Russell (1985). In sun1, practical goals can be clearly attained, whereas theoretical goals are openended. Consider, for instance the search for the origin of the universe. In this case it is not clear whether we will ever find a definitive answer. Or think of the explanation of what holds the physical world together. Those broad inquiries seem to be open-ended because we do not know where to stop or whether we will ever be satisfied with the current results and answers to those questions. The problem with this distinction is that theoretical goals in general need not be open-ended. Consider the goal of inquiries in science today. All inquiries are designed in such a way that within a certain period of time the inquiry has to come to an end. In principle, theoretical inquiries can be finished. Still, this distinction is helpful for understanding what Rorty has left out by his narrow characterization of goals.

26

The question that matters to us pragmatists is not whether a vocabulary possesses meaning or not, whether it raises real or unreal problems, but whether the resolution of that debate will have an effect in practice, whether it will be useful. We ask whether the vocabulary shared by the debaters is likely to have practical value. (Rorty 2007, p. 34)

This general pragmatist line of thought affects the relation between justification and truth, as Rorty thinks himself: Pragmatists think that if something makes no difference to practice, it should ITlake no difference to philosophy. This conviction makes them suspicious of the philosopher's emphasis on the difference between justification and truth. For that difference makes no difference to my decisions about what to do. If I have concrete, specific doubts about whether one of my beliefs is true, I can only resolve those doubts by asking if it is adequately justified - by finding and assessing additional reasons pro and con. I cannot bypass justification and confine my attention to truth. Assessnlent of truth and assessment of justification are, when the question is about what I should believe now, [...] the same activity. (Rorty 1995, p. 281)

In short, Rorty argues that justification and truth are indistinguishable in social contexts, and justification is the epistemic goal in the sense that Rorty supports. In this context it is worthwhile to reconsider Rorty's dismissive attitude toward the traditional idea that our beliefs somehow correspond to the world. We do not need to reconsider the first part of Rorty's statelnent; more interesting for us is the second part - especially what he says about the utility that beliefs should have. Rorty says:

I read Sellars and Brandom as pragmatist, because I treat psychological nominalism as a version of the pragmatist doctrine that truth is a matter of the utility of a belief rather than of a relation between pieces of the world and pieces of language. (Rorty 1998, p. 127)

The puzzle here is how Rorty could endorse the standard of "the utility of a belief' when, at the same time, beliefs, for him, are not related to the world. In order to complete the picture of how Rorty thinks about justification, consider the following statement as well:

27

For, once again, the only criterion we have for applying the word "true" is justification, and justification is relative to an audience. So it is also relative to that audience's light - the purpose that such an audience wants served and the situation in which it finds itself. This means that the question "Do our practices of justification lead to truth?" is as unanswerable as it is unpragmatic. It is unanswerable because there is no way to privilege our current purposes and interests. (Rorty 1998, p. 4)

Here is an example to illustrate that Rorty's audience-relativity of justification is problematic. Suppose I claim that the local branch of the Deutsche Bank in Constance is open this Saturday. I have read in the newspaper that the director of the bank has decided to open on Saturdays so that the customers of the other local bank, which went bankrupt, will be attracted to the Deutsche Bank to do their banking. And suppose I want to check whether the article in the local newspaper was right and I go to the bank, because it is on my way to the grocery market. By entering the bank and seeing that full bank operations are going on, I can verify the report of the newspaper and claim that I have grounds to justify believing that the bank is open on Saturday. So in what sense do the standards of the audience condition the process of justification? Whether I am justified in believing that the bank is open depends solely on whether it is the case that the bank is open or not. In general, I think that Rorty commits a simple mistake concerning the practical value of beliefs. True beliefs guarantee more practical success than false one. From the pragmatic point of view, we will prefer true belief over mere justified belief?7 Relevant true beliefs give us a guarantee for successful action, whereas merely justified belief lacks such a guarantee. This does not preclude that there could be cases in which, for some reason, a merely justified (but false) belief has more utility than a true belief. I want to defend the claim that true beliefs have, in general, higher utility in comparison to false beliefs, mere justified beliefs or suspension of belief

27 Loewer (1980, 1993) provides an excellent discussion of how truth has practical value~ and he gives a good overview of how decision theory is helpful in that context.

28

against Rorty's dismissal of this sense oftruth.

28

Suppose that I usually walk to university, but that on a very rainy day I want to take the bus. I acquire a bus schedule, but fail to notice that it is out of date, and that a new schedule has supplanted it. Let us also suppose that the new time table differs a lot from the old one. By looking up the times in the old schedule I form a justified belief, but it either might not result in successful action or result in successful action only by chance. The crucial point that the example is supposed to make is: there is quite a difference in practice whether I act on merely justified belief or whether 1 rely on true belief when I act. The difference in outcomes is notable: if our initial aim is to have justified belief then this aim requires less than the truth goal. For the former aitn it is enough to look up in any time schedule (even the old ones) because by looking up the times we are justified to believe what is there written about them. But when we aim at truth only the justification leading us to truth is the right justification. In general, false beliefs are not likely to have practical value. If we were to unpack the utility of belief in cases such as that I have outlined above from Rorty's point of view, then the ironical result is that truth has more practical value than mere justified belief.

28 I discuss in Chapter 3 several counter-examples which are designed to show that false beliefs can have high practical utility.

29

3. DAVIDSON ON TRUTH AND EPISTEMIC GOALS Rorty's deflationary approach to truth is paralleled by Davidson's, who Rorty acknowledged as an influence. Davidson also declares that it is meaningless, philosophically speaking, to speak of the "pursuit of truth" or to say that "truth is the goal of inquiry,,?9 Davidson and Rorty agree that truth cannot be a target in the practical sense. The only thing we can achieve when pursuing an inquiry is to increase our confidence in holding particular beliefs as true. A full reconstruction of Davidson's objection against the truth goal is complicated, because his argument is never worked out in detail in anyone essay, but must be pieced together from the several passages in Davidson's later work that express his denial of truth as an epistemic goal. But those passages contain statements where the reader is left to reconstruct the framing argument that would lend them plausibility. I will first present some citations, and then will introduce some background about Davidson's theory of truth. In the following Davidson refuses that truth can be a goal. Davidson states in a reply to Pascal Engel: But I do not find it adds anything to say that truth is a goal of science or .anything else. We do not aim at truth, but at honest justification. (Davidson 1999b,p.461)

This is how Davidson expresses his refusal to countenance truth as a goal in more detail: Since it [truth] is neither visible as a target nor recognizable when achieved, there is no point in calling truth a goal. Truth is not a value, so 'the pursuit of truth' is an empty enterprise unless it means only that it is often worthwhile to increase our confidence in our beliefs, by collecting further evidence or checking our calculations. From the fact that we will never be able to tell for certain which of our beliefs are true, pragmatists conclude that we may as well identify our best researched, most successful beliefs with the true ones, and give up the idea of objectivity. (Davidson 2005, p. 6-7) 29 See Davidson (I 999a, p. 17). But Davidson denies being a pragmatist.

30

One very unwelcome consequence of Davidson's view is that we could not indicate which of our beliefs are actually true. Davidson recognizes -this difficulty, but he thinks that he must make a choice between two wellestablished intuitions about the nature of truth. The first intuition is that truth is objective, and the second can be summarized under the heading that we can be sure when we have attained truth~30 Davidson opposes the second intuition as pointless, because truth cannot be pursued in a practical sense. Therefore any so-called truth norm must be empty, due to the 'fact', as Davidson thinks, that truth cannot be a goal.31 Before we continue, we will need to consider some background on Davidson's theory of truth. Davidson's central assumption, which distinguishes him from many other truth theorists, is that he claims truth to be unanalyzable. Here is an illuminating passage from his article 'The Folly of Trying to Define Truth': For the most part, the concepts philosophers single out for attention, like truth, knowledge, belief, action, cause, the good and the right, are the most elementary concepts we have, concepts without which (I am inclined to say) we would have no concepts at all. Why then should we expect to be able to reduce these concepts definitionally to other concepts that are simpler, clearer and more basic? We should accept the fact that what makes these concepts so important n1ust also foreclose on the possibility of finding a foundation for them that reaches deeper into bedrock." (Davidson, 2005, p.20)

So, for Davidson, truth is a primitive concept. According to Davidson, the task of defining the "general concept of truth" is incomplete, because the concept of truth is not definable in terms of other, more fundamental concepts. The methodological consequence, if we accept Davidson's view, is that we cannot think of truth as having any properties. According to Davidson, the philosophical tradition was wrong in the assumption that truth could be analyzed to yield n10re fundamental concepts. 32 30 Engel (1999) discusses that issue in Davidson's philosophy. 31 See Davidson (2005, p. 7), where he affiliates himself with Rorty. 32 Such a primitivism about truth was also held by G. E. Moore and B. Russell (see

31

This conclusion does not affect Davidson's thinking that truth is important. In his John Dewey Lectures Davidson states: Truth is one of the clearest and most basic concepts we have, so it is fruitless to dream of eliminating it in favor of something simpler or more fundamental. (Davidson 1990, p. 314)

For Davidson, truth has an important, property-like function in human cognition and Davidson gives to it a central place in his theory of interpretation :33 Without a grasp of the concept of truth, not only language, but thought itself, is impossible. Truth is important, then, not because it is especially valuable or useful, though of course it may be on occasion, but because without the idea of truth we would not be thinking creatures, nor would we understand what it is for someone to be a thinking creature. (Davidson 2005, p. 16).

Davidson expresses two attitudes towards truth. First, truth is important, and second, talk of the "pursuit of truth" is pointless. It is difficult to make Davidson's two attitudes consistent with each other. On the one hand, truth is dispensable for inquiry, because truth, not being a value, cannot be a goal; yet, on the other, it is fundamental for the mental life of human beings. 34 We have a grasp of truth but that grasp is not enough to make truth one of our goals. How does Davidson defend his position? I contend that Davidson offers basically two interconnected arguments. The first argument looks as follows:

Sosa 2001, pp. 653-655). 33 In this context we should reconsider the central role of truth in Davidson's Principle of Charity and for our practice of belief attribution (see Baumann 1997). Ironically, Davidson has undercut the possibility of indicating whether the majority of beliefs in our belief system are really true. Sober (1994) provides a insightful discussion about Davidson's a priori argument for his majority claim and its shortcomings. 34 Rorty states in a similar manner: "You cannot have language without rationality, nor either without truth" (Rorty 1995, p. 284).

32

(D 1) If truth is an epistemic goal (or an aim of inquiry), then it is an epistemic concept. (D2) Truth is not an epistenlic concept (because it is not definable via particular properties).

(Conclusion 1): Truth is not an epistemic goal. (Conclusion 2): We aim at best at honest justification.

Davidson's argument falls short of explaining why the truth goal must presuppose an epistemic concept of truth. This assumption is only plausible if one accepts that the only characterization of a goal must be in terms of its feasibility. And, indeed, one of the motivations for speaking about epistemic theories of truth is to interpret the truth goal as a fixed goal, which then gives us an endpoint for the process of inquiry. 35 The first argument is thus not strong enough in itself to support the second conclusion. So we have to fall back on a second argument. Here, we look at the pattern of goal-nlaking in terms of intentions and action. 36 If we take this as the paradigm for all goal making and seeking, then Davidson would be right to say that we can only aim at justification because it is a realizable goal. Or, in other words, justification is directly related to action. (D3) If there is an epistemic goal of inquiry then it must be involved in action. (D4) Justification can be a goal of our action. (Conclusion) Justification is our epistemic goal.

Note that Davidson is vulnerable to the charge that the truth goal does not have to presuppose an epistemic concept of truth. Davidson's worry is that this would niean that the truth aim can't be neutral with respect to the concept of truth. Epistemologists in general attempt, however, do employ a neutral concept of truth. After all, epistemologists employ a very simple view of truth, which stems fronl the folk usage of the concept of "fact" (see for instance Bonjour 2002, pp. 32-38, Chisholm 1977, Ch. 2, Lehrer 2000, pp. 26-31). 36 Buekens (2007) has proposed such a reconstruction for Davidson.

35

33

The second argument is a reformulation of the argument of Rorty's which we discussed before. My criticism of Rorty applies as well to Davidson: On the one hand, Davidson's position rules out the possibility of the existence of epistemic goals. On the other hand, Davidson argues that there is an alternative epistemic goal. Yet it is unclear in what sense the actionrelated conception of justification can play an epistemic role. Rorty and Davidson are interested in such a concept, but they do not elaborate it. Thus far in this chapter we have presented Donald Davidson and Richard Rorty's objections to the notion that truth is a goal. The discussion so far has been critical, in that we are showing how certain confusions concerning the characteristics of epistemic goals have arisen, and what they lead to. The plan in the next part is to develop a positive account of how truth can be an epistemic goal. In the following I will give evidence that the concept of truth as a goal has a systematic role in epistemology.

4. THE TRUTH GOAL AS A REGULATIVE IDEAL In what sense can truth be an epistemic goal? The short answer is: truth is an epistemic goal because truth provides a regulative ideal. This idea accommodates the intuition that truth is in a minimal sense objective. 37 Just consider Aristotle's well-known definition of truth: To say that what is is not, or that what is not is, is false; but to say that what is is, and what is not is not, is true. (Aristotle, Metaphysics, 1011 b25)

Michael Lynch gives a less formal description: When we say something true, the world is as we say it is. And when we believe truly, the world is as we believe it to be. It is the way the world is that matters for truth, not what we believe about the world. (Lynch 2004, p. 11)

Truth is objective, because what is true is independent from what people take to be true. Our beliefs have to fit how the world is, and not the other way around. Basically, Lynch recommends that we look at what role truth

37 See Lynch (2004, pp. 10-12).

34

38

plays for our belief systems. He supposes the following: If I notice that I hold a false belief then I ought to give it up because I have missed the truth goal in that particular case. I can revise beliefs on the grounds that particular beliefs have led to falsehood. By discarding false beliefs we proceed towards the stage where all the beliefs in our belief system are 39 true, which is the ideal. The truth goal does not prescribe a definite state that we have to reach, or even give us a rigorous methodological content into which we can fit all inquiries. Generally, we can make progress by discarding one false belief after the other without having a fixed goal as a 40 guiding point. The truth goal is a goal of exclusion. But this is only one part of the truth goal. The other part is to believe the truth and only the truth. 4 \ The positive insight in conceiving truth as a regulative ideal is that it provides a means for asserting indirect control over our belief system. But this control depends not on direct action, but rather on reflection with regard to what epistemic standards we are using to comply with the exigencies of our particular inquiry - the unity of the whole process being bound up with the concept of an ideal. The advantage of understanding epistemic goals as epistemic ideals is that we sidestep the confusions related to the meaning of "goal", to which Rorty and Davidson have pointed us. Since the term 'goal' is often understood as 'personal goal', it becomes mysterious how truth can be a goal. Therefore it is clear that the truth goal cannot be literally a goal as, for instance, a goal in a game, at which we aim in order to win; but we can understand truth as literally an ideal. In contrast to Davidson and Rorty, we shall allow a technical usage of the term "goal" in the sense that we have certain epistemic ideals. By making the equivalence between goal and 38 See Lynch (2004, p. 12). This is a pivotal point for understanding truth as a theoretical goal. See also Blackburn (1980), Engel (2002, Ch. 5), Railton (1994, pp. 71-75), Owens (2003), Velleman (2000, Ch. 11). 39 We take it, by definition, that we think that all our beliefs are true. 40 This formulation is too one-sided, because there are two main norms that can be derived from the truth goal: (a) avoiding error, (b) holding truths. So defining the truth goal as a goal of exclusion overlooks the second norm: holding truths and actually pursuing truths. See Chapter 4 for further discussion of how those two norms are related. 41 In Chapter 3 and 4 I discuss the conditions on which we value truth.

35

regulative ideal, I think we can resolve all the confusions that cause Rorty and Davidson to discard truth as a goal of inquiry. In the following, I discuss how we can understand epistemic goals as epistemic ideals. 5. CHARACTERISTICS OF EPISTEMIC IDEALS Hilary Kornblith (2001) advocated the idea of epistemic ideals. It makes a good fit with his emphasis on epistemic obligations, which are entailed by those ideals. Kornblith presented his notion of an idealsderived set of epistemic obligations in contrast to Richard Feldman's account of epistemic obligation, which is based on what he calls a "role ought". There are, Feldman claims, certain conventions or obligations that determine roles. For instance, in teaching, according to Feldman, a teacher is committed to certain oughts related to his social role as a teacher. The role ought implies more than what the job description in a contract prescribes. It expresses the sum of what others in a given culture expect from a teacher as well. Our expectations of a teacher are relative to our idea of teaching excellence, even though in practice we might not witness any teacher who actually enacts all aspects of excellent teaching. The role ought of teaching excellently is not negated by those cases. Feldman wants us to construe the process of believing what we ought to believe in analogy to the teacher's case. Kornblith's objection to role oughts is: they are not necessarily tied to intrinsic goods. Human beings may well be pressed into social roles that harm them. Prostitutes are pressed into their role. The same applies to slaves or to child labor. These are all cases in which excellent performance may be an ideal, but it is an ideal of an oppressive social role. The difference between the oughts of social roles and ideals is that we attribute to the latter intrinsic value or a certain type of intrinsic goodness. Taking up the thread of intrinsic value, Kornblith modifies Feldman's concept of the role ought by inversing the relationship between the epistemic and the social role. Epistemic ideals can define social roles, but social roles cannot define epistemic ideals. From our discussion so far, we can see that there is a distinct advantage to shifting from the notion of practical goals - such as the

36

I

practical goal of getting to Larisa or Berlin - to the notion of a regulative ideal as a goal. The central advantage of the idea of epistemic ideals is that it captures the broader range of relationships between human beings and goals. Ideals are responsive to human capacities, because they play a guiding role outside of any expectation that, through some process of responsible inquiry, they'will be realized. Ideals can be appealing to human beings, but ideals are not constituted by personal acceptance: 42 Also, an ideal can be independent from personal interests in a way that practical goals can't be. That is, we can never subordinate the truth goal under personal or subjective preferences. In contrast to this, however, ideals need to reflect human capacities, because an ideal that cannot be appealing to any believer won't function as a guide to human belief-formation. This characteristic prevents us from introducing supererogatory epistemic ideals that only gods could fulfill. In this chapter, we have moved from examining the criticisms made about the truth goal as an endpoint of practical action to the goal as an ideal which we attempt to live up to. This construction of the truth goal obviates the criticisms of Davidson and Rorty, which, as we have shown, are directed towards another, shallower model of goals. Davidson and Rorty, who are both motivated by the desire to embed the epistemic value structure into our research projects by discarding the impossible goal of truth, end up involving themselves in a number of paralogism concerning objectivity, truth and justification. By making the move to the goal as ideal, we can dislniss the objection that truth is not an episten1ic goal while at the same time accommodating the insight that it is not a goal that is necessarily achieved. Quite the contrary, truth is a viable epistemic goal because it constitutes a regulative ideal for our belief system.

42 I follow closely Kornblith's (2001, p. 138) characterization of ideals.

37

3. The Value of Truth Once we have established that truth is an epistemic goal, and that epistelnic goals are ideals, we have laid the groundwork for investigating the epistelnic value of truth. Recall that neither side in the conflict between epistemic value monists and pluralists object to the value of truth as such. Rather, the dispute is about the primacy claim of epistenlic value monism. 43 For epistemic value pluralists, truth may well be very important; it is sitnply not fundamental. Both camps attempt to prove (or disprove) that truth is fundamental by clarifying the conditions under which we value truth. Our aim in this chapter is to clarify what, in general, the value of truth is relative to, if anything. With "the value of truth" I mean of course 44 the value of true beliefs. One thing is clear: We do not value truth unconditionally. Just think of trivial truths: Counting the leaves of the tree in front of my house cannot be seen as an inquiry that results in significant truth. The same applies to learning by heart the contents of a telephone book. We do not value trivial truths, because they do not have any relevance for us. Even if someone appreciates them, we see this person as being irresponsible for using his resources on such trivial truths instead of devoting himself to more substantial ones. True beliefs qua being true are not automatically interesting for. us. A common phrase used to disparage trivial truths is that they are a 'waste of time' - the idea behind this being that, given a finite life span and pressing practical and moral tasks, one has to economize and concentrate on truths that have a personal value in one's lifestyle. There is a second reason for not valuing truth unconditionally: other values, practical or moral ones, may override it. All things considered, we can sometimes be legitimately ignorant of truth, because the value of truth can be outweighed by moral or personal values. Suppose an intruder enters 43 Truth theorists, across all positions, are concerned with the value of truth. See for instance Horwich (2006) or Lynch (2004). Kvanvig (2003, Ch. 1) argues for the stance that truth has intrinsic value, while he defends epistemic value pluralism. 44 I will leave aside the question whether the possession of the concept of truth has any value for us.

39

my house and forces me to believe that he is not in the house by pointing a real gun to my head, then it is legitimate to believe whatever the intruder forces me to believe in order to avoid being shot. There are two tasks that are set before us in this chapter. The first is to counter the skeptic's doubt that the truth has any intrinsic value at all among the many things we do value. 45 Or, relatedly, the skeptic might argue that other things are promoting our personal goals. The short answer to the skeptic is that truth has value at least in a practical sense. 46 I will develop this response in more detail in the first part of this chapter. The second task poses the question whether or answer to the skeptic is all there is, or whether we value truth for something else besides its positive practical consequences. Is there a non-instrumental value of truth?47 It is clear that we want to know certain truths because we are curious about certain subjects. But does this allow us to infer that our belief that truth has intrinsic value is correct? In the following section we will proceed with the first task. 48

45 This is actually a variation of the multiple-goal argument. In short, it is stated that we cannot value truth because we already value so many other things. This argument has some force concerning the intrinsic value of truth. On the descriptive level we can actually wonder whether all persons in this world accept that truth has intrinsic value. 46 Truth is in general helpful to obtain what we want and what we desire. If we have any desires and we want to fulfill those desires we will rely on truth in order to do so (see Kornblith 1993, Zagzebski 2005). 47 I say "non-instrumental value" because I want to have a broad category into which both intrinsic and final value fits. See RabinowiczJRoennow-Rasmussen (1999), as well as their (2003) paper, for a detailed discussion of their thesis that all intrinsic values are final values. 48 Stich has introduced three different arguments against the value of truth. Two of them are relevant for our discussion. See Stich (1990, Ch. 6).

40

1. DOES TRUTH HA VE PRACTICAL VALUE? Truth is beneficial towards fulfilling our practical goals. William James captures this idea in his slogan "truth pays" .49 This view of the value of truth is not exclusively restricted to the pragmatist theory of truth. See for instance Horwich, who defends a deflationary theory of truth. He states: True belief facilitates successful behavior. (Horwich 1990, p. 46)

Loewer expresses the same idea in a slightly different terminology:

[...] acquiring true beliefs relevant to a decision problem prior to acting is valuable to the agent. (Loewer, 1993, p. 272)

We can sumn1arize the practical value of truth in a more general version: (VT) True belief acquired cost-free is always beneficial for human decisionmaking.

49 William James' theory of truth is notoriously misunderstood, because it is wrongly assumed that the pragmatists treat truth as equivalent to successful action. This connection puts their theory in a dubious light. Pragmatists themselves are responsible for such crude identifications, because we can find slogans like "Truth is what works", advertising their pragmatist theory of truth. This slogan can be seen as paradigmatic for the tension between the project of defining truth and the project of explaining the value of truth. From the historical point of view, especially William James seemed to encourage a crude instrumentalist understanding of his theory of truth. He was understood as saying that we define the truth of a belief by the goodness of its practical outcomes. James used sketchy slogans like"An idea is 'true' so long as to believe it is profitable to our lives", II We cannot reject any hypothesis if consequences useful to our life flow from it", "If the hypothesis of God works satisfactorily in the widest sense of the word, then it is true". Russell describes his understanding of James' theory as follows: "I find great intellectual difficulties in this doctrine. It assumes that a belief is 'true', since it is only after we have decided that the effects of the belief are good that we have the right to call it 'true'" (see Russell [1946]/1967, p. 771). See Fuhrmann (2006) for a wellelaborated reconstruction of James.

I use the term 'cost-free' since a responsible believer would reflect whether he wants to spend part of his resources in order to get a certain required information. 50 This means that we have to have some criteria by which to prioritize some data over other data relative to the resources we are investing in our inquiry, with the low bound on this scale being data that we attain almost effortlessly. Seeing the tree in the garden when I look out the window does not 'cost anything': I don't need to take resources from another inquiry to simply see what is in the garden. Different situations of inquiry will require different expenditures of time and resources. Suppose I desire to know the birth date of chancellor Adenauer, and the only way to look up this information is to go to the public library. I will then have to calculate the value of finding out the truth, here, as against the expense and time of taking the bus to the library. I will have to consider the value of doing this against other important tasks I have to do, some of which are more urgent to fulfill. When, for instance, a researcher calculates the size of the data set of a research project, she has to think about how many persons are to be asked to fill out her sample. The more persons are asked, the more representative is the data set, but the more persons are questioned, the more expensive will be the inquiry. Given finite resources, the researcher has to figure out how best to spend those resources that are available to her. 51 Epistemologists like to presume that the costs of an inquiry fade, so to speak, into the background, and have no bearing on the believer's belief system. Thus, they treat the believer's belief system as something that has been obtained without greater efforts. Since this is the general assumption, I, too, will take it for granted in the rest of the chapter. Additionally, I assume that a believer is able to distinguish relevant information from non-relevant. For the sake of simplicity I assume that human beings have the ability to distinguish between what is relevant to know for them and what is not. 52 50 In this chapter, I do not differentiate between true belief and information. Thus I presuppose that information contains truth as well. 5 I From this I don't want to draw the conclusion that every belief represents an opportunity cost. But in certain situations we have to weigh the costs for an inquiry. 52 We can indicate relevance very clearly by translating it into expected utility. We

42

In the next section we will see that (VT) is strong enough to respond to various counter-examples. These counter-examples are helpful for understanding the actual scope of (VT). I will argue that the following counter-examples rely on an incorrect understanding of (VT).

2. COUNTEREXAMPLES Stephen Stich (1990, p. 123) introduced the first example into the literature. He argues that there might be other doxastic policies that are optilnal in satisfying our desires than the truth-centered doxastic policy. 53 Doxastic policies are policies of forming our beliefs according to a central goal. Stich proposes that these policies can also be related to other goals. Here is his counter-example: Example 1: Harry is a business consultant who travels several days a week by airplane in order to visit his custonlers. One day he believes mistakenly that his flight departs at noon. When he arrives at the airport his flight has already left. So he missed it due to his false belief. But he is very fortunate, because the flight crashed and all passengers died in the crash.

Stich presents a case in which having a true belief is at first sight not favorable to satisfying our desires. In this case, Harry's goal of visiting customers would be rendered impossible by being killed in a plane crash. Thus, the goal was made more likely of achievement than less by the failure to remember the time correctly. I think that we should be clearer in the details than Stich. He construes his example such that the first desire (reaching the airplane on time) will be frustrated, but the more general assign a greater value to data that has a greater probability of being accurate than that data with less. In other words, the better confirmed an piece of information is, as we can express it, the greater is its expected utility and therefore its practical value. In the context of the calculation of expected value we take it for granted that "degrees of belief closer to the truth imply expected values closer to objective ones" (Horwich 1990, p. 46). 53 It is hard to reconstruct Stich's main argument from his writings because we can reconstruct it in three different ways (see Alston 1996, Chap. 8; Goldman 1999, Chap. 3; and Loewer 1993).

43

desire (staying alive) will be realized. Stich's general strategy is the following: With the counter-example in place, Stich thinks that we must deny that true beliefs have instrumental value in all possible cases. My response against Stich's objection is: Stich has the burden of proof, because he wants to install an alternative doxastic policy. Even if Stich can indicate such an alternative doxastic policy, he has to prove that this alternative policy is better with regard to desire-fulfillment than the truth-centered policy. Here lies the defect of Stich's account: how can this policy be better if it does not enable us to make reliable predictions? Because without prediction, practical success isn't possible. Stich faces an unwelcome consequence of his example: What should Harry do in the future if he wants to follow the policy favored by Stich? In fact, there is nothing Harry can do. If we made a general rule from this incident, it would be: If you want to avoid dying in an airplane crash, you should rely on false beliefs from now on. But we know this isn't true. The problem Stich faces is that when we want to learn more about a certain subject, our object is to know the truth about it. For Harry that would mean, perhaps, learning the ratio of airplane crashes to safe flights, because as a business consultant Harry has to travel by airplane very frequently. Yet, this is not information that is going to do much for Harry, in as much as he is a passenger. Perhaps it would be prudent for Harry not to keep too tight a schedule of visits to his clients, since both missing the plane and the plane crashing have shown that all kinds of accidents can, at the very least, delay his client visits. But he can't indefinitely delay the visits. In sum, Stich does not have a sensible recommendation that is optimal to fulfilling Harry's desire to survive and to do his job. It would, in fact, be a superstitious belief in Harry to believe that his missing the flight caused the flight to crash. This would, indeed, be both a false belief and a harmful one.

44

3. FALSE BELIEF WITH HIGH UTILITY The second example was presented by Nicholas Rescher (1975). Rescher aims to show that false beliefs can have instrumental value. Example 3: Mr. Smith believes that he has a disease X. He therefore takes medication Y and believes that medication Y will help him against disease X. .However, his belief that he has disease X is n1istaken, and he actually has disease Z without knowing it. Incidentally medication Y helps against disease Z. By taking the medicine Y, Mr. Smith restores his health fully.

Rescher supports a different argunlent than Stich. Stich attempted to argue that there nlight be alternative doxastic policies one could adopt which would be optimal to satisfy one's desires. Rescher, by contrast, aims to shed doubt on the connection between utility and true belief. He concludes that, from the point of view of realizing practical value, we cannot rely on any absolute distinction between false and true belief; the implication being that true beliefs do not have as such a practical value. In his example, there is a false belief that has high utility. Notice, however, that Rescher sets his example up so that this particular false belief stands isolated from the rest of the nlany other truths in Mr. Snlith's belief system. How is this possible? What are the background beliefs Mr. Smith takes for granted when he decides to take the medicine? We nlight think of the following implicit background belief: "Take medication Y because it helps to fight your disease" or "Taking the right medicine is helpful" or "Medication Y treats whatever disease I have".54 Those propositions are true, and they are logically implied by Mr. Smith's false belief. Thus, in this example, the false belief does not have high utility by itself. Its high utility stems from other, true beliefs of Mr. Smith's belief system. But the example so far is not described in all the necessary details 55 for us to use it as a guide to solve this issue.

54 That response goes back to Loewer (see Loewer 1980, p. 374). 55 How is it possible, for instance, to track down the utility of one false belief? Or, in

45

Let us consider what would be the consequences if Rescher's argument were right. What would Mr. Smith's recommendation be if false belief had the same practical consequences as true belief? If, that is, he recommended medication Y to someone who truly had disease X, wouldn't that lead to ruinous consequences? If so, having a true belief about medication Y accrues a utility that the false belief can't match. All of which points to the question: how can the fulfillment of our desires work if we obtain by luck what we actually have attained? Just consider if it would be right for us to advise someone to rely on false beliefs because of their alleged practical utility.

4. VALUE OF MEANS-END BELIEFS The third example stems from Chase Wrenn (2001). He attempts to show that the scope of (VT) is limited. He does not doubt the validity of (VT), but he thinks that it cannot be used to defend the value of truth in general, because only belief with a certain structure has actual practical value. Example 3: The goal of a baseball catcher is to catch the ball before it hits the ground. A successful strategy is for the catcher to run towards the ball at a certain angle in order to shorten the distance between himself and the possible landing point of the ball. This strategy helps saving time in comparison to strategy B. Strategy B would require the catcher to first calculate where the landing point is and only thereafter to run for the ball. 56

According to Wrenn, (VT) is in trouble because it explicitly relies on means-end beliefs for human decision-making. If a true belief should have any practical value, this value arises only if there is in an means-end belief formed. In the baseball case there is no time to form means-end beliefs. There are two points that help us construct a defense in response to other words, how can we measure the degree of utility of a false belief? 56 Wrenn construes the example in such a way that the player does not have the time to form a belief, because in order to stay in the game he must react to the current play.

46

Wrenn. First, his counterexample does not affect (VT) because his case describes a situation where the agent cannot acquire a cost-free belief. Second, it is not clear whether the value of truth comes exclusively by way of identifying n1eans-ends beliefs. Let us proceed in by looking at these points in succession. The effort of calculating, as strategy B requires, takes up so much valuable time in the playing situation that it would make it adoption hazardous - it would lead to more failures than following the first strategy. Therefore the catcher cannot accept this strategy. But we can imagine that ·the catcher actually does take the tin1e to reflect where the ball will land. As with Rescher's example, Wrenn's example gains by singling a situation out from the background in which knowledge is acquired. We cannot exclude the possibility, for instance, that before the game the catcher prepares to run according to the hitting habits of each hitter. Why should the catcher not appreciate each information about the hitter he or she faces in the next game? The coach might inform the catchers about the strategies of each hitter. But more importantly, Wrenn introduces a case where the acquisition of certain beliefs is not cost-free. Therefore, his case is not a counterexample to (VT), because it does not apply to (VT). With Wrenn's example in mind, however, we can think about a more central point concerning the actual scope of (VT): Is (VT) too restrictive with respect to the condition on which we value truth? I would argue that (VT) is not too restrictive, nor that we value true belief beyond its actual relevance, based on two reasons. First, contra Wrenn, it misconstrues (VT) to think that it entails that we only value means-end beliefs exclusively. Therefore (VT) is not too restrictive. Closely connected to the first reason is the more interesting argument that we value all our true beliefs because there is no strict correlation between successful action and particular true beliefs. In the following, I will explain both reasons in more detail. Means-end beliefs have the form "means M will help person P to achieve end E". Maxim (VT), Wrenn thinks, requires the catcher to form his beliefs prior to his action, e.g., "I must run to the right corner of the field in order to catch the ball before it will hit the ground". And to repeat: The catcher Inust follow a far more efficient strategy with respect to the playing situation. It is unclear, however, why only explicit means-end

47

beliefs should be able to deliver the practical value. While it is true that the catcher does not have the time to calculate and determine where the ball will land, Wrenn assumes too much when he supposes that (VT) requires the catcher to form means-end beliefs. Goldman (1999, p. 74) states that ifour belief "means M will promote my end E" is true, and if the agent accepts means M to promote E, then the agent will attain his goal. 5 ) This is why we might think that (VT) requires us in every situation to value means-end beliefs in particular, because those beliefs so often prove explicitly helpful in achieving our goals. If , Goldman were right, we would only value means-end beliefs (while presumably reducing the set of means-end beliefs to the set of means-end beliefs relevant for action). Can this view possibly be right? I think this misunderstands the motivation for forming means-end beliefs. We can do so with the idea that they are generally conducive to goal realization, but we don't have to form beliefs which are directly effective to attain the corresponding goals. Consider the following example: I form the belief "traffic lights and crosswalks are good for crossing the street safely", so if I want to cross the street safely I will make use of traffic lights and crosswalks. According to Goldman and Wrenn, I must value this particular belief because it is helpful for realizing the goal of crossing the street. But there is a problem with clinging to this belief as my only means to cross the street: I can cross the street safely wherever I want, as long as there is no car passing by. Crossing-the-street-wherever-I-want-belief does not apply to streets in calm neighborhoods, because presumably in those neighborhoods I can safely cross the street without any traffic lights or crosswalks. In that situation we have a means-end belief which is superfluous concerning my practical decision making, but that might still have practical value in other situations. The proposal that we particularly value means-end beliefs is too restrictive if it takes all the situations that can be associated with the goal and applies a one-size-fits-all ends-means belief. I would say that (VT) is purposefully vague as to which particular true belief we should adhere in order to cash out its practical value. It does 57 I do not find this description plausible, because even if the agent is well prepared some shortcoming in the world can prevent the agent from attaining his goal. Or we might think of a case where the agent's knowledge is incomplete, even though she is well prepared and well informed about how to attain goal E.

48

not imply that anyone n1eans-end belief lTIonopolizes the goal, but rather suggests a best practice, given that certain conditions are met. Thus, (VT) reflects the fact that certain facts will possibly significant for practical decisions and does not mean that we can indicate in every situation which belief in my belief systen1 is to be followed for practical success. Consider a friend who informs me that his mother's company is looking for someone who has the academic degree and specialization I have. Having this tip is very helpful, because thereby I know where I can apply for an open job for which I qualify. I value this particular information because I value having a career opportunity. But the tip doesn't lead me directly to the position. Obviously, in order to obtain the job, I still will have to write a good application and go through the application process. Assume, though, that I do all this and obtain the job, which counts as practical success. In retrospect I can reflect on which of my beliefs were especially helpful. However, given the limits of my knowledge, especially in a situation that depends upon the motives of other people, it is difficult to indicate all the helpful true beliefs which are responsible for the given success. For instance, perhaps a better qual ified candidate decided not to apply for the job. This is a fact I may well never know anything about. I think that at a certain level, our belief system as a whole, rather than some particular beliefs, can be relied on to help us make well-informed decisions for making us take the time and preparation to hand in an application that truly reflects the level of our skills. I think that the previous examples show that Goldn1an and Wrenn are mistaken in their reading of (VT). Their discussion, however, has led us to a more crucial question: What value do truths have which are clearly not relevant for decision-making or for our personal goal-fulfillment? Barry Loewer has indicated the limited scope of the argument for the practical value of truth: It does not show that truth in general has any value. It applies only to beliefs concerning the states of nature characterizing one's decision problems and only when seeking the truth will result in learning for certain which state of nature obtains. So far I have not shown that there is reason to value the truth of any other propositions. Also, the argument does not show that one has an overriding reason to seek the truth. (Loewer 1993, p. 275)

49

Loewer is right: (VT) does not establish a value for all the propositions that we hold and that we should value. But once the believer has accepted that truth is the fundamental epistemic value, certain personal values are overridden by the value of truth. Therefore the value of truth has overriding force in a limited sense. According to a strict reading of (VT), beliefs which do not have practical value do not have any value at all. For this reason there seems to exist a va~gap between (a) the set of truths that we value already due exclusively to their being true and (b) the set of truths which we value because of their practical output. Presumably the former is larger than the latter set. In the following section I will develop a reading of (VT) that is capable of accounting for this value gap. The holistic assumption that showed us why the Goldman-Wrenn view is not plausible also provides a satisfactory explanation of why we value true beliefs without actual practical value.

5. KNOWING ABOUT RELEVANCE IN THE FUTURE In this section, I will present a modification of (VT) that allows us to exclude the strict reading. Is there a criterion of relevance that tells us which truths are clearly relevant for our decisions? And even more important, how do we know whether certain of our true beliefs will be relevant for future decisions, when we do not know all the plans we will have in the future? It would seem to follow that we cannot say which beliefs are relevant for supporting our future plans and goals. The basic idea I shall propose to close the value gap is the following: We care about many truths over and above the ones that have direct relevance for our decision-making in anyone situation at anyone time because we do not know for sure which of the beliefs in our belief system might be helpful for future decisions. So, in order to be prepared effectively, it is helpful to have as many accurate and relevant beliefs as possible. This position relies on the fact that we cannot determine in advance the relevance of each of the beliefs we have. It is clear that we treat beliefs which are clearly relevant (in relation to our current lifestyle)

50

differently from beliefs without obvious relevance. From this it does not follow that we may be totally careless with regard to all the beliefs whose future relevance we do not yet know. One reason is that we are aware that we can be mistaken about future relevance relations. Another reason is that since we cannot isolate our beliefs from each other, it is hard to single out which particular true belief was responsible for a certain practical success. One might derive from this an objection against the practical value of truth.. The objection would be the following: If we cannot indicate the practical relevance of certain true beliefs, how can we argue for the view that we value true belief because of its practical relevance? But I am not arguing that we cannot in general relate particular true beliefs with particular outcomes of practical success. Quite the contrary: In the jobapplication case, knowing the fact that there is an open position in the firm of my friend's mother is clearly helpful because it occasions me to write my application and to apply. I have claimed, however, that it is impossible to make a spread sheet that matches all of our true beliefs in our belief system with their practical relevance because we cannot single out our beliefs from the belief system. But still, there is a certain subset of beliefs which we could clearly match up to their practical relevance. The rest of my belief system is still valuable, as I take it, because those true beliefs might be useful for future plans and decision-making. 58 Furthermore, we would not throw out a true belief just because we could not identify its practical pay-off. 'rhis result is actually positive, because we can establish the following requirement concerning the pursuit of truth: Try to have at all times an accurate and comprehensive belief system, because you can never predict the complete contour of all situations for which you imagine your practical goals, and hence, you can't predict which items in your belief system might turn out to be profitable. "Accurate" means that we should attempt to avoid as many errors as possible, and "comprehensive", that we should avoid holding defective beliefs, because such a belief system would keep us from being practically successfu1. 59 58 I take the holistic assumption about my belief system for granted. We cannot put our beliefs to the test one by one. 59 In Chapter 5 I will argue against the idea that certain beliefs are minor from an

51

There is a second reason for being worried about a strict reading of the practical value of truth for it too closely identifies truth and utility, and thus weakens the idea that truth has a fundamental epistemic value. As long as there is no overriding personal or moral reason for not seeking the truth, we like to think that the truth itself provides us with a sufficient reason to seek for it. So far, we have established the first task: Truth is practically valuable. Does truth also have intrinsic value? As we have seen in our discussion of the strict reading of (VT), we value truths even if their direct practical value is not presently evident. This indicates where the intuition about the intrinsic value of truth comes from. This will lead us, in the following section, to use a thought-experiment of Robert Nozick's to discuss why we think that we value truth for non-instrumental reasons. 60 One could also be dissatisfied per se with the conclusions we arrived at in responding to the first explanation task because even having established the practical value of truth, one might argue that we are still far away from having stated something relevant about the epistemic value of truth. 6. OUR PREFERENCE FOR TRUTH Here is Nozick's thought-experiment to show that the value of truth is intrinsic. 61 Consider the following choice between two different worlds: The first world is our real world, where we can be wrong and our desires can be frustrated, because the world can be different from what we thought it was. So, in the first world "truth hurts". In the second world all our desires are fulfilled by a machine, so whatever I desire or dream of in the second world will become true. As a result of this our dreams or wishes absolute standpoint, because there is no tolerance principle that regulates what is a minor false belief and when we are allowed to hold one. So, getting ahead of ourselves a bit and taking into account this factor, we value having many truths, because one day those truths might be important by indicating how to attain some goal we do not yet have. 60 I use the term "non-instrumental" because I think it is more neutral than the term "intrinsic". 6 I Lynch (2004, p. 15) also considers a version of Nozick's machine example. See Lynch (2004, p. 18) for a variation which originally goes back to Russell.

52

and the world around us agree perfectly. In comparison, the two worlds differ in the direction of fit. In the first world our beliefs and our desires must be adapted to the world, whereas the second world is a one-to-one realization of our desires. What would be our preference? We prefer world lover world 2, I take it, because World 1 represents 'real life'. In our choice of world 1 we show that we do not desire truth simply for practical reasons, because world 2 is the place where the satisfaction of our desires works perfectly to make our beliefs true. World 1 corresponds to our notion that truth is a property of the real world that we can't change; and we would prefer to keep this arrangement over the desire-satisfying dream-made-true-world. But if we spurn the second world because we want to have truth, then it must be that at some level we prefer truth on non-instrumental grounds. This clearly supports our intuition of the intrinsic value of truth. 62 The view about the intrinsic value of truth is an optimistic one, which is contested by the reserved view in the literature. 63 The problem we face now is how to justify this intuition. We are looking for a second explanation concerning the value of truth. It consists in explaining why the view that truth has intrinsic value is the correct view, assuming that Nozick's thought-experiment nicely captures our intuition that the alternative of a world in which the truth doesn't have intrinsic value has something repugnant about it. We can exclude the idea that by granting it intrinsic value, we value truth unconditionally; because all things considered, there are other, more important goods. Any argument that proves the view of truth's intrinsic value correct 62 It would be interesting, and would perhaps complicate our response, if we could forget the fact that World 2 is a wish-fulfillment dream world once we are born into this world, or whether we would be aware of the fact that the second world is only a dream world. 63 For instance, Ernest Sosa says: "Truth mayor may not be intrinsically valuable absolutely, who knows?" (Sosa 2007, p. 72). I think we can read out of this quotation an insecurity about how to proceed with the concept of intrinsic value. It would be good to know whether he denies the existence of intrinsic values in general or is just skeptical about whether truth could be seen as having intrinsic value. Or does he think that the whole project is not fruitful?

53

has to introduce a so-called pro-attitude,64 like care for truth, curiosity for truth, desire for truth, which connects the believer with truth in a noninstrumental way. The practical value of truth derives from the value of the desired object. The intrinsic value of truth is constituted by the believer's directly wanting, desiring, caring for the truth. What are the arguments which prove that truth is intrinsically valuable? There are three arguments: (1) the desire-for-truth argument, (2) the curiosity argument, (3) the argument that truth is a constitutive good. Let us proceed in the given order. 7. THE DESIRE ARGUMENT For Horwich, truth is an intrinsic good, as justice or beauty is. The intrinsic value of truth comes from our having a desire for truth. He states: (H) It is desirable to believe what is true and only what is true. 65

The aim of (H) is to express a universal requirement to seek truth. Horwich hopes that with establishing the intrinsic value of truth we will obtain "epistemological import", as well as clear explanation why lying is wrong. Horwich thinks that from the intrinsic value of truth we can infer more about the issue of lying. I suspect Horwich uses the term "value of truth" ambiguously, because the case of lying concerns the moral value of truth and not primarily the epistemic value of truth. I will not concentrate on what epistemological import Horwich believes can be obtained from his argument (even though that would be clearly interesting). How can we understand (H)? First, I understand (H) as a normative claim, since the corresponding factive claim is not adequate. Many people actually do not have a desire to attain truth in various situations, and from 64 Here I follow a tradition in value theory which goes back to Brentano. In that tradition "the concept of intrinsic goodness is analyzable in terms of the worthiness of some attitude" (see Roennow-Rasmussen/Zimmermann 2005, p. XIX). 65 See Horwich (2006, p. 347). He thinks about a further elaboration of (H), but in the end he claims that (H) can be stated in its general form (see Horwich 2006, p. 349).

54

the existence of those persons who actually do have a desire for truth in all situations we cannot infer the existence of a general desire for truth. Having a desire for truth is the pro-attitude that can explain why we intrinsically value truth. It sounds reasonable that we ought to desire truth. But how shall we do that according to Horwich? Consider the following example, which shows that even if we desire holding truth in lnany cases we might still desire being ignorant in other cases. Suppose there is an old man who has discovered the first faint evidence of having cancer. 66 But this old man has decided to remain ignorant about his health state, because he thinks, all things considered, that it is better for him not to know. He subordinates truth, here, to the desire to live an undisturbed life, which he takes to be one of the highest goods. This good overrides the value of being informed properly about his health state. This does not imply that the old man is ignorant toward truths in general. Suppose the old man was in general interested in the pursuit of truth, but he opts out with respect to his health. Let us imagine that the old man is a cancer specialist himself, and as long as he was responsible for his patients, he was willing to put in any effort to know the truth about the state of health of each patient. In respect to his profession, the old man followed (H), in order to help his patients. But concerning his personal health state he prefers to be ignorant and he accepts the consequences. The old man case poses a problen1 for the desire-for-truth argument in that it sets up a conflict of values. Even though the old man generally accepts the value of truth, there is an area where he opts out because there is a stronger desire. How can we judge the case appropriately and save the desire for truth argument? We might simply call the old man irresponsible for adopting an attitude of willful ignorance, especially in the case where the old man could have saved his life by checking earlier on whether he had cancer or not. I reject this strategy because I think that it cannot teach us anything about the intrinsic value of truth. Another strategy is to check whether the old n1an had an overriding reason that justified his ignorance. 66 Papineau has introduced this illustrative case (see Papineau 2003, p. 14).

55

We can rethink the old man's situation along the following lines: We can understand the phrase "it is desirable" as an ought-to-desire that is not strong enough to influence the actual decisions of the old man. The question then is what we should infer from the decision to override the value of truth. Of course, it would be different if the old man knew the truth and actively lied to himself - here, he doesn't know the truth yet, and shuts down the inquiry that would allow him to know it. There are two interesting results we can draw from that observation. First, it is a problem for the epistemic imperative, 'ought-to-desire', that it is not effectual in all possible decision processes. This means that we have to revise the oughtto-desire because of its alleged weakness. But there is something wrong with such a strong result, which would seem to effect all of our oughts in as much as no ought could regulate all possible decisions. Even moral oughts or legal oughts are not strong enough to regulate all decision processes. Second, we can argue that the ought-to-desire is not affected by the old man's case because the old man actually did have a good overriding reason. So we value truth as long as its value is not overridden by more important personal values with which it comes into conflict. Still, the argument doesn't say that the truth gains its value in relation to these other values - it still preserves its intrinsic value, then, despite the fact that it is sometimes overridden. Actually we need cases as the old man's in order to point out again that truth has a standing value that remains as long as it is not overridden by another value. Moral values are clearly legitimately overriding because moral values do override the value oftruth.67 However, the interesting question is whether this standing value of truth in the hierarchy of desires is evidence that the truth has intrinsic value. Remember, here, our previous claim that we have to adduce some kind of pro-attitude to understand the non-instrumental nature of the truth's value. Thus, I would argue that the value we discover in the old man's case, which stems from desire, isn't the correct basis for intrinsic value. We still have to find the right pro-attitude that matches our general assumptions 67 Jonathan Kvanvig for instance has proposed that I can trade off the value of truth for the value of staying alive, given the situation that a mugger forces me at gunpoint to believe something false.

56

about the value of truth. The desire argument fails to fulfill this task, because it has inconsistent presuppositions: On the one hand it presupposes that we need to have a desire for truth, but on the other hand it appears that this desire should be taken as a general desire in such a way that we value truth unconditionally. Here are two further examples which show that desire might not be the right pro-attitude. First example: hnage a wife who has learned that her husband was unfaithful. Would we say that she has desired to know this truth? I would think that it is better for her to know the truth now, but we would not say that she actually desired to know it. In what sense is it better for the wife to desire the truth (if we assume that the wife doesn't find it desireable that her husband be unfaithful)? In this case it is still good to know the truth, but it is unclear in which value category to put the case. The central point I want to infer from this case is that the value of truth does not go hand in hand w'ith a desire for the truth in all cases. Let us proceed to the second example, which I think is an even clearer one. The second example is a nledical case again. 68 Suppose you consult a doctor because you are afraid that you might have a terrible disease. And let us say that there are two options in this case: One, you have the dreadful disease and there is no treatment for you. Second, you do not have it and no treatlnent is necessary. The doctor makes a highly reliable test, because he is curious to know the outconle of the test. What is your attitude concerning the results of the test? You desire that the test comes out negative. There is, however, no utility connected with that outcome of the test. And the positive outcome of the test does not have any utility either, because there is nothing you can do about this dreadful disease. You will not take the outcome of the test to be a valuable thing, but still you desire the test to be negative. This example shows that value and desirability fall clearly apart. Therefore we should not assume that if there is value then there must be a desire for that value. Let us proceed to the second argument, the one about curiosity.

68 Gilbert Harman (2000, pp. 146-147) discussed this enlightening case.

57

8. THE CURIOSITY ARGUMENT Curiosity constitutes a positive attitude toward truth, where we are not interested in the practical pay-off of the truths which we are pursuing. At first sight, the curiosity argument looks convincing. 69 Curiosity is a state of being interested without paying attention to the utility of what we are curious about. If we assume that a person who is inquiring for a truth is curious, we can explain why he or she values the truth they hold as a result of the inquiry. But how helpful is the curiosity argument? As we have seen, prima facie, curiosity gives us the right attitude. Curiosity, however, does too much: it can justify the pursuit of totally trivial truths, because curiosity can lead us to insignificant results. We can be curious about many things (the lives of pop stars, trivia about athletes, the mobile telephone market in Africa) which might be totally useless in our lives. Or even worse: we can be curious about a subject without making any attempt to attain true beliefs about it. I can be curious about playing an instrument, but this curiosity is such a faint attitude that it will never motivate me to take music lessons. The central problem for the argument from curiosity is the following: if curiosity were the paradigm attitude needed to explain the value truth has for us, then trivial truths would also be valued by us. Thus trivial truths would automatically obtain intrinsic value in the moment they became the object of someone's curiosity. In sum, if curiosity is the right attitude there are no grounds left to blame anyone for collecting insignificant truths, or satisfying one's idle curiosity concerning trivial news. If a schoolboy is curious to count the leaves of the tree in the garden instead of doing his mathematics homework, we would not consider the curiosity an excuse. We would blame the boy for having wasted his time without learning something cognitively valuable. One response to this problem is to distinguish between different kinds of curiosity with respect to the goal of finding intellectually interesting things. The non-idle kind of curiosity is directed at intellectually 69 This line of thought is defended by many epistemologists. See, for instance, Alston (2005, p. 31), Goldman (1999, p. 3) and Lynch (2004, pp. 15-16).

58

i

interesting questions in contrast to the idle kind, which pursues trivial truths or truths that are none of the inquirer's business. How helpful is this distinction with respect to the trivial-truths objection? With curiosity as the main driving attitude in place we cannot establish any requirement to pursue significant truths. The argument puts us in a difficult situation: If we accept curiosity as the right attitude toward truth with which to explain how truth can have intrinsic value then we have to accept that trivial truths can be the goal of our curiosity as well. This would go against my claim that a good theory of the non-instrumental value of truth should also be able to explain why inquiry after pointless or totally useless truths is invidious. I think the main lesson we derive from the curiosity argument is related to the last two examples, which work against the desire argument. Note that we can generalize from the two examples and shed doubt on the possibility of explaining the value of truth by a pro-attitude in general, in as much as it must be son1ething extraneous to the attitude that explains the division between trivial and significant truths. If we can transfer our arguments fron1 desire cases to cases involving curiosity then it would be clear that pro-attitudes are not generally constitutive for the value of truth. 9. THE ARGUMENT THAT TRUTH IS A CONSTITUTIVE GOOD There is another proposal, which states that we care for truth because it is a constitutive good. For Lynch, truth is part of a constitutive good, and he thinks that is why we care for truth in a non-instrumental way. Lynch describes his approach as follows: Asking why truth is worth caring about will tell us what makes truth good; it will help us understand why truth matters. (Lynch 2004, p. 120)

So, our care for truth provides the reason for thinking that truth has noninstrumental value. Lynch's hope is to pinpoint personal reasons for caring about truth as the source of its non-instrumental value. His proposal differs [roln the curiosity argument and the desire argument, because here the nature of the relation between the believer's attitude and truth is more

59

complex. The basic idea for Lynch's argument is that "constituent goods are extrinsic but non-instrumental goods" (Lynch 2005, p. 334). Lynch writes: caring about truth is deeply connected to happiness. This is because our lives go better when they are lived authentically and with integrity. And authenticity and integrity are both, in different ways, connected to truth. (Lynch 2004, p. 120)

Lynch's "strategy is to show that truth and our caring about it, much like love, but unlike, say money, are constituent parts of human flourishing" (Lynch, 2005, p. 334). It appears that truth should be put into the vicinity of moral intrinsic goods (if we accept taking human flourishing in the domain of moral values), which will show us its family resemblance to these goods. Presumably, the list of constituents of human flourishing and happiness is long. I want to focus on two constituents that Lynch advocates: the first is integrity, the second sincerity. I shall summarize Lynch's arguments in a joint argument. 70 The argument is intended to illuminate how truth is a good because it is a constituent for integrity and sincerity: (Ll) If intellectual integrity (and sincerity) is a constitutive good, then so is caring about truth as such. (L2) Intellectual integrity (and sincerity) is a constitutive good. (L3) Therefore, caring about truth as such is a constitutive good. (L4) If caring about truth as such is a constitutive good, then truth as such is worth caring about for its own sake. (Therefore), truth as such is worth caring about for its own sake. (Lynch 2004, p. 136)

70 Lynch introduces the arguments separately. But since both have the same structure, I evaluate them as one (see Lynch 2004, p. 157, for the argument about sincerity). Lynch also provides a third argument along lines similar to the arguments I discuss here (see Lynch 2004, p. 127), as well as a fourth separate argument about the social value of truth in law-making (see Lynch 2004, Ch. 10).

60

In short, the argument states that caring for truth as intrinsic value is an essential element of integrity and sincerity as well. The crucial question for us here is: Does Lynch succeed in convincing us that this is the correct path to establishing truth as an intrinsic value? His ainl is to indicate the source of the non-instrumental value of truth, and he starts by highlighting the connection between truth and what should nlatter most to us. There is no reason to doubt that truth plays an important role in attaining happiness, sincerity and integrity. But the question is whether attaining happiness, sincerity and integrity presuppose that we value truth intrinsically. This claim looks prima facie hard to defend, because we can also attain happiness while just valuing truth instnlmentally. We can even attain faked and short-tenn happiness by accepting falsehoods that deceive us about how the world is. I can be happy because I am deceiving myself that I will go out on Friday, while the truth is that I cannot afford going out. Lynch's argulnent seems vulnerable to the same objections as we brought to the desire argument: we care for n1any things and prima facie it seems that our care for holding true beliefs is in conflict with other things we care about. 71 Lynch recognizes this problem: "So even if we do value truth for its own sake, that is irrelevant to our everyday concerns as human beings" (Lynch 2004, p. 120). In everyday life we do not care about many truths. Lynch replies to this objection in the following way: I may not desire to believe what I do in fact believe~ or something I could without that showing that I don't care about believing all and only what is true, or the truth as such. I can care~ for example~ about living a virtuous life without having a separate desire~ for each and every morally right action, to do that action. So the fact that I don't desire to know how many threads there are in my shirt doesn't mean I don't care about believing the truth as such. (Lynch believe~

2005~p.332)

What is Lynch's actual reply? As we have seen before in our discussion of (the desire argument, the claim is that we care for truth as such, even though 71 Kornblith has nicely observed that there is no general desire for truth. Given all the things we can desire it looks as if the desire for truth is one desire among many others.

(it

in everyday life we also care about many other things. So Lynch states that' the intrinsic value of truth is untouched by the fact that we care for other i things too. As I have stated beforehand, I agree that truth has a standing value, but I cannot see how this would provide evidence for the view that i truth has intrinsic value. Thus, whatever it is what makes truth an intrinsic value seems to lie outside of it. So far, we have seen that the curiosity argument, the desire argument and the argument that truth is a constitutive good cannot explain why the thesis of the intrinsic value of truth is correct. 72 All three arguments have one deficit in common: Truth is an intrinsic good concerning the epistemic value domain, but all three arguments attempt to locate the intrinsic value of truth outside of the epistemic value domain. So all three arguments do not pay sufficient attention to the fact that the concept of intrinsic value is a relative concept. We can indicate what is an intrinsic value only given a certain value domain or domain of given candidates for goods. Concerning all value domains or all possible goods it seems to be impossible to indicate what is intrinsically valuable. In sum, it is clear why we think that truth has a standing value: it derives primarily from its practical value, which has been proven in the past and the present, and which we can inductively believe will continue to be the case in the future. So far I have not said anything in detail about whether we value believing truths more than avoiding error. Do we value both equally? This will be the central topic of the next chapter.

72 One option I have not considered so far is what I call primitivism about the value of truth. By "primitivism" I mean that we maintain the existence of an intrinsic value of truth, but are unable to analyze further what this intrinsic value consists in. We might justify this view by stating that this unanalyzability is an indication that we have reached the ultimate value foundation. Methodologically this primitivism is not very attractive, because it leaves unanalyzed what we want to explain.

62

i

4. Requirements of the Truth Goal In Chapter 2 we characterized the truth goal as a regulative ideal. In that context, the truth goal was exclusively defined by the exclusion of error. This assumption requires further elaboration because it is only half of the story. The exclusion of error alone is too negative a condition, because this requirement would prevent us from the trial and error that goes into expanding our belief system and thus from believing many truths. In addition, it is unclear what epistemic value the avoidance of error actually has. The aim of this chapter is to give a definite answer to the question of how the truth goal can be appropriately expressed by showing which defensible requirements can be inferred from the truth goal. I will argue that the traditional view, which divides the truth goal into two goals, that of believing all true beliefs and that of avoiding false beliefs, fails to provide a defensible requirement. It fails to provide a motive internal to the truth goal that would justify its division; and it does not recognize that normal believers are looking for significant truths. Before I present my objections, I will lTIOre fully characterize what I take to be the traditional view. After that I discuss what I call the "objection of the overarching goal", which Wayne Riggs (2003a) has presented against epistemic value monism.

1. THE TRADITIONAL VIEW Alvin Goldn1an writes about the value of believing the truth: Whatever the exact threshold for belief may be (greater than 0.50), believing a truth carries more veristic value than suspension ofjudgment; and suspension of judgment carries more veristic value than disbelief. [...] Thus, the intuitive rank-ordering of veristic value is confirmed: True belief is preferable to suspension of judgment, which is preferable to false belief (error). (Goldman 2001, p. 36)

If Goldman is right about the ordering among believing, disbelieving and

63

belief suspension, then the avoidance of error cannot be the only requirement, because believing the truth has a higher epistemic value than the suspension of belief. Additionally, Goldman's view stands in opposition to what I call the traditional view. The traditional view commingles two opposing requirements: (a) believing (all) truths/ 3 and (b) avoiding error. In sum, it attempts to avoid two extremes, "believing nothing" and "believing anything". The most noted consequence of the traditional view is that we have to trade off believing truth with avoidance of error in certain situations. William James' version of the traditional view entails precisely this dual requirement. James states: We must know the truth; and we must avoid error-these are our first and great commandments as would-be knowers; but they are not two ways of stating an identical commandment, they are two separable laws. (James 1969, pp. 203204; emphasis in original)74

James is right to separate the requirement of believing the truth from the requirement of avoiding error. Those requirements are clearly distinct, because avoiding error is not another way of attaining truth. It is important to note that James' own discussion is directed at the question of truth and certainty.75 Keith Lehrer characterizes the traditional view in greater detail. His statement is worthy to quote in full: Truth seeking, as William James noted, incorporates a double interest in truth. One is the interest in accepting all that is true, and this interest makes adventurers of us. To accept all that is true, we must accept a great deal, and, in the process, we will most likely accept much that is false. This conflicts with another interest in truth. That of accepting only what is true. This makes skeptics of us. To accept only what is true, we must refuse to accept a great deal, and, in the process, we shall most likely refuse to accept much that is true. Since we probably cannot achieve both of these exalted goals, we vary between being adventurers and skeptics as we focus our attention first on the goal of accepting all that is true, and then on the goal of accepting only what is true. (Lehrer 1981, p. 39; emphasis in original) 73 A weaker reading of this requirement would be to believe as many truths as possible. 74 Quoted from Riggs (2003a, p. 342). 75 This question should be treated separately from our task, albeit it is an important epistemological question.

64

Lehrer clearly sees the conflict inherent in the two requirelnents once we specify the truth goal. To use his image, the believer serves two masters at the san1e time by following her interest in truth. And from that Lehrer infers that we are driven by two conflicting motivations. The first one is supported by what he calls optimism, the second by pessimism. I think Lehrer has struck on the traditional view and its perils. Yet I believe Lehrer's terminology doesn't bear up to our critical scrutiny - and by extension, the traditional view doesn't bear up to our critical scrutiny. I have three arguments to show why this should be the case. First, believing all truths cannot be a defensible requirement because this requirement is not just optimistic, it is utopian, considering the sheer quantity of propositions there are and the quantity yet to be discovered. 76 What believer can believe all true propositions? The traditional view wrongly assumes that our interest in truth requires us to believe all existing truths. Defenders of the traditional view might respond that my reading of their position is too strong and that the traditional view actually is based on a weaker reading (of believing not all truths, but as many truths as possible). However, I think the traditional view tends towards the stronger reading in as much as it puts aside the relevance issue. It is not only the truth as such which we care about, it is also truth for something. One cannot, by fiat, decide to bypass the relevance issue, for in so doing one distorts the entire line of argument. In short, my argument is that a defensible requirement of the truth goal needs to be explicit about the relevance issue. As we have seen in Chapter 3, we value truth, but we do not value it unconditionally. Our interest is in significant truths. We can neglect the requirement of believing all truths as clearly supererogatory, which cannot possibly give any guidance to the believer. Second, the traditional view assumes that the expansion of our belief systelu is done at the price of raising the risk of error. It is stated that the believer needs to be an "adventurer" in order to expand her belief system. 76 Assuming this scenario is right, our discussion leads to other questions. A second question would concern knowability: When and where would we be in a position to know that we now believe all truths? On a practical level, we III ust suppose a realistic picture of the capacities of the believer.

65

This is a false assumption, because there exists at least one case in which we can expand our belief system without necessarily acquiring false beliefs. Just consider the telephone book case. We can extend our belief system by learning by rote a phone book. We assume here that the phone book is currently updated (and it will be, if someone moves away and so on). So the phone book provides new beliefs in a reliable way; I can thus extend my belief system without any greater risk of error. I do not claim that the traditional view is mistaken about the picture of the adventurer. [ claim, however, that it overlooks the fact that there a several ways of expanding one's belief system in a very controlled and reflective way in which we can be certain not to acquire a false belief. So the expansion of our belief system need not be an adventure. Third, the pessimistic requirement Lehrer refers to does not automatically follow from our interest in truth, because the avoidance of error does not lead us to believe anything positive. Rather, it leads us to minimize our belief system. In extreme cases, the requirement of avoiding error might even lead us to a total suspension of belief. 77 Above all, we need to qualify what is meant by "error" or "mistake" in the traditional view. It would seem that the traditional view must presuppose that mistakes are always serious mistakes in that this view doesn't have any criterion with which to sort through relevant truths. But not all errors we make are serious errors, and I think that harmless errors are not epistemically relevant, since they are not even worthy of remembering. I may, as with an example I used earlier, mistake the date of Adenauer's birthday, but this mistake has no bearing on my life. I could, of course, make it have a bearing by betting on it - but the very fact that I can make a truth more important shows us that there is a scale upon which truths sit, as it were. Mistakes are really only worthy of remembering if there is a 'systematic' source of errors. This, however, presupposes truth, because in 77

66

We might interpret Descartes's Meditations in such way. Taking a closer look at what systematic role doubt plays in Descartes's epistemology, an alternative reading of Descartes's epistemology is possible. According to that reading, the role of doubt is to indicate extremely certain truths which can provide a foundation for the believer's belief system. Broughton (2002) gave this analysis of the role of doubt. Such a reading would be adaptable to epistemic value monism because Descartes seems to trade off ordinary belief against very secure truths.

i

order to learn about our mistakes we have to know in which cases we have been mistaken. In sum, the traditional view, riven by the conflict between two extreme poles of optimism and pessimism, of believing all truths and the terror that into that infinity there will creep an erroneous belief, does not picture the normal believer's real interest in truth. Our interest in truth can be nicely analyzed by focusing on what conditions we actually value. The traditional view needs revision because it could not derive a defensible requirement from the truth goal. Before we can continue to seek a more defensible requirement, we need to overcome another confusion of the traditional view with respect to the expansion of our belief system.

2. BELIEVING MORE BY RISKING LESS?78 I claim in this section that the traditional view is confused about the risk of error in expanding our belief system. It presupposes that any expansion of our belief system will raise the risk of error. I will argue that this assumption is not right. The increase and decrease of risk primarily depend on the certainty of the routine a believer follows in the expansion of her belief system, and on how reliable the environment is in which she follows it. So, expansion cannot be a problem at a general level. This false assumption about the expansion of our belief system has the consequence that expansion seen1S to be generally problematic. Why is this so? We might consider cases where the believer's environment is an 'episten1ically friendly' environn1ent. Consider the following situation: A person travels to a foreign city and after leaving the train station she asks the next person in the street how to get to her hotel. Suppose that the pedestrian in the street informs our traveler ina reliable way. Given the reliability of the source the expansion of the belief system is not prima facie problematic. The risk of error depends on the degree of unreliability 79 of the source we draw on to expand our belief systen1. In other words, 78 I have chosen the title with a friendly nod to a paper by Luc Bovens and Eric Olsson (see Bovens/Olsson 2002). 79 We can modify the previous case in order to represent a more ordinary case: This

67

where the source is reliable, there is objectively no risk in depending on that source. If we do not have any previous information concerning the reliability of the source, the believer runs the risk of relying on an unreliable source. Given the reliability of the source, the expansion can have positive results, as Bovens and Olsson write in the context of the question whether the coherence of a belief system is conducive to its truth: [... ] acquiring more beliefs need not decrease the probability of the whole, and hence need not increase the risk of error. In fact, more beliefs can mean an increase in the probability of the whole and a corresponding decrease in the risk of error. (Bovens/Olsson 2002, p. 137; emphasis in original)

In contrast to the traditional view, Bovens and Olsson develop a different picture of the risk of error when we add new beliefs to our belief system. We do not need to consider their technical framework to approve the point Bovens and Olsson make. Only under certain conditions do we find that expansion automatically increases the risk of error. Suppose, for instance, that we live in the evil-demon world. This point undermines the crucial premise of the trade-off between believing nothing and believing all truths, because it makes it plausible that the expansion of our belief system does not in general increase the risk of error. Let us consider the following case, which illustrates that a believer in a rather friendly environment has a chance of expanding her belief system without greater risk. Suppose there was a jewel robbery. One of the suspects is Dunnit. Inspector Rob is the detective on this case. The inspector needs to obtain more information about what happened in the jewelry store and who was involved in the crime. The why-questions concerning the case are pushing Rob towards gaining new knowledge in order to be able to explain what happened. Under what condition will inspector Rob accept new beliefs? He finds several eye witnesses. How does Rob decide which information to accept and which to reject? The time our traveler uses a modem mobile phone with a navigation system. By looking up the way, she adds new beliefs to her belief system. So, if the navigation system is working properly, our traveler is likely to reach her destination. The risk of error is pretty low and the traveler can accept these new beliefs.

68

inspector's goal is to subject every new belief to the test of how it fits into his evolving narrative of the robbery in order to obtain a better explanation of what happened and, most importantly, who the culprit is. The inspector is not interested in telling a thrilling story, but an accurate one he can use to recommend prosecution of a suspect. He must have reliable information that will enable him to draw the right conclusions to that end. In the beginning of the investigation the inspector has to consider several hypotheses. There are two possible conclusions from the evidence that he has collected: (HI) Dunnit has committed the robbery, a conclusion which is pointed to by several pieces of evidence of incriminating evidence; and (H2) Dunnit is not the culprit, because the evidence for (HI) is can be otherwise explained. The concrete evidence which Rob has collected is (EI) to (E5). All the interviewees who gave testimony seem in Rob's judgment to be sufficiently reliable. Here are the hypotheses from which Rob has to conclude what happened in the jewelry shop. (EI) The bank teller of Dunnit's bank saw Dunnit depositing a huge amount of cash. Dunnit is unemployed and has never brought huge amounts of cash to the bank before. (E2) Dunnit was placed at the crime scene in a car by a bystander in the street. (E3) Dunnit owns a gun similar to the one that seems to be the one used in the crime scene according to experts who have analyzed the security video tape. The police has found the gun in Dunnit's apartment. (E4) Dunnit has a criminal record, and his probation officer tells Rob that Dunnit grew up in an orphanage and later made a criminal career. Against these testimonies, Dunnit has an alibi. According to an old won1an who claims to be Dunnit's grandmother, (E5) Dunnit had tea with her at the time the robbery was going on. Her flat in the suburbs is a one-hour drive fron1 the robbery scene. Furthennore, this woman's theory is that the Italian mafia has an interest that Dunnit goes back to jail. Without (E5), all the evidence Rob has collected directs him toward the conclusion that Dunnit committed the robbery. Without his alibi (E5), there is a good explanation available of what happened at the crime scene. If the inspector can eliminate (E5) there seems to be no obstacle to the conclusion that Dunnit is the culprit. Assume that the inspector learns next week that the grandmother has been convicted before for giving fake alibis

69

for money. Dunnit actually has no family. This is the additional evidence Rob seeks that the counter evidence to (HI) can be rejected. Thus, it seems there is an air-tight case against Dunnit. Rob has rightly collected all the information available during his investigation. Instead of confirming the evidence (E I) to (E4), Rob went on to check whether (E5) can be true, since (E5) is inconsistent with the testimony of the witnesses and the match of the make of the gun, although it is consistent with (E4), in that Dunnit's criminal associations in the past could possibly have led someone who knows police procedure to set him up. Given the evidence available to Rob, after he has found out about the false testimony of the fake grandmother, the best explanation for him is that Dunnit committed the robbery.80 It is clear that the handling of testimonial sources poses problems once we want to indicate necessary and sufficient conditions for testimonial knowledge. My argumentative aim was a modest one: I have aimed at establishing a case where in a reasonably stable environment, the expansion of my belief system need not raise the risk of error, if the inquirer has the cognitive capacities to filter the available information. So far I have given my reasons for rejecting the traditional view, which aU turn about the idea that the traditional view casts too skeptical an eye on the possibility of increasing knowledge. Beyond the cases I have adduced from ordinary life, there is, of course, the growth of science, which seems to systematicaUy contradict the traditional view. We have made progress in knowing about the natural world; we have not regressed. 81 This in itself should make us suspicious of the traditional view's claim. In the foUowing section I shaU look for a defensible 80 We have to be careful about stating that the best explanation available will necessarily lead to truth. We might expect, however, that the best explanation is good enough. Indicating the best explanation available for the evidence is a tool in what Peter Lipton has called "plausibility monitoring". Evidence (E5) is in the light of the other hypotheses quite implausible. I find it attractive to apply inference to the best explanation to the case of knowledge from testimony. But we do not need to go into the details of such an application. See Lipton (2008) for an insightful discussion. 81 Fahrbach (forthcoming) points out that 80% of scientific work has been done after 1950.

70

requirement for the truth goal. 3. LOOKING FOR A DEFENSIBLE REQUIREMENT How can we replace the supererogatory requirement the traditional view has endorsed with a more reasonable one? The following two proposals are supposed to overcome the tension between avoiding error and believing the truth. (S) It is good to believe p, if p is safe. 82

This condition draws on Sosa's proposal that the safety of beliefs is the defensible requirement for the truth goal. The safety of belief p is defined such that we would believe p only if p were true. Or in different words: "A belief is safe provided it would have been held only if (most likely) p" (Sosa 2007, p. 25). (S) has the virtue of incorporating the twin requirements of believing the truth and avoiding error. Following from (8), as I take it, the trade-off which the traditional view presupposes will not take place. So, is (S) a defensible requirement? Even though (S) follows the right motive, it is troubled in the following sense: We best fulfill (S) by adding trivial truths to our belief system. Thus, learning the telephone book by heart would be a good procedure to expand one's belief system. This is not satisfactory. There is a second proposal that incorporates both requirements: (R) For all propositions p, the truth goal consists of amassing a large body of beliefs with a favorable truth-falsity ratio. 83

The positive component of (R) is that it brings together the two requi~ements the traditional view has separated. The idea is the following: Once I anl able to identify the truth-falsity ratio, I am in principle in a position to give a statistical account of which beliefs in my belief system 82 See Sosa (2001, p. 50), as well as Sosa (2007, Ch. 2). 83 Here I draw in parts from the formulation of Latus (2000, p. 29). The idea of the favorable ratio between truth and falsity goes back to Alston (1985).

71

are most likely to be true and which ones are false. And for this reason I am in a position to exclude the false beliefs from my belief system. Even though I think the motivation for stating (R) goes in the right direction, it suffers from a crucial problem: How can we account for (R) on quantitative grounds? This question brings us to two separate problems: First, how can we count our beliefs one by one? Second, what is a favorable ratio of truth to falsity if we are able to count our beliefs? Both questions, it seems to me, assume the possibility of counting our beliefs. This I consider a doubtful project, as we don't have any rule to follow for quantifying over beliefs. As a direct consequence we have to give up the idea of an acceptable ratio of truth to falsity as well, because identifying a positive ratio would require counting our beliefs. We can, however, keep the positive insight of integrating the avoiding-error requirement and the believing-truth requirement, while giving up the idea of creating a specific truth-falsity ratio. Marian David gives such a proposal. He states: (T) For all propositions p, if p is true then S believes p, and S believes p only if p is true. (David 2001, p. 158)

David nicely incorporates the requirement of avoiding error by stating that we should believe the truth and only the truth. But I think that the "all propositions"-clause needs some clarification. Does (T) already presuppose the old requirement of believing all truths that exist? The question remains whether (T) requires us to believe too much, an issue we have already visited. If this is right then David would have to explain how his account differs from the traditional view. As I have emphasized throughout, my claim is that we are not interested in all the propositions there are; moreover, the distinction between important, relevant beliefs and insignificant, irrelevant beliefs should be reflected in our theory of epistemological value. Therefore, I think it is more appropriate to state that prima facie we prefer holding significant true beliefs. 84 84 David rejects this proposal. But I think his dismissal comes too quickly because

72

(TS) Concerning all propositions p, S believes the proposition p if p is true and significant, and S believes p only if p is true.

(TS) intentionally leaves open the question of what counts as a significant belief. Whether a "significant" belief must be subjectively or objectively significant is still open to further elaboration. Here is a short characterization of in what sense significance is supposed to work in (TS): Significance is important for ruling out trivial-truth cases such as those we looked at in Chapter 3. Additionally, it should rule out the truism cases, where we securely attain truth by forming tautologies but do not therefore 85 acquire any new information about the world. The advantage of (TS) is that it also explains where the motivation for believing the truth comes from: the motivation is packed into the significance requirement. In the expectation of learning something significant, minor mistakes are acceptable (if our monitoring and detecting capacities are working properly). Therefore I take (TS) to be the best expression of the defensible requirement we can derive from the truth goal. In the following section we will see that once we have a defensible requirement for the truth goal, epistemic value pluralists lose their argument that we need to balance the requirement of believing truth and avoiding error.

we need the significance condition. See David (2001, p. 159). 85 It is not only truth we want, but informative truth about the world. At this point we have to be careful to distinguish (S) from the alleged epistemic goal of understanding how the world works. Steven Luper-Foy presents such a cognitive goal in the context of what we might strive for as cognitive and rational human beings. See Luper-Foy (1990, p. 50).

73

4. RIGGS'S ARGUMENT ON BALANCING THE TRADITIONAL REQUIREMENTS Riggs (2003a) has argued that we are in need of introducing an overarching goal to make our knowledge seeking globally meaningful. He claims that the overarching goal can't be truth, as this would create a circular solution to the tension between the two requirements that are traditionally inferred from the truth goal. But, so far, he has not come up with a satisfactory description of this overarching goal. 86 For Riggs, the traditional view of the requirements stands in conflict with epistemic value monism. The pivotal assumption in Riggs's argumentation can be summarized in the following way: (RST) Any theory of epistemic justification and knowledge needs to spell out the weighting issue between the requirement of avoiding error and the requirement of believing the truth with respect to the varying degrees of justification.

Epistemologists have not taken care of this task so far, Riggs thinks. 87 For him, this issue has been ignored in current epistemology, because, so far, theories of epistemic justification "assume a solution to it without argument" (Riggs 2003a, p. 345). Given the traditional view of our requirements, it is clear that (RST) applies to any epistemological theory. What is the weighing issue expressed in (RST)? It seems that the weighing issue concerns the situations where the balance between the requirements of believing the truth and avoiding error depends on how we adjust the standards of epistemic justification. If we have high standards for epistemic justification, for instance, we will stoke the fear of believing falsehoods and make the avoiding-error requirement dominant over believing truth. With somewhat lower standards of epistemic justification 86 He gives various proposals, with which he is not satisfied (see Riggs 2003a, pp. 349-351). He seems to lean towards "understanding" as the name of the overarching goal, but his sketch of what that means is, as he admits, fuzzy. 87 See Riggs (2003a, p. 347).

74

88

we lessen the need for weighing the two requirements. Very low standards "invert the weighing" toward the goal of believing the truth, Riggs thinks. So with "weighings" Riggs means the different states of the balance between avoiding error and believing the truth. The crucial issue for Riggs with these weighings is which adjustment between the two of them is the correct one. If this reconstruction is right, it becomes clear why Riggs thinks that the "weighing issue" requires us to introduce an overarching goal, which is able to regulate it. In general, Riggs argues for the view that there is no fixed hierarchy between our two requirements, and that adjustment depends on our comfort zone. Riggs has introduced a comparison from economic theory intended to illustrate his view. The example is about trade-offs, taken from the 89 business world. Consider two factories that produce widgets. Both factories belong to one firm and this year's management retreat proposes two targets for the next year: (Target 1) Both factories are supposed to produce as many widgets as possible. (Target 2) There must be no defective widget on the market. It is clear that both targets are in conflict with one another because Target 1 will have the consequence that a higher percentage of widgets will be defective. 90 If Target 2 were to obtain high priority, the total production of widgets would decline since the reduction of the amount of widgets will lower the amount of defective widgets.

In order to meet their targets, both factories get a new quality inspector: Quality inspector A is very eager to meet Target 1. Quality inspector B differs from A. He is an anxious type of person. For him, Target 2 has priority. During the year, both quality inspectors receive a letter from the CEO, announcing that the work of both inspectors will be evaluated at the end of the year. The priorities of the qual ity inspectors wi II clearly affect the output of the factories. The second factory, with the anxious quality inspector, 88 We can neglect the possibility that Riggs is interested in the equal status of our requirements. 89 See Riggs (2003a, pp. 343-345). 90 Under ordinary conditions the firm might outweigh the losses from the defective widgets by producing and selling more widgets. But Riggs excludes this option.

75

produces only one widget over the whole year, but one without defects. Thus, the second factory fulfills the second target, but completely misses the first one. By contrast, the first factory has a high output of widgets, but five percent of the production is defective. The first factory meets well the first target. Which factory will obtain the approval of the CEO? Which factory has fulfilled the two given targets appropriately? Riggs has designed this case as one where no easy solution to the weighing issue can be given. I think, however, that Riggs does not pay enough attention to how certain details of his case affect its illustrative power. Let us consider that the overall goal of the factories is maximizing the firm's profits. Then it becomes questionable how the second factory has contributed to this goal. Note that I assume that there is an overriding goal. It does not seem plausible to me that the second factory can be seen as an economically working factory. From that I conclude that Riggs's initial description of the case cannot be right, because there is no diverging weighing between high quantity of widgets and low rate of defective widgets, unless Riggs maintains that the production and the sale of one widget could outweigh the costs of running the factory for an entire year. 91 If that is right, then the initial construal of Riggs's example is undermined and the widget case does not form an analogy to what Riggs attempted to show. Without an explanation of why the second factory is equally successful, Riggs's example proves the contrary point: Any target or requirement that violates the overall goal of the firm cannot be taken as a reasonable requirement. 92 In other words, the two requirements in play cannot have the same significance, because the second requirement is likely to lead to unprofitable results for the firm. So, the task of quality control to keep the quantity of defective widgets low is a supplementary requirement to the requirement of producing widgets.

91 Riggs (2003, p. 349) accepts this overall goal, but he comes to a different conclusion. 92 The firm might even maximize its profit when there are defective widgets on the market. As long as someone still buys the widgets, the firm can make a profit. But the firm should aim to diminish defective widget, for defective products could easily cause profit to collapse in the future, and erasing present gains.

76

What goes wrong when we follow the logic in Riggs's example? Riggs confuses the position on the scale of priorities of the two requirements with the ultimate goal the firm actually has. The firm has only one ultimate goal. Why is there a need to doubt the correctness of the thesis that believing significant truths has priority over avoiding error? Common sense, or convention, should give us enough adjustment space to make for our overarching goal - believing significant truths - while pulling us back if we commit too many errors - the latter of which is a precautionary rule, not a goal. The weighing issue depends on the traditional view, which presupposes indefensible requirements. As we have seen, Riggs's problem with adjusting between standards - the weighing issue - can be ignored once we have derived defensible requirements from the truth goal that reflect the overarching goal of acquiring significant true beliefs. Riggs takes his case from the practical world of business decisions to emphasize that trade-offs do exist in practice. Yet the inspectors of his factory, by having no standards to which to refer - neither standards of industrial practice nor rules of thumb on ROI - commit themselves to arbitrary interpretations of the directive they do have. Thus, what starts off as an interesting argument for a trade-off between the requirements of believing truth and avoiding error, ends up failing in its goal: it does not have the illustrative power he supposes, as it arbitrarily excludes just those real rules in the practice of business that govern competing firms. I conclude that there is no weighing issue as proposed by Riggs. If I am right, then the objection of the overarching goal can be met by my (TS) condition. Riggs mistake is to press for an argument where no argument need be given.

5. CONCLUSION I have claimed that we can define the truth goal by identifying a defensible requirement deriving from the truth goal that emphasizes significant truths, rather than all truths. Furthermore, I have shown that the traditional view of the requirement is confused, because it presupposes that both believing all truths and avoiding all errors are equally important. By adopting the significant-truth requirement, there is no divergent ordering

77

between avoiding error and believing truth. Believing truth is prior to error avoidance. The latter is a precautionary rule, not a goal. Once we see that the traditional requirements are invalid, attempts to create a goal problem that derive from the traditional view, such as Riggs's notion that epistemologists need to introduce an overarching goal that balances the tensions between two mentioned requirements, is also invalid.

78

5. The Primacy of the Truth Goal In this chapter I discuss what I call the objection of alternative epistemic value. It is argued in the literature that we sometimes attribute epistemic value to beliefs even when those who hold these beliefs have not attained truth. This would ilnply that there exists a second source of epistemic value outside of the truth goal. This would contradict epistemic value monism, which tnaintains that other epistemic goals as having 93 merely instrumental status, because they exist in order to obtain truth. Thus, if there are alternative epistemic values, than all epistemic goals would not be truth-related, and epistemic value monism would fall. We will consider three cases. As I will show, it will turn out that none of them supplant the truth primacy claim. Therefore, all epistemic goals must be truth-related. To defeat the prin1acy claim, epistemic pluralists have to find only one case in which an alternative epistemic goal does not have merely 94 instrumental value. Or, to put it differently, if there is an acceptable case where the truth goal is not satisfied, but some other epistemic goal, then the primacy clainl is defeated. The epistemic pluralists' main argument against the primacy of the truth goal is as follows: (A) If truth is the primary episten1ic goal (or value), then all other epistemic goals (or values) are related to truth. (B) Not all epistemic goals are related to the truth goal.

93 The instrumental relation between epistemic values in the traditional form is discussed in terms of epistemic justification and truth (in short, the truth connection). The truth connection unifies the various theories of epistemic justification such as coherence, foundationalist theories of justification, or reliabi lism. See for instance Lehrer/Cohen (1983), Cohen (1984), Conee (1992), Sturgeon (1991). See Olsson (2005) for a thorough treatment of the problem of the truth-conduciveness of coherence. See David (2001/2005) and Kvanvig (2005) for the current debate. 94 See Pritchard (2007 Q, 2007 b, § 7).

79

(Therefore) Truth is not the primary epistemic goal.

In this form, the argument puts the burden of proof on epistemic value pluralists to show one non-truth related epistemic goal which has epistemic value. 95 The case must exhibit some recognizable cognitive achievement that is independent of truth and yet still has distinct epistemic value. In the current literature, three cases have been proposed in support of (B). The first case concerns the epistemic value of understanding. The second case is a version of the evil-demon case, which is supposed to prove that being a responsible believer is detached from truth. The last case concerns Jonathan Kvanvig's proposal of the technical concept of empirically adequate belief. All three cases are helpful in clarifying the primacy claim of epistemic value monism. In this chapter, I will evaluate all three cases and, as far as possible, give a reply from the epistemic value monists' side.

I. THE VALUE OF UNDERSTANDING Very recently, epistemologists have become interested in the epistemic value of understanding given the problem of the value of knowledge. 96 Jonathan Kvanvig (2003) has been a prominent advocate for the distinctive epistemic value of understanding. The starting-point of his book, The Value of Knowledge and the Pursuit of Understanding, is the search for the distinct value of knowledge over mere true belief. Kvanvig works through various routes to the defense of this distinctive value and finds no single route, no conclusive argument. He concludes that we cannot complete this venerable search; the project that started with the 95 I see the primacy claim of epistemic value monism as a prima facie claim. That means, the claim is valid as long as there is no striking counter-example. 96 Simultaneously, we can observe efforts of expanding the scope of traditional epistemology. Kvanvig (2005, pp. 286-287) calls for the expansion of epistemology in terms of expanding the range of cognitive achievements within epistemological theorizing. See also Axtell (2008a).

80

Meno has come to a dead end. This suggests to Kvanvig that we should 97 shift our epistemological focus to the phenomenon of understanding. Kvanvig proposes a revisionism concerning our epistemic values, because 98 understanding is introduced as a bearer of epistemic value. Kvanvig's own account of understanding is factive. However, on Kvanvig's reading, what distinguishes understanding from knowledge is that: Understanding requires the grasping of explanatory and other coherencemaking relationships in a large and comprehensible body of information. One can know many unrelated pieces of information, but understanding is achieved only when informational pieces are pieced together by the subject in question. (Kvanvig, 2003, p. 192)

How can understanding provide the case epistemic value pluralists want to establish? Admittedly, understanding is a good candidate for two reasons: it is a broad cognitive phenomenon, and its relation to truth seems to be mediated by the subject's activity in "piecing" together disparate bits of information. In summary, epistemic value pluralism needs to defend the following claims: (a) There is an account of understanding that is not truth-related (= understanding without holding true beliefs). (b) The account of understanding in (a) has epistemic value.

In the following, I shall provide examples that are supposed to illustrate aspects of understanding that the epistemic value pluralist finds important. Thereafter, I will argue that (a) and (b) are both false.

97 See Kvanvig (2003, Ch. 8). See also Kvanvig (2009). 98 See Pritchard (2007b). Pritchard has introduced the term 'revisionism' with respect to the value-of-knowledge problem. Revisionists think that something that must not be knowledge has distinctive value besides true belief.

81

2. TWO CASES OF UNDERSTANDING In the following quotation, Catherine Elgin, like Kvanvig a prominent advocate of understanding as an alternative goal of inquiry, gives two examples to introduce what she means by understanding: To understand the Athenian victory [in the battle of Marathon] involves more than knowing the various truths that belong to a suitably tethered comprehensive, coherent account of the matter. The understander must also grasp how the various truths relate to each other and to other elements of the account. She should also be able (and perhaps be aware that she is able) to use that information-to reason with it, to apply it, perhaps to use it as a source of working hypotheses about related matters. Someone who knows geometry, for example, knows all the axioms, all the major theorems, and how to derive the major theorems from the axioms. You can acquire this knowledge simply by memorizing. But someone who understands geometry can reason geometrically about new problems, apply geometrical insights in different areas, assess the limits of geometrical reasoning for the task at hand, and so forth. Understanding something like the Athenian victory is not exactly like understanding geometry, since the applications and extensions are more tentative, the range to which insights can reasonably be applied is more restricted, the evidence for a successful application is empirical (and may be hard to come by), and so on. But in both cases understanding involves adeptness in using the iriformation one has, not merely an appreciation that things are so. Evidently, in addition to grasping connections, an understander needs an ability to use the information at his disposal. (Elgin 2007, p. 35; emphasis added)

Let us consider the second example first. There, understanding implies expertise. 99 The understanding of geometry depends not only on memorizing certain axioms, but on being able to apply those axioms to different contexts. According to Elgin, we would say that the student who can't apply axioms in different areas or reason about new problems, but who can nevertheless repeat the geometrical axioms, lacks understanding of the subject. Understanding requires, then, not only a knowledge of the axioms, but a knowledge of how to apply the axioms properly in a given 99 Another account, for instance, would be the account of empathy with others. Consider the phrase "I can understand your decision" or "I now understand your reason for doing so". These phrases refer to the situation that 1 can participate in another person's inner life.

82

case, or in relation to new problems. Elgin is right that in ordinary language, we use the term understanding in the applicative sense when we refer to education in a certain expertise, or to expert knowledge. In the context of evaluating the abilities of experts, we use phrases like "He understands his physics" or "He understands microbiology very well" .100 The first case is n10re mysterious. Let's say that I wonder why the battle of Marathon occurred. Presumably this is a question about the reasons the Athenians and the Persians had to fight each other. For Elgin, knowing the mere facts about the battle is not sufficient to obtain the understanding that would answer the question, why did the battle of Marathon occur. In order to obtain understanding of this complex historical case, the subject needs to organize the body of information. For her, something must be added apart from the facts about the battle to explain the reasons for the battle. This something, she thinks, is understanding. Elgin's account of understanding is holistic. Where does the epistemic value of this holistic version of understanding come from? The answer is that understanding is a different cognitive state from other propositional attitudes, like knowledge or belief. Elgin describes the difference as follows: 'Understanding' is a better term for the epistemic achievement that concerns us here. Not being restricted to facts, understanding is more comprehensive than knowledge ever hoped to be. We understand rules and reasons, actions and passions, objectives and obstacles, techniques and tools, forms, functions, and fictions, as well as facts. We also understand pictures, words, equations, and patterns. Ordinarily these are not isolated accomplishments; they coalesce into an understanding of a subject, discipline, or field of study. (Elgin 1996, p. 123)

In another publication, Elgin defines understanding as a cognitive-success term. IOl It seems that success in understanding consists in having a 100 Concerning physics, Elgin writes: "Understanding physics is not merely or mainly a matter of knowing physical truths. It involves a feel for the subject, a capacity to operate successfully within the constraints the discipline dictates or to challenge those constraints effectively" (Elgin 1996, p. 123). 101 See Elgin (2007, p. 34).

83

comprehensive belief system or a comprehensive body of information. This provides also an argument for the independence of the goal of understanding: Since understanding aims at comprehensiveness, it has a distinct value in comparison to knowledge and belief, Elgin and others think. In sum: If understanding has a distinctive value, then we can treat understanding as an independent epistemic goal. This opens up the possibility that we can satisfy the goal of understanding while we fail to attain truth. Zagzebski, who has also considered the claims of understanding, emphasizes that it has epistemic value because we cannot be wrong about it. She writes: Understanding has intemalist conditions for success, whereas knowledge does not. Even when knowledge is defined as true justified belief and justification is construed internalistically, the truth condition for knowledge makes it fundamentally a concept whose application cannot be demonstrated from the inside. Understanding, in contrast, not only has internally accessible criteria, but it is a state that is constituted by a type of conscious transparency. It may be possible to know without knowing that one knows, but it is impossible to understand without understanding that one understands. (Zagzebski 2001, p. 246)

In short, Zagzebski's argument is that understanding and knowledge differ because knowledge cannot fulfill the transparency condition satisfied, on the other hand, by understanding. In sum, the school of epistemic value pluralists I am considering here assume that understanding, as they conceive it, has epistemic value because: 1) it aims at comprehensiveness; and 2) it constitutes an internal state of the believer which differs from knowledge. In the following sections, I want to argue that either understanding has epistemic value because we have a good or appropriate explanation as a foundation for it, or understanding accommodates minor false beliefs, as Elgin has proposed. Elgin's account of understanding does not, however, have any epistemic value. Before we come to the main argument, we have to clarify the role explanation plays in our debate.

84

3. EXPLANATION AND EPISTEMIC VALUE Understanding and explanation are closely related in the following sense: A good explanation increases our understanding. The better the explanation that we have, the better is our understanding. My argument presupposes that explanation is clearly truth-related, because any appropriate explanation must describe the conditions of the event to be explained. And an explanation cannot count as appropriate without getting the description of the initial conditions right. lo2 In contrast, epistemic value pluralists think that explanation can be detached froln truth. Roughly speaking, epistemic value pluralists rely on the assumption that the explanatory power of a theory is detached from the theory's truth. 103 For Riggs, for instance, it is possible to trade off the truth of a theory for its explanatory power, and vice versa. This assumption only makes sense within an anti-realistic framework for science. But such a view must deny the traditional view that there is a close relation between explanation and truth. Quine describes the relation between explanation and truth as follows:

We see [...] that there can be mutual reinforcement between an explanation and what it explains. Not only does a supposed truth gain credibility if we can think of something that would explain it, but also conversely: an explanation gains credibility if it accounts for something we suppose to be true. (Quine/Ullan 1978, p. 120)

Instead of referring to an explanation and a fact to be explained, we can talk more briefly about 'the theory' and 'the fact'. The better fact and theory fit together, the better it is for both. In the following, I want to show that the epistemic value we attribute to understanding is derived from its providing appropriate explanations. Consider a case from history again, but let us modify the example. Consider the case of two students, both studying the political history of the 102 I don't want to claim that there is only one unique explanation of an event. 103 See for instance Riggs (2003 b, p. 219).

85

United States of America. The first student has learned all the important facts and dates of the political history of America, but this student cannot provide any deeper explanations about the relations between those facts. Consider a second student, who has deliberately not learned as many facts by heart, but instead worked mostly on deepening his understanding of America's political history. His focal interest is the relation between social history and political history. The key question for the second student is, how did American presidents improve the living situations of the middle and lower classes? He knows where to look up the facts when he needs to verify them for his argument. We would say that the second student achieves more, cognitively. The second student's method is more praiseworthy from the epistemic point of view, because he is trying to be the better source from which to obtain knowledge about the political and social history of America. The first student can only offer knowledge about facts. Thus far I can accept Elgin's reference to our value intuition, but I think that her criticism can be met once we dispense with the assumption we criticized in the last chapter that is, necessarily bonding epistemic value and the knowledge of all facts. As we saw, we can modify epistemic value monism so that the inquiry about which we are concerned is always about significant truths. Taking the traditional approach as her target, Elgin thinks that epistemic value monists must find the first student praiseworthy simply because he has so diligently accumulated historical facts; but a modified epistemic value monism would have no problem judging between the two students. We can deduce from Elgin's line of argumentation about a certain familiar image of epistemic value monism, one in which the monist must necessarily defend as praiseworthy any and all acts of fact gathering. From her point of view, understanding does have a significant advantage: it separates the trivial from the significant, the irrelevant from the relevant. But if, in fact, as we are contending, we can pack significance into our truth goal as an internally emergent characteristic, then much of the motivation for understanding falls. We do, in fact, have an epistemic value monism that is able to distinguish and praise or dispraise cognitive

86

achievements that lie beyond the level of facts. 104 To return to our comparison: it appears that we prefer the student who obtained a holistic understanding of the matter over the student who has merely learned the relevant facts by heart. Why do we value the 'sophisticated' student? This student is able to give complex, connected explanations for events in American history. Or, to take Elgin's example, the sophisticated student is able to give us a well-argued explanation of why the battle of Marathon did occur. If this is right, we have inlplicitly introduced another requirelnent: Our understanding depends on the appropriateness of the explanations it yields. We value the person with deeper understanding because he provides more appropriate explanations. I infer from this that the epistemic value in play is derived from the value of explanation. In the following section, I will discuss Elgin's proposal that the value of understanding has its source in understanding's being 'tolerant' toward minor mistakes and minor false beliefs.

4. HOW IS UNDERSTANDING POSSIBLE WITH MINOR FALSE BELIEFS? In what sense does understanding, on Elgin's account, have epistemic value? Elgin thinks that her conception of understanding has the benefit of being tolerant toward minor false beliefs. Elgin argues that there are occurrences in our cognitive life that have positive epistemic status despite the absence of truth. So she thinks that the epistemic value of her conception of understanding lies in the fact that it is tolerant toward holding minor false beliefs. Let us call Elgin's account of understanding "E-understanding". That is, understanding can take place even when we have minor false beliefs. The difference between E-understanding and knowledge, Elgin thinks, is that knowledge is purely factive whereas E-understanding allows minor errors. The central point of my evaluation is: if E-understanding has 104 See Chapters 3 and 4 for detailed discussion under which conditions we value truth. Again, trivial truths should not be valued as such.

87

epistemic value, then we can derive a normative requirement that helps us to better define that epistemic value contained in E-understanding. Taking this relation between epistemic values and normative requirements for granted, it is obvious, I think, that Elgin commits a mistake by postulating an epistemic value which is incapable of generating an epistemic requirement. Here is my argument against Elgin's claim: (A) If E-understanding has epistemic value, then we could derive a tolerance requirement for E-understanding. (B) There is no tolerance requirement.

(Con) E-understanding does not have epistemic value.

We can support premise (A) in the following way: If E-understanding has value because it is tolerant toward false beliefs then we need certain requirements to attain this value that comes with E-understanding. The critical point I want to make here is that Elgin should indicate when and for what reason we have to be tolerant towards minor mistakes. If she is not able to spell out the exact requirement for her account then this affects her value claim. The tolerance requirement is necessary for giving us guidance in which cases false beliefs are negligible and therefore do not block understanding. The requirement should answer the following questions: To what degree is one allowed to hold minor false beliefs? What are minor false beliefs? We might indicate what a minor false belief is by saying that it is harmless for the stability of our belief system and that it might even promote our understanding. But we face the problem that we cannot generalize from this situation. In this situation we allow an exception to the rule that we should stick to the truth and to nothing else. 105 This also affects the central components of E-understanding. Elgin stated that E-understanding is a cognitive-success term, and she thinks that

105 And if we cannot generalize, then it is questionable whether we have obtained a normative requirement.

88

E-understanding differs from such things as astrology. Or, to put it differently, her account excludes astrology from providing understanding "because it lacks suitable tether" (Elgin 2007, p. 35). In order to qualify an understanding that takes in some false beliefs from one that takes in too many, Elgin has to give us a sense of introducing a threshold concerning the amount offalse beliefs that still allow E-understanding, and justify it as a matter internal to, and generated by, understanding. This brings us back to the question that should bedevil Eunderstanding: why must we prefer accepting minor false beliefs over simply suspending belief? Or how can we prefer E-understanding if there is an appropriate explanation within the canon of truth-seeking knowledge for the event we are interested in? Consider the following story: Suppose that my car did not work last week because the radiator was frozen. Suppose further that I am gullible. I receive two different explanations of what happened from two different persons. One is a mechanic I meet on the bus, to whom I describe my situation. The mechanic speculates that it might be the radiator. It is winter, and I might not have put in anti-freeze. The mechanic does not want to say for sure without having seen the car. I note that the mechanic was not very confident about his explanation and therefore I do not consider it further. The second person is the secretary of my boss. She is a strict physicalist (with a strong sense of humor). She loves to invent ghost stories in order to make fun of esoteric and pantheistic world-views. Moreover, this secretary likes to take advantage of my credulity. She makes up an explanation, states it with much confidence and thus makes it seem plausible to me. Her fake explanation is that I have not sufficiently worshiped the ghost in the car. So the ghost went on strike and therefore the car did not work. I take the fake explanation to be right, because it seems to be more comprehensive. Or, in other words, I think if I accept the fake explanation then I can learn more about the physical world and how ghosts are actually running it. Later, the secretary reveals to me that she was joking and she endorses the mechanic's explanation, because that explanation is able to explain the case of my broken car appropriately.

89

What particular epistemic value does the fake explanation have, which might have some entertainment value but is definitely wrong? Under normal conditions I should not have considered it. Where is the distinctive value that Elgin proposes lies in accepting false beliefs that look comprehensive at first sight? I do not think that Elgin gives a straight answer to those questions. Additionally it appears to me that Elgin's notion of understanding is too liberal in the sense that it has problem to demarcate scientific from non-scientific explanations. The upshot of my argument is this: In the car case, we have two explanations. As long as we do not have the appropriate explanation, we have to consider all the alternative explanations, but we should be aware that all alternative explanation can count as scientific. The believer in the car first allows both explanations. But once he has accepted the mechanical explanations for car repair, he is unlikely to accept ghost explanations for car repair in the future. When we are engaged in an inquiry to find an appropriate explanation for a phenomenon, at every stage certain explanations lose their value in competition with other explanations as they prove to be premised on false claims, or to be outside the realm of the paradigm of explanations we are using, or as they fail to succeed in experiment. This is, of course, what we would expect from the notion that truth is the goal of inquiry. From a methodological point of view our preference for the adequate explanation also expresses where the epistemic value has its source. Elgin would perhaps object that I am being unfair to the purpose of Eunderstanding, which is supposed to apply to such things as non-verbalized practices and scientific innovations. Maybe she thinks of something else which is praiseworthy in E-understanding. Maybe she thinks that the subject in the car case is praiseworthy because he was looking for insights. This response relies on the idea that understanding coincides with a sort of awareness or comprehension which belongs to the understander himself. I want to show that this response is not viable. By that I hope to show that understanding cannot escape the problems of the kind of knowledge from which it is distinguished by the subject's synthesizing awareness of the facts he wants to understand. For that reason we cannot say that understanding has distinct epistemic value.

90

Duncan Pritchard (2009a) has proposed the following case (DPl): I COOle home and my house has burnt down. I want to understand what happened, so I ask a fireman who is on the property, and he informs me that it looks like faulty wiring caused the fire. This information gives nle a sense of the primary cause of my house burning down, which improves my understanding of what has happened. Coming back to Elgin's possible reply, we might think that E-understanding is valuable because a certain type of awareness is involved. Putting Elgin's view together with Zagzebski's characterization of understanding as fulfilling a transparency condition, what we have in this example is a distinct epistemic value arising from the fireman's remark, which made me see the cause of the fire. I generally understand, then, the whole sequence involved in the event of my house burning down. Is this claim true? Let us consider the following modification of Pritchard's story. In a different version (DP2), I come home, see that nlY house has burned down, ask a fireman on the property what happened, and am told that the faulty wiring in the house caused the fire. However, unbeknownst to me, the man dressed as a fireman is actually a man going to a costume party in a fireman's suit, who was simply a spectator of the fire. As I learn later - perhaps when I am gathering infonnation for my insurance claim - the person was not actually well informed about the fire, and answered as he did because he assumed that most house fires are caused by faulty wiring. In (DP2) it is true that faulty wiring caused the fire. I was lucky that I obtained the right infornlation, which helped me 'see' the whole sequence of the fire and which I could even use practically. But it turns out that I was initially relying on merely accidentally correct testimony. What do we learn about the transparency condition with regard to the modified case? Does Iny understanding in this case have any epistemic value after I have learned that I acquired it in an unreliable way? I understand why the house burned down, but I acquired that information in an accidental way. I think that this case shows that understanding is not luck-proof, which is not in general bad. The story sheds a dubious light on the distinction between knowledge and understanding, as Zagzebski has argued for. In sum, I have shown that understanding does not have an additional epistemic value, but

91

rather accrues the epistemic values it has by guiding us towards the truth. The supposedly truth independent cognitive activity caught by understanding - of simplifying, synthesizing information, performing a task, etc. - all turn out to be valuable insofar they are based on the truth, and valueless insofar as they aren't.

5. AN EVIL-DEMON CASE The second case seems to bring weight on the side of epistemic value pluralism. Consider an evil-demon world in which there is an evil demon that turns all beliefs of the believer into false ones. In this world the only things that are safe from the demon's deception are internal states. This means, the deceived believer can only rely on her internal intention to help her to have an adequate belief system. Now consider a different world, a benign-demon world, where the believer can behave irresponsibly, but a benign demon fixes his beliefs in the end such that he will always end up with true beliefs. Let us say that if we prize epistemic responsibility, then our preference between the two worlds would go to the first one because there the believer is virtuous, while in the second world the believer is just lucky to hold true beliefs. The believer in the first world wants to pursue knowledge, but is unlucky enough to live in an environment where her beliefs never correspond to the world. The choice is between the true belief of the second world (which is obtained irresponsibly, without effort or method) and responsible belief (which is not obtained because of an extrinsic force). This is an interesting case, because epistemic value pluralists seem to obtain what they need to defeat the instrumentality thesis: The agent satisfies an alternative aim while missing the truth goal, and the believer is not to blame for having violated an epistemic requirement. The argument that epistemic pluralists attempt to establish is the following: 106

106 The argument in this form is inspired by a discussion with Trent Dougherty. But my version of the argument is different from his. 1 simplified it a little in order to keep the methodological question out in the first place. The list of defenders of responsible believers is long: See Foley (1993, 2005), Greco (1993), Montmarquet

92

(Responsible belief argument) (PI) We value the responsible-but-false-beliefs world over the irresponsiblebut-true-beliefs world. (Therefore) we value (in general) responsible beliefs more than true beliefs.

With the preference as expressed in (PI), a general choice is determined. That is, in all cases we value responsible belief more than true belief irresponsibly acquired. The case brings into play two different views of the criteria by which we award merit for epistemic responsibility, and hence, two slightly conflicting immediate responses from the epistemic value monism perspective. The first response is that the agent in the evil demon world has the intention to seek the truth, and follows it up as she can. She has followed fundamental epistemic principles in a world in which they don't count. 107 The upshot is: Good motives are necessary for good epistemic performance. lOS The second response supports the idea that we should suspend our judgment that the agent should be accorded positive epistemic status because it is far from clear what the agent has achieved .. Iy. 109 cognitive So what is the epistemic value monist's final response? A good reply IIO to the evil-demon case is this: On what basis does the deceived believer still form her beliefs? Supposing that the believer was snatched out of a normal world and transplanted into the evil-demon world, we would presume that she transposes the beliefs she had in the normal world to the demon world. If that is the case, we do not have a clear-cut choice between the believer's cognitive responsibility over the truth (which is structurally inaccessible), because the important beliefs for the deceived believer are (1993), Riggs (1997). 107 See for instance Foley (2001, pp. 37-38), Greco (1993, pp. 416-419). Thanks to Trent Dougherty for making me aware of Greco's work. 108 See Fairweather (200 I). He gives a nice overview of the discussion in the virtue epistemology camp that supports the idea that good motives are necessary for good cognitive performance. 109 See for instance Goldman (1999, Ch. 3). 110 The initial response would be to doubt the outcome of the thought experiments, because there is intimate relation between responsibility, even moral one, and truth. That response clearly begs the question.

93

the old ones that were attached to truth. III This reply also gives us a good explanation of why we also think that the deceived believer is still responsible, even though her attempts to form true beliefs do not succeed. The belief system of the deceived person works correctly because it rests upon the set of old beliefs. 112 The epistemic value pluralists have overlooked the details that can explain why the belief system is still operating in this hostile world. Epistemic value pluralists can still modify their thought-experiment in order to undercut the first reply. Consider the case that we have to choose whether we want to be born in the evil-demon world or in the benigndemon world from something like Rawl's 'original position' - that is, we chose with a "veil of ignorance", having no personal knowledge of past epistemic success to guide us. ll3 What is changed with this modification? First, the believer does not carry with him any old beliefs from the normal world. Therefore it seems that we have made a clear distinction between responsibility and truth. Second, epistemic value pluralists think that if we prefer the first world it is because the believer in this evil-demon world is virtuous in her intentions compared to the second person. 114 The former believes that belief should be gained in a responsible way and therefore we should credit her. We should focus on the question why the evil-demon thoughtexperiment makes us think that responsibility and true belief diverge. It is peculiar that we should attribute a positive epistemic status to the deceived person. On the one hand we should credit the deceived person for being a responsible believer; on the other we should be indulgent with respect to the outcomes of her inquiries, because the evil demon is responsible for it. This strikes me as unconvincing. III Given this scenario, we can even speculate that the consistency or, more generally, the accuracy of the belief system of the deceived believer is based on the set of old beliefs. Without that set it becomes questionable how the deceived believer can have a stable belief system at all. 112 Duncan Pritchard made me aware of this option. 113 Trent Dougherty has proposed this modified version. 114 Epistemic value pluralists take it for granted that providing more details about the case will not change much. They think that if we are given these options it clearly seems that we prefer responsibility over truth.

94

Suppose that the evil demon is lazy and only interferes when the performance of the deceived person does not result in false beliefs. To put it differently, the demon allows the deceived person to behave epistemically badly. This would give the deceived person the "right" to perform epistemically badly in the evil-demon world as well, while being responsible for those performances. This modification of the evil-demon case has the following upshot: The deceived person is responsible for her performance when her performance results in a false belief, as long as the demon has not intervened. Filling in the details, epistemic value pluralists fail to detach responsibility from truth, here. liS For this reason, epistemic value pluralists do not succeed in making their case. Epistemic value monism can respond to the new evil demon in the following way: Initially, the thought-experiment brought us to the conclusion that the deceived person is not responsible for any falsehood she believes because the evil demon interfered, but is responsible for her intention, which is good. Thus, the deceived person is praiseworthy because she is a responsible believer. But things look different if we allow the believer to operate when she is mistaken, and let the demon interfere for to distort her true beliefs. The demon is only accountable for turning the true beliefs into false beliefs, while the believer herself is responsible for the false beliefs, which she has acquired without being meddled with. This response profits from the fact that the evil-demon case is not well described concerning the question who causes which shortcomings. I 16 In other words: the evil demon must only intervene if Kathy does not behave according to his plans. In sum, the truth goal nonetheless plays a role even in that scenario, even though it is a negative one. Let us now turn to the third case.

115 The linkage between responsibility and truth will again be a topic in Chapter 6. There I give a systematic argument that shows the linkage between epistemic responsibility and truth. 116 I am indebted to Christoph Kelp for pointing out this option (see Kelp 2008). He has produced a similar example with the evil demon world that shows the problems with the claim that safety, in Sosa's sense, is a necessary condition for knowledge.

95

6. EMPIRICAL ADEQUACY

The third case originally stems from the philosophy of science, specifically from the field of confirmation theory. Jonathan Kvanvig attempts to make a connection between van Fraassen's concept of empirical adequacy and the concept of belief. Van Fraassen (1980) suggested that empirically adequate belief, rather than truth, is the criterion for the acceptance of a theory. For him, we accept a theory because it is empirically adequate, not because it is true. Following van Fraassen's lead, Kvanvig thinks that we can replace true belief by empirically adequate belief. This concept is supposed to introduce a second epistemic goal that is independent of truth, created by drawing an analogy between the concept of belief formation and the concept of an empirically adequate theory. Kvanvig gives a description of how he wants to understand empirically adequate belief: An empirically adequate theory is one that will never be revealed to be false in virtue of some false prediction it makes about the course of experience, and so no matter what use we make of our beliefs, we will be just as well off employing empirically adequate beliefs as employing true ones. (Kvanvig 2003, p.39)

Kvanvig's attention to constructive empiricism probably stems from van Fraassen's claim that we will never attain truth without empirical adequacy. This can be made to look as if van Fraassen were defending the claim directly opposed to epistemic value monism, that is, if we take it that empirical adequacy is a goal always comes prior to truth in testing our theories. I 17 We do not need to go into a detailed exegesis of van Fraassen's position here. Our debate is with Kvanvig. The problem is that Kvanvig's strategy is to take the alternative epistemic goal to be the primary goal without showing that the truth goal is missed, because he missed to give a plausible case.

117 Van Fraassen, by contrast, defends alternative monism.

96

Instead, his intent is to prove that empirical adequacy is independent of truth. But Kvanvig rules out a pluralism of epistemic goals, which leaves us then with a monism in which inquiry is always ultimately aimed at empirical adequacy.1l8 Thus, Kvanvig's picture goes like this: empirical adequacy of scientific theories is independent of truth, but functions as the Jl9 primary epistemic goal as regards the confirmation of theories. For all intents and purposes, Kvanvig has simply replaced the monism of truth by what we might call "empirical adequacy monism". I shall make a further critical remark. Kvanvig is obviously confronted with the problem of whether he can in all cases replace true belief by empirically adequate belief. But there are often cases where the propositions involved seem not to involve the empirical at all, like those concerning mathematical propositions. 1fT believe that two plus two is four then I can ascertain the truth of that belief by considering the laws of arithmetic. The crucial point is that the formation of this belief is independent of empirical adequacy. To be consistent, Kvanvig must embrace a psychologism about the foundations of mathematics, a position that has not fared well, historically. To sum up: All three cases, I take it, cannot establish what epistemic value pluralists need. Therefore I conclude that truth has endured the test of these objections: we can safely conclude that it is our primary epistemic goal in the prima facie sense. That is, as long as there is no counter120 In the last example, we can further accept the primacy claim is valid. section of this part I want to discuss an argument which I call the "multiple-goal argument". This argument aims to defeat the primacy claim as well.

118 In Chapter 6 1 discuss alternative monism in greater detail. 1191 owe this observation to Wolfgang Spohn. 120 At this point we should distinguish between ultima facie and prima facie claims. The former claim is the stronger one, because it commits epistemic value monists to giving an actual proof for the truth-relatedness of every alternative epistemic goal. Whereas the latter claim is valid as long as there is no good counter-example available.

97

7. FINAL REMARK In this section I discuss the multiple-goal argument, which will round up how one might argue against the primacy-of-truth claim. The argument is the following: (The multiple-goal argument) Truth is not the primary epistemic goal because we have so many different goals.

Richard Feldman, for instance, uses this argument. He writes on the question whether believing all existing truths can be our goal: One idea commonly extracted from this is that our epistemic goal is to believe all and only truths. But if this implies that all people actually have the goal of believing all truths, then I doubt that it is true. People's goals are a varied lot, and I doubt that all people have this one. Moreover, believing all truths is obviously an unattainable goal. We simply can't believe all the truths. Furthermore, attaining it is not desirable. (Feldman 2004, p. 182; emphasis added)

Feldman uses the multiple-goal argument to scupper the idea that believing all truths can have any epistemic value. 121 Asking someone to believe all the truths there are (and only the truths) is for Feldman an exaggerated requirement. He is right, but from that it does not follow that one must endorse the multiple-goal argument. Richard Rorty uses the multiple-goal argument in a slightly different version: [Inquiry] has many different goals, none of which has any metaphysical presuppositions. (Rorty 1995, p. 297)

We can reformulate Rorty's argument in one sentence: Truth cannot be our primary epistemic goal, because we have many other goals. So the primary of the one goal, truth, seems to be swamped by the plurality of the other 121 It is unclear whether there is any philosopher who has defended the view that believing all truths is a factual goal of all human beings. Usually this maximalist claim is spelled out as "believing as many truths as possible".

98

goals we can think of. This argument is ambiguous, however, in so far as the "because" linking the claim to the conclusion is not quite clear. Does the argument imply, that our resources for attaining our goals are so limited that we don't have time to inquire after many significant truths? Or does the argument exploit the fact that human beings might have difficulties in choosing among several goals? It is clear that inquiry can serve many different goals, a fact never denied by epistemic value monists. Consider research on cancer. A successful theoretical inquiry may end in providing the framework for proper treatment. However, no cancer researcher would ask for research funds by saying that he was uninterested in the truth, just in finding a proper treatment for cancer. Nor would his actual research about cancer not rely on the pursuit of truth, even if the goal of his behavior is to add to the social welfare by curing cancer. Remember, our debate is about epistemic values; epistemic values are, of course, not the only values. How should that affect our debate?

The multiple-goal argument brings with it another unwelcome consequence: If we were to accept the argument, then knowledge could not be the goal of inquiry either, because there are so many different goals. Justification too could not be our primary epistemic goal. And we could go on: The goal of confirming a scientific theory or of having a wellconfirmed theory would be undermined, because it would be swamped by the many different (practical) goals. Read like this, the multiple-goal argument is clearly unsound, because it would sweep away all epistemic goals. There is something in the argument that may seem initially convincing, but that cannot be right in the end. I think the correct diagnosis of the multiple-goal argument is that it misconceives the fact that human beings have time to engage in theoretical inquiry among other activities in their time budget, like feeding themselves, amusing themselves, having love affairs, etc. What this means is that there may be many practical and emotional goals in a life; but if there is epistemic activity that takes the form of inquiry, truth will be the goal of that inquiry, even in the midst of the inquirer's many different practical goals. There's nothing to prevent a philosopher from working on a theory of truth while also pursuing many practical goals beside his

99

research. So we can say in a general way that theoretical goals are compatible with practical goals. The compatibility between those goals is not necessarily threatened by the multiple-goal argument, because practical goals need not interfere with the truth goal. Practical goals clearly can override theoretical goals. But as long as the truth goal is not challenged by practical goals, it stays intact. In sum, I have shown the various deficiencies in the alternativeepistemic-value objection, which make them not so attractive as they may first appear. 122 In particular, the notion that understanding should be a stand alone epistemic goal has attracted much philosophical attention, much of which derives from casting the epistemic value monism position as one in which all truths, unselectively, are given as the goal of our inquiries. Once, however, we see that a criterion for sorting significant and insignificant or irrelevant truths can be derived from the truth goal, the attraction of understanding becomes much less. In the next chapter I will discuss more systematically whether alternative monism provides any striking objections.

122 The discussion in this chapter has benefited from invaluable comments by Duncan Pritchard, Trent Dougherty, Wolfgang Spohn and the audience of my talk at the Graduate Conference "Understanding and Knowledge" at the University of Edinburgh in November 2007.

100

6. Alternative Monism In this chapter I shall evaluate another challenge to epistemic value monism, one which we briefly touched on in Chapter 5. This challenge consists in replacing truth by another goal as the primary epistemic goal, which is something like what Kvanvig seems to be after with 'empirical adequacy'. I call this move "alternative monism". It keeps the monistic ~tructure of epistemic goals, but puts in place of the 'truth' some other claimed fundamental epistemic goal. This is what I call the objection fronl alternative monisnl. In short, the objection is: truth is not the primary epistemic goal, because another goal is the primary epistemic goal. What alternative goal can take this position? We have seen how an alternative monism can be sketched out in the domain of confirmation theory. There, justified belief is taken to be a more fundamental epistemic goal than truth. Or, to put it in other words, a believer who holds a justified belief has already obtained a positive epistemic status. 123 For the alternative monist, forming one's belief in a rational way is the goal that replaces the truth goal. 124 In the literature, rationally formed belief is called "rational belief' .125 The adjectives "rational" and "responsible" are synonymous here. Strictly speaking, beliefs themselves cannot be rational or responsible; it is the believer who is rational or responsible. "Rational" and "responsible" also describe a property that we attribute to the believer when we ask how he acquired his beliefs. As we will show, there are two different proposals for spelling out those standards of responsibility: first, a believer is responsible if he grounds his belief on the available evidence. I call this answer the

123 I use the term "positive epistemic status" in the sense that a believer obtains positive epistemic status if he has obtained something of epistemic value. Greco (1993, p. 415) has presented a more detailed characterization of this term (see also Plantinga 1988). 124 I leave out Wayne Riggs's position since his conception is too broad to fall within the conception of epistemic rationality. He argues that wisdom is the primary epistemic goal (see Riggs 2003b). 125 See, for instance, Miller (1995).

101

"evidentialist standard".126 Second, a believer is responsible if he forms his believe in accordance with his own intellectual standards. I call this the "egocentric-rationality standard" .127 I argue in this chapter that the evidentialist standard and the egocentric-rationality standard both support alternative monism. Then I will go on to argue that alternative monism is false, because alternative monism is not capable of generating an alternative goal that is more fundamental than the truth goal.

I. ALTERNATIVE MONISM What are the motives behind alternative monism? Epistemic value monism is committed to the following value thesis: If truth is our primary goal then all other goals have merely instrumental value, insofar as they help to attain that primary goal. This line of thought does not accommodate the intuition which the evil-demon case is designed to show. There, we concluded that a person could obtain positive epistemic status just by acquiring a justified belief. If this is true then justified belief is valuable in itself. This affects the structure of epistemic goals. If there is an epistemic value, such as justified belief, that does not have instrumental value with respect to truth, then we have generated an independent goal within the epistemic domain. Goldman describes the situation as follows: An obvious challenge to veristic unitarianism [in our terminology, epistemic value monism] arises from the fact that, on everyone's theory, justified belief is a distinct state of affairs from true belief but a preeminent example of an epistemically meritorious or valuable state of affairs. Unless the admittedly distinct state of justified belief can be shown to have its value in some derivate fashion from the value of true belief, veristic unitarianism is in trouble. If justificational status has to be posited as a value entirely autonomous and independent of truth, it looks like pluralism wins the day. Or perhaps justifiedness could even replace true belief as the core epistemic value. (Goldman 2001, pp. 31-32)

126 See Conee/Feldman (2004). 127 Foley (1987,1993,2001) has defended such an account.

102

While Goldman uses a different terminology from ours, his argument is helpful in analyzing the issues here. If, as we all agree, we can differentiate between the state of holding a justified belief and the state of holding a true belief, then why shouldn't the state of holding a justified belief have a value independent from whether that belief is true? Alternative monism and epistemic value pluralism share common ground here. 128 The late Roderick Chisholm endorsed a rather enigmatic form of alternative monism by re-describing the role of epistemic justification: I have previously written, incautiously, that one's primary intellectual duties are to acquire truth and to avoid error. What I should have said is that one's primary intellectual duties are to believe reasonably and to avoid believing unreasonably. (Chisholm 1986, pp. 9G-91; emphasis in original)

Chisholm indicates the existence of a more fundamental duty than that of seeking the truth, and therefore seems to point us to a more fundamental epistemic goal (reasonableness) than the truth goal. We should reformulate the basic argument in a semi-formal way, since the objections to the instrumental-value assumption will occupy us in the following chapters as well: (VI) If there is an epistemic value that is not merely instrumental valuable with respect to truth, then there exists another (stronger version: an even more) fundamental epistemic goal. (V2) Justified belief has non-instrumental value. (Conclusion) There exists another (stronger version: an even more) fundamental epistemic goal.

Again, epistemic value pluralism and alternative monism partly share common ground. The difference between the two positions is that the latter must defend the stronger version of (VI), since, as a monism, it supposes that the truth actually serves the alternative goal. Alternative monism introduces an alternative epistemic goal that is supposed to be more 128 Although it should be noted that alternative monism has the longer history. I am grateful to Trent Dougherty for making me aware of this fact.

103

fundamental than the truth goal. 'Fundamental' means in our context: If a goal is fundamental, it can't be reduced to another epistemic goal. Alternative monism attempts to introduce a fundamental goal that is not reducible to truth. I will call this "the minimal epistemic goal":

(Minimal epistemic goal) A believer fulfills the minimal epistemic goal if he forms his beliefs rationally.

The main idea of the minimal epistemic goal is a formal idea: the agent satisfies the minimal epistemic goal by following the criteria for rational belief formation. I call it minimal in the sense that it is more fundamental than the truth goal. Alternative monists need to defend a strong version of justification internalism. Presumably, the believer can more efficiently attain the goal of forming his beliefs rationally than would be the case under the truth goal. The standards of rational belief formation leading to correct beliefs are directly controllable by, or at least accessible to, the believer, whereas the actual obtaining of truth or knowledge is less . accessible to the believer. 129 So far we have not said anything specific about what requirements are involved in forming beliefs rationally. In the following sections I will propose and discuss two of them, both of which incorporate the minimal epistemic goal.

2. EVIDENTIALISM AND RATIONAL BELIEF The first proposal is from Conee and Feldman. They take the evidentialists' doctrine as the central component of any type of epistemie justification and justified belief. 130 Roughly, the doctrine states that we should not believe anything without basing our belief on evidence. I will 129 See, for instance, Wedgwood (2002, p. 355). 130 See their co-authored book Evidentialism (OUP 2004). Feldman (1988, 2001, 2002) has published several papers specifically discussing what epistemic obligations are. Those papers are relevant for our discussion as well.

104

call the norm to believe in accordance with the evidence "the evidence norm". Conee and Feldman have described this doctrine in much greater detail : (EJ) Doxastic attitude 0 toward proposition p is epistemically justified for S at t if and only if having 0 toward p fits the evidence S has at t. (Conee/Feldman 2004,p.83)

It appears that Conee and Feldman think that (El) is sufficient for all types of epistemic justification, because we obtain a justified belief if we have evidence for it. This is an interesting issue which I can not deal with 131 here.

There are several terminological differences between (El) and our rough sketch of the evidentialists' doctrine. The concept of belief was replaced by the technical term "doxastic attitude toward a proposition". Additionally, Conee and Feldman left unspecified how they understand the 'fitting relation' between the evidence and the proposition and it is unclear how to specify 'attitude D toward p' except that it happens in time. Neither did they further qualify what counts as evidence. We can leave these questions about (El) aside, however, since it is more to our purpose to understand how Conee and Feldman evolve the alternative goal of rational belief formation as the primary goal. (El) gives a requirement for how to form our beliefs rationally. We form a belief rationally if we follow (El), and we might obtain true belief by following (El). Additionally, (El) provides a criterion for when beliefs are formed rationally. Following the evidence is a central characterization of rational belief formation. Feldman states: The current view [that is, evidentialism] is that reasonable beliefs are epistemically valuable and that following one's evidence is a perfect means to getting valuable beliefs. (Feldman 2004, pp. 185-186) 131 Another issue in (EJ) is what 'fits' is supposed to mean in this context. We will sidestep this issue as well. (EJ) faces the central problem of transferring justification to propositions that are entailed by the proposition initially justified. So (EJ) allows justification to be transferred without further checking if the old evidence also fits q, r, s and so on (if q, r, s are implied by p). However, I will not go into the details of this problem.

105

Feldman defends the idea that having a justified belief has intrinsic value. 132 There are two ideas that motivate Feldman's account. The first is that any person who follows (EJ) should be credited. And second, Feldman thinks that the truth goal sets standards too high for the ordinary believer. Here is a statement of the first idea. The following passage is worth quoting in full: First, suppose the person follows his evidence. [...] this person believes as he ought. But according to the earlier view [that is, truth monism] his beliefs, being false, lack epistemic value whereas according to the current view [that is, evidentialism] they are epistemically valuable. Second, suppose that the person does not follow his evidence. In that case, both views imply that he does not believe as he ought, but the earlier view implies that he is, by luck, achieving epistemically valuable beliefs. In both cases, I find the implications of the current view more appealing. I don't see anything epistemically good about the person who irrationally gets true beliefs. I don't think that it would be correct to say of him that he's achieved epistemic excellence, even though he's done it in an irrational way or merely by luck. Rather, I think he's failed epistemically, not only because he doesn't believe as he ought but because he does not have rational beliefs. (Feldman 2004, p. 186)

Feldman thinks that the motives for acquiring beliefs are important for our epistemic evaluation as well. If the agent has intended and actually acquired justified belief, this is praiseworthy, Feldman thinks. As we have seen in our discussion of the evil-demon case, many epistemologists think that we do not evaluate a belief system just in as much as it is true, but that we want to know about the believer's path to these truths. If, as Feldman describes them above, the truths are acquired by luck, or in an epistemically irresponsible way, we do not credit the status of the believer. So Feldman shares the intuition epistemic value pluralists have brought forward in the evil-demon case. The second source of motivation concerns limited human capacities. The best we can do epistemically, according to Feldman and Conee, is to 132 Conee agrees with Feldman, because he refers to the same conditions used by Feldman in his paper "The Ethics of Belief'. See the afterword to Conee's paper "The Truth Connection" (2004, pp. 254-258). Feldman and Conee defend a specification of the requirement of forming beliefs rationally: (V3) "When adopting (or maintaining) an attitude toward a proposition p, a person maximizes epistemic value by adopting (or maintaining) a rational attitude toward p" (2004, p. 257).

106

follow our evidence or "gaining the doxastic attitude that fits one's evidence" (Conee/Feldlnan 2004, p. 89). So Conee and Feldman translate "trying one's best episten1ically" as trying to have a doxastic attitude fitting the evidence. For them, the truth goal is exaggerated, because it would require every subject to believe all existing truths. 133 In the next section I shall investigate whether (El) provides a general norm. I will put (El) to the test to see whether the minimal epistemic goal does set fundamental standards. My suspicion is that the minimal epistemic goal lacks the characteristic of a fundamental goal, because it does not give us general guidance, as the truth goal does.

2. IS THE EVIDENCE NORM A GENERAL NORM? In this section, I will argue that (EJ) is context-dependent and therefore can't be a general guide. This result will become obvious when we attempt to apply the evidence norm to a realistic case. Let the situation be the following: A wife who has been happily married for several years finds a snapshot of a woman in the pocket of her husband's shirt. Both the wife and the husband agree that their marriage is founded on mutual trust. Recently, however, their relationship has been rockier. What ought the woman to believe in respect to the photo? Is this evidence of her husband's relationship with another woman? The question is ambiguous, since there are several oughts in play, here. We can differentiate the oughts by correlating them to the goals they serve. 134 For instance, the woman may hold the goal of happiness, and feel that she would be less happy if she entertained the thought of her husband's faithfulness. For this practical reason she could believe that the photos in his pocket is there by accident. Thus she could dismiss the pang of worry 133 Conee and Feldman understand Roderick Chisholm in such a way (see Conee/Feldman 2004, p. 88). Feldman argues that believing all truths cannot be an epistemic goal: "Moreover, believing all truths is obviously an unattainable goal. We simply can't believe all the truths. Furthermore, attaining it is not desirable" (Feldman 2004, p. 182). 134 Moser (1985, p. 221) also proposes such a reformulation of "ought" with respect to epistemic, moral and prudential goals (see also Chrisman 2008).

107

about the photo as silly in the face of her old belief about her husband's faithfulness. Let us leave those moral and personal grounds aside. We can interpret our case in two purely 'epistemic' ways. In the first interpretation, the wife comes to the conclusion that her husband may be seeing another woman, because she takes the photo as evidence for this new belief. The recent state of the relationship serves to adjust and reinforce the interpretation of the new evidence. The new evidence has brought into the open what has bothered the woman about the rocky phase of the marriage. In the second instance, the woman excludes her worries about the recent state of the relationship and ignores the evidence of the photo, since it is not sufficient to hang any particular interpretation upon. There are various possible explanations of how the photo could have gotten in the pocket of the shirt. In this case, the wife registers the photo but sticks to her old belief. In this case, the old evidence - the accumulated experience of her life with her husband - outweighs the new evidence - the photos in the pocket of the shirt. The wife in the second interpretation is not irrationally denying the photo exists. Rather, it does not infer anything about her husband that would make her change her mind about the what the old evidence infers. I claim that in both interpretations only epistemic reasons are in play. Additionally, in both cases the wife holds a justified belief, because she believes in accordance with the evidence. We can explain the difference between the two cases on the grounds that in each one the woman determines what counts as evidence of her husband's behavior for her. The evidentialists' doctrine in (EJ) does not provide constraints such that we can see what counts as 'evidence for' in general. The wife can decide according to her own comfort level, her own anxieties, her own mood what will count as good evidence for her. This result sheds a dubious light on the evidence norm because we can't use it to foreclose on contradictory outcomes: the belief that results from the first interpretation contradicts the belief that results from the second interpretation, but each is equally rational. According to Conee and Feldman, the wife would be subjectively justified in believing that her husband is cheating on her, and she would

108

also be subjectively justified in believing that he is faithful. In both interpretations, the wife has gained her belief in accordance with the available evidence. If this is right, Conee and Feldman face an unwelcome consequence: How can the evidence norm give good guidance if it allows two opposite interpretations of one and the same case? The suspicious wife case quite convincingly shows that believing in accordance with the evidence is a context-dependent process. Therefore, the evidence norm cannot serve as a general epistemic norm. In comparison, the believe-the-truth norm is universal. 135 In the case of the suspicious wife, it would be best for her to suspend her belief concerning her husband's infidelity as long as there are no more decisive grounds. If the wife wonders, is my husband cheating or not, her question is not aimed at finding evidence per se, but about whether he is in fact cheating. The truth is her goal. This is what would have been needed in order to complete the epistemic evaluation in the cases discussed above. So truth does enter the picture in order to allow a complete epistemic evaluation. In sum, the evidence norm is not a general norm, since evidence is evidencefor something. What it is for can't be regulated by, or read off of, the evidence alone. What we ought to believe in all possible cases is what the evidence tells us - some approximation of the truth. Conee and Feldman might respond the following in defense of their account: The suspicious wife did not follow the evidence norm appropriately, since there was not enough evidence available to believe anything in particular at all. This response would require the wife to believe at most only in the possibility of the infidelity, because one should not believe on the basis of insufficient evidence. This response can only be viable if Conee and Feldman can introduce a threshold above which we have sufficient evidence to support our belief in a given case. Conee and Feldman, however, do not provide such a threshold. In sum, (El) faces a situation where two opposing attitudes toward the evidence are possible. Timothy Williamson has developed another case that questions 135 This statement precludes so-called relativism about truth. See Kolbel (2002) or McFarlane (2005) for a defense of such a theory of truth.

109

evidentialism. For him, it is rational to avoid evidence in certain circumstances. I bring this case intor our discussion as well, since it reinforces my proof that the evidence norm can't be a general epistemic norm. Here is the case Williamson suggested to Feldman: If a person finds a drug or a machine that can erase memories from his brain and arranges to be immersed in a sensory deprivation tank, he'll have very little evidence regarding anything. By believing very little, he'll then be highly rational according to evidentialist standards. (Feldman 2004, pp. 189-190)

Williamson's line of argument differs from mine. Williamson thinks that in this case the evidentialists' doctrine invites agents to form few beliefs, or to suspend judgment period. In doing so, they would fulfill (EJ), even though we would balk at calling such a strategy rational. (EJ) seems to encourage us to make our belief system artificially smaller. What is Feldman's answer to Williamson's case? He says that how the believer came to put himself in a state of evidential deprivation had nothing to do with our epistemic evaluations, since this is "irrelevant to this epistemic fact" (Feldman, 2004, p. 190) of having a justified belief. That means in short that we need not know how the justification has come about. This seems odd, since the whole brunt of evidentialism would seem to emphasize how important it is that the believer came to obtain justified belief in a responsible way. Thus, in the case of a believer who obtained truths in an irresponsible fashion (for instance, through luck), we are to withhold credit, but now, with respect to the Williamson case, the generic history perspective should not be taken into account. This strikes me as · 136 contradlctory. 136 Williamson's case allows a different reading as well. We might think of a very weak believer who decides to use a sensory deprivation tank in order to save his fragile belief system from collapse. If we give Williamson's story this twist, the agent is rational with respect to his clear and honest self-evaluation. This ought to be credited. lowe this idea to a suggestion by Christoph Kelp. Evidentialism faces a two-tiered task if the evidence norm is supposed to be a general norm. Evidentialists have to explain (I) when it is appropriate to neglect evidence or to be indolent towards it, and (2) when it is appropriate to rely on the given evidence. In connection with the second task, evidentialists must indicate how much evidence is enough evidence.

110

In conclusion, I have attempted to show that the evidence norm that is explicated in (EJ) does not generally provide the guidance for inquiry we want, because, as in the case of the suspicious wife, it gives contradictory outcomes, both of which fulfill the norm, that it can't judge between. In comparison, believing in accordance to the truth norm avoids such cases since whatever the truth is, you have to believe the truth. If it is not clear what the truth is at a given time in a given circumstance, it is reasonable to suspend one's belief and decide whether or not to investigate further. Evidentialism faces a two-tiered task if the evidence norm is supposed to be a general norm. Evidentialists have to explain (1) when it is appropriate to neglect evidence or to be indolent towards it, and (2) when it is appropriate to rely on the given evidence. In connection with the second task, evidentialists must indicate how much evidence is enough evidence. In sum, if the evidence norm were derived from a minimal epistemic goal and it happens that the derived norm is not general, then this makes the status of the alleged minimal epistemic goal doubtful. So far I have argued that evidentialism attempts to introduce a more fundamental epistemic goal. I have shown that this minimal epistemic goal is not the primary epistemic goal since the evidence norm is not general. In the next section I will introduce another proposal for how the minimal epistemic goal could be further explicated. 3. FOLEY ON RATIONAL BELIEF Richard Foley's epistemological theory is quite distinct from others. 13 ? He wants to construct a theory of epistemic rationality that would embed it in the context of a general theory of rationality. This introduces a certain tension in Foley's position, which nevertheless is treated by many epistemologists as a species of truth monism because, in several passages of his work, he declares that he believes the truth and avoiding error to be the central epistemic goal. Marian David has made a similar observation when he states:

137 See Foley (1987, 1993,2001).

III

As Foley himself emphasizes repeatedly, our primary intellectual goal is to believe the truth. And this goal is built into the notion of egocentric rationality by definition. But, as Foley also emphasizes, when we reflect about egocentric rationality, we realize that there is no particularly good reason for thinking that being egocentrically rational will promote the goal of believing the truth one bit. One cannot help but feel that there is something odd about this situation. (David 1996, pp. 951-952)

For Foley, a justified belief is formed by following what he calls "egocentric rationality". Egocentric rationality provides the standards for justified belief (independent of the truth goal). Interestingly, Jonathan Kvanvig gives Foley's work a different twist. He seems to reconstruct Foley as an epistemic value pluralist:

Though Richard Foley wholeheartedly endorses the standard epistemic goal of getting to the truth (now) and avoiding error (now), his theory doesn't require this teleological aspect [that is, the instrumental relation between justification and truth]. On his theory, a belief is justified if and only if it conforms to one deepest epistemic standards. These standards are epistemic principles one would endorse given as much time to reflect as is needed to reach a stable point of view. Once we have such internal standards playing a theoretical role, real deontologism can emerge, for it is in conforming to these standards that one's epistemic duties are satisfied. (Kvanvig 2005, p. 294)

Kvanvig presents Foley as an epistemic value pluralist because for Foley, the truth goal is just one goal besides others. Justified belief has an independent epistemic value, since justification is not directed at truth. If Kvanvig's interpretation is right, Foley defends an incoherent position. My reconstruction focuses on what Foley calls egocentric rationality. 138 In the course of this section we will see that the more details we obtain about egocentric rationality, the more it will appear that Foley supports alternative monism. 139 In a second step, I will check whether we can derive epistemic norms from Foley's account of egocentric rationality. 138 I do not claim to give a complete picture of Richard Foley's complex theory. I think, however, that Foley has made various statements in his later work which support alternative monism. 139 See Foley (200Ia, 2004, 2005, 2006). Foley (1994) contains his answer to the question of what we ought to believe.

112

Thereafter I shall put those derived norms to the test again.

4. WHAT IS EGOCENTRIC RATJONALITY? What is a justified belief for Foley? The answer we have seen above is: we justify our beliefs in accordance with our deepest epistemic standards. Foley explains this in more detail: 1 am to make up my mind by marshalling my intellectual resources in a way that conforms to my own deepest epistemic standards. If) conduct my inquiries in such a way that 1 would not be critical of the resulting beliefs even if 1 were to be deeply reflective, then these beliefs are rational for me in an important sense, an egocentric sense. There are various ways of trying to spell out exactly what this amounts to, but for our purposes here the details can be left open. The basic idea is if 1 am to be egocentrically rational, I must not have internal reasons for retraction, ones whose force I myself would acknowledge were 1 to be sufficiently reflective. (Foley 1994, p. 148)

What is the role of the subjective component? And, more importantly, what is Foley's motivation for giving this subjective component such a central position in his theory? The short answer is: For Foley, rational belief is independent of truth because in the case of the evil demon the believer is able to believe rationally while being prevented from attaining truth. 140 A second component of rational belief, Foley thinks, is that we should form beliefs in the light of our own reason. 14 ! For Foley, rationality can't be identified with "adhering to the opinions, rules, practices, or presuppositions of one's community or tradition.,,142 He uses examples from the history of science to show that the "rebels" and the "mavericks" held extreme beliefs that

140 Foley (200 I, p. 38) says about such cases: "It deprives you of knowledge, but one of the lessons of such thought experiments is that being thoroughly deceived does not also automatically prevent you from being rational in one important sense, a sense that is closely associated with how things look from your perspective." 141 Foley (1993, p. 106) thinks that in general we form many beliefs automatically. But in cases where our reflection is needed, this process should be guided by one's own standard of rationality. 142 See Foley (2001, p. 44).

113

turned out to be true. The basic idea is the following: a belief is justified qua being widely held in a community. Extreme beliefs can be rational beliefs (even if they are in extreme opposition to mainstream). Taking this stance, it is clear why Foley comes to following conclusion:

not still the the

Being rational in this sense is instead a matter of making oneself invulnerable to intellectual self-criticism to the extent possible, of living up to one's Own deepest intellectual convictions and standards. It requires one to have opinions and to use faculties, methods, and practices that are capable of withstanding one's own most severe critical scrutiny. Accordingly, one can be rational in this sense even if one is a brain-in-a-vat whose opinions are massively unreliable, and likewise one can be rational even if one's opinions, methods, and assumptions are massively at odds with those of one's community, era, and tradition. (Foley 2001, p. 39)

For Foley, the believer as rational thinker is strong enough to maintain his idiosyncratic belief system and his perspective on the world if he believes the world - whether that of the evil demon or that of tradition - is irrational. The greater problem he sees as coming from the believer's social environment: [...] our primary intellectual threat is not that of chaos, and our primary intellectual need is not for advice about the most fundamental matters of intellectual outlook. We cannot help but be largely guided by our intellectual inheritance on these matters. The primary threat is rather that of intellectual conformity, and our primary need is for intellectual autonomy. Little in life is more difficult than resisting domination by one's intellectual environment. It is all too easy for us to be intellectual lemmings. (Foley 1994, p. 160)

So the individual develops his intellectual autonomy by strengthening his ability to reflect in accordance with his own deepest epistemic standards. Where is the halting point for one's own reflection? As we know, reflection-is potentially endless, whereas believing something is a definite outcome of a process. Foley believes that we can reach a state of "reflective stability,,143 by 143 See Foley (1993, p. 99).

114

being invulnerable to self-criticism. 144 This state, he thinks, has a distinct epistemic value, because the believer has formed his beliefs in accordance with his egocentric rationality. Foley provides another reason for taking egocentric rationality as a standard. He objects that reliabilism doesn't provide us a priori with a guide to which methods are reliable and which are not. 145 For this reason he rejects the reliabilist idea that we form a belief rationally if we decide in accordance with n1ethods that reliably produce true belief. I do not want to evaluate Foley's objection here, but rather note this as another motivation for'taking egocentric rationality as one's standard. At this point it is helpful to understand how Foley thinks that egocentric rationality differs from Descartes's project of pure inquiry.146 For Foley, unlike Descartes, the cogito gives us no guarantee of having obtained a truth. Egocentric rationality doesn't tell us whether we should be optimistic about this missing truth guarantee, since, we are still able to obtain truth; or pessimistic, because, lacking this certainty, we must abandon all hope of finding the truth-preserving reasons. Foley stands between these two attitudes since, according to him, we can't rely on the facts to tell us whether we have attained truth or not; but he is optimistic enough to create an evaluative standpoint which is accessible to the believer and which gives him intellectual autonomy. Foley describes this as follows: [...] egocentric rationality brings with it no guarantees of truth or likely-truth, and as result it brings with it no guarantees that rational people with access to the same information will agree with one another. Why, then, should we be interested in egocentric rationality? Because we are interested in having beliefs that are accurate and comprehensive and because by being egocentrically rational we will be pursuing this end in a way that by our own light seems effective. To be sure, this involves a leap of intellectual faith. (Foley 2003, p. 190)

How does Foley's egocentric rationality regulate our belief formation? Or, 144 See especially Foley (2001, Ch. 2). 145 See Foley (1994, p. 148). See also Foley (1985). 146 See Foley (1993, pp. 98-99, 122-123).

115

what is it to believe something on the grounds of egocentric rationality? We should be able to regulate our beliefs even if we do not have fixed a foundation. To see this, consider, again, the suspicious wife. Again we have two possible outcomes to her story - one being her belief that her husband is cheating, the other her belief that he isn't - and we want to test how egocentric rationality would guide her way. In the first version, she believes in the new evidence and, after careful reflection on the plausibility of the photo being deposited there by chance, she comes to the conclusion that her husband is cheating on her. In the second version, she does not take the new evidence to give her a certain inference that her husband is cheating. In both versions of the story the wife has formed her belief rationally, because she can reach the point of being invulnerable to self-criticism. In the first case she concludes: her husband is not faithful. In the second: that there is no reason to belief that her husband is not faithful. Since the wife does not need to justify her conclusion to anyone else, she can think for herself that she has good reasons for her conclusion, and after a certain period of time her belief is invulnerable to her self-criticism. This is not convincing: Foley's account of egocentric rationality provides subjective justification in both cases. Once the truth is on the table, in one of the cases the woman will find that she has not been justified to believe what she did believe. What guiding role can egocentric rationality play if it provides justification where it should counsel one or the other conclusion? In sum, Foley's alternative fails in the same way that evidentialism fails: both lack a requirement with guiding force. In other words, both lack a normative goal - which is what the alternative project sets out to give us. As long as truth is not disclosed, we have no normative sense of what the suspicious wife (epistemically) ought to believe. Alternative monists attempted to introduce a more minimal epistemic goal upon which we can build the inquiry for truth. But this proposal doesn't work precisely because it lacks the truth goal. As long as the truth is not tracked down, it is rational to suspend one's belief and wait until the evidence gives one guidance toward the truth. We cannot derive this basic epistemological move from

116

either evidential ism or egocentric rationalism. This explains why alternative monism is mistaken, because truth is so fundamental a goal that it cannot be replaced by any other epistemic goal. My intention was to focus on the purely epistemic side of the case. In sum, I think we must vote against giving primacy to the so-called minimal epistemic goal because the norms which we can derive from the requirement of forming our beliefs rationally are not general, and can't give us general guidance when we need it. Lacking this goal-likeness, we find it more reasonable to say that evidence or rationality, however construed, is instrumental to the truth, rather than vice versa. On the other hand, the demise of these alternatives might seem to aid and comfort epistemic value pluralism camp. At least, the failure of alternative monism does not rule out the lessons drawn by epistemic value pluralism from the evil demon example, because believing responsibly might still be an additional fundamental goal besides truth. In the following chapter we will, accordingly, plunge into the value problem of knowledge.

117

7. The Value of Knowledge Epistemic value pluralists have quite forcefully pronloted the suspicion that primacy of truth endorsed by epistemic value monism can't be the case because knowledge provides an additional source of epistemic value, which is unrelated to truth. Let us call this "the dual-value objection". If this suspicion is correct, truth will not monopolize the place for fundatnental epistemic values. One suggestion is that knowledge itself is a primary epistemic goal. Knowledge can matter to us in the same way as truth matters to us; maybe it matters even more than truth. The semantic core of the word "epistemic" refers to knowledge. If we take this semantic hint seriously, we face a peculiar situation: how can truth be the primary epistemic goal if the very term 'epistemic' refers to knowledge? Shouldn't the very existence of the knowledge goal be a threat to epistemic value monism? The main purpose of this chapter is to answer that question. I argue that, unless there is something more to the goal of knowing than knowing the truth, the knowledge goal is not problematic for epistemic value monism. What, after all, would inquiry for inquiry's sake be about? In addition, since knowledge entails truth, the two goals are not necessarily different in kind. This chapter has two main parts. In the first part we will take up two accounts directed against there being a separate knowledge goal in inquiry. Let us subsume these accounts under the label "revisionism" .147 Revisionism suggests a revision of the order of our epistemic goals. If revisionism is right then epistemic value monists don't need to worry about the threat of the knowledge goal. Revisionism defends a very strong form of epistemic value tnonism. In the second part of this chapter we will evaluate the dual-value objection that is supposed to show that epistemic value monism is badly affected by the existence of the knowledge goal.

147 Duncan Pritchard introduced the term "revisionism" (see Pritchard 2007a).

119

1. REVISIONISM AND THE KNOWLEDGE GOAL In this section I discuss revisionism, the position that knowledge cannot be our epistemic goal. I focus on two different constructions of the revisionist position. The first is Marian David's. For him, the concept of knowledge is theoretically impotent. By contrast, Mark Kaplan's revisionism maintains that it is sufficient to attain justified true belief, despite the fact that what we want to attain is knowledge. 148 I discuss the two explanations in the given order. For David, the knowledge goal should not be given credit in epistemology, because it does not have a viable theoretical function. Marian David defends a meta-theoretical perspective for the relation between the knowledge goal and the truth goal. He writes:

Invocation of the truth goal serves primarily a theoretical need, a need that arises from the overall structure of epistemology. The knowledge-goal would not serve this need. As far as epistemology is concerned, the knowledge-goal is theoretically impotent. (David 2001, p. 153)

David concedes that knowledge is the central epistemic concept; however, the main component of knowledge is justification. Since epistemic justification, David thinks, is clearly truth-related, truth satisfies the theoretical needs of epistemology. So, making the classic monist's move, David concludes that the central epistemic means, epistemic justification, is directed at truth, which must be the primary epistemic goal. David thinks that the concepts of belief and truth should not be the primary concern of epistemologists, because they are not primarily "epistemic concepts" .149 According to David, epistemic concepts must therefore be grounded in non-epistemic concepts, because the concept of belief and the concept of truth serve as the conceptual foundation for 148 Sartwell (1991, 1992) argues along similar lines. He maintains that true bel ief is already sufficient for attaining the knowledge goal. I do not consider Sartwell here, because his proposal is quite complicated (see Hofmann 2005 and LeMorvan 2002, 2008). 149 See David (2001, p. 153).

120

epistemic concepts. This leaves epistemic justification as the central epistemic concept. I call this the anchor thesis: David attempts to ground the central epistemic concept in non-epistemic concepts. This thesis argues for a division of labor between truth theorists, who construct our concept of truth, philosophers of language, who analyze the concept of belief, and epistemologists, who focus on the concept of epistemic justification. 150 David thinks we have to anchor epistemic concepts in non-epistemic concepts in order to avoid circularity in the analysis of our concepts. How does the anchor thesis affect the relation between truth and the knowledge goal? David's answer is:

Although knowledge is certainly no less desirable than true belief, the knowledge-goal is at a disadvantage here because it does not fit into this picture [of the anchor thesis]. Invoking the knowledge-goal would insert the concept of knowledge right into the specification of the goal, which would then no longer provide an independent anchor for understanding epistemic concepts. In particular, any attempt to understand justification relative to the knowledge-goal would invert the explanatory direction and would make the whole approach circular and entirely unilluminating. After all, knowledge was supposed to be explained in terms of justification and not the other way around. This does not mean that it is wrong in general to talk of knowledge as a goal, nor does it mean that it is bad epistemology to invoke the knowledge-goal as part of the theory of knowledge because it is quite useless for theoretical purposes: The knowledgegoal has no theoretical role to play within the theory of knowledge. (David 2001, p. 154; emphasis in original)

David is right in pointing out the meta-theoretical 'fact' that all central epistemological concepts are linked to truth. Additionally, he states that it is hard to understand how the knowledge goal can be a fundamental goal inside a theory about how we justify knowing what we claim to know, given the conceptual order which the anchor thesis presupposes. David's anchor thesis has two problems, however: First, the metatheoretical justification of the primacy of the truth goal does not have much force, because epistemic value pluralists have a different view of the ISO See David (200 I, p. 153).

121

theoretical needs of epistemology; other criteria are central for epistemological theories. What is prima facie wrong about that? Epistemology covers a wide range of cognitive phenomena. Therefore, it seems plausible that there are many different epistemological theories, which serve different theoretical needs. Second, David's anchor thesis is obscure. David is right in pointing out that actual research in epistemology over the last decades has mainly concentrated on the theory of epistemic justification. So what? Unless David's point is that the division of labor tells us something deep about truth, belief and epistemology. The anchor thesis is obscure if we want to interpret it as a methodological thesis. The methodological thought behind the anchor thesis is the following:

(M) Any philosophical discipline needs to ground its central concepts on other concepts, which are in turn the objects of study in different philosophical disciplines.

The general directive behind the anchor thesis contains three steps: (a) identify the central concepts of your discipline; (b) indicate how those concepts are interrelated with other concepts of a neighboring discipline; (c) declare that we avoid the circularity threat by grounding epistemic concepts in non-epistemic ones. This is the anchor thesis translated in a prescriptive language. Avoiding circularity in your conceptual analysis is prima facie useful. How do we know that certain philosophical concepts actually do have the properties they are said to have? We loosely call a concept epistemic due to its theoretical function in epistemology, but that is not enough to indicate a special kind of property that would distinguish epistemic concepts only. In the end the directive (M) is absurd, since there is nQ meta-philosophy that organizes the different anchoring tasks, because no authority exists in philosophy that regulates the division of conceptual tasks. This holds even if the subdivision of conceptual labor is how David actually described it. This does not entail that the subdivision is justified. Still, there are many people who would prefer that an inquiry end in knowledge and not mere true belief. This preference is more substantial

122

than David admits. This brings us to the second explanation for revisionism. Mark Kaplan (1985) has proposed a revision of the analysis of knowledge and thus also of the value of knowledge. He argues that justified true belief is good enough, because in most cases we obtain justified true belief. Instead of aiming at or wanting to attain knowledge, he says, we should aim at justified true belief, which is equally valuable because we attain it regularly. So Kaplan concludes that we should revise the order of our epistemic goals. If he is right then the knowledge goal is replaced by justified true belief as the central epistemic goal. Revisionism is attractive because of its simplicity.151 We need to separate, however, the question of the attainability of an epistemic goal from its importance. A goal is not important in proportion to its attainability. Quite the contrary, in many cases the things which are harder to attain for us are esteemed more by us for a number of reasons: they may have more scope, they may allow us explain things on more levels, they may address more objections, they may be more enduring, etc. So it might be the case that we value a goal even more which, because of its complexities, we are less likely to obtain. In sum, unlike the alternative monists we examined in chapter 6, who are motivated by questions arising from how we bestow credit for epistemic responsibility upon believers (one road to understanding the primacy claim), the revisionists in this chapter go down the other route, worrying about the internal order that differentiates between the instrumental and the fundamental. Revisionists start with the question of what means are available to us for attaining our epistemic goals. They arrive at the conclusion that the only available epistemic means are the ones that conduce to truth. They infer from this that only truth can be our fundamental epistemic goal. It is legitimate to ask about the availability of means for obtaining our epistemic goals, once we have accepted talk of goals in this context. Yet, we cannot infer that knowledge is not an epistemic goal merely because other epistemic goals can be more easily 151 It is clear that a truth-centered epistemology is more elegant because of the simpler theory construction. See for instance Hofmann (2007, p. 148) who points out that this is an advantage of revisionism.

123

attained. With this result, we have to proceed to the question whether the knowledge goal poses a threat to epistemic value monism after we have disposed of the revisionist objections. As I see it, the mere existence of the knowledge goal is no trouble for epistemic value monism. To see how the knowledge goal is actually problematic we have to consider a further premise. This is the premise that knowledge constitutes an additional source of epistemic value, which cannot be reduced to truth. This brings us to what I call "the dual-value objection". 2. EPISTEMIC VALUE MONISM AND THE VALUE PROBLEM Quite recently, epistemologists rediscovered the value problem of knowledge. JS2 To solve this problem, epistemologists had to explain why knowledge is more valuable than true belief. ls3 Wayne Riggs (2002a) has brought forward an argument against the instrumental-value assumption of epistemic value monism. He argues that epistemic value monists have failed to solve the value problem. Riggs thinks that "knowing that p is valuable in part because constituents of this state of affairs are themselves valuable" (Riggs 2002, p. 95; emphasis in the original).154 For Riggs, knowledge has intrinsic value. If this is right, then epistemic value monism faces the question whether truth is the only source of epistemic value. The basic idea behind Riggs's argument is: If knowledge contains intrinsic epistemic value then there is

152 See Pritchard (2007a, 2007b) and especially Kvanvig (2003). Plato introduced this problem in the Meno. Elsewhere in the literature the value problem of knowledge is introduced as a particular problem for reliabilism. The problem can be stated in the form of the following question: Why should reliably acquired true beliefs be more valuable than merely true beliefs? See, for instance, Jones (1997), Kvanvig (2003) and Zagzebski (2004). Olsson (2007) gives a reply to the swamping problem from the reliabilist perspective. 153 At this point we do not consider another line .of argument concerning the value of knowledge: Knowledge might have a higher practical value than true belief because knowledge is more stable than true belief. Olsson (2007) and Hawthorne (2006, Ch. 1) have explored this argument. 154 See also Riggs (2002a, 2002b, 2003).

124

something else besides truth that possesses fundamental epistemic value. Therefore, epistemic value monism cannot provide the correct standards for doing value-driven epistemology. Riggs's argument rests on two pivotal conceptual presuppositions: First, knowledge is more valuable than mere true belief. It is important to note that Riggs assumes that with the value of knowledge consists of the idea that knowledge must have intrinsic value. Second, any additional epistemic goal provides a distinct source of epistemic value. Particularly, Riggs assumes that knowledge is intrinsically valuable. Logically, this means that the value cannot be derived from the value of truth. Taking these presuppositions into account, it ought to be clear, at least prima facie, that epistemic value monism might not able to accommodate both presuppositions. Here are both presuppositions in Riggs's wording: (AI) For any p, S's knowing that p is more epistemically valuable than S's merely having the true belief that p. (Riggs 2002a, p. 89)

Presumably, we prefer knowing p over believing p because we would attribute to knowing a higher epistemic status than to merely believing truly. The second presupposition is: (A2) The ultimate sources of all specifically epistemic values are the goods specific to our cognitive pursuits. (Riggs 2002, p. 89; emphasis in original)

(A2) limits the range of epistemic goals. Unfortunately, Riggs does not determine what exactly he means with 'cognitive pursuit'. Do emotions, for instance, belong to our cognitive pursuit? So, could empathy for the fate of other human beings be part of our cognitive pursuit? Since Riggs's general punch line is clear, we can in the following assume that truth and knowledge are the main goals of most of our cognitive pursuits. If (A2) is true, then Riggs only needs one further epistemic goal because then there is another source of epistemic value which cannot be reduced to truth. I reconstruct (A2) as the claim that each epistemic goal constitutes a distinct epistemic value, and in our case knowledge has intrinsic value. Furthermore Riggs attributes the following instrumental-value claim to

125

epistemic value monism:

(VK) [T]he epistemic value of knowing that p derives entirely from the specifically epistemic goods of having true beliefs and avoiding false ones, and some of this value must be additional to the epistemic value that can be derived from the same sources for merely believing p when p is true (and not believing p when p is false). (Riggs 2002a, p. 89)

In a semi-formal way we can reformulate Riggs's argument as follows: (R I) (A I) and (A2) are true. (R2) (VK) contradicts (AI) and (A2).

(R3) (VK) is false. [R2] (R4) If(VK) is false, then value instrumentalism is false. (Therefore) Value instrumentalism is false. [R3, R4]

Riggs is committed to a certain value principle that contradicts epistemic value monism. 155 In his analysis of knowledge, he presupposes that his account must explain why knowledge is allegedly more valuable than true belief. (VK) is "false" only in the context of the two other principles. In sum, Riggs attempts to refute epistemic value monism on the basis that it relies on the wrong value principle. In the following I will discuss Riggs's value assumptions, because these principles do the major lifting in Riggs's argument. I will argue that (AI) can be fulfilled by epistemic value monism. Additionally, I will argue that knowledge does not have intrinsic value. At least, this cannot follow from the value problem.

He writes: "In the following section, I will argue that VK is false" (Riggs, 2002, p. 89). Unfortunately, he does present any textual evidence who has subscribed to (VK).

155

126

3. KNOWLEDGE IS MORE VALUABLE In this section I argue that epistemic value monism can explain why knowing has a higher epistemic value than true belief. Matt Weiner has argued that the concept of knowledge is desirable to have because it unites nlany epistemic useful functions. 156 The concept of knowledge combines other concepts like those of truth, justification, and possibly even stable or safe belief. He calls knowledge "a Swiss army knife concept". The knife integrates many different instruments that could each be a separate device, but are handier when combined in one. We could and may have a separate can opener, a separate knife, a separate spoon and so on; but in certain situations, we prefer to have all the instruments together. So, if we compare the Swiss army knife with a simple knife, then it is obvious that we find the Swiss army knife more desirable because of its multi-functionality. By analogy, we find knowledge more desirable or more valuable because it integrates several epistemic concepts which are central to us, in contrast to Riggs's claim that epistemic value monists cannot explain the value of knowledge. It is important to emphasize that epistelnic value nl0nists are capable of fulfilling (A 1) without presupposing that knowledge has intrinsic value. If knowledge is a concept that functions like a Swiss army knife then this does not indicate that the Swiss army knife has intrinsic value as such. It is just lTIOre valuable in comparison to a one-function tool. If we prefer knowledge over true belief, then we cannot infer that knowledge has intrinsic value as such. It is important to see the difference between comparative and intrinsic value - than knowledge has comparative value doesn't mean it has intrinsic epistemic value. Riggs argues that, since the knowledge goal is intrinsic, we can point to a source of epistemic value independent from the value of truth. My response to (A2) is derived from my response to (A 1). Riggs hasn't made a case for his claim that knowledge has intrinsic (rather than comparative) value. 156 See Weiner (2009).

127

Riggs makes a questionable assumption about the existence of a second value domain beyond that of truth. The crucial assumption that supports Riggs's argument is that with every important or primary epistemic goal there is another source of epistemic value. This means that with the knowledge goal we obtain another fundamental epistemic value domain. However, it is unclear where this second epistemic value domain comes from and how it is constituted independently of truth. Thus far, I have discussed one version of the dual-value objection. In the following I discuss a proposal that attempts to explain what constitutes this second epistemic value domain. 4. THE VIRTUE-THEORETIC CONCEPTION OF KNOWLEDGE In this section I introduce the virtue-theoretic conception of knowledge, largely as advocated by Ernest Sosa. Knowledge is, according to this conception, true belief produced by the competence of the agent. A belief counts as knowledge, if, 1) it is true, and 2) if the belief is acquired by the exercise of a competence of the agent. This conception is relevant for our discussion because it supports the dual-value thesis. Sosa introduces two examples for illustration. The first example is about archery, the second comes from ballet. Both examples are intended to highlight the structure of dual evaluation. Let us consider the archer first. We evaluate the performance of an archer with respect to the result, his hitting or missing of the target, and whether the competence of the archer was involved. Sosa wants to draw an analogy between the archer making a good shot and the believer obtaining a truth. We should evaluate the performance of the archer not only with respect to the result, but also to whether his own competence was involved. If the shot were good only by sheer luck we would not praise the archer, since we praise only competence. It is crucial in such games of skill that the result of the performance is caused by the competence of the agent.

128

The second case has the same evaluative structure. I57 Suppose we see a splendid performance by a ballerina. But later we learn that she was drunk and unable to control her body movements. Then we would retract our praise for her performance because that performance was not due to her ability. In analogy, we attribute knowledge to the agent only if we are sure that it resulted from her performance due to his competences. In the performance case we have success through artistic competence; in the analogy, we have true belief due to cognitive competence. I58 The virtue-theoretic conception of knowledge can explain why the dual-evaluation takes place and why it might be the case that the value of truth cannot exhaust all possible epistemic values. In the following I will discuss Sosa's proposal of a solution to the value problem. This proposal is significant for us, because Sosa defends the dual-evaluation thesis. Sosa's explanation for why knowledge is more valuable than mere true belief is that the agent obtains value by acquiring true beliefs largely through her cognitive competence. We therefore have to ascribe additional epistemic value for the involvement of her cognitive competence. Sosa writes:

One part at least of the solution to the value problem lies in a point central to virtue epistemology: namely, that the value of apt belief is no less epistemically fundan1ental than that of true belief. For this imports a way in which epistemic virtues enter constitutively in the attainment of fundamental value, not just instrumentally. Virtues are thus constitutive because the aptness of belief is constituted by its being accurate because con1petent. (Sosa 2007, p. 87; emphasis in original)

157 In Sosa's terminology, the evaluation takes place with respect to what Sosa calls "the AAA structure". Beliefs are evaluated for their accuracy, their adroitness and their aptness. The accuracy of a belief is its truth, its adroitness contains the manifestation of an epistemic virtue or competence. Its aptness evolves because of competence (see Sosa 2007, Ch. 2). 158 There is a small difference between archery and ballet in that the former is a competitive sport, which we evaluate by counting the points corresponding to the marks on the target, while ballet is evaluated in aesthetic categories. This will become relevant later.

129

Sosa introduces another epistemic value, which has its origin in the agent's epistemic virtue. Besides the acquisition of truth, it is the manner in which the truth was acquired, and whether it involved the agent's competence, which is important. If Sosa is right, the knowledge goal has two aspects: (a) Obtaining the truth because of (b) the competence of the agent. The latter component is supposed to establish a second track of epistemic evaluation. This second track evaluates how the truth was obtained. 159 In the following, I discuss the question whether the dual-evaluation thesis is tenable. I will evaluate the examples that intended to support Sosa's claim and I will argue that none of them supports the dualevaluation thesis, because either the second track of evaluation can be reduced to the value of truth or the second track of evaluation clearly lies outside the epistemic value domain. Afterward I will make a third critical remark on the dual-evaluation thesis. DePaul and Grimm (2006) have introduced another example from sports. Ski-jumping is used to illustrate how standards of double evaluation are supposed to function. At first sight, ski-jumping is a more useful example than archery, since it is evident that ski-jumpers are not only evaluated by the length of the jump, but also with regard to the style of their flight and the professionalism of their landing. To be effective for epistemic value pluralists, the sports analogy needs to show not only that several standards of evaluation are in play, but also, more importantly, that the second value in play must not be reducible to truth. Given this assumption, however, the ski-jumping example is not adequate, because evaluation of ski-jumping is intended to determine the winner of the ski jump competition. This means that even though the jump is evaluated under several standards, in the end all those sub-standards add up to one score. This final score determines the winner of the competition. I take it that any competitive sport supports monistic standards of evaluation due to the fact that we know at the end of the competition who has won, who is second, who third, and so on. Therefore, I think that epistemic value monism can tum any example from competitive sports in favor of the 159 For Sosa this second value is not reducible to the value of truth.

130

I

position that one goal counts in the end, and all others are instrumental to it. 160 This has the consequence that we can reduce any given additional value (e.g., style of the performance and so on) to an instrument for giving us an analogy to the truth. Let us consider the Sosa's ballerina exalnple. The ballerina case is the fall-back case for epistemic value pluralists, because ballet is not a competitive sport where a definite score gives us an analogy of one goal. Again, we credit the ballerina for her performance once we can see that it is due to her abil ities. Returning to the ballerina case, we should think about what kind of value is being attributed to her. The ballerina case seems to me to involve us in a distinction between aesthetic and epistemic values that goes back to the Kantian theme of purposiveness without a purpose. We evaluate her performance with regard to many criteria, such as her artful movement, but in the end it seems to me that we have changed the value domain from epistemology to aesthetics. The example contains a certain value in that, certainly cognitive activity of a type is involved; we must rethink whether the case still fits well into the epistemic value domain. Defenders of the dual-evaluation thesis might accept my objections but still insist that the dual-evaluation thesis is tenable, despite the fact that examples in the literature are not supportive. This response is based on the so-called preference test between obtaining truth with one's own involvement and without it. Sosa sees that we always want to obtain truth by our performance:

We prefer truth whose presence is the work of our intellect, truth that derives from our own virtuous performance. We do not want just truth that is given to us by happenstance. (Sosa 2003, p. 174)

160 Let's consider football, for instance, which is a sport where the score often does not represent the play on the field. After the game, the manager and the coach often reflect about the situation that the score of the game does not represent how the teams were actually playing. I take it that we do evaluate how soccer results developed, but this second evaluation does not change the final score.

131

This passage is instructive because Sosa proposes a preference test to show that we always opt for acquiring truth in a virtuous way. If that is right, then dual evaluation must be taken for granted. For Sosa it is part of human life that humans want to achieve something on the basis of their own abilities. 161 In the following I introduce a case where the agents involved are not concerned about the second value and we cannot blame them for not caring. Here is the case: Consider a law department at one of the highestranked universities in the country and imagine that the head of department has an open position in one of his research projects. The head of department intends to hire a post-graduate student. In a faculty meeting he hears Professor Meyer, the professor of criminal law, speak of one of the recently graduated students. Meyer has read, as every year, the list of the graduating students with their grades. And the best student this year has obtained an outstanding result. Her final grade was way above the average. Two or three law professors are involved in the conversation. All the professors express their intention to make offers to that student. None of the professors, however, can say what character the student has and how she actually obtained the excellent results in the exams. All the professors participating in the conversation rely on Meyer's report being accurate. Meyer himself, however, obtained his information by quickly scanning the list of the graduating candidates.

I refer to Sosa's (2003, p. 161) writing because in some passages he presents a mixed account between epistemic value monism and epistemic value pluralism. Here is a longer passage which supports my judgment: "Truth-connected epistemology might grant the value of truth, of true believing, might grant its intrinsic value, while allowing also the praxical extrinsic value of one's attributably hitting the mark of truth. This praxical extrinsic value would reside in such attributable intellectual deeds. But in addition to the extrinsic praxical value, we seem plausibly committed to the intrinsic value of such intellectual deeds. So the grasping of the truth central to truth-connected epistemology is not just the truth that may be visited upon our belief by happenstance or external agency. [...] What we prefer is the deed of true believing, where not only the believing but also its truth is attributable to the agent as his or her own doing" (Sosa 2003, pp. 174-175; emphasis in original). If Sosa were a straight epistemic value pluralist then he would avoid terms like "truth-connected epistemology".

161

132

Our head of department follows the conversation interestedly. He is informed that a very qualified job candidate is on the market. He knows that many of the faculty n1embers will offer the student a job. We should add to this story one decisive, external factor which influences the decision of the head of department: His reputation will rise if the excellent law student will sign up with him and not with the others. After the meeting, the head of department writes the student an email and asks her to apply for his open position. The student applies the next day and her CV is excellent. He offers her the job and she finally accepts the job offer. On the weekend, the head of department tells his wife about his new hiring. I-Iis wife is puzzled why he hired the student without any job interview, and she thinks that he should not have relied on the exams grade alone when deciding for her. Is the wife is right? What does the head of department know about the student? How does he know whether she will fit in his team and whether she fulfills the job requirements? Suppose the wife of the head of department asks him why he thinks that it was a good decision to hire the student without any job interview. He might reply: "Who would not want to have the best student? All colleagues would have liked to hire her as well." And, more decisive for us: "She is the best student who graduated this year. What else do I need to know?" 162 The crucial point we learn from this case is that, contrary to Sosa's account, none of the professors is interested in attaining truth in what Sosa would call a virtuous way, because only truth counts. The head of department accepted the information without further checking, and Meyer just read off the result from the list of the best students. We would give no intellectual credit to this way of going about obtaining a cognitively valuable result. And no-one involved in this case is interested in obtaining intellectual credit by attaining the truth in an epistemic virtuous way. This undercuts the standard reply of virtue epistemologists in testimonial cases, because neither the receiver of the knowledge nor the giver of the true information is interested in the way they acquired the truth. 163 162 lowe Robert Kowalenko for framing this case in the way I argue for it now. 163 John Greco, for instance, wants to save his epistemic credit theory fronl this counter-example by saying that the knowledge-receiving person does not display

133

Meyer just read this fact off from a list as he would read off the price from one of his shopping receipt. More crucially, Meyer and the head of department are not to blame for how they acquired it. If the intellectualcredit intuition were as profound as virtue epistemologists think then this situation would have to be evaluated differently. I conclude that we do not always prefer acquiring the truth in an intellectually credit-worthy way, because it is just the truth that counts in the described cases. If this is right, then epistemic value pluralists are mistaken in thinking that the preference test will always favor true belief that was acquired in a responsible way. Still, epistemic value pluralists might respond the following to the law student case: Professor Meyer attains truth in a virtuous way because his reading competence was involved in the acquisition of the information. This is a trivial response, however, as any cognitive act would then count as virtuous in this way. Meyer's everyday actions - remembering a price from a shopping receipt - would then receive our praise. The trivialization of praise for competence would, at that point, entirely debase the whole notion of praise. Sosa's examples and others generated to make the same point may teach us a lot about the evaluation of competence, but they are not sufficient to prove the dual-evaluation thesis. Therefore the dual-value objection fails. Either we can reduce the additionally given value to truth, or it is clear that the second value is not epistemic in nature. In the next section I want to introduce the so-called epistemic credit theory, which is based on Ernest Sosa's virtue epistemology. I have chosen to discuss this theory because it contributes also to the dual-evaluation objection.

5. THE EPISTEMIC CREDIT THEORY The basic assumption of epistemic value pluralists is the dualevaluation thesis. Again, this thesis claims that it is not only important to evaluate whether the believer or knower has obtained truth, but also how any greater cognitive achievement. Greco argues that the cognitive achievement was obtained by the knowledge-giver (see Greco 2007).

134

i

she has attained it. Cases from sports were used to illustrate how such credit attributions are supposed to work. Note that our appreciation of the performance of a baseball player evolves once we have a more complete sense of the skills involved in hitting, fielding, running, etc. Plus we have a sense of how these performances fit into teamwork. In this section I will discuss a variant of virtue epistemology in which the dual-evaluation thesis is integrated into a framework of epistemic norms. This framework is called "epistemic credit theory" .164 The basic idea of credit attribution is that we credit a baseball player if his performance is due to his abilities. We refrain from giving credit when we learn that a good result occurred due to mere luck. This idea is supposed to work for knowledge attribution as well. Greco nicely describes the main insight of the theory: But one of the central functions of knowledge attributions is to give credit for true belief. When we say that S knows p, we imply that it is not just an accident that S believes the truth with respect to p. On the contrary, we mean to say that S gets things right with respect to p because S has reasoned in an appropriate way, or perceived things accurately, or remembered things well, etc. We mean to say that getting it right can be put down to SIS own abilities, rather than to dumb luck, or blind chance, or something else. (Greco 2003, p. 116)

Greco's account draws on the following thought: We would not attribute knowledge to someone if this person had acquired truth accidentally. To see how credit attribution is supposed to work in detail, recall the initial context of attribution, which is moral blaming. Greco draws on the work of the moral philosopher Joel Feinberg and his account of moral blaming. According to Feinberg, we morally blame someone if three conditions are fulfilled. We blame a person for an action only if (a) this person commits a morally faulty action, (b) the faulty action can be ascribed to the person we are evaluating, and (c) the faulty action reveals the faulty moral character 165 of the person doing it. The crucial point for Feinberg in attributing blame 164 See Greco (2003), Riggs (2002, 2006) and Sosa (1991, 2003) for more details on the epistemic credit theory. In this chapter I focus mainly on John Greco because he has developed an explicit account of norms with is intended to fit within epistemic credit theory. 165 I draw here from Greco's description (see Greco 2003, p. 120).

135

is that blame implies causal explanations. To blame a person for setting a house on fire we need to indicate that the bad consequence was caused by an action of that person. 166 Greco sees credit-giving as the counterpart to blaming. We give moral credit for morally good action and we give moral blame for morally bad action, as described above. So blaming seems to consist in giving negative credit. At this point it becomes clear why virtue epistemologists are interested in moral blaming, because it is assumed that morally blameworthy action is generally caused by moral faulty character. Greco transposes the credit attribution schema from moral philosophy to epistemology.'67 A believer S obtains intellectual credit, as Greco uses the term, for believing the truth of a proposition p only if (G I) believing the truth regarding p has intellectual value, (G2) believing the truth regarding p can be ascribed to S, and (G3) believing the truth regarding p reveals S's cognitive reliable character.

If the analogy between credit attribution in moral philosophy and in epistemology holds, and if credit is the counterpart of blaming, then cognitive virtues are a necessary correlate of the acquisition of knowledge. Good deeds are the expression of a morally good character and bad deeds express the bad character of the agent. The same line applies to epistemology as well: the person who acquires true belief in a responsible way is of a good cognitive character. So (Gl) and (G3) simply express what virtue epistemologists endorse in general. More commentary is needed on (G2). First of all, Greco endorses such a strong analogy between acting and believing that standards for the evaluation of action apply to belief as well. This is the premise on which his analogy between blaming in moral philosophy and credit attribution in epistemology is based. The attribution of epistemic credit is supposed to work analogously to the moral case. Therefore we must always investigate 166 Feinberg takes it that the language that expresses the consequence of an action can always express the cause. For instance, the sentence "She killed her neighbor" can be translated in the sentence "She caused her neighbor's death". 167 See for more detail Greco (2003, p. 123).

136

how the agent or believer obtained a given belief. This completes Greco's analysis of knowledge. Greco states that S knows p "only if S's reliable cognitive character is an important necessary part of the total set of causal factors that give rise to S's believing the truth regarding p" (Greco 2003, p. 123). 1680 n the one hand, Greco wants to defend the dual-evaluation thesis; on the other hand, the attribution of credit depends on the identification of causal explanations. This is peculiar, in that, properly seen, the identification of causal explanations is the result of a truth-seeking inquiry. Greco is blind to the fact that his causal explanations have epistemic value as well. If causal explanations have epistemic value, then this value depends primarily on truth. A scientific theory that gives us an adequate causal explanation can be seen as a true theory because it describes how the world actually functions. I refer to this point because this opens up the possibility for another objection against the tenability of the dual-evaluation thesis. With his reliance on causal explanations, Greco introduces the possibility of reducing the second track of evaluation to the first track which backs nlY argumentation. The second track is thus reducible to truth. This is instructive to see how dominant truth in the epistetuic value domain is. To sum up, the value clainl of epistemic value monism only concerns the value domain of epistemology. Once we have defined the fundamental value in that domain, all other values in that domain have merely instrunlental value. There are two possible readings of the value claim. The weak reading is: "Fundamental" means that the value of truth cannot be reduced to another value. The strong reading states: Truth is the fundanlental epistemic value, because any possible epistemic value can be derived from truth. Riggs's objection is primarily directed against the strong claim. Epistemic value monism in general is not committed to the strong claim. The weaker claim also has the advantage, vis a vis that the argument 168 Greco actually has two formulations. I am not going to consider the first one because I think that it is already entailed in the second. The first formulation contains the requirement that the acquisition of the belief is due to the reliable cognitive character of the believer. What relation, apart from a causal relation, can there be between the belief and the believer?

137

of the epistemic value pluralists, that the latter are left with the burden of proving that there is a fundamental epistemic value outside of the truth. Knowledge, as we have seen, is not a candidate.

138

8. Conclusion Traditional epistemology has presupposed the centrality of truth without giving a deeper analysis. 169 It was necessary to rethink the longaccepted platitude that truth is our primary epistemic goal, once several objections about epistemic value were formulated. The whole debate is instructive for understanding how the epistemic value domain is structured. 170 I started out by clarifying the basic assumption about truth as an epistemic goal. As a condition of this assumption, I showed that we are justified in using the terminology of epistemic goals. Then I focused on the question under which conditions we value truth, because if we accept that truth is an epistemic goal, then we are likely to accept that truth is valuable. I claimed that we do value significant truth, not all truths. And finally, I showed that the logic behind epistemic value conflicts concerns whether other goals are instrumental to the truth, or fundamental and thus apart from the truth. If it can be shown that a goal, however well developed, ultimately functions and gains its value from serving the truth, I claim that it is instrumental. If the same logic could be proven for the truth - that it serves some other goal - then we would have an alternative monism or pluralism. Chapters 2 and 3 prepared the general basis for the debate between epistemic value monism and epistemic value pluralism. It is clear, however, that both camps agree at least that truth is an epistemic goal and that truth is valuable. As noted earlier, the central issue of the debate concerns the primacy claim of the truth goal. Epistemic value pluralists attempt to introduce several arguments for the conclusion that there are several equally 169 Conceptual truths, like the one that epistemic justification refers to truth, or the platitude that "bel ief aims at truth", should be the starting-points for an analysis, not end points of a philosophical fiat. I70 I could not investigate here how epistemic values and skepticism go together. The question about the primacy of the truth goal might develop quite differently in the context of skepticism. See for the treatment of skepticism within a value-driven epistemological framework Pritchard (2007a, pp. 101-102) and, in greater detail. Pritchard (2008).

139

important epistemic goals besides truth. In Chapters 3-7 I have discussed whether the following objections are fatal for epistemic value monism: (01) Objection of the overarching goal: The truth goal entails two opposing requirements and we need an overarching goal in order to balance them. (Chapter 4) (02) Alternative-epistemic-value objection: We can obtain epistemic value without attaining truth. (Chapter 5) (03) Alternative-monism objection: There is a goal more fundamental than truth. (Chapter 6) (04) Dual-value objection: There is more than one fundamental epistemic value. (Chapter 7)

In going through these arguments, I wanted to show that epistemic value monism is the optimum well-suited framework for considering questions of epistemic value. I argued that epistemic value monism is capable of meeting the objections listed above. The objection of the overarching goal contains the idea that the two main requirements derived from the truth goal are in opposition to each other. Either we accept the requirement of believing truths, or we follow the requirement of error exclusion with the result that we best suspend believing altogether. I dismissed (01) because this objection was based on indefensibly burdening the truth goal with the requirement of believing all truths, or making all truths the end of inquiry. The main point of (02) was that there are cases where the believer obtains epistemic value while failing to attain truth. I rejected (02) because all three counter-examples were defective. From that, I concluded that the prima facie claim of epistemic value monism is sound.l?l J 71 Another project could have taken up the question of evaluating epistemic goals

like coherence or having explanatory power with respect to in what way they are conducive for truth. In this paper, I criticized the general usage of the word "epistemic goal" in epistemology, because coherence and explanatory power and so on are means to attain the truth goal. So I use the term "epistemic goal" in a

140

(03) proposes that we are mistaken in thinking that truth is the most fundamental goal, because there is another, even more fundamental goal. I showed that the arguments for the given alternative goal lead us back to the truth goal. They are not more fundamental. (04) has its origin in the value problem of knowledge. Epistemic value pluralists see the value problem as fatal for epistemic value monists. If they want to explain the distinct value of knowledge, as opposed to truth, then they are in need to find an additional source of epistemic value apart from truth. If there were such an additional source then epistemic value monism would appear to be defective, because its main value assumption would be false. I have argued against (04) that the knowledge goal as such does not pose a problem. It is clear that knowledge is an epistemic goal, albeit revisionism attempted to prove the contrary. The knowledge goal is only a threat to epistemic value monism if it comes in train with the assumption that knowledge has intrinsic value. I argued against this claim by saying that we cannot infer that knowledge has intrinsic value fronl the fact that knowledge is more valuable than true belief. On this score, I argued for an epistemic value nl0nism that overcomes the value problem while maintaining intact the central claim about truth.

It was instructive to clarify whether these objections are effective in the context of the controversy over the positive contribution of epistemic value monism to value-driven epistemology. I presented, however, a couple of arguments intended to show that epistemic value monism is capable of providing such a contribution.

restricted way. Truth and knowledge are epistenlic goals.

141

9. Appendix: Overcoming the Problem of Epistemic Relativism In this chapter I want to focus on the problem of epistemic relativism. 172 Epistemic value monism may face difficulties in responding to the problem of epistemic relativism. I mainly argue that epistemic value monism can sidestep the problem, because it is not con1mitted to a monism concerning our epistemic methods, standards and norms. To repeat: epistemic value monism consists in the claim that truth is our primary epistemic goal. This does not imply a further commitment to a particular set of epistemic norms and epistemic standards, as long as truth is their goa1. 173 In the following I want to explain what the problem of epistemic relativism is. After that I explain how epistemic value monism can overcome that problem. The problem of epistemic relativism arises because it appears that we have a plurality of epistemic standards on which we justify our beliefs. This can be illustrated by the following case of disagreement. Consider the story in the Gospel of Marc where one disciple reports that Jesus provided food to several thousand pilgrims by dividing three loaves bread and five fishes. 174 We can doubt that this report is true, because it is physically in1possible to feed so tnany persons with such a limited amount of bread and fish, which cannot be endlessly divided. So, am I justified in believing that the report is false? Suppose a Christian believer accepts the report as 172 Epistemic value pluralists differ from epistemic relativists in that the former accept the importance of truth, even while they deny the primacy of the truth goal. 173 I suppose that epistemic value monists may be committed to opting for the most efficient means of attaining truth. At least it would be peculiar if we kept using a certain means which is less conducive to truth than others. From this there follows a rationality requirement of using those norms or standards which are actually conducive to truth. 174 Pritchard (2009b) has given a similar example about believing what is written in the bible. He takes the bible case as a case of faultless disagreement. In contrast, I take it that in the bible case no faultless disagreement takes place, because the Christian believer does not draw his justification fronl any epistemic framework at all.

143

true and a secular believer rejects it. The disagreement between the secular believer and the Christian believer is whether the latter is justified in believing the report in the bible to be true. With respect to the standards of the secular believer we cannot be justified in believing the report from the bible, because it is physically impossible. The Christian thinks that his belief in the report is justified because he has gotten it from what he thinks is a rei iable source, the bible. 175 How does the secular believer know that he is justified in believing what he believes? The problem expressed in the example is that we are unable to prove that the standards we take for granted are better than the standards of another, divergent view, for instance, the view of the Christian believer. Let us call this issue "the problem of epistemic relativism". The problem of epistemic relativism is derived from three assumptions that lead us to the conclusion that we have a plurality of epistemic standards: (AI) We cannot attain a justified belief without relying on a certain epistemic framework. (A2) There are at least two epistemic frameworks that are independent from each other. 176 (A3) If there are at least two independent epistemic frameworks (or more) then both (respectively, all of them) are equally good for attaining knowledge. (A I) is quite plausible, because when we evaluate whether a belief is justified then we must presuppose certain standards to which the belief should comply. In the following section I will discuss the "no-permissiveness account". This account argues that (A2) is false because questions about competing standards can be solved in a definitive way.

175 At this point we sidestep the issue that there are two usages of the word "truth" in play in the example. The secular believer refers to ordinary facts, whereas the Christian believer might think about a theological use of the word "truth". We will neglect this possibility here. 176 (A2) is similar to what Boghossian has called "the equal validity thesis" (see Boghossian 2006, p. 2).

144

1. THE NO-PERMISSIVENESS ACCOUNT Paul Boghossian (2006) and John Pollock (2003) are representatives of what I call "the no-pernlissiveness account". The no-permissiveness account claims that we can settle disagreements about epistemic standards and norms by referring to absolute epistemic facts. So, I call it "the nopermissiveness account" because it implies that there is no choice over our epistemic norms since statements involving our epistemic nornlS are either true or false. Boghossian expresses this thought in the following passage: Against this backdrop, the thesis of the objectivity of reasons can be stated as the claim that there is an objective fact of the matter which epistemic principles are true, and, consequently, which sets of rules a thinker ought to employ to shape his beliefs, if he is to arrive at beliefs that are genuinely justified. (Boghossian 2001, p. 2)

Pollock defines epistemic norms in the following way: Norms are general descriptions of the circumstances under which various kinds of normative judgments are correct. Epistemic norms are norms describing when it is epistemically permissible to hold various beliefs. A belief is justified if and only if it is licensed by correct epistemic norms. (Pollock 2003, p. 192)

So Pollock and Boghossian introduce the statement that a belief is justified if the principles involved in its justification are objectively correct. Objective justification adheres to correct epistemic norms and these norms are correct because they are true. For them it is not sufficient that a belief is justified by any epistemic principle, but only by the true principles. This means that epistemic principles and epistemic norms are truth-apt. This clainl affects the problem of epistemic relativism because the nopermissiveness account provides a direct response to it. Paul Boghossian has given a reconstruction of the argument for epistemic relativism. The argument for epistemic relativism goes as follows: I??

177 See Boghossian (2006, p. 74).

145

1) If there are absolute epistemic facts about what justifies what, then it ought to be possible to arrive at justified beliefs about them. 2) It is not possible to arrive at justified beliefs about what absolute epistemic facts there are. 3) (Therefore) There are no absolute epistemic facts. 4) If there are no absolute epistemic facts, then epistemic relativism is true. (Therefore) Epistemic relativism is true.

Boghossian evaluates the given argument as follows: "The argument is evidently valid; the only question is whether it is sound" (Boghossian 2006, p. 74). For Boghossian premise 1 expresses the central intuition of epistemic relativism. His strategy is to block the intuition of competing standards by showing that there are absolute or universal epistemic facts. Therefore we are in a position to arrive at objective justification. He accepts premise 4 for the sake of the argument. 178 In contrast to that, Boghossian takes premise 3 as an outright falsehood, because he believes that epistemic facts do exist. Finally, he concludes that the argument for relativism is unsound. So, in Boghossian's argumentation the term "absolute epistemic facts" does the heavy lifting. We might be interested in what those absolute epistemic facts consist in. I understand them as 'facts' that make our decisions about standards and norms right or wrong. Do such facts about justification really exist? It is clear that facts in everyday life are seen as being universal. What kind of facts are those about epistemic justification? How can those meta-facts actually exist? There is an epistemological question related to that ontological question: If those facts of justification exist, how can we access them? I doubt that we can answer these questions in a satisfactory way. Therefore I argue in the following that we should dismiss the no-permissiveness account. Let us consider another case of disagreement which helps us to sharpen our intuitions on Boghossian's facts about justification. Here is a case which shows that if facts about justification were to 178 We shall pay more attention to premise 4 later.

146

exist, then our response to the person in the example would run differently: Gary is a student in the introductory epistemology class, and in the middle of the session Gary cOlnes up with a sort of provocative question. 179 Gary states the fo Howing:

These philosophers we've been reading seem to agree that there are certain standards of belief, standards we should follow even when they lead us to conclusions we don't like. They spend all their time disagreeing about exactly what these standards are, but they just seem to assume that we'll want to follow theIn. Suppose I don't? What can they say to nle? (Railton 1997, p. 54)

What can epistemologists say if he says he does not want to follow the standards they are proposing? This question is important because if epistemic norms are based on merely hypothetical grounds then Gary can easily question those norms. He can dodge the standards of epistenl010gists by claiming that these epistemologists prescribe goals which he does not share. But we have seen that the truth goal sets an ideal, and ideals are independent from personal goals. So the first answer to Gary is that it would be irrational or irresponsible to neglect the central epistemic norms, because those norms have a valuable function. Additionally, truth sets a certain epistemic ideal we should attain. There is little disagreement on this point among epistemologists. Therefore we nlight point out to Gary that there is a certain common ground despite the fact that epistemologists disagree about the standards of epistemic justification. Does Gary have good reasons for deviating from epistemic standards? It appears not, since he does not present nlore background information to support his question. Therefore he cannot get away with the objection that general epistelnic standards are not standards subject to his personal approval. 180 This case is instructive: If Boghossian were right in his clainl about the existence of universal epistemic facts we would be in a different 179 Railton (1997) develops the story also with respect to practical reasons, and there Gary sits in the introductory class to action theory. 180 This also sheds light on the nature of epistemic norms: epistenlic standards are not a matter that anyone nlust endorse personally.

147

position to meet Gary's challenge. We could then easily respond to Gary by pointing out to him the fact he missed. Therefore I think that we should dismiss the terminology of absolute epistemic facts. This does of course not mean that I give in to epistemic relativism. I agree that the argument for epistemic relativism is unsound, but I think there are other grounds for showing that one of the premises of the argument for epistemic pluralism is false. In contrast to Boghossian I take it that we can obtain justified belief in the right way without postulating the existence of absolute epistemic facts. 2. EPISTEMIC VALUE MONISM EPISTEMIC RELATIVISM

AND

THE

PROBLEM

OF

It is clear that there many ways to justify one's belief, but we shall distinguish between standards which are directed at attaining truth and other, non-epistemic standards, which are directed at non-epistemic goals. We can convincingly exclude radical alternatives for justifying our beliefs. In the following I want focus on the case of disagreement concerning Jesus feeding three thousand people with three loaves of bread and five fishes. I will argue that the dispute here is not about an issue within the epistemic framework, since the Christian believer is arguing for a truth outside it. Does the Christian believer have his own independent epistemic framework? The Christian believer cannot maintain that the story in the Gospel according to Mark actually occurred as described. It is clear that the case must be taken as a sort of illustration that Jesus was capable of achieving the unthinkable. My remark is just intended to point outthat the type of justification involved in justifying the Christian believer is not epistemic. The reasons for sticking to such a report are primarily religious ones. If this is right, then the conflict between the two believers is resolved because the Christian believer cannot claim anymore that the principles supporting his beliefs are epistemic principles. This case cannot count as a genuine case of disagreement because only one of the believers is arguing within the framework of correct epistemic norms. It is not clear, as I said earlier, whether the Christian believer adheres to epistemic norms when he

148

sees his belief about the report in the bible as justified. Using this case, we can ask whether (A2) is true. We contend that (A2) is false because there are never any cases of two independent epistemic frameworks - one will always turn out to be operating under another framework, say of revealed truths. So far, it appears that (A2) is not sound, and if (A2) fails then (A3) becomes senseless, because if there are no two independent epistemic frameworks then we do not need to think about whether one of them is better than the other. This gives us a new diagnosis of the dispute, which, contra the relativists, is not about determining the absolute correctness of a particular set of epistemic norms, but about the existence of a conflict between epistemic frameworks at all. Once we have discarded (A2), it is easy to see that epistemic value monism can accept a pluralism of episten1ic norms. In sum, epistemic value monism provides a twofold response to epistemic relativism: First, we can rule out that there are radical alternatives to our epistemic frameworks. Second, it is incorrect to infer that epistemic relativislTI is confirmed by the reference to absolute epistemic facts. 181

181 This is in principle the same answer as Wesley Salmon gave to Thomas Kuhn (see, for instance, Saln10n 2005, p. 61). Kuhn asked for objective criteria for making decisions over theories. Since there are no such objective criteria, Kuhn inferred that the decision-making of scientists is determined fully by non-episten1ic reasons.

Bibliography Adler, J. E. (2002), Beliefs Own Ethics, Cambridge, Mass.: MIT Press. Alston, J. (1985), 'Concepts of Epistemic Justification', The Monist 68: 5789. (2005), Beyond "Justification": Dimensions of Epistemic Evaluation, Ithaca, NJ: Cornell University Press. Axtell, G. (2008a), 'Expanding Epistemology. A Responsibilist Approach', Philosophical Papers 37: 51-88. Axtell, G.lCarter, A. J. (2008b), 'Just the Right Thickness: A Defense of Second-Wave Virtue Epistemology', Philosophical Papers 37: 413-434. Baumann, P. (1997), 'Sind die meisten unserer Meinungen wahr?', Logos 4: 116-136. Beckerlnann, A. (2001), 'Zur Inkoharenz des Wissensbegriffs. PUidoyer fUr eine neue Agenda der Erkenntnistheorie', in: Zeitschrijt fur philosophische Forschung 55: 571-589. (2002), 'Lasst sich der Wissensbegriff retten? Replik auf die Kritiken von Peter Baulnann, Thomas Grundmann und Frank Hofmann', in: Zeitschrijt fur philosophische Forschung 56: 586-594. Bennett, J. (1990), 'Why is Belief Involuntary?', Analysis 50: 87-107. Blackburn, S. (1980), 'Truth, Realism, and the Regulation of Theory', Midwest Studies in Philosophy 5: 353-371. (1998), 'Wittgenstein, Wright, Rorty and Minimalism', Mind 107: 157181. -

(2005), Truth: A Guide for the .Perplexed, London: Allen Lane.

Boghossian, P. (2001), 'How Are Objective Epistemic Reasons Possible', Philosophical Studies 106: 1-40. (2006), The Fear of Knowledge. Constructivism, Oxford: Clarendon Press. -

Against

Relativism

and

(2007), 'The Case Against Epistemic Relativism', Episteme. A Journal

151

ofSocial Epistemology 4: 49-65.

Bonjour, L. (1985), The Structure of Empirical Knowledge, Cambridge: Harvard University Press. (2002), Epistemology. Classic Problems and Contemporary Responses, Lanham: Rowman & Littlefield Publishers. Bovens, L.lOlsson, E. (2002), 'Believing More, Risking Less: On Coherence, Truth, and Non-Trivial Extensions', Erkenntnis 57: 137-150. Brady, M. (2009), 'Curiosity, Interests and the Value of Truth', in: Epistemic Value, eds. A. Haddock, A. Millar and D. Pritchard, Oxford: Oxford University Press. Brandom, R. (2000), 'Insights and Blindspots of Reliabilism', in: idem, Articulating Reason. An Introduction to Inferentialism, 97-122, Cambridge, Mass.: Harvard University Press. Brendel, E. (1999), Wahrheit und Wissen, Paderbom: Mentis. Broughton, J. (2002), Descartes's Method of Doubt, Princeton: Princeton University Press. Buekens, F. (2007), 'Truth is not a Goal of Inquiry: Some Davidsonian Considerations', Talk at The Value of Knowledge Conference, Amsterdam, August 2007. Chisholm, R. (1977), Theory of Knowledge, Englewood Cliffs, NJ: Prentice Hall. (1986), 'The Place of Epistemic Justification', Philosophical Topics 14: 85-92. Chrisman, M. (2008), 'Ought to Believe', Journal ofPhilosophy 105. Clifford, W. K. F([1877]/1993), 'Ethics of Belief, in: The Theory of Knowledge. Classical and Contemporary Readings, ed. L. P. Pojman, 502505, Belmont, Calif.: Wadsworth. Cohen, S. (1984), 'Justification and Truth', Philosophical Studies 46: 279295. Conee, E. (1992), 'The Truth Connection', Phenomenological Research 52: 657-669.

152

Philosophy

and

Conee, E.lFeldman, R. (2004), Evidentialism, Oxford: Oxford University Press. David, M. (1996), 'Working Without A Net: A Study of Egocentric Episten1010gy', Philosophy and Phenomenological Research 56: 943-952. (2001), 'Truth As The Epistemic Goal', in: Knowledge, Truth, and Duty: Essays on Epistemic Justification, Virtue, and Responsibility, ed. M. Steup, 151-169, Oxford: Oxford University Press. (2005), 'Truth As The Primary Epistemic Goal: A Working Hypothesis', in: Contemporary Debates in Epistemology, eds. E. Sosa and M. Steup, 296-312, Oxford: Blackwell. Davidson, D. (1990), 'The Structure and Content of Truth' (The Dewey Lectures 1989), Journal ofPhilosophy 87: 279-328. (1996), 'The Folly of Trying to Define Truth', The Journal of Philosophy 93: 263-278. (1999a), 'Is Truth A Goal Of Inquiry? Discussion with Rorty', in: Donald Davidson. Truth, Meaning and Knowledge, ed. U. M. Zeglen, 1719, London: Routledge. (1999b), 'Reply to Pascal Engel', in: The Philosophy of Donald Davidson, ed. L. E. Hahn, 460-461, Chicago: Open Court. (2005), 'Truth Rehabilitated', in: idem, Truth, Language, and History, 3-17, Oxford: Clarendon Press. DePaul, M. (2001), 'Value Monism in Epistemology', in: Knowledge, Truth, and Duty. Essays on epistemic Justification, Responsibility, and Virtue, ed. M. Steup, 170-183, Oxford: Oxford University Press. (2004), 'Truth Consequentialism. Withholding and Proportioning Belief to Evidence', Philosophical Issues 14: 91-112. DePaul, M.lGrimm, S. (2006), 'Review Essay on Jonathan Kvanvig's The Value ofKnowledge and the Pursuit of Understanding', in: Philosophy and Phenomenological Research 74: 498-514. Dewey, J.lRussel, B. (1985), Dewey and Russell. An Exchange, ed. S. Meyer New York: Philosophical Library. Douglas, H. (2000), 'Inductive Risk and Values in Science', Philosophy of

153

Science 67: 559-579. Dummett, M. (1959), 'Truth', Proceedings of the Aristotelean Society 59: 141-162. Elgin, C. (1996), Considered Judgment, University Press. -

Princeton, NJ: Princeton

(2004), 'True Enough', Philosophical Issues 14: 113-31.

(2007), 'Understanding and the Facts', Philosophical Studies 132: 33.42. (2009), 'Is Understanding Factive?" in Epistemic Value, eds. A Haddock, A. Millar and D. H. Pritchard, Oxford: Oxford University Press. Ellis, B. (1988); 'Solving the Problem of Induction Using a Values-Based Epistemology" British Journal for the Philosophy ofScience 39: 141-160. Engel, P. (1999), 'The Norms Of The Mental' in: The Philosophy ofDonald Davidson, ed. L. E. Hahn, 447-460, Open court, La Salle, Illinois. -

(2002), Truth, Chesham: Acun1en Press.

Fairweather, A. (2001), 'EpistemicMotivation" in: Virtue Epistemology. Essay on Epistemic Virtue and Responsibility, ed. A. Fairweather and L. Zagzebski, 63-81, Oxford: Oxford University Press. Fahrbach, L. (forthcoming), 'How the Growth of Science Ends Theory Change, Synthese. Feldman, R. (1988), 'Epistemic Obligation', Philosophical Perspectives 2: 235-256. (2001), 'Voluntary Belief and Epistemic Evaluation', in: Knowledge, Truth, and Duty: Essays on Epistemic Justification, ed. M. Steup, 77-92, Oxford: Oxford University Press. -

(2003), Epistemology, Englewood Cliffs: Prentice Hall.

(2004), 'The Ethics of Belief, in: Conee, E./Feldman, R., Evidentialism, 166-195, Oxford: Oxford University Press. Firth, R. (1978), 'Are Epistemic Concepts Reducible to Ethical Ones?', in: Values and Morals, eds. A. Goldman and J. Kim, 215-229, Dordrecht: Reidel.

154

Foley, R. (1985), 'What's Wrong with Reliabilism', The Monist 68: 188202. (1987), The Theory of Epistemic Rationality, Cambridge, Mass: Harvard University Press. (1993), Working Without a Net: A Study of Egocentric Epistemology, New York/Oxford: Oxford University Press. (1994), 'What am I to Believe?', in: Naturalism, eds. S. J Wager and R. Warner, 147-162, Notre Dame: University of Notre Dame Press. (200 1a), Intellectual Trust in Oneself and Others, Cambridge: Cambridge University Press. (200 1b), 'The Foundational Role of Epistemology in General Theory of Rationality', in: Virute Epistemology. Essay on Epistemic Virtue and Responsibility, eds. A. Fairweather and L. Zagzebski, 214-230, Oxford: Oxford University Press. (2003), 'SkepticislTI and Rationality', in: Epistemology. An Anthology, eds. E. Sosa and J. Kim, 182-191, Malden: Blackwell Publishing. (2004), 'The Trial Separation Between the Theory of Knowledge and the Theory of Justified Belief, in: Ernest Sosa and his Critics, ed. J. Greco, 59-71, Malden: Blackwell. (2005), 'Is Justified Belief Responsible Belief?', in Contemporary Debates in Epistemology, eds. E. Sosa and M. Steup, 313-326, Oxford: Blackwell. Fuhrmann, A. (2006), 'Is Pragmatist Truth Irrelevant To Inquiry?', in: Knowledge and Inquiry: Essays on the pragmatism ofIsaac Levi, ed. Eric Olsson, 32-49,Cambridge: Cambridge University Press. Fumerton, R. (2001), 'Epistemic Justification and Normativity', in: Knowledge, Truth, and Duty: Essays on Epistemic Justification, Responsibility and Virtue, ed. Matthias Steup, 49-61 ,Oxford: Oxford University Press. Gibbard, A. (2005), 'Truth and Correct Belief, Philosophical Issues 8: 338-350. Goldman, A. (1992), Liaisons. Philosophy Meets the Cognitive and Social

155

Sciences, Cambridge, Mass.: MIT Press. (1999), Knowledge in a Social World, Oxford: Oxford University Press. (2001), 'The Unity of Epistemic Virtues', in: Virtue Epistemology. Essay on Epistemic Virtue and Responsibility, eds. A. Fairweather and L. Zagzebski, 30-48, Oxford: Oxford University Press. Greco, J. (1990), 'Intemalism and Epistemically Responsible Belief, Synthese 85: 245-277. (1993), 'Virtues and Vices of Virtue Epistemology', Canadian Journal ofPhilosophy 23: 413-432. (2003), 'Knowledge as Credit for True Belief, in: Intellectual Virtue. Perspectives from Ethics and Epistemology, eds. M. DePaul and L. Zagzebski, 111-134, Oxford: Oxford University Press. (2007), 'Discrimination and Testimonial Knowledge', Episteme. A Journal ofSocial Epistemology 4: 335-351. Grimm, S. (2001), 'Ernest Sosa, Knowledge and Understanding', Philosophical Studies 106: 171-191. (2006), 'Is Understanding a Species of Knowledge?', British Journal for the Philosophy ofScience 57: 515-535. Grundmann, T. (2002), 'Warum wir Wissen als einen wichtigen Begriff der Erkenntnistheorie behalten sollten - Eine Antwort auf Ansgar Beckermann', Zeitschriftfur philosophische Forschung 56: 118-124. Haack, S. (1976), 'The Pragmatist Theory of Truth', British Journalfor the Philosophy ofScience 27: 231-249. Harman, G. (2000), Explaining Values and Other Essays in Moral Philosophy, Oxford: Clarendon Press. Heil, J. (1984), 'Doxastic Incontinence', Mind 93: 56-70. Hempel, C. G. (1942), 'The Function of General Laws in History', Journal ofPhilosophy 39: 35-48. Hetherington, S. (2001), Good and Bad Knowledge. On Two Dogmas of Epistemology, Oxford: Oxford University Press.

156

Hofmann, F. (2005), 'Epistemic Means and Ends: In Defense of some Sartwellian Insights', in: Synthese146: 357-369. (2007), 'Wahrheit und Wissen. Einige Oberlegungen zur epistemischen Normativitat', in: Zeitschri/t fur philosophische Forschung 61: 147-174. Horwich, P. (1990), Truth, Oxford: Basil Blackwell. -

(2006), The Value of Truth, Noils 40: 347-360.

Huber, F. (2008), 'Assessing Theories, Bayes Style', Synthese 161: 89-118. James, W. (1956), The Will to Believe and other Essays on Popular Philosophy. Hun1an Immortality. Two Supposed Objections to the Doctrine, New York: Dover Pub. Jones, W. (1997), 'Why Do We Value Knowledge?', in: American Philosophical Quarterly 34: 423-439. Kaplan, M. (1985), 'It's Not What You Know', in: Journal of Philosophy 82: 350-363. (1994), 'Epistemology Denatured', in: Midwest Studies in Philosophy 19: 350-365. Kelly, T. (2003), 'Epistenlic Rationality as Instrumental Rationality. A Critique', Philosophy and Phenomenological Research 66: 612-640. Kelp, C. (2008), 'Knowledge and Safety', Journal for Philosophical Research 33. Kolbel, M. (2002), Truth Without Objectivity, London: Routledge. Koppelberg, D. (2005), 'Zum Wert von Wissen: Das Menon-Problem', in: Logical Analysis and History of Philosophy. Focus: History of Epistemology, eds.U Meixner and A. Newen, 46-56, Paderbon: Mentis. Kornblith, H. (1983), 'Justified Belief and Epistenlically Responsible Action', Philosophical Review 92: 33-48. -

(1993), 'Epistemic Nonnativity', Synthese 94: 357-376.

(2001), 'Epistemic Obligation and the Possibility of Internalism', in: Virtue Epistemology: Essays on Epistemic Virtue and Responsibility, eds. A. Fairweather and L. Zagzebski, 231-248, Oxford: Oxford University Press.

157

-

(2002), Knowledge and Its Place In Nature, Oxford: Clarendon Press.

Kutschera, F. (1982), Grundfragen der Erkenntnistheorie, Berlin: de Gruyter. Kvanvig, J. L. (1992), The Intellectual Virtues and the Life ofthe Mind. On the Place of the Virtues in Epistemology, Savage, MD: Rowman & Littlefield Publishers. (2003), The Value of Knowledge and the Pursuit of Understanding, Cambridge: Cambridge University Press. (2005), 'Truth is Not the Primary Epistemic Goal', in: Contemporary Debates in Epistemology, eds. E. Sosa and M. Steup, 285-96, Oxford: Blackwell. (2009), 'Responses to Critics', Epistemic Value, eds. A. Haddock, A. Millar and D. H. Pritchard, Oxford: Oxford University Press. Lackey, J. (2007). 'Why We Don't Deserve Credit for Everything We Know', Synthese 158: 345-361. Lackey, L.lSosa, E. (eds.) (2006), The Epistemology ofTestimony, Oxford: Oxford University Press. Lehrer, K. (1981), 'A Self-Profile', in: Keith Lehrer, ed. R. J. Bogdan, 3104, Dordrecht: Reidel Pub. (2000), Theory of Knowledge, Boulder: Westview Press (Second Edition). Lehrer, K.lCohen, S. (1983), 'Justification, Truth, and Coherence', Synthese 55: 191-207. LeMorvan, P. (2002), 'Is mere True Belief Knowledge?', in: Erkenntnis 56: 151-168. (2008), 'Epistemic Means and Ends: A Reply to Hofmann', in: Synthese 162: 251-264. Lipton, P. (2008), 'Alien Abduction. Inference to the Best Explanation and the Management of Testimony', Episteme. A Journal of Social Epistemology 4: 238-251. Loewer, B. (1980), 'The Truth Pays', Synthese 43: 369-379.

158

-

(1993), 'The Value of Truth', Philosophical Issues 4: 265-280.

Luper-Foy, S. (1990), 'Arbitrary Reasons', in: Doubting, eds. M. D. Ross and G. Ross, 39-55, Dordrecht: Kluwer. Lynch, M. (2001), 'A Functionalist Theory of Truth', in: The Nature of Truth. Classical and Contemporary Readings, ed. M. Lynch, 723-749, Cambridge Mass.: MIT Press. (2004), True to Life: Why Truth Matters, Cambridge, MA: The MIT Press. -

(2005), 'Replies to Critics', Philosophical Books 46: 331-342.

Maitzen, S. (1995), 'Our Errant Epistemic Aim', Philosophy and

Phenomenological Research 55: 869-876. McFarlane, J. (2005), 'Making Sense of Relative Truth', Proceedings ofthe Aristotelean Society 105: 321-339. Miller, R. (1995), 'The Norms of Reason', Philosophical Review 104: 205245. Misak, C. (1991), Truth and the End of Inquiry. A Peircean Account of Truth, Oxford: Clarendon Press. Montmarquet, J. A. (1987), 'Epistemic Virtue', Mind 96: 482-497. (1993), Epistemic Virtue and Doxastic Responsibility, Lanham: Rowman & Littlefield Publishers. Moser, P. (1985), Empirical Justification, Dordrecht: Reidel. (2005), Against Coherence: Truth, Justification, Oxford: Oxford University Press. Olsson,

-

E.

Probability,

and

(2007), 'Reliabilisln, Stability, and the Value of Knowledge', in:

American Philosophical Quarterly 44: 343-355. Owens, D. J. (2003), 'Does Belief have An Aim?', Philosophical Studies 115: 283-305. Papineau, D. (2003), 'Nonnativity and Judgement', in: idem, The Roots of

Reason, 6-37, Oxford: Oxford University Press. Pierce, C. S. (1955), The philosophical Writings of Peirce, ed. J. Buchler, New York: Dover Publications.

159

Piller, C. (1991), 'On Keith Lehrer's Belief in Acceptance', Grazer philosophische Studien 40: 37-61. Plantinga, A. (1988), 'Positive Epistemic Status and Proper Function', Philosophical Papers 2: 1-50. (1992), 'Justification in the 20th Century', Philosophical Issues 2: 4377. Pollock, J, (2003), 'Epistemic Norms', in: Epistemology. An Anthology, eds. E. Sosa and J. Kim, 192-225, Malden: Blackwell Publishing. Pritchard, D. (2004a), 'Epistemic Deflationism', The Southern Journal of Philosophy 42: 1-32. (2004b), 'The Epistemology of Testimony', Philosophical Issues 14: 326-348. (2007a), 'Recent Work on Epistemic Value', American Philosophical Quarterly 44: 85-110. (2007b),'The Value of Knowledge', Stanford Encyclopedia of Philosophy (Winter 2003 Edition), (ed.) E. Zalta, URL . -

(2007c), What is a Thing Called Knowledge, London: Routledge.

(2008), 'Radical Scepticisn1, Episten1ic Luck and Epistemic Value', Supplement to the Proceedings ofthe Aristotelean Society 82: 19-41. (2009a), 'Knowledge, Understanding and Epistemic Value', in Epistemology (Royal Institute ofPhilosophy Lectures), (ed.) A. O'Hear, 1943, Cambridge: Cambridge University Press.

-

(2009b), 'Defusing Epistemic Relativism', Synthese 166: 397-412.

Quine, W. V. (1986), Reply to Morton White, in: The Philosophy ofW V. Quine, eds. L. E. Hahn and P. A. Schilpp, 663-665 , La Salle, Ill.: Open Court Publ. (1990), Pursuit of Truth, Cambridge, Mass.lLondon: Harvard University Press. Quine, W. V.lUllian, J. S. (1970/1978), The Web of Belief, New York: Random House.

160

Rabinowicz, W.lRoennow-Rasmussen, T. (1999), 'A Distinction in Value: Intrinsic and For its Own Sake', Proceedings of the Aristotelian Society 100: 33-49. (2003), 'Tropic of Value', Philosophy and Phenomenological Research 66: 389-403. Railton, P. (1994), 'Truth, Reason, and the Regulation of Belief, Philosophical Issues 5: 71-93. (1997), 'On the Hypothetical and Non-Hypothetical in Reasoning about Belief and Action', in: Ethics and Practical Reasoning, eds. G. Cullity andB. Gaut, 53-79, Oxford: Clarendon Press. Rescher, N.lVinci, T. C. (1975), 'On the Relevance of the Pragmatic Utility of Beliefs, Review of Metaphysics 28: 443-452. Riggs, W. (1997), 'The Weakness of Strong Justification', Australian Journal ofPhilosophy 75: 179-189. (2002a), 'Beyond Truth and Falsehood: The real Value of Knowing that p', in: Philosophical Studies 107: 87- 108. (2002b), 'Reliability and the Value of Knowledge', in: Philosophy and Phenomenological Research 64: 79-96. -

(2003a), 'Balancing our Episten1ic Goals', Nous 37: 342-352.

(2003b), 'Understanding 'Virtue' and the Virtue of Understanding', in: Intellectual Virtue: Perspectives from Ethics and Epistemology, eds. L. Zagzebski andM. DePaul, 203-226, Oxford: Oxford University Press. (2007), 'Why are Epistemologists so Down with their Luck', Synthese 158: 329-344. (2008), 'The Value Turn in Epistemology', in: New Waves in Epistemology, eds. V. F. Hendricks and D. Pritchard, 300-323, Houndslnill: Palgrave Maclnillan. (2009), 'Understanding, Knowledge and the Meno Requirement', in: Epistemic Value, eds. A. I-Iaddock, A. Millar and D. H. Pritchard, Oxford: Oxford University Press. Roennow-Rasmussen, T.lZimmermann, M. J. (2005) (eds.), Recent Work on Intrinsic Value, Dordrecht: Spinger.

161

Rorty, R. (1986), 'Pragmatism, Davidson and Truth', in: Truth and Interpretation, ed. E. LePore, 333-355, Oxford: Blackwell. (1995), 'Is Truth a Goal of Inquiry? Davidson vs. Wright', Philosophical Quarterly 45: 281-300. (1998), Truth and Progress. Philosophical Papers. Cambridge: Cambridge University Press.

Volume 3,

(2005), 'True to Life: Why Truth matters', Philosophy and Phenomenological Research 71: 231-239. .Rorty, R.lEngel, P. (2007), What's the Use of Truth?, ed. P. Savidan, New York: Columbia University Press. Rosen, G. (2007), 'The Case against Epistemic Relativism: Reflections on Chapter 6 of Fear of Knowledge', Episteme: A Journal of Social Epistemology 4: 10-29. Rosenberg, J. (2002a), 'How Not To Misunderstand Peirce - A Pragmatist Account of Truth', in: What is Truth?, ed. R. Schantz, 283-298, Berlin: Walter DeGruyter. -

(2002b), Thinking about Knowing, Oxford: Clarendon Press.

Salmon, W. (2005), Reality and Rationality, eds. P. Dowe and M. H. Salmon, Oxford: Oxford University Press. Sartwell, C. (1991), 'Knowledge is Philosophical Quarterly 28: 157-165.

True

Belief,

in:

American

(1992), 'Why Knowledge is Merely True Belief, in: Journal of Philosophy 89: 167-180. Sober, E. (1994), 'The Primacy of Truth-Telling and the Evolution Of Lying', in: idem, From a Biological Point of View, 71-92, Carrlbridge: Cambridge Press University. Sosa, E. (1991), Knowledge in Perspective, Cambridge: Cambridge University Press. (1993), 'Epistemology, Perspectives 7: 1-16. -

162

Realism,

and

Truth',

Philosophical

(2001a), 'For the Love of Truth?', in: Virtue Epistemology. Essays on

Epistemic Virtue and Responsibility, eds. A. Fairweather and L. Zagzebski, 49-62, Oxford: Oxford University. Press. (2001 b), 'Epistemology and Primitive Truth', in: The Nature of Truth. Classic and Contemporary Perspectives, ed. M. P. Lynch, 641-662, Cambridge, Ma: The MIT Press. (2003), 'The Place of Truth in/Epistemology', in: Intellectual Virtue: Perspectives from Ethics and Epistemology, eds. L. Zagzebski and M. DePaul, 155-179, Oxford: Oxford University Press. (2004), 'Reply to Zagzebski', in: Sosa and His Critics, ed. J. Greco, 319-322, Malden: Blackwell. (2007), A Virtue Epistemology: Reflective Knowledge and Apt Belief, Oxford: Oxford University Press. Spohn, W. (2000), 'Wo stehen wir heute mit dem Induktionsproblem?', in: Erfahrung- und UrteilskraJt, ed. R. Enskat, 151-164, Wlirzburg: KonigshausenlNeumann. (2002), 'The Many Facets of the Theory of Rationality', Croatian Journal ofPhilosophy 2: 247-262. Stich, S. (1990), The Fragmentation of Reason. Preface to a pragmatic theory ofcognitive evaluation, Cambridge, Mass.: MIT Press. Stocker, M. (1982), 'Responsibility Especially for Beliefs', Mind 91: 398417. Sturgeon, S. (1991), 'Truth in Epistemology', Phenomenological Research 51 : 99-108.

Philosophy

and

Unwin, N. (1987), 'Beyond Truth. Toward a New Conception of Knowledge and Communication', Mind 96: 299-317. van Fraassen, B. (1980), The Scientific Image, Oxford: Clarendon Press. Velleman, J. D. (2000), The Possibility of Practical Reason, Oxford: Oxford University Press. Warfield, T. (2005), 'Knowledge Perspectives 19: 405-416.

from

Falsehood',

Philosophical

Weatherson, B. (2003), 'What Good are Counterexamples?', Philosophical

163

Studies 115: 1-31. Wedgwood, R. (2002), 'Intemalism Explained', Phenomenological Research 65: 349-369.

Philosophy

and

Weinberg, J. M. (2006), 'What's Epistemology For? The Case for Neopragmatism in Normative Metaepistemology', in: Epistemology Future, ed. S. C. Hetherington, 26-47, Oxford: Oxford University Press. (2007), 'Moderate Epistemic Relativism and our Epistenlic Goals', Episteme. A Journal ofSocial Epistemology 4: 66-92. Weiner, M. (2009), 'Does Knowledge Matter?', in: Epistemic Value, eds. A Haddock, A. Millar and D. H. Pritchard, Oxford: Oxford University Press. Williams, B. (1973), 'Deciding to Believe', in: idem, Problems of Self. Philosophical Papers, 1956-1972, 136-151, Cambridge: Cambridge University Press. Williams, M. (1980), 'Coherence, Justification and Truth', Review of Metaphysics 34: 243-272. (1986), 'Do We (Epistemologists) Need A Theory of Truth?', Philosophical Topics 54: 223-242. Wrenn, C. (2001), Truth and the Normativity of Epistemology, Ph.D. Dissertation, Washington University. -

(2006), 'Epistemology as Engineering', Theoria 72: 60-79.

Zagzebski, L. T. (2000), 'From Reliabilism to Virtue Epistemology', in: Knowledge, Belief, and Character, ed. G. Axtell, 113-122, Langham: Rowman & Littlefield Publishers. (2001), 'Recovering Understanding', in Knowledge, Truth, and Duty: Essays on Epistemic Justification, Responsibility, and Virtue, ed. M. Steup, 235-251, New York: Oxford University Press. (2003a), 'The Search for the Source of Epistemic Good', in: Moral and Epistemic Virtues, eds. M. Brady and D. Pritchard, 13-28, Malden, Mass: Blackwell. (2003b), 'Virtues of the Mind', in: Epistemology. An Anthology, eds. E. Sosa and J. Kim, 452-467, Malden: Blackwell Publishing.

164

(2004), 'Epistemic Value Monism', in: Sosa and His Critics, ed. J. Greco, 190-197, Malden: Blackwell. (2005), 'Epistetnic Value and the Pritnacy of What We Care About', Philosophical Papers 33: 353-377.

165