Pragmatic Reasons: A Defense of Morality and Epistemology

Pragmatic Reasons This page intentionally left blank Pragmatic Reasons A Defense of Morality and Epistemology Jerem...

Author: Jeremy Randel Koons

17 downloads 649 Views 1MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Pragmatic Reasons

This page intentionally left blank

Pragmatic Reasons A Defense of Morality and Epistemology Jeremy Randel Koons Georgetown University

© Jeremy Randel Koons 2009 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6-10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The author has asserted his right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2009 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN-13: 978–0–230–57696–4 hardback ISBN-10: 0–230–57696–6 hardback This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. 10 9 8 7 6 5 4 3 2 1 18 17 16 15 14 13 12 11 10 09 Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne

To my parents Don and Margaret Koons, and to Mark Nelson whose teaching ﬁrst sparked my interest in philosophy

This page intentionally left blank

Contents

Acknowledgments

viii

1 Introduction

1

Part I Pragmatism and Normativity 2 Pragmatism, Causal Explanation, and Normative Facts

13

3 Pragmatism and Rationality

46

4 Pragmatism, Normativity, and Relativism

92

Part II Pragmatism and Morality 5 Pragmatism, Interests, and Morality

121

6 Pragmatism, Freedom, and Responsibility

152

Part III Pragmatism and Epistemology 7 Pragmatism and Epistemology

189

8 Pragmatism, Internalism, and Externalism

234

Notes

263

Bibliography

280

Index

288

vii

Acknowledgments This manuscript owes its present form to the penetrating criticism of a number of readers. Mark Lance, Margaret Little, Mark Nelson, Steve Kuhn, Gregg Osborne, David Alm, Henry Richardson, Mark Murphy, and Muhammad Ali Khalidi read and provided generous feedback on portions of the manuscript. Various chapters were read at the Auburn Philosophical Society, the Alabama Philosophical Society Conference, the 32nd Conference on Value Inquiry, Society for Exact Philosophy, and the Philosophy Department at the American University of Beirut; I am grateful to these audiences for their input. Also, a grant from the Hewlett Foundation allowed me to take research leave from the American University of Beirut during the Spring 2004 semester, during which time I wrote a portion of the manuscript. My greatest intellectual debt is owed to Mark Lance, who ﬁrst introduced me to the Sellars/Brandom school of contemporary pragmatism; and Margaret Little, who (along with Mark) helped me clarify my thinking on the nature of normativity. They will not agree with all of the points set forth in this book, but without their inﬂuence there wouldn’t even be a book. Of course, any argumentative lapses in this work are to be chalked up to my account, and not theirs. Finally, my deepest gratitude is due to my wife Lucy, whose love and support sustained me through this project.

viii

1 Introduction

This book is concerned with two projects: First, it seeks to develop a novel and viable version of pragmatism, one that is free from the relativist tendencies one tends to associate with pragmatism (especially contemporary pragmatism). Second, it seeks to provide a justiﬁcation for two types of normative discourse (morality and epistemology), and shows that moral (and epistemic) constraints are rational constraints. I wish to show how this novel version of pragmatism can provide an answer to the question ‘Why be moral?’ (as well as the under-asked question ‘Why subject myself to epistemic constraints?’). Recent years have demonstrated a resurgence of interest in pragmatism. Entire journal issues have been devoted to the topic of pragmatism,1 and dozens of books on pragmatism have emerged, covering topics ranging from the classical pragmatism of Dewey, James and Peirce to ‘neo-pragmatists’ such as Rorty and Putnam. However, despite pragmatism’s current popularity, it is still viewed by many as suspect. Pragmatism has not been able to shake its associations with relativism and subjectivism. Why does pragmatism have such associations? One reason is that some traditional pragmatists write as though they had a commitment to relativism or wanted to reduce truth to mere agreement. David Bakhurst writes: Another potential obstacle to realism lies in pragmatists’ tendencies, when explaining the idea that true beliefs are those that survive the trials of inquiry, to collapse the concept of truth into that of agreement. Peirce himself writes that “the opinion which is fated to be ultimately agreed to by all who investigate is what we mean by the truth, and the object represented in this opinion is the real.”2 Here it seems that, “at the end of inquiry,” human agreement determines 1

2

Pragmatic Reasons

truth. But such a conclusion surely does violence to the idea that our beliefs are accountable to a reality that is as it is independent of how we take it to be.3 Other pragmatists like William James and F.C.S. Schiller argue for the mutability of truth. James writes, “The truth of an idea is not a stagnant property inherent in it. Truth happens to an idea. It becomes true, is made true by events. Its verity is in fact an event, a process: the process namely of its verifying itself, its veri-ﬁcation.”4 Many contemporary pragmatists, such as Richard Rorty, embrace rejection of objectivity, and treat this rejection as an essential part of pragmatism. Rorty famously writes, “A liberal society is one which is content to call ‘true’ whatever the upshot of such [free and open] encounters turns out to be.”5 He rejects philosophy’s concern with objectivity writing that an ideal liberal society would eliminate such Enlightenment ideals as objectivity and rationality: “In my view, an ideally liberal polity would be one whose culture hero is Bloom’s ‘strong poet’ rather than the warrior, the priest, the sage, or the truth-seeking, ‘logical,’ ‘objective’ scientist.”6 Although Rorty rejects objectivity, he also denies that he is a relativist, but in a way that many seekers of objectivity ﬁnd discomforting: there is no such thing as the “relativist predicament” . . . there [is] no higher standpoint to which we are responsible and against whose precepts we might offend . . . There [is] no such activity as scrutinizing competing values in order to see which are morally privileged. For there [is] no way to rise above the language, culture, institutions, and practices one has adopted and view all these as on a par with all the others.7 Another reason behind pragmatism’s association with relativism may be pragmatism’s historical concern with what is useful, which has (rightly or wrongly) led people to interpret pragmatism as relativistic or not concerned with a properly objective sort of truth. For example, Peirce writes: Truth is neither more nor less than that character of a proposition which consists in this, that belief in the proposition would, with sufﬁcient experience and reﬂection, lead us to such conduct as would tend to satisfy the desires we should then have. To say that truth means more than this is to say that it has no meaning at all.8

Introduction

3

Wiggins, commenting on this passage, notes, “This sounds awful,”9 and “This is the kind of statement that has given pragmatism such a bad name.”10 And indeed it does sound bad.11 William James is pilloried for writing things like, “‘The true,’ to put it very brieﬂy, is only the expedient in the way of our thinking, just as ‘the right’ is only the expedient in the way of our behaving.”12 Even fellow pragmatist Peirce was critical of James and Schiller, writing, It seems to me a pity they should allow a philosophy so instinct with life to become infected with seeds of death in such notions as that of the unreality of all ideas of inﬁnity and that of the mutability of truth, and in such confusions of thought as that of active willing (willing to control thought, to doubt, and to weigh reasons) with willing not to exert the will (willing to believe).13 The ﬁnal parenthetical clariﬁcation is obviously a reference to James’s famous essay ‘The Will to Believe.’ I want to argue that though pragmatism is right in being concerned with what is useful (read what serves our interests), it is wrong in identifying truth with usefulness. This latter move relativizes truth, and plays into the hands of critics of pragmatism. I will argue that not only can pragmatism justify our moral and epistemic discourse, it can do so in a way that does not lead to a relativist conception of morality and epistemology. Suspicion regarding the existence and status of moral facts is a thread running through the history of philosophy. Hume, in A Treatise of Human Nature, writes: Take any action allow’d to be vicious: Wilful murder, for instance. Examine it in all lights, and see if you can ﬁnd that matter of fact, or real existence, which you call vice. In which-ever way you take it, you ﬁnd only certain passions, motives, volitions and thoughts. There is no other matter of fact in the case. The vice entirely escapes you, as long as you consider the object. You never can ﬁnd it, till you turn your reﬂexion into your own breast, and ﬁnd a sentiment of disapprobation, which arises in you, towards this action.14 Writing two centuries later, John Mackie says, “If there were objective values, then they would be entities or qualities or relations of a very strange sort, utterly different from anything else in the universe.”15 Gilbert Harman, in an inﬂuential piece,16 argues that moral facts are

4

Pragmatic Reasons

not invoked in the best causal explanation of any uncontroversial phenomenon and hence do not exist. Curiously, although these same arguments can be applied to epistemic (as well as moral) facts, few philosophers have done so. I have argued elsewhere17 that this double standard is untenable, that traditional arguments against moral realism apply equally well against epistemological realism, and that moral and epistemic facts stand and fall together. I will not pursue such an argument in this book (except to touch on it brieﬂy in Chapter 2), but instead focus on giving a pragmatist justiﬁcation and elaboration of moral and epistemic discourse. Many pragmatists have conceded (as they ought to) that normative facts (including moral facts) do not exist in the world, and that normative terms do not denote natural kinds or real relations. However, these pragmatists argue that to concede this is not to equate such terms with ‘phlogiston’: that is, it is not to say that such terms should be expunged from the language as resting on a false theory of what the world contains. Rather, they argue that it is not the function of normative terms to denote natural kinds or causally explanatory facts. The key notion here is that these pragmatists wish to focus on the role that normative terms such as ‘refers,’ ‘knows,’ ‘immoral,’ and ‘true’ play within the language. Again, their role in the language is not to denote natural kinds. Brandom, for example, denies that the phrase ‘x refers to y’ denotes a causal relation between a word and an object; rather, he reconstructs the function of ‘refers’ in terms of anaphoric chains within the language.18 Similarly, Mark Lance and John O’Leary-Hawthorne deny that meanings are natural facts or causal relations (or sets of dispositions, etc.). Rather, they focus on the role that meaning ascriptions play in the language, and argue that to say ‘x means y’ is to endorse a norm, not to refer to any naturalistic fact about meaning.19 And in this book, we will describe the role that moral and epistemic terms play in the language, while conceding that sentences containing these terms are not justiﬁed by appeal to causally efﬁcacious or explanatorily potent natural facts about morality or justiﬁcation. Thus, by arguing that normative terms play important roles within the language, we open the possibility that what justiﬁes our use of such terms is their usefulness. In particular, I will argue in this book that morality and epistemology are justiﬁed because these normative practices satisfy certain human interests. Thus, we connect with traditional pragmatism by connecting with what is useful (i.e., with what serves our interests).

Introduction

5

But wasn’t this connection with usefulness that got pragmatism into trouble in the ﬁrst place? This is true; but traditional pragmatists often get into trouble by equating usefulness with truth, or making some similar connection. This is where this book aims to promote a novel version of pragmatism. The focus in this book is on practices, a practice being a set of rules. The practice of morality is therefore the set of rules constituting morality. Pragmatic reasons justify the practice of morality, consisting of the set of moral rules; but these pragmatic reasons cannot ‘inﬁltrate’ the practice to justify individual actions. Individual actions can only be justiﬁed by appeal to the rules constituting the practice; you may not appeal directly to pragmatic reasons to justify individual actions. Thus, pragmatic reasons can justify the rule ‘It is prima facie wrong to kill innocent humans,’ but you cannot justify killing a particular innocent human on the grounds that it is useful to do so. Individual actions can only be justiﬁed by appeal to the rules constituting the practice. Thus, we have a two-level system: pragmatic reasons justify our moral and epistemic practices; and these practices in turn justify particular actions and beliefs. Thus, the practice serves as a buffer between pragmatic reasons and individual actions and beliefs. This two-level version of pragmatism is still pragmatist (in that it justiﬁes our practices in terms of interests) yet allows for objective truth (in that truth is not deﬁned in terms of usefulness or what works). Again, pragmatism is often associated with relativism and subjectivism, and one of the novel aspects of this book is that it gives a version of pragmatism that is neither relativistic nor subjective, but which instead leaves plenty of room for objectivity. These notions are further elaborated in Chapters 2 and 4.

Chapter summaries Chapter 2 seeks to motivate pragmatism about morality and epistemology. It begins with a discussion of the charge (stemming from Gilbert Harman’s book The Nature of Morality) that normative claims cannot be veriﬁed by observation, and that normative facts do not ﬁgure in the best causal explanation of any non-controversial physical phenomena. I concede that Harman is right, but ask whether this is sufﬁcient justiﬁcation for jettisoning all moral and epistemological terminology from our language. The answer, I conclude, is ‘No’; pragmatic reasons can serve to justify continued participation in types of discourse (such as moral and epistemic discourse) that do not serve the role of causal explanation.

6

Pragmatic Reasons

The project of seeking to justify morality and epistemology pragmatically raises its own worries, such as the worry that this project conﬂates theoretical justiﬁcation (Is a belief theoretically justiﬁed?) with practical justiﬁcation (Is it useful to hold this belief?). I respond to these worries by developing a two-level version of pragmatism according to which pragmatism justiﬁes certain practices (consisting of sets of rules), but pragmatic reasons are not allowed to inﬁltrate the practice and justify individual claims within the practice. Thus, for example, a belief is justiﬁed if it is formed and maintained in conformity with the rules of the practice (which will be normal epistemic rules), but a belief cannot be justiﬁed by appeal to pragmatic considerations (i.e., on grounds that it is useful to hold this belief, or that it makes the believer happy, etc.). The strategy developed in Chapter 2 is that pragmatic considerations justify a set of rules, and individual actions are justiﬁed by appeal to these rules, and not by appeal to pragmatic considerations. This strategy allows us to give a genuinely pragmatist account of morality and epistemology, while denying that truth is mere usefulness and maintaining the connection between truth and objectivity. However, this version of pragmatism is itself vulnerable to a version of J.J.C. Smart’s ‘rule-worship’ objection against rule utilitarianism. In Chapter 3, I argue that this objection rests on a dogma that rationality must be understood atomistically, in terms of isolated agents and individual, discrete actions. Against this dogma, I defend the claim that rationality must often be understood cooperatively and in terms of strategies. In developing this answer, this chapter makes clear how acting morally is rational, and refutes those who claim that when morality and self-interest conﬂict, it is always rational to choose according to self-interest. (However, this chapter also shows how other consequentialist systems, like that of Singer, err in thinking that morality almost always trumps self-interest and places unrealistic demands on agents.) Pragmatism is often associated with relativism, an association the present account seeks to escape. In Chapter 4, I argue that reasons (in particular, prudential reasons of the sort that justify moral rules on our pragmatist account) display a ‘timelessness.’ When combined with David Wiggins’s and John McDowell’s account of truth as excellence of reasons, this feature of reasons allows us to conclude that moral and epistemic truth need not be conceived of relativistically. I will further argue that morality and epistemology have certain features that lend themselves to a social practice account of normative discourse. This account allows us to explain normativity in a way that does not offend against naturalism, and gives us a conception of morality, epistemology,

Introduction

7

and interests that need not be relativistic. The decision as to whether to conduct our moral and epistemic practices relativistically turns out also to be a decision to be made on pragmatic grounds. Chapter 5 turns to a discussion of the connection between morality and our interests. Pragmatism asserts that morality is justiﬁed by appeal to our interests, namely, the ends that constitute human ﬂourishing. It is these interests that justify participation in moral (and epistemic) practices, and which contribute to the speciﬁcation of moral (and epistemic) norms. But one might object to such an account, claiming it takes for granted a notion of interests that is immune from and prior to any sort of moral evaluation; justifying (say) morality by appeal to something which is already morally ‘loaded’ is circular and question-begging. I concede that interests are morally loaded, but argue that even if this is the case, interests may still be used to justify moral principles; this sort of circularity is, I argue, permissible. The discussion of moral norms in Chapter 5 leads to a discussion of moral responsibility and freedom. I argue in Chapter 6 that a pragmatist approach sheds light on the traditional problems of free will and moral responsibility. Free-will discourse serves as an adjunct to our moral discourse: that is, it divides actions into those for which the agent should be held morally responsible, and those for which the agent should be absolved of responsibility. The distinction between free and unfree actions turns on the question of whether a particular type of action can be altered through education and application of sanctions. Actions belonging to the type that is amenable to such alteration are free, and those that do not belong to this type are unfree. Thus, free-will judgments help enforce our cooperative moral strategy by identifying those actions that ought to be sanctioned or rewarded, thereby reinforcing moral behavior. Interestingly, the criterion that a pragmatist uses to decide which actions are free and which ones are not turns out to be the same criterion a hard determinist would use to decide which actions should be subject to sanctions or rewards and which actions should not. We will thus be able to argue that even hard determinists are committed to the pragmatist version of compatibilism. Hence, pragmatism allows us to see that from every perspective in the free-will debate— libertarian, compatibilist, and hard determinist—humans are free; free will and responsibility ascriptions are inescapable. After responding to several objections to this thesis, Chapter 6 extends the discussion to the issue of epistemic responsibility. I conclude that while familiar arguments against doxastic voluntarism (the thesis that our beliefs are under our voluntary control) undermine the idea that we

8

Pragmatic Reasons

are responsible for our beliefs, we can still evaluate the epistemic performance of agents and exclude bad performers from certain discussions and inquiries. Chapter 7 turns to a discussion of a pragmatist epistemology. Epistemology is justiﬁed by our interest in explaining and understanding the world. It emerges that this interest in fact justiﬁes two distinct but related epistemic practices—our practice of evaluating individual agents, and our practice of justifying theories or beliefs. That our epistemic practice has these two dimensions is illustrated by the fact that we make judgments of the following form: ‘Astrology as a system has been soundly refuted, but given her upbringing and community, her belief in astrology is perfectly rational.’ The latter half of this judgment reﬂects our concern with an individual agent’s epistemic performance, which is important, for example, when issues of culpable ignorance arise. The ﬁrst half of the judgment reﬂects the fact that often we do not care whether an individual agent has performed well or badly; rather we want to know what it is that is rational to believe. In this chapter, I develop these two different dimensions of epistemic evaluation, and show how pragmatism helps us ﬂesh out these two evaluative perspectives. I further argue that many (perhaps most) traditional epistemological theories do not have the resources to capture the distinction between these two evaluative dimensions, and that this represents a signiﬁcant advantage that pragmatist epistemology has over these other theories. In Chapter 8, I argue and conclude that this two-dimensional pragmatist approach to epistemic evaluation—the two epistemic practices justiﬁed by pragmatist considerations, sketched in Chapter 7—sheds light on such important issues as the internalism–externalism debate in epistemology. In particular, the two evaluative perspectives explored in Chapter 7 correspond to internalism and externalism in epistemology: the pragmatist evaluation of individuals’ epistemic performance corresponds to internalism, and satisﬁes the intuitions that have traditionally motivated internalist epistemologies, whereas judgments about what it is that is rational to believe correspond to externalism, and this perspective satisﬁes the intuitions that motivate externalist epistemologies. Thus, Chapter 8 argues that adopting a pragmatist view of epistemology allows for a reconciliation of internalism and externalism in epistemology. I will also argue that the traditional connection between justiﬁcation and truth has been misunderstood, and I present a new way of understanding the connection that sheds light on the internalism–externalism debate.

Introduction

9

A ﬁnal note: this book is not intended as a historical exposition of pragmatism; nor do I make every attempt to trace the historical connections between (say) my account and the philosophies of the early American pragmatists. In places where there is a direct connection between my view and that of a historical pragmatist, I will discuss this connection; but I am more interested in developing a viable version of contemporary pragmatism than I am in proving that the theory in this book belongs to the pragmatist genealogy. Those who are familiar with pragmatism will recognize in this book many elements central to that theory: fallibilism, the connection of our philosophical practice with our interests, the emphasis on social practice, and so forth. I rest content with that.

This page intentionally left blank

Part I Pragmatism and Normativity

This page intentionally left blank

2 Pragmatism, Causal Explanation, and Normative Facts

The term ‘epistemology’ refers to two rather different practices.1 Construed broadly, it is our actual epistemic practice. That is, it is the practice common in every ﬁeld from chemistry to philosophy, of evaluating beliefs as justiﬁed or unjustiﬁed, scientiﬁc methods as rational or irrational, etc., and of evaluating and revising not only those beliefs, but the standards of evaluation themselves. Construed more narrowly, it is explicit theorizing about epistemology in the broad sense. This theorizing can be the attempt to construct a set of formal conditions for the justiﬁcation of beliefs (e.g., explanatory coherence among beliefs; proper inferential relation to foundational beliefs; proper causal genesis, à la reliabilism, etc.); or it can merely be an attempt to uncover some of the features of justiﬁcation and knowledge, without any commitment to the possibility of formal theorizing. Epistemology in the narrow sense is in essence our practice of theorizing about epistemology in the broad sense, about our practice of epistemology. We can be committed to the viability of epistemology in the broad and the narrow sense, and think that we can make true theoretical claims about justiﬁcation and knowledge, without necessarily thinking that we can construct a formal theory of justiﬁcation. Similar comments can be made about the term ‘morality’: this term can be construed broadly, as our practice of praising certain acts and traits and condemning others; or it can be construed narrowly, as our practice of theorizing about morality in the broad sense. Again, a philosopher might be committed to morality in the narrow sense without thinking that we can construct a formal theory (such as utilitarianism or Kantianism) of right and wrong.2 In both the epistemic and moral cases, a person can make moral and epistemic judgments without understanding the theoretical underpinnings of these judgments. That is, one can practice morality and epistemology 13

14

Pragmatism and Normativity

in the broad sense without themselves having a grounding in theoretical morality and epistemology. Although most philosophers are convinced of the viability of epistemology in this broad sense, many are somewhat less sanguine about the prospects for normative moral discourse (again, in the broad sense). Some of this skepticism stems from an inﬂuential objection raised by Gilbert Harman in the ﬁrst two chapters of his 1977 book The Nature of Morality. According to Harman, moral facts neither ﬁgure in the best causal explanation for any observations nor are reducible to facts which are explanatory. Harman infers from this that there are no moral facts, and that some form of moral nihilism must therefore be true. While Harman’s explanatory requirement (as it has been named) has caused a good deal of worry concerning moral discourse, fewer philosophers have worried about its ramiﬁcations for epistemological discourse. Unfortunately, philosophers who use epistemological discourse in the broad sense—and this is the sense with which I will be concerned in this chapter—are in no position to demand that moral facts serve a causal role in explaining observations, as epistemological facts serve no such role either.3 If there are no causally efﬁcacious moral or epistemological facts, then what justiﬁes our continued use of moral and epistemological terms? After all, a discovery that there are no unicorn facts in the world, capable of causally impacting our cognitive mechanisms, should surely cause us to be eliminativists about unicorns—that is, to claim that there simply are no unicorns, and that statements to the contrary are false (or meaningless). There is, however, an alternative to eliminativism about morality and epistemology. I would like to suggest, rather, that those who are committed to epistemological discourse must concede that pragmatic reasons can serve as a bulwark against eliminativism, even if the discourse in question does not provide genuine causal explanations. The argument of this chapter, then, has two parts. First, I am presenting an equality argument: I am arguing that epistemology and morality stand and fall together. Each is as robust (or feeble) as the other. Thus, those philosophers—such as Harman, Mackie, and Ayer—who believe in a robust epistemology but who are eliminativists or nihilists about morality have it badly wrong.4 The second part of this chapter is an account of what, precisely, justiﬁes our continued participation in the normative practices of morality and epistemology. Again, I will argue that pragmatic reasons justify our continued participation in both practices.

Pragmatism, Causal Explanation, and Normative Facts

15

I hope, too, that this chapter helps to clear up some other worries philosophers have about the status of moral discourse. For, eliminativism and full-blown moral realism are not the only possible stances to hold regarding moral discourse. Far more interesting and important are the intermediate positions vis-à-vis moral discourse held by legions of philosophers. These are the views of philosophers such as Blackburn, Gibbard, Wright, and others, who endorse a sort of ‘quasirealism.’ According to them, moral utterances can be true, and the true ones even have some form of authority over us; but these moral utterances are not as robustly true as, say, true scientiﬁc utterances. But there is a residual unease in adopting this middle position. ‘Quasi-realism’ (if I may hijack Blackburn’s term) brings with it the worry that moral discourse is second best; it may still be alive, but it only limps along. But the conclusion of this chapter should alleviate some of these concerns. After all, if I am right, then moral discourse is on par with something as robust as epistemological discourse. Epistemological discourse is the very thing philosophers spend their life employing; it is their stock in trade. After all, what philosophers do is advance and evaluate arguments; philosophers try to tell us what it is that is rational to believe about ethics, about the meaning and reference of words, about mental states. Without epistemology, philosophy as we know it grinds to a halt. Science does the same: how can science succeed if there is no distinction between science and pseudoscience, rational and irrational research methods, justiﬁed and unjustiﬁed scientiﬁc beliefs, and so on? I will have much more to say on this topic later in the chapter. Let us begin.

The explanatory requirement Harman begins his criticism of moral realism by asking the question, “Can moral principles, like scientiﬁc ones, be tested against the world?” The ﬁrst response given to this question is that we do, indeed, observe moral properties, in the sense that we can have a perceptual experience (say, hoodlums setting a cat on ﬁre) and form (non-inferentially) the judgment, ‘That is cruel.’ Harman writes: If you round a corner and see a group of young hoodlums pour gasoline on a cat and ignite it, you do not need to conclude that what they are doing is wrong; you do not need to ﬁgure anything out; you can see that it is wrong.5 But do we perceive moral properties in the same way we perceive scientiﬁc properties and objects? Many point out that we perceive moral

16

Pragmatism and Normativity

properties only because we have been trained to do so. But this holds of scientiﬁc properties, as well: it is fairly well-accepted in philosophy these days that there is no such thing as an uninterpreted observation language.6 There is no theory-neutral observation language, and theories are the sorts of things that are taught and learned. However, Harman thinks he has another tool for prying apart moral and scientiﬁc perception: the explanatory requirement. It is a current tenet of epistemology that we are justiﬁed in positing the existence of an entity if this entity is invoked in the best explanation of some accepted phenomenon. For example, say a scientist sees a vapor trail in a cloud chamber. The best explanation for this observation is that a proton just went by; therefore, we get to be realists about protons. Now let us consider the case of the cat. What do we need to invoke to explain our observation that igniting the cat was cruel? Certain psychological facts, facts about upbringing, etc., enter into the equation, but, Harman argues, cruelty need not. The best explanation of the observation does not invoke cruelty or posit moral properties; thus we do not get to be realists about morality. Clearly, Harman is on to something. I am willing to concede that moral facts can never causally explain non-moral facts. This is actually a matter of dispute; some argue to the contrary.7 My concession also cuts somewhat against the pragmatist grain, as many pragmatists (instead of conceding the premise of objections like Harman’s) reply by trying to blur the fact–value distinction.8 However, I wish to take the weaker position and argue that even if normative facts do not ﬁgure in the best explanation of any observations, we still need not be eliminativists about morality and epistemology. If causal explanatory impotence is a sufﬁcient reason to jettison morality, then much else ends up going with it. I will argue that if we take Harman’s requirement seriously, we will have to jettison much discourse we should be reluctant to discard.

Morality and causal explanation To be an explanatory eliminativist about a particular set of entities or sentences is to deny that those entities or sentences are genuinely explanatory.9 Do moral facts enter into causal explanations? Harman suggests not, and without giving an argument, I am willing to concede that they do not. While we might casually employ moral facts in workaday explanations (‘The overseer’s cruelty ultimately led to the slaves’ revolt’), these moral facts are eliminated in a more penetrating analysis (‘The overseer’s treatment of the slaves caused them great suffering,

Pragmatism, Causal Explanation, and Normative Facts

17

which in turn bred rampant resentment among the slaves, leading to their revolt’). Moral facts certainly do not appear in the most penetrating causal explanations (i.e., scientiﬁc explanations).10 There is an important misunderstanding which might arise at this point. Many will claim that moral facts do enter into some causal explanations: they enter into rational explanations. Whether rational explanations are causal is controversial. However, even if rational explanations are causal, moral facts enter into these explanations in a way that is consistent with the causal impotence of these facts. Let us suppose that rational explanations are causal. I assume the reader has an intuitive idea of the shape of rational explanation: these explanations explain agents’ actions in terms of reasons. Moral facts ﬁgure in some rational explanations. Thus, though not all rational explanations involve moral properties, some such explanations will: ‘I returned the lost wallet because it was the right thing to do.’ It would be misleading, though, to think that the action’s rightness caused me to return the wallet. Harman, for one, would deny that this is the case. Furthermore, I am inclined to agree with him. I returned the wallet not because it was the right thing to do, but because I believed that it was the right thing to do. Sellars writes that “obligation enters into the causal order only as an element in the intentional object of a mental act,” that is, “via facts of the form Jones thinks (feels) that he ought to pay his debt.”11 Thus, moral properties and facts enter only into rational explanations, and only as the objects of intentional states. I think anyone will concede that this degree of causal involvement is not sufﬁcient to vindicate moral realism; anything can serve as the object of an intentional state. After all, it might well be true that Jones believes that his chronic headaches are caused by the tiny unicorns which dance on his forehead every night while he sleeps, and this belief may even cause Jones to take certain actions (e.g., grease his forehead so the unicorns lose their footing); so the belief is causally efﬁcacious. Even if this belief attribution is true, we shall still be eliminativists about tiny dancing unicorns (I hope). Thus, even if rational explanations are causal, the involvement of moral facts in such explanations is not sufﬁcient to vindicate moral realism. Of course, not all explanations are causal. Consider two examples: ﬁrst, there are inﬁnitely many primes, because for any number n, if you take the ﬁrst n primes, multiply them together and add 1, the result is a new prime; and repeating this procedure will reveal ever-larger prime numbers. Another example is this: ‘We do not know who will win the election because given the sample size of our study, its margin of error

18

Pragmatism and Normativity

was greater than the predicted vote differential.’12 Both of these seem to be non-causal and also genuinely explanatory—they not only give reason to believe the claim in question, but show why the claim is true. And as van Fraassen, for example, has convincingly argued, explanations are those propositions that answer ‘why’ questions. An important category of non-causal explanation is the category of normative explanations. When we give a normative explanation, we explain how an action or feature alters or contributes to a particular normative characterization of an action, person, or object. For example, while watching (American) football I might say, ‘His crossing the 35-yard line while carrying the ball made it ﬁrst down.’ The football player’s crossing the 35-yard line does not cause the ﬁrst down; rather, it changes the normative status of the players in the game. True, his crossing the 35-yard line initiates certain causal chains which result in certain actions that accompany the awarding of a ﬁrst down—movement of the down markers, etc.—but the arrival of a ﬁrst down is itself a normative occurrence, not a causal one. In the moral case, I might say, ‘That was cruel because it caused her pain.’ Her pain does not cause cruelty; nor does the cause of her pain cause cruelty. The ‘because’ in this explanation is a normative, not a causal one. The same is true in epistemology, even in the versions that rely on causal chains: if I say that belief B is justiﬁed because it was caused in the right way, I am asserting that the belief, not the justiﬁcation, was caused. Rather, the justiﬁcation is an additional, normative fact that is not caused by the belief, nor by the cause of the belief.13 If this sort of epistemological theory is right, a belief’s justiﬁcation is explained by that belief’s causal etiology, but this is a normative sort of explanation, not a causal one. Examples of normative (and hence non-causal) explanation can be found in any normative endeavor, from epistemology to morality.14 To sum up, we have seen there are causal and non-causal explanations. Moral facts ﬁgure in certain non-causal explanations, to wit, in some normative explanations. Morality may enter the causal order through rational explanations, but only as the object of an intentional act or state. Rational explanations in which morality seems to play a more robust causal role (e.g., ‘I returned the wallet because it was the right thing to do’) are more perspicuously rendered (e.g., ‘I returned the wallet because I believed it was the right thing to do’) to reﬂect the causal impotence of moral facts. Thus, I am willing to concede the premise of Harman’s objection: morality does not enter, in any interesting way, into causal explanations. That is to say, I concede causal explanatory eliminativism with

Pragmatism, Causal Explanation, and Normative Facts

19

regard to moral discourse. Does this imply doctrinal eliminativism about morality, that is, the claim that all moral utterances are false, or that they are neither true nor false? If it does, then epistemology goes with it, because this causal impotence is echoed by normative epistemological discourse.15 That is to say, epistemological properties are equally impotent. I argued above that beliefs can be causes. However, justiﬁcation and warrant are not themselves causally efﬁcacious. All the same moves which can be made to deny that moral properties are causally efﬁcacious can be made to deny that epistemological properties are causally efﬁcacious. So we can say, ‘I believe P because P is justiﬁed,’ but just as in the moral case, the real causal explanation is ‘I believe P because I believe that P is justiﬁed.’ The same move can be made for prudential explanation too (‘I did Y because it was optimal’ becomes ‘I did Y because I believed it was optimal’). So if causal explanatory impotence entails doctrinal eliminativism, then epistemology and rationality are eliminated as well. Epistemic facts are causally impotent even if epistemology is naturalized, as writers such as Alvin I. Goldman have attempted to do. For these philosophers justiﬁcation or knowledge consists in the appropriate sort of causal connection to natural facts. But notice that it is the natural facts that are causally efﬁcacious; the epistemic facts are themselves causally impotent. The epistemic facts consist of certain causal relations, but causal relations do not themselves cause anything; it is the relata, not the relation, that are causally efﬁcacious. For example, suppose on a pool table, the cue ball strikes the eight ball and causes it to roll into the corner pocket. Thus, a causal relation obtains between the two pool balls. Notice, though, that the causal relation is not itself causally efﬁcacious (if it were, I suppose perception of causal relations would be unproblematic, and Hume would never have become a skeptic about them); rather, the relata (the cue ball and the eight ball) can be causes, the relation cannot. So epistemic facts are causally impotent even if epistemology is naturalized. Furthermore, Harman’s eliminativist move can be applied here: if your explanation for why S believes that P is S’s belief that P was caused in a certain way, then you have a complete and naturalistic explanation of S’s belief that P. That P is justiﬁed (or any other normative epistemic ‘fact’) is otiose; it is not part of the best, most pared-down explanation of why S believes that P, and so can be eliminated. This is parallel to the move on the moral side: once you have analyzed the observation of the cat being set alight, you will realize that ‘cruelty’ is not something that is part of the explanation of your observation. Nor is

20

Pragmatism and Normativity

‘justiﬁcation’ an ineliminable part of the explanation for S’s believing that P (unless you simply reduce by ﬁat ‘cruelty’ or ‘justiﬁcation’ to the set of observed facts). We will discuss reductionist theories of morality and epistemology later in this chapter, although I argue in my article ‘An Argument Against Reduction in Morality and Epistemology’16 that such theories fail. Clearly, we may colloquially employ phrases in which moral and epistemological facts appear to serve a causal explanatory role. Why did Selim Woodworth fail to rescue the Donner party? Because he was no damned good.17 Why do you believe that the passenger pigeon is really extinct? Because the best reasons support this conclusion. Harman’s point (although he only makes it about morality) is that the moral/epistemological part of the explanation is eliminable, and does not appear in a more penetrating natural-scientiﬁc explanation. We can thus see that the cost of adopting the explanatory requirement would be high: we would be deprived of epistemological discourse as well as moral discourse. We will see in a moment how such pragmatic considerations are relevant in the realism–antirealism debate.

Eliminativism and pragmatic reasons We have so far concluded that causal explanatory eliminativism obtains with regard to morality and epistemology. In other words, neither moral nor epistemological facts are genuinely causally explanatory. What does this show about morality and epistemology? I think it shows at most that ontological eliminativism obtains with regard to these types of discourse (although I will neither insist on this conclusion nor argue for it; I am willing to concede it for argument’s sake). To be an ontological eliminativist about a particular subject matter is to claim that certain nounlike words or phrases belonging to that subject matter do not refer. Most of us are ontological eliminativists about unicorns and phlogiston; many are ontological eliminativists about holes, shadows, and the like. Thus, I am willing to concede something like the following entailment: if K-talk is not involved in the best causal explanation of any uncontroversial phenomenon, then K-talk does not refer.18 The mistake is thinking that failure to satisfy the explanatory requirement entails doctrinal eliminativism. Recall, from a few pages ago, that doctrinal eliminativism holds that none of the (supposedly assertoric) sentences belonging to a particular subject matter is ever true, either because they are all false, or because they are neither true nor false. Consider the

Pragmatism, Causal Explanation, and Normative Facts

21

sentence, ‘The average American mother has 2.3 children.’ It is uncontroversial, I take it, that there is no entity to which this sentence refers; there is no such thing as the average American mother. Indeed, I imagine that there is no causal explanation in the language of austere physics that makes appeal to this entity. So we have an example of a sentence that has the following two properties: the sentence uses a non-referring nounlike expression (which expression fails to satisfy the explanatory requirement), and the sentence is uncontroversially true. Hence, causal explanatory eliminativism does not entail doctrinal eliminativism. In other words, K-talk can be true even if it does not refer.19 (Examples can be multiplied indeﬁnitely. Consider ‘John has a big ego.’ We may concede that this phrase conveys information, and can be true, even if we are ontological and causal explanatory eliminativists about egos.)20 Epistemology and pragmatic reasons What, then, is the correct test for doctrinal eliminativism? I want to argue that they are ineliminable because they are an essential part of a larger, successful pragmatically justiﬁed project. Let me start by discussing Thomas Kuhn’s article ‘The Essential Tension’ (1959). In this article, he notes that we typically think that one of the chief virtues of a successful scientist is the ability to engage in so-called ‘divergent thinking’: the ability to look at obvious facts without necessarily accepting them; the ability to actively imagine unique possibilities, instead of merely interpreting facts in the way taught in textbooks and by scientiﬁc authorities. Kuhn acknowledges the importance of divergent thinking; without it, there would be no scientiﬁc revolutions. Ptolemaic astronomy would never have been discarded in favor of Copernicanism; oxygen would never have dislodged phlogiston; and Einstein would not have supplanted Newton. But very little of science is revolutionary; ‘normal’ science is more akin to ‘puzzle solving’ than to revolution.21 And in normal science, Kuhn emphasizes, ‘convergent’ thinking is more useful than its divergent counterpart. [N]ormal research, even the best of it, is a highly convergent activity based ﬁrmly upon a settled consensus acquired from scientiﬁc education and reinforced by subsequent life in the profession. Typically, to be sure, this convergent or consensus-bound research ultimately results in revolution. Then, traditional techniques and beliefs are abandoned and replaced by new ones. But revolutionary shifts of a scientiﬁc tradition are relatively rare, and extended periods of convergent research are the necessary preliminary to them.22

22

Pragmatism and Normativity

The textbooks of an era indoctrinate scientists-in-training with the current paradigms. Kuhn notes that in the latter half of the 19th century textbooks stated unequivocally that light was wave motion. Before that, books on optics were equally unequivocal in their assertions that light consisted of particles. And of course now textbooks state (again, unequivocally) that light exhibits properties of both waves and particles. It may seem odd that textbooks and scientists are so dogmatic in asserting theories that ultimately prove false, but Kuhn argues that science would not have achieved its current lofty stature were it not for its emphasis on convergent thinking. The history of theories of light does not . . . begin with Newton . . . From remote antiquity until the end of the seventeenth century there was no single set of paradigms for the study of physical optics. Instead, many men advanced a large number of different views about the nature of light . . . As a result, a new man entering the ﬁeld was inevitably exposed to a variety of conﬂicting viewpoints; he was forced to examine the evidence for each, and there was always good evidence . . . This earlier mode of education was obviously more suited to produce a scientist without prejudice, alert to novel phenomena, and ﬂexible in his approach to his ﬁeld. On the other hand, one can scarcely escape the impression that, during the period characterized by this more liberal educational practice, physical optics made very little progress.23 This preparadigm or preconcensus period in optics is echoed in virtually every other discipline, from dynamics to physiology. And further, the study of the preparadigm periods of these other disciplines strongly suggests that “without a ﬁrm consensus, this more ﬂexible practice will not produce the pattern of rapid consequential scientiﬁc advance to which recent centuries have accustomed us.”24 Without a theory to which one is committed, one does not know what phenomena are signiﬁcant, and what problems are worth-solving (or perhaps indeed what the problems are in the ﬁrst place). Further, one must be committed to a theory to undertake the serious work which is required to extend, deepen (and ultimately overthrow) the theory. “Who, for example, would have developed the elaborate mathematical techniques required for the study of the effects of interplanetary attractions upon basic Keplerian orbits if he had not assumed that Newtonian dynamics, applied to the planets then known, would explain

Pragmatism, Causal Explanation, and Normative Facts

23

the last details of astronomical observation? But without this assurance, how would Neptune have been discovered and the list of planets changed?”25 How does any of this bear on the project of epistemology? The answer is this: epistemic discourse is one of the tools that enable a consensus to arise and be maintained, and is therefore one of the tools that speed scientiﬁc progress, which is something we value. The job of epistemology (in the broad sense) just is to sort beliefs into those that are to be believed, and those that are not. For a physicist working in 1800, it was rational to believe that Newtonian physics was correct. It was rational to carry out research projects using Newtonian techniques, and it would have been irrational to attempt to use Aristotelian physics to account for planetary motion. A physicist who incorporated unjustiﬁed techniques (such as those of Aristotelian physics) would have been sanctioned by partial or complete exclusion from the profession of physics.26 Those who were adept at applying Newtonian techniques to the study of planetary motion were rewarded by inclusion. Epistemic evaluation can be discursive or practical. In other words, we can call a methodology irrational or a belief unjustiﬁed, or we can carry out our scientiﬁc practice in a way that favors this methodology, or that presupposes the falsity of that belief. And when our 18th-century professional physicists applied Newtonian techniques to the study of planetary motion, they were practically (though not discursively) endorsing these techniques. Similarly, the exclusion of the Aristotelian physicists from their ranks was a form of practical evaluation; it was an action that practically judged belief in Aristotelian methods to be unjustiﬁed. This practical evaluation is crucial to scientiﬁc progress. As Kuhn pointed out, if we do not practically favor a particular paradigm over all others, science does not progress rapidly. Rapid scientiﬁc progress requires practical epistemic evaluation, which can help enforce uniformity of methodology— the crucial element in scientiﬁc advancement, according to Kuhn. Hence, epistemology serves the pragmatic end of hastening scientiﬁc progress. It might be objected that epistemic practice, even in the broad sense, is not the sort of thing that can be implicit in scientiﬁc practice. Epistemology in this sense refers to the labeling of beliefs as justiﬁed and unjustiﬁed, the labeling of methods as rational and irrational, and this labeling must be explicit (spoken, written, etc.). Hence (goes the objection), something other than epistemology does the primary work of ensuring scientiﬁc consensus.

24

Pragmatism and Normativity

But evaluative beliefs, such as those of epistemology, are surely the sort of thing that can be implicit in practice. For example, a person may ﬁrmly believe that lying is wrong, even if she never utters the sentence ‘Lying is wrong.’ Her belief can be expressed by her constant truth-telling, especially in those cases where lying would clearly be more expedient for her. Alternately, her belief in the wrongness of lying could be manifested by her feelings of guilt when she does tell a lie. Similarly, my belief that the dishwasher will wash my dishes is implicit in my actions (placing dirty dishes in the washer and turning it on) even if I do not utter the sentence ‘My dishwasher will wash my dishes’ and do not use the belief in a bit of explicit internal practical reasoning. We can evaluate or believe on particular occasions without saying anything, even in foro interno. Our evaluation or belief is explicit in some cases, and implicit in others. In fact, I imagine that the majority of our practice occurs at the level of the implicit—we only proceed to the level of explicit discourse when we need to inform, educate, or argue.27 So is it with epistemological evaluation. Our belief that double-blind studies are more reliable than chicken bones is implicit in our decision to perform a double-blind study instead of visiting a practitioner of bone-reading. Thus, epistemology is a crucial element of scientiﬁc progress: without convergent practical epistemological evaluation implicit in scientiﬁc practice, science advances only slowly. We value rapid scientiﬁc advancement. Hence, epistemology as a practice is justiﬁed pragmatically.28 Epistemic discourse serves other uses. At the level of explicit discourse, it can allow for revision of our scientiﬁc and other sorts of truth-seeking practices, thereby serving an auxiliary role to such practices. For example, I might suggest to you that instead of eating peyote in an attempt to contact the spirits and ﬁnd out from them whether the sun orbits the earth, you might read the Bible and ﬁnd the answer there. ‘Why should I do that instead of eating peyote?’ you inquire. I reply that reading the Bible is more likely to lead you to the truth on this matter, or that the Bible contains true things whereas peyote just makes you hallucinate. In other words, I say something that has the same practical signiﬁcance as the epistemological sentence, ‘Consulting the Bible is a more rational method of investigating scientiﬁc matters than is the consumption of peyote.’ I am communicating to you, either practically or discursively, my belief that consulting the Bible is a more (epistemically) sound method than is eating hallucinogens. Consider another use of epistemic evaluation. It allows me to know what sort of practices confer commitments and entitlements on me. For

Pragmatism, Causal Explanation, and Normative Facts

25

example, if you tell me that Stephen Hawking is a competent physicist, I know that I am entitled to believe his pronouncements on physics. Furthermore, I am under an obligation or commitment to defer to Hawking on questions of physics. Again, this sort of evaluation can be discursive (you can say ‘You are justiﬁed in believing what Stephen Hawking says about physics’), or practical (suppose that, as a matter of practice, you defer to him on questions of physics because you recognize him as an authority).29 We will have much more to say about the pragmatic import of epistemic evaluation in Chapters 7 and 8. The point that epistemic evaluation can be implicit in practice rather than being necessarily explicit allows us to extend our argument beyond science, and demonstrate the pragmatic necessity of epistemic evaluation in everyday life. We have already seen some of the roles implicit epistemic evaluation can play in our everyday inquiry. However, there is a more radical sense in which epistemic evaluation is implicit in our practice. To enter into rational discourse in the ﬁrst place is to exhibit practically a division of beliefs into those that are to be believed and those that are not; this follows from the fact that in such discourse we are presented with competing claims, and take ourselves to be entitled to choose between them on some basis.30 Even to deny that we can choose between two beliefs—to endorse skepticism—is to take an epistemic stance. Thus, our epistemic practice consists of a range of implicit practices of evaluating beliefs, endorsing some and rejecting others. Our very ability to engage in rational discourse depends on such a practice. Thus, there is a powerful pragmatic argument justifying our epistemic practice.31 We will return to our pragmatic construal of epistemology in Chapter 7. In the meantime, let us turn to the role of pragmatic reasons in justifying morality.

Moral eliminativism and pragmatic reasons It appears, then, that our epistemic practice is pragmatically justiﬁed. Just as particular scientiﬁc beliefs are justiﬁed by being part of a larger, successful explanatory project, our individual epistemic judgments are justiﬁed by virtue of belonging to a larger, pragmatically undertaken project. This is an important result, for (assuming we are not yet eliminativists about epistemology) it allows us to appeal to pragmatic considerations in defending a discourse against doctrinal eliminativism (recall that doctrinal eliminativism is the thesis that the allegedly assertoric sentences of a discourse are all false, or neither true nor false). And, of course, in the absence of explanatorily potent epistemic and moral

26

Pragmatism and Normativity

facts, this is exactly the sort of considerations we must appeal to to justify our epistemic and moral practice. Certainly, not all pragmatic justiﬁcations are created equal, just as not all explanatory justiﬁcations are created equal (phlogiston fails; oxygen succeeds). Not all sorts of discourse are worthy of being pursued. The goal ‘to maintain the political power of the clergy’ would not be a particularly good reason to continue pursuing a certain type of discourse. However, the goal ‘to facilitate human interaction’ might be an excellent goal, as is ‘to explain the natural world.’ The point is, a question of whether a type of discourse is robust and worthy of pursuit is not a question that can be decided on a priori grounds, merely by gesturing to natural facts. This is a question which is decided on an a posteriori basis, on the basis of reasons (among them pragmatic reasons). We will argue in Chapter 5 that pragmatically structured practices such as morality and epistemology should be structured to serve our fundamental human interests, a concept we will ﬂesh out at that point in the book. I should also note that not all means to pragmatically justiﬁed ends are created equal, either. While evaluating the intentions of a potential suitor is probably worth doing, consulting a psychic is not an effective means to this end. It does not serve the pragmatic end we want it to serve. Moral discourse, however, clearly seems to facilitate human interaction. It is an effective means to the pragmatically justiﬁed end of interacting safely and proﬁtably with our fellow humans. Epistemological discourse seems to be an effective tool for advancing our scientiﬁc and other truth-seeking practices. Thus, as both morality and epistemology are effective means to worthwhile ends, we should continue to practice both. I do not think that I can over-emphasize the role of pragmatic reasons in questions of doctrinal eliminativism. Not only is it wise to consider pragmatic reasons in these debates, but our actual practice recognizes the importance of such reasons. Consider an example of how pragmatic reasons argue in favor of continuing a particular practice.32 The rules of games such as basketball are human inventions; no one argues (I hope) that they are answerable to basketball facts in the world. It is important to note, though, that no one argues on these grounds that we ought to quit playing the game of basketball. What would be the point of quitting? (Note that when we ask what the point of a move would be, we are asking a pragmatic question.) Would we be better off not playing basketball? Of course not. Notice how out-of-order, and indeed bizarre, we would regard a Harman-style attack on the practice of basketball (basketball rules do not explain observations, are not reducible to natural facts,

Pragmatism, Causal Explanation, and Normative Facts

27

etc.). Such objections are beside the point, because explanation of the natural world is not part of the aim of basketball. The point of basketball is to have fun, and perhaps also to get exercise. Basketball achieves the pragmatic ends at which it aims, and as these ends are themselves acceptable ones (remember, not all pragmatic ends are worthy of pursuit), we have good reason to continue the practice of basketball. The case of basketball is analogous to the moral case: moral facts do not stand in causal, explanatory relation to natural facts, but that is not the point of moral discourse, anyhow. Moral discourse serves an end other than explanation, and we would not be better off eliminating morality. Thus, Harman-style challenges are just as inappropriate to our moral practice as they are to our practice of basketball. Criticizing morality because it is a poor explanatory tool is like criticizing a wine glass because it is a poor hammer. Explanation is not morality’s purpose. And again, by talking about the purpose or point of a discourse, we are discussing pragmatic questions. I would argue that morality survives the a posteriori challenge. After all, would we be better off not playing the moral game? It seems likely that the answer is ‘No.’33 Just as the rejection of epistemology would be tantamount to giving up the ability to sort beliefs between those to be believed and those not to be believed, so would the elimination of morality be tantamount to giving up the ability to sort actions between those to be done and those not to be done. As pragmatic considerations can defend theories against doctrinal elimination, and as moral discourse acquits itself admirably on pragmatic grounds, I should think that we have an excellent reason to be doctrinal realists about moral discourse. Thus, the justiﬁcation I offer for morality is a sort of ‘decynicized Hobbesianism.’34 We have interests—selﬁsh and other-regarding—and these interests of ours are best met if we enter into a cooperative arrangement: morality. I will, of course, discuss the pragmatic justiﬁcation of morality in more detail in Chapters 3 and 5; a more detailed discussion of pragmatism and epistemology will have to wait until Chapters 7 and 8. One might argue that, given the course of my argument, I do not have the option of grounding morality and epistemology in what are, ultimately, prudential35 reasons.36 After all, prudential reasons—reasons about why I ought (rationally) to perform a given action—are also causally impotent. Thus, one might conclude that moral, epistemic, and prudential reasons are all equally hardy—none is more, or less, robust than the others. But (goes the objection) how can I maintain both that (a) morality and epistemology are grounded in prudential reasons, and

28

Pragmatism and Normativity

that (b) prudential reasons are no hardier or more secure than moral and epistemic reasons? I think we need to distinguish between (1) a discourse’s being hardier or more secure than another and (2) a discourse’s being logically prior to another. Prudential reasons are logically prior to moral and epistemic reasons, but moral and epistemic reasons are no less hardy than the prudential reasons out of which they arise. Consider an analogy: suppose we want to prove a logical or mathematical theorem T2. As a preliminary step, we need to prove another theorem T1; T1 is an essential element in the proof of T2. Thus, T1 can be said to be logically prior to T2 in the sense that the proof of T2 rests on T1. But we should not infer that T1 is better-grounded than T2; both theorems are certain and necessary. So, logical priority does not translate into greater justiﬁcatory robustness. In any case, the thesis of the chapter is that morality and epistemology are on a par; as both rest on prudential reasons, my thesis is not harmed if it turns out that prudential reasons are bettergrounded than moral or epistemic ones. For the reasons just rehearsed, though, I do not think prudential reasons are better-grounded. However, this objection raises more serious worries. The argument of this chapter is that even though moral and epistemic facts are causally impotent, it is nevertheless reasonable to regulate our behavior by moral and epistemic norms. But if prudential reasons are equally impotent, on what basis is it reasonable to regulate our behavior by prudential norms? Ultimately, this is not a question that can be answered. If we wish to provide reasons for why someone should regulate her behavior with prudential reasons, then we shall fail. Ultimately, the question ‘Why should I do X, which is in my self-interest?’ is one that can only be answered from within the perspective of prudential reasons; no external reason can be given. In this way, the current account rests morality and epistemology on prudential reasons, while acknowledging that the latter cannot be justiﬁed. Some philosophers will no doubt claim that we ought to continue to use moral and epistemological talk, but with our ﬁngers crossed, so to speak. That is to say, we should continue to use this discourse at the same time acknowledging that the sentences of this discourse are not really true. But if this is all that doctrinal eliminativism amounts to, then its practical import is limited. If doctrinal eliminativism about moral discourse concedes that we ought to continue practicing moral discourse, then doctrinal eliminativism does not represent a threat to the robustness of our moral practice. A philosophical critique of a particular practice is not particularly damaging if, in the end, it concludes that we ought to carry on pretty much as before. And in any case,

Pragmatism, Causal Explanation, and Normative Facts

29

morality and epistemology would be equally deﬂated, so the thesis of my chapter—that morality and epistemology are equally objective— remains undamaged. We will, however, return later in the chapter to the charge that pragmatism entails that moral and epistemic truth is somehow different from other sorts of truth. Harman’s objection, thus, is off the mark because the point of moral and epistemic discourse is not to explain (and as I noted above, when we talk about the point of a practice, we are talking in pragmatic terms). Harman’s objection might be relevant if directed at a discourse which offered allegedly scientiﬁc explanations, but we need not regard this as the pragmatic end served by moral and epistemic discourse. And we ought not adopt Harman’s explanatory requirement as a test for doctrinal eliminativism because the pragmatic cost of such a move would be too high. We would have to abandon every practice which served an end other than explanation. Morality, epistemology, and other areas would be eliminated. Would we be better off then? I have argued that we would not be better off if we did not practice epistemology; and I will assume that we are better off practicing morality than not. Thus, we ought to continue practicing morality and epistemology. There is an important lesson to be gleaned from all of this. A few pages back, I wrote that as Harman’s explanatory requirement was unsuitable to serve as a criterion for doctrinal eliminativism, I needed to propose a criterion of my own, so as not to open the ﬂoodgates to all sorts of unacceptable discourse (astrology, tea leaf reading, etc.). It now seems clear to me, however, that no such criterion can be suggested which will apply to all cases. We cannot say a priori what the correct connection is between acceptable sorts of discourse and the world. Rather, in each case, scientiﬁc and pragmatic reasons will apply in varying degrees, with scientiﬁc reasons in general serving a subsidiary role to the pragmatic ones. In astrology, scientiﬁc reasons might tell us that the pragmatic goals of astrology (namely, accurate prediction of the future) are not attainable by study of the stars and planets, but science cannot evaluate the goals of astrology. The same applies in morality: science might tell us whether moral practice serves the goal of pleasant coexistence, but it cannot evaluate the goal itself. Scientiﬁc explanation is itself a project we carry out for pragmatic reasons (thirst for knowledge, instrumental control over the world, etc.).37 We cannot say in abstraction from individual cases what precise role will be played by scientiﬁc and pragmatic reasons, or what the precise mix of the two will be. Thus, questions about doctrinal eliminativism are not ones that can be addressed a priori or according to a formal set of rules, but ones that must be answered

30

Pragmatism and Normativity

as part of an ongoing debate concerning the appropriateness, objectivity, and usefulness of various types of discourse. It is true that scientiﬁc claims generally stand in some causal relation to natural facts, but this causal relation is only explanatory and justiﬁcatory in the context of a larger, successful explanatory project, a project whose undertaking is warranted not by scientiﬁc reasons, but by pragmatic ones. The fact that a discourse stands in some causal relation to a certain class of natural facts does not in itself explain why that discourse is more worthy of pursuit than other ones that have different goals. At this point it might be objected that the pragmatic goals of morality and epistemology are not equal. Morality is concerned with peaceful, ﬂourishing coexistence or some such, but epistemology aims at truth. This seems to set the practice of epistemology apart from other endeavors; it seems to give epistemology a metaphysically necessary justiﬁcation. This is no mere pragmatic justiﬁcation at all (goes the objection). Hence, epistemology and morality are not on a similar footing at all; epistemology has a more secure foundation. This objection fails. First, I will note that many have argued that epistemology is not aimed merely at truth.38 After all, they argue, if we only cared about truth, then we could merely sit by and calculate arithmetic tables all day long, thereby providing ourselves with an unlimited supply of true (indeed, necessarily true) beliefs. What this example shows is that truth is too thin of a goal to guide epistemology all by itself. We care about more than just truth; we want simplicity, explanatory coherence, instrumental control, predictive accuracy, and so forth. This is why I do not think our epistemic practice can be given the metaphysical grounding suggested by the objection; truth is not the only goal of epistemology,39 and the goals of epistemology are no more worthy of pursuit than are the goals of morality. More important, though, there is no reason to suppose that the interest promoted by epistemology is somehow more important than the interests promoted by morality. I will argue in Chapter 5 that morality is a set of practices aimed at promoting our interests, which I deﬁne as the ends which promote human ﬂourishing. These interests are life and health, afﬁliation with others, and so forth. These interests are fundamental constituents of human ﬂourishing, and there is no reason to pick out one interest (the interest in truth and understanding) as somehow more important and more legitimate than the others. Nor is there any reason I can see for singling out the practice aimed at promoting this single interest as somehow more justiﬁed than practices aimed at promoting other interests. Thus, the above objection fails

Pragmatism, Causal Explanation, and Normative Facts

31

to demonstrate that epistemology is somehow more justiﬁed or more important than morality.

Perception and normative facts We are now in a position to respond to another worry raised by my account. The worry is that if moral and epistemic facts are not causally efﬁcacious, then we cannot perceive them, except perhaps with a special sense of moral or epistemic intuition.40 Let us brieﬂy discuss why this conclusion is mistaken. (For simplicity of phrasing, I will discuss the issue with respect to moral reasons, etc., as repeating the phrase ‘moral and epistemic’ ad nauseam is a bit awkward.) As moral properties do not appear in scientiﬁc explanations, it is easy to draw the mistaken inference that they are only perceivable by some special faculty above our ﬁve senses. But we need not be driven to this conclusion. Indeed, Harman himself gives us an account of how moral perception might work. We observe actions and institutions, and react to them as we have been trained to do. We have been trained to classify certain actions as cruel and others as kind. Thus, when we see the hoodlums lighting the cat, we respond as we have been trained to respond: with the belief that the action in question is cruel. Thus, the reaction in question is “due [not] to the actual wrongness of what you see [but is] simply a reﬂection of your moral ‘sense,’ a ‘sense’ that you have acquired perhaps as a result of your upbringing.”41 I will concede, then, that moral properties are non-natural properties. Non-natural need not be equated with supernatural or non-physical, however. Clearly, early non-naturalists (such as G.E. Moore) used ‘non-natural’ in this sense. But latter-day nonnaturalists, such as John McDowell and David McNaughton, embrace a more plausible conception of non-naturalism. To say that a predicate is non-natural is to say that it divides objects in a way that does not cut nature at its joints. For example, a non-naturalist might say that the division of actions into moral and immoral actions reﬂects social convention, rather than some underlying physical property of these actions. McDowell argues that our division of actions into right and wrong essentially involves an evaluative disposition, and those lacking this disposition will be unable to grasp the contours of the predicate ‘morally right,’ because morally right actions are not uniﬁed by any naturalistically speciﬁable features. We will see (in Chapters 3 and 5) that morally right actions are those actions that are sanctioned by a set of rules that best serve our interests. Thus, the property of moral rightness is delineated with reference to our interests; there may be no underlying

32

Pragmatism and Normativity

naturalistic feature that uniﬁes all instances of morally right action. So to say that a property is non-natural is not, again, to say that it is supernatural or non-physical; it is merely to indicate that the property is in some sense ‘artiﬁcial’ or perhaps social in nature. Thus, the concession that moral properties are non-natural amounts only to the claim that moral observations are not the result of causally explanatory moral facts impinging on our perceptual apparatus, but are instead explained by one’s moral sensibility. Indeed, although moral properties are non-natural, non-natural properties are properties of natural objects, perception of which is (relatively) unproblematic. We have been taught to classify certain natural items as chairs, certain others as lions, and yet certain others as cruel acts. And though cruelty and warrant may be causally impotent, cruel actions and warranted beliefs are not. Thus, we may causally interact with actions and beliefs. We are trained to group together certain actions as cruel and certain beliefs as justiﬁed, but this training applies to all predicates, social, non-natural, and natural. So in the case of all types of predicates, a natural object acts on us causally, and we classify it as we have been trained to do so. Hence, there is no more difﬁculty with moral perception than there is with the perception of mountains, chairs, stars, or blades of grass. Whether an item is classiﬁed as a natural kind or a social kind has no bearing on what faculties are required to perceive it. Even if one holds (as does McDowell) that one needs a special sort of training, or a particular affective nature, to be able to perceive immoral actions as immoral actions, this again does not show that moral properties are supernatural or non-physical. The point is merely that moral predicates are a way of grouping causally efﬁcacious items (actions, social arrangements, etc.), so the fact that moral predicates are nonnatural predicates (and therefore perhaps not invoked in the best causal explanation of any non-controversial phenomenon) does not present any special difﬁculties of perception. Let us conclude this section with a summary of our progress so far. We started with the claim that morality was eliminable because of its explanatory impotence. I conceded this impotence, but noted that epistemology displayed the same sort of impotence, so epistemological realists and quasi-realists are in no position to demand that moral facts ﬁgure in scientiﬁc explanations. I then argued that questions of eliminativism cannot be settled a priori, but must be settled on a case-by-case basis, and that pragmatic decisions must play a crucial role in the decision. In light of this conclusion, my account of moral truth and the pragmatic justiﬁcation of moral discourse can be seen as an a posteriori

Pragmatism, Causal Explanation, and Normative Facts

33

justiﬁcation of moral practice. I conclude, then, that Harman’s standard is too rigorous, and ought not be applied in philosophy. However, to claim that morality and epistemology are justiﬁed by pragmatic reasons is not to win the battle against doctrinal eliminativism. After all, some might object to the notion of pragmatic justiﬁcation in the ﬁrst place (e.g., on the grounds that it conﬂates practical with theoretical justiﬁcation). Thus, we must now turn to a discussion of worries about pragmatic justiﬁcation.

Practical vs theoretical justiﬁcation Probably the most difﬁcult objection to the pragmatic justiﬁcation of a practice is that it conﬂates theoretical and practical justiﬁcation.42 Theoretical justiﬁcation justiﬁes the content of a belief, whereas practical justiﬁcation justiﬁes the act of holding a belief. The worry is that pragmatic reasons cannot justify the content of a belief (how could they?), and if we allow them to justify beliefs practically, then we run the risk of holding practically justiﬁed, yet false, beliefs. I will respond to this worry as follows: ﬁrst, I will show why pragmatic reasons cannot be used to justify individual beliefs. I will next show the constraints that must be placed on the pragmatic justiﬁcation of practices. Pragmatic justiﬁcation and individual beliefs Consider an individual belief that, although false, might be justiﬁed pragmatically. [S]uppose I am an Auburn football fan believing that Auburn will win the Alabama game this year. Is this belief justiﬁed? Theoretically, no; a comparative analysis of games played by the two teams this season makes this all too clear. My belief is just wishful thinking. But practically, maybe yes. After all, I might need to believe that Auburn will win the Alabama game to be able to maintain any interest in yet another mediocre season, and to get enjoyment from watching the team’s other games. So then it is a good thing that I have the belief—and if I did not, I would have a reason to acquire it.43 But surely this type of belief justiﬁcation is not permitted. Fortunately, I need not concede that it is permitted. When I say that we might pragmatically justify practices, I am not thereby saying that we may appeal to pragmatic considerations when justifying individual beliefs within the practice. Indeed, the reasons that serve to justify the practice itself

34

Pragmatism and Normativity

are not allowed to ‘inﬁltrate’ the practice and justify individual beliefs within that practice; the practice ceases to function if that is allowed to happen. Simply put, the goals of the practice are not met if you allow the pragmatic justiﬁcation of the practice to trickle down to the level of individual acts or beliefs. If this were to happen, pragmatic considerations would be applied directly to individual acts and beliefs. Epistemology, for example, would no longer be able to serve its purpose. If pragmatic reasons (and not just epistemic reasons) were allowed to justify individual beliefs, then the link between justiﬁcation and truth would be severed, and our knowledge-seeking practices would cease to be truthoriented. This would hamstring our pursuit of knowledge—we could no longer take justiﬁed beliefs as likely to be true, we could never be sure that our background assumptions and theories were supported by evidence rather than by merely pragmatic reasons, and so forth. It would be impossible to pursue any truth- or knowledge-seeking practice in any meaningful way, if we did not keep pragmatic reasons out of our epistemic discourse. Morality would suffer a similar fate. For example, if I promised to do X, then (if pragmatic considerations are applied directly to actions rather than practices) whether I ought to do X is determined wholly by pragmatic considerations, and the fact that I promised to do X has no particular relevance. You can in no way depend upon me to carry through with my promise. Thus, there can be no reliance on promises or contracts, if the pragmatic justiﬁcation is allowed to trickle down to the level of individual actions. The need to create a barrier between consequentialist considerations and deliberation regarding particular actions was seen clearly by Rawls in his article ‘Two Concepts of Rules.’ In this article, he was trying to establish the superiority of rule utilitarianism over act utilitarianism. With act utilitarianism, utilitarian considerations justify individual actions. Unfortunately, as act utilitarianism is plagued by free-rider problems and the like, it in fact fails to maximize utility. This is where rule utilitarianism comes into play: utilitarian considerations justify the adoption of a practice (consisting of a set of moral rules), but individual actions are justiﬁed by appeal to the rules of the practice, and not by appeal to considerations of utility.44 Here is Rawls: There are obvious utilitarian advantages in having a practice which denies to the promisor, as a defense, any general appeal to the utilitarian principle in accordance with which the practice itself may be justiﬁed. There is nothing contradictory, or surprising, in this:

Pragmatism, Causal Explanation, and Normative Facts

35

utilitarian (or aesthetic) reasons might properly be given in arguing that the game of chess, or baseball, is satisfactory just as it is, or in arguing that it should be changed in various respects, but a player in a game cannot properly appeal to such considerations as reasons for his making one move rather than another. It is a mistake to think that if the practice is justiﬁed on utilitarian grounds then the promisor must have complete liberty to use utilitarian arguments to decide whether or not to keep his promise. The practice forbids this general defense; and it is a purpose of the practice to do this. Therefore what the above arguments [against utilitarianism] presuppose—the idea that if the utilitarian view is accepted then the promisor is bound if, and only if, the application of the utilitarian principle to his own case shows that keeping it is best on the whole—is false. The promisor is bound because he promised: weighing the case on its merits is not open to him.45 Thus, in acknowledging that pragmatic considerations can justify practices, we need not concede that pragmatic considerations can justify individual beliefs any more than we need to concede that the baseball or chess player may appeal to utilitarian considerations to justify a particular play or move, or any more than a utilitarian needs to concede that a person may appeal directly to considerations of utility to justify a particular action. What this objection fails to recognize, then, is the difference between the pragmatic justiﬁcation of a practice and the pragmatic justiﬁcation of an individual belief. The former is permitted; the latter is forbidden. We will return to this point in detail in the next chapter, where we will see the rational superiority of rule-based, as opposed to act-based, reasoning. Pragmatic justiﬁcation and practices The critic is not silenced. He will concede the following: the pragmatist is not forced to concede that we may pragmatically justify individual beliefs. But the critic thinks that pragmatic justiﬁcation of entire practices is equally problematic. It seems as though the pragmatist must admit that we can use pragmatic considerations to justify entire practices, even though we know that these practices are theoretically unjustiﬁed. For example, someone might be a doctrinal eliminativist about religious discourse: she might reject the Christian doctrines of sin, salvation, heaven and hell, and so forth.46 Yet the same person might think that we ought to continue promoting religious belief, and encouraging people to believe the doctrines of the Christian religion.

36

Pragmatism and Normativity

For instance, such belief might (so goes the reasoning) encourage good behavior among people, and provide them with an additional motivation to behave well. Thus, we have an example of a practice about which (a) one is a doctrinal eliminativist, but (b) she thinks that we are pragmatically justiﬁed in continuing to participate in this practice. This objection (that pragmatic reasons cannot really justify practices) is right about something: pragmatic reasons cannot license us to believe false things about the world. In other words, pragmatic reasons cannot commit us to objects or causal powers that do not exist. The pragmatic justiﬁcation of religious discourse fails precisely because it requires that we come to believe something false about the world. That is, if religion is to serve the function of keeping people in line, these people cannot be eliminativists—doctrinal, ontological, or causal—about religious discourse: they must really believe in heaven, hell, God, miraculous powers, and the like if religious discourse is to serve the role envisioned above. Further, as we noted above, our implicit epistemic practice of sorting beliefs into those to be believed and those not to be believed fails if the sorting is based on prudential rather than epistemic reasons. And note that in justifying a religious practice in the way envisioned by the critic, we would in fact be justifying an abuse of our epistemic practice: we would be sorting beliefs we knew to be false into the category of things to be believed; we would be doing violence to our epistemic practice. There are a couple of worries the critic might have about this response. First, a critic might present the pragmatist with a familiar dilemma: ‘When you say that we cannot pragmatically justify false beliefs, is this an externalist or an internalist constraint? If the constraint is externalist, then it will face the familiar problems of externalism. But if it is internalist, then won’t it be possible for a community to epistemically justify, say, religious belief even if this belief is false?’ The answer to this objection is that the constraint is more internalist than it is externalist. We will return at length in Chapters 7 and 8 to the pragmatic conception of epistemology, but we can make a few brief comments here. First, a theory or belief can be assessed relative to any epistemic background. Thus, a particular religious theory T might be justiﬁed relative to a community’s epistemic situation, whereas we might judge not only that our epistemic situation is superior to this community’s, but also that T is not justiﬁed according to our lights. So to say that a false theory like T might be justiﬁed is not to say that it is justiﬁed relative to every epistemic standpoint. Thus, epistemic evaluation is in some sense relative—it is relative to an agent’s (or community’s)

Pragmatism, Causal Explanation, and Normative Facts

37

information and epistemic standards. Epistemic constraints need not lead inerrantly to the truth: a constraint is an epistemic constraint so long as its purpose is to achieve the truth; to require this constraint to actually achieve the truth is to attempt to return epistemology to the days when justiﬁcation was thought to require certainty. These points will be further elaborated and defended in Chapters 7 and 8. However, a more serious worry arises from this objection. One might concede that, in general, epistemic constraints must be aimed at the truth, or else our epistemic practice cannot achieve its goals. But what about a ‘compartmentalized’ epistemology? That is, couldn’t we have truth-apt epistemic norms in most areas, but set aside an area (say, religion) where we believe what is pragmatically useful to believe? As I indicated above (and will argue for in Chapter 7), a constraint does not count as an epistemic constraint unless it is aimed at the truth. And so pragmatic constraints on belief would not be epistemic constraints. But why not have certain areas of belief where belief is governed by non-epistemic constraints? There are several reasons against doing this. First, the strategy is epistemically ad hoc, exempting as it does sets of beliefs from our epistemic practice (whose purpose is belief regulation). Second, the strategy would be unstable for many or most people: knowing that a certain set of propositions were held not because there is evidence for them, but for merely pragmatic reasons, would undermine many people’s belief in these propositions. In response to this latter point, it is conceivable that in some communities, the population would not know that the justiﬁcation was only pragmatic. In such communities, the beliefs might be dictated by a set of leaders who did not hold the beliefs, but who instead convinced the population of this community of these propositions merely for pragmatic reasons. Thus, the leaders might participate in a version of the ‘noble lie.’ This strategy also seems unstable. The citizens in the community would have to take the beliefs purely on faith; any reﬂection among the members of the community might destroy their belief in these false propositions. But there is a limit to how long a community’s epistemic inquiry can be held in check. Take the example of Galileo and the Catholic Church. Galileo was, by all accounts, a devout Catholic. And yet his investigation into non-religious matters had the effect of contradicting the Church on matters of Church doctrine. Galileo’s improvements on the telescope, and the resulting astronomical observations that he made (particularly the orbiting moons of Jupiter and the phases of Venus) were instrumental in establishing the truth of the

38

Pragmatism and Normativity

Copernican heliocentric model of the solar system, which contradicted Church teaching. Thus, the Church’s authority was not enough to maintain consensus, as investigation of phenomena only indirectly related to Church teachings had the effect of proving certain Church doctrines false, and persuading others of the falsity of these teachings. I am not suggesting, of course, that the leaders of the Catholic Church did not believe the teachings they disseminated; I am merely using this example to illustrate the instability of authority as a method of belief ﬁxation. To conclude, then, there are not only moral reasons against perpetrating a deceptive set of beliefs, but there is reason to believe that such ‘top down’ dissemination of beliefs is not a strategy that can maintain longterm stability. People will reason for themselves, and will in many cases come to reject what they previously believed purely on authority. Peirce made similar comments about the instability of any method of belief ﬁxation that allowed “a belief . . . [to be] determined by any circumstances extraneous to the facts.”47 The above-described method of ﬁxing a belief in the mind of the public most closely resembles Peirce’s method of authority, by which An institution [is] created which shall have for its object to keep correct doctrines before the attention of the people, to reiterate them perpetually, and to teach them to the young; having at the same time power to prevent contrary doctrines from being taught, advocated, or expressed. Let all possible causes of a change of mind be removed from men’s apprehensions. Let them be kept ignorant, lest they should learn of some reason to think otherwise than they do.48 As Peirce notes, this method is inherently unstable. Peirce notes that “no institution can undertake to regulate opinions upon every subject. Only the most important ones can be attended to, and on the rest men’s minds must be left to the action of natural causes.”49 But this will prove the downfall of this method. First, these ‘natural causes’ will lead people to form opinions which may contradict some aspects of ofﬁcial doctrine. After all, beliefs do not exist in isolation, but have inferential relations to each other; and if people are allowed to form opinions on some topics freely, then eventually they will form opinions that are in conﬂict with the ofﬁcially promulgated view. Further, as Peirce notes, some in that culture will note that other cultures have different opinions and beliefs, and such people “cannot help seeing that it is the mere accident of their having been taught as they have, and of their having been surrounded with the manners and associations they have, that has caused them to

Pragmatism, Causal Explanation, and Normative Facts

39

believe as they do and not far differently.”50 The problem is that many such people are such that “when they see that any belief of theirs is determined by any circumstance extraneous to the facts, [they] will from that moment not merely admit in words that that belief is doubtful, but will experience real doubt of it, so that it ceases to be a belief.”51 And of course, for Peirce, doubt is that state of irritation which “causes a struggle to attain a state of belief.”52 This struggle is inquiry. David Wiggins, commenting on “The Fixation of Belief,” writes: Peirce holds, once some people are led by unregulated convictions to reject that which is ofﬁcially prescribed for general belief, more and more others will come to think that their own adherence to this or that approved opinion may be owed to “the mere accident of having been taught as [they] have.” Where people are already tending toward doubt, the beliefs that they think they owe to this source are bound to come adrift.53 The key thing to note here is, as Wiggins writes, that “By its nature, belief is a touchy, uncomplacent condition of the mind, a disposition which will not and cannot stay around on just any terms . . . [B]elief that p, once challenged, is a state which needs, on pain of extinction, to see itself as a state not ‘determined by circumstances extraneous to the facts [concerning whether or not p].’”54 Peirce writes that it is a tautology that belief aims at truth: “[W]e seek for a belief that we shall think to be true. But we think each one of our beliefs to be true, and, indeed, it is mere tautology to say so.”55 So once we realize that our belief that p does not originate in the fact that p, but in facts extraneous to the truth or falsity of p, then doubt arises. And Peirce claims that even “in the most priestridden states” some individuals will make just this realization, if the method of authority is used for the ﬁxation of belief, leading (as Wiggins notes) to the spread of doubt and the destruction of the method of authority. For those readers who are less sanguine than Peirce about the inherent instability of an authority-based belief structure, I should further note that there are (pragmatically justiﬁed) moral reasons against perpetrating such a fraud as the so-called noble lie. There is a justiﬁcation for moral principles forbidding deception and manipulation; the former offends against our interests in knowledge and autonomy and the latter, against our interest in autonomy. We will return to the topic of morality, and the relation between moral rules and our particular interests, in Chapters 4 and 5.

40

Pragmatism and Normativity

So if our pragmatically justiﬁed epistemic practice forbids us to pragmatically justify false beliefs about the world, does this constitute an objection to the pragmatic justiﬁcation of morality? The answer is ‘no.’ Moral discourse does not commit us to false beliefs about the world, because moral discourse need not be a descriptive discourse in the ﬁrst place. Moral discourse need not attempt to describe world, or the moral facts therein. There are several alternatives to viewing normative discourse as purely descriptive. For example, instead of describing, moral utterances might legislate or endorse.56 To issue a normative (e.g., moral) utterance is to endorse our current practice, or to call for its revision. In other words, normative utterances legislate, they need not describe. Thus, if I say ‘Setting cats on ﬁre is cruel,’ I am endorsing our practice of condemning hoodlums who set cats on ﬁre. Or, if our practice does not currently condemn the igniting of cats, I might say ‘Setting cats on ﬁre is cruel,’ thereby calling for the revision of our practice; I am legislating. Coupled with a deﬂationary account of truth, moral discourse can then be truthapt without being descriptive. (We will see shortly that morality might well have a descriptive element too, which should hopefully make the reader feel more at ease about the prospects of moral discourse being truth-apt.) Another possibility is contained in the expressivist literature. Gibbard argues that “to call something rational is to express one’s acceptance of norms that permit it.”57 It is not to state that one accepts such a system of norms; rather, it is to express one’s acceptance. Both of these accounts—the norm-legislative and the norm-expressivist account— offer a way of explaining how morality can be non-descriptive. Again, expressivists have a story to tell about how this discourse can be truth-apt, can embed in inferences, and so forth. Pragmatic reasons therefore justify our making moral and epistemic claims, and they justify our endorsing and condemning certain actions or practices. But they do not commit us to the existence of any mythical causes or objects—I argued for this much when I argued for causal explanatory eliminativism and ontological eliminativism about moral and epistemic discourse. Indeed, moral and epistemic discourse need not even commit us to the existence of properties such as cruelty or kindness or justiﬁcation that inhere in certain actions or beliefs or practices—such properties might only be the reﬂection of our moral and epistemic practice. The practice is prior to these properties, and ascribing these properties to such actions is merely an elliptical way of legislating or endorsing or expressing commitment to moral and epistemic rules.58

Pragmatism, Causal Explanation, and Normative Facts

41

Let us brieﬂy examine how morality and epistemology, though primarily legislative, could have a descriptive component which is parasitic on the legislative component. Terms have (to use Dummett’s useful terminology) criteria and consequences of application. The criteria of application for a sentence are those conditions that license utterance of that sentence. For example, both ‘A’ and ‘B’ are criteria of application for the sentence ‘A v B’. Once we license utterance of a sentence, we become entitled to that sentence’s consequences of application. For example, ‘A v B’ has, among its consequences of application, ‘∼(∼A & ∼B).’ The criteria of application for a moral term (such as ‘cruel’) will in general be a mix of natural and normative facts (the fact that it caused X pain, the fact that X had done nothing to deserve this pain, etc.). It might be that the criteria of application for a particular moral predicate follow a pattern (although this pattern might only be expressible using moral language; McDowell, for one, forcefully argues this point). That is, the criteria of application for a particular moral term might resolve into some kind of property. If this is so, then we can countenance the existence of moral properties. But the important thing, for the pragmatist, is that these properties are parasitic upon our interests, and the rules we legislate to help promote our interests. Thus, to state it crudely, our moral obligations do not arise from the existence of moral properties; rather, these properties arise from the existence of moral obligations (which obligations arise as a set of cooperative strategies for promoting our interests, as we will argue in the next chapter). So unlike religious practice, which cannot be pragmatically licensed because the practice commits us to the existence of God, angels, and mysterious causal powers, the moral practice commits us to no mythical entities or causes. It is merely a pragmatically justiﬁed way of dividing up the world; it is a way of dividing actions into those that are to be done and those that are not to be done. As long as we control the consequences of application of any terms that we introduce, a practice may be both descriptive and pragmatically justiﬁed. So, as long as ‘P is cruel’ does not have as a consequence of application anything like ‘God exists,’ or ‘Miracles can happen,’ the predicate ‘is cruel’ does not seem problematic. The problem with our pragmatically justiﬁed religious practice was precisely that it did commit us to such beliefs. If moral and epistemic properties arise in the parasitic way outlined above, then moral and epistemic claims would combine both legislative (or expressive) and descriptive elements. To call an action cruel, for example, is to characterize it in a certain way, and to say that there is at

42

Pragmatism and Normativity

least a prima facie reason not to perform this action. The former is the descriptive element of the moral claim, and the latter is the legislative or expressive element. Moral discourse, then, is legislative or expressive in that its primary purpose is to endorse an action or practice, condemn an action, or call for the revision of a practice. It is also descriptive in that it ascribes properties to actions—it divides actions into those that are to be done, and those that are not to be done. As this division of our actions into these two categories (and perhaps others—in addition to permissible and forbidden, there are the categories of obligatory and, perhaps, supererogatory) does not commit us to any mythical objects or causal powers, there is nothing objectionable about this way of dividing up the world. Consider the matter this way. It is in our interest to encourage certain behaviors and discourage others, and this is what justiﬁes morality. Morality has a normative element in that moral utterances concern what one ought to or ought not do. But if an action is wrong, it is wrong in part because of its descriptive features. Thus, although moral utterances and claims possess a non-descriptive, normative element (as described by a legislative or expressivist account of normative discourse), they also might have a descriptive element in that a certain moral claim (such as ‘That was cruel’) might entail (or at least prima facie imply) certain descriptive facts about the world (such as ‘That action caused suffering’). So morality might have normative and descriptive elements; and these elements might be connected in the way I have suggested above: the descriptive element is parasitic on the normative or legislative aspect. So pragmatic reasons may not commit us to misdescriptions of the world. They may not be used to justify individual beliefs; nor can they justify us in adopting practices that misdescribe the world. Pragmatic reasons may justify us in pursuing scientiﬁc discourse (whose goal is in part accurate description, explanation, and understanding of the world); they may justify us in pursuing moral and epistemic discourse [which involves both a normative (e.g., legislative or expressive) element and a descriptive element, via the introduction of new nonnatural (i.e., social) kind of terms]; but they cannot justify us in adopting a practice that misdescribes the world, as the above-mentioned religious discourse would. Does this mean I have abandoned a commitment to the truth-aptness of moral and epistemic discourse? By claiming that these types of discourse is not primarily descriptive, am I denying that moral and epistemic claims are candidates for truth and falsity? No; moral and

Pragmatism, Causal Explanation, and Normative Facts

43

epistemic claims can be true or false. I merely claim that what it takes for a moral or epistemic sentence to be true is different from what it takes for a purely descriptive sentence to be true. A descriptive sentence, to be true, must be descriptively adequate to the world; it must accurately describe features of the world. A moral sentence, by contrast, has different truth requirements59 (say, if it is supported by the best reasons in a practice which is an effective means to advancing certain of our interests—the interests that give rise to our moral practice).60 So for a purely descriptive (say, scientiﬁc) sentence to be true, it must accurately describe the world. If the sentence is ‘Electrons have a negative charge,’ then for it to be true, the world must contain electrons and these electrons must have a negative charge. On the other hand, for the sentence ‘Setting cats on ﬁre is cruel’ to be true, it need not be the case that it accurately describes anything; such a rule must merely be part of a pragmatically justiﬁed practice of behavior legislation. Moral and epistemic sentences must ‘jump through a different set of hoops’ than those jumped through by scientiﬁc sentences to be counted as true. Thus, if we pragmatically justify our moral and epistemic practices, we are not committing ourselves to a misdescription of the world, as we would be if we pragmatically justiﬁed a set of religious doctrines. Rather, we are committing ourselves to a primarily legislative or expressive (and only minimally descriptive) practice. Is normative (moral and epistemic) truth then a different kind of truth from scientiﬁc truth? If the requirements for normative truth are different from the requirements for scientiﬁc truth, if moral and scientiﬁc truths have to jump through different hoops to be counted as true, then are there multiple kinds of truth? Truth is a large topic, too large for me to do justice to here. For now, I must rest content to show the following: from the fact that normative discourse and science have different ‘truth requirements,’ it does not follow that normative truth and scientiﬁc truth are different things. If we have a predicate (say, a truth predicate) T, the fact that p and q must meet different requirements for T to be predicated of them does not mean that T is not univocal. Consider an analogy:61 Imagine two people and the requirements they must fulﬁll to become voters in the United States. Margaret was born in the United States to American parents; Marguerite, a recent immigrant to the States, was born outside of the United States to parents who were not US citizens. Marguerite must jump through different hoops to be allowed to vote; registration will be much harder for her to acquire than it will be for Margaret. Presumably, Marguerite will ﬁrst have to become a naturalized citizen, or at least a resident alien—they must fulﬁll different

44

Pragmatism and Normativity

requirements to become voters. But ‘registered voter’ is not equivocal; if both Margaret and Marguerite register to vote, ‘registered voter’ means the same thing in both cases. It is not as though ‘registered voter’ is predicated of Margaret and ‘registered voter∗’ is predicated of Marguerite. They have achieved the same thing (registration), even though Marguerite, a foreign citizen, had to fulﬁll different requirements to achieve it. So we should not infer from the fact that normative discourse and science have different requirements for being counted as true that normative truth and scientiﬁc truth are different. Thus, normative sentences can be true; but by committing ourselves to moral and epistemic practice we are not committing ourselves to a misdescription of the world. This is because morality and epistemology are primarily legislative or expressive (and only minimally descriptive) practices; and hence the fact that it is not primarily descriptive is not an objection to it, not any more than it is a criticism of a spoon that it does not cut. In any case, what if moral and epistemic truth is different from scientiﬁc truth? What is the practical import of such a conclusion? As I noted at the beginning of the chapter, some philosophers will no doubt continue to claim that we ought to continue to use normative talk, but with our ﬁngers crossed, so to speak. That is to say, we should continue to use this discourse at the same time acknowledging that the sentences of this discourse are not really true. But as we have noted earlier in the chapter, if this is all that doctrinal eliminativism amounts to, then its practical import is limited. If doctrinal eliminativism about normative discourse concedes that we ought to continue practicing moral discourse, then it does not represent a threat to the robustness of moral practice. A philosophical critique of a particular practice is not particularly damaging if, in the end, it concludes that we ought to carry on pretty much as before. In any case, the idea to be developed in the next chapter is that moral and epistemic ‘oughts’ are essentially ‘oughts’ of instrumental rationality. So moral and epistemic truth would essentially be as robust as truths of instrumental rationality. To the extent that philosophers hold prudential reason to be objective, morality and epistemology will inherit that objectivity.

Summary and conclusion Rarely do philosophers argue that we ought to jettison moral or epistemic language altogether. As Robert Black writes, “From Hobbes onwards philosophers embracing ontological skepticism about values have advocated not the abandonment but rather the reinterpretation or

Pragmatism, Causal Explanation, and Normative Facts

45

morality.”62 Black himself is one of the few exceptions, arguing that we should in fact do away with both moral and epistemic language. In general it seems that if one has decided that some feature of the world that one thought was there—God, or the luminiferous ether, or the moral law—isn’t there after all, then one should take advantage of this discovery and stop talking as though it were . . . In the case of morality, the skeptic should embrace amoralism; in the case of induction, the skeptic should conclude that most of our everyday beliefs are not justiﬁed in the way in which we customarily think.63 In this chapter, I conceded that normative (e.g., moral and epistemic) facts are not involved in the best causal explanation of any noncontroversial phenomenon. However, I argued that this conclusion does not justify us in jettisoning (as Black would have us do) these types of normative discourse, because causal explanation is not the point of such discourse. Rather, I argued that we can appeal to pragmatic reasons to justify continued participation in practices that do not play the role of causal explanation. (Over the remainder of this book, we will attempt to construct a viable morality and epistemology using the notions of cooperative rationality and fundamental human interests.) Next, we saw that a two-level practice approach offered the best account of normative discourse, an account that does not conﬂate practical and theoretical justiﬁcation or descend into the unhappy consequences of act-consequentialism. But is this two-level approach viable? It has been argued that it is not. Smart’s rule-worship objection against rule utilitarianism might be taken to show that a practice approach is irrational. Let us turn to this important issue. In answering Smart, we will develop a novel conception of rationality, which challenges the dogma that rationality must be understood atomistically, in terms of isolated agents and individual, discrete actions. This conception of rationality will not only help us answer the rule-worship objection against consequentialism; it will also enable us to explain precisely how submitting to moral and epistemic constraints is rational. Thus, the next chapter offers a new answer to the age-old question ‘Why be moral?’ (as well as an answer to the rather neglected question ‘Why should I submit to epistemic constraints?’).

3 Pragmatism and Rationality

In Chapter 2 I argued that as moral and epistemic facts do not ﬁgure in causal explanations, we must turn to pragmatic reasons to justify continued participation in these practices. What I hope to demonstrate in this chapter is that it is prudentially rational to submit to moral and epistemological constraints. Certainly this has been argued before, at least with regard to moral constraints. However, previous attempts to prove that moral constraints are rational constraints have typically run into various difﬁculties. In this chapter, I will argue that these difﬁculties have arisen out of certain shared misunderstandings regarding the nature of rationality. First, too many philosophers have thought that the individual action is the basic unit of rationality. That is, in questions of means–end rationality, it has too often been thought that we must examine and evaluate individual actions, and determine each isolated action’s rationality in terms of how well that individual action promotes the end in question. I will argue that this is a mistake, and that often strategies, not individual actions, are the basic units of rationality. That is, the rationality of individual actions often cannot be determined in abstraction from the strategy they constitute. Second, philosophers have too often thought of rationality as individualistic. That is, philosophers have thought that we can only evaluate the rationality of actions performed by individuals. I will argue that this, too, is a mistake, and that there is a viable (and indeed indispensable) notion of cooperative rationality. Rationality is often ineliminably cooperative, and failure to recognize this fact has led to many puzzles and confusions in philosophy. Let us begin.

46

Pragmatism and Rationality

47

Pragmatism and rule worship So, we have seen that worries about a conﬂation of practical and theoretical justiﬁcation are mitigated by a move to a practice conception of rules, which will not allow pragmatic considerations to justify individual beliefs. On our pragmatist account, these practices are structured to promote our interests. On an appropriate understanding of rationality, it is rational to follow the rules constituting these practices. What do I mean when I say it is rational to follow these practices? I mean that it promotes our interests to do so. Simply put, it is a case of means–end rationality. We have certain interests, and rationality consists of acting to promote these interests. As it turns out, many of these interests are best-promoted by following cooperative strategies. This is what the rest of this chapter aims to demonstrate. Having deﬁned ‘rationality’ in purely means–end terms, let me pause to assuage some worries. The account here need not deny that there is a notion of constitutive rationality. In other words, I am not denying that we can reason about what our ﬁnal ends are. If there is some notion of constitutive rationality, then we could use that to establish our rational interests. Morality and epistemology, then, would be strategies aimed at promoting these interests.1 Thus, even if there is a viable notion of noninstrumental rationality, morality and epistemology are still understood in terms of instrumental rationality: they are strategies aimed at promoting these interests, whether or not these interests are pre-rational, or constituted by non-instrumental rationality. An important worry remains to be met, though. It seems as though our two-level strategy (in terms of practices and actions justiﬁed by these practices) is vulnerable to an old objection, the rule-worship objection raised by Smart against rule utilitarianism. Rule utilitarianism comes in many different versions, but the basic idea is that we should not test individual actions by the principle of utility; rather, we should test rules (or sets of rules) by this principle. Thus, the set of rules, observance of which (or promotion of which, etc.) would maximize utility is the correct set, and an action is right if it conforms to a correct rule. The charge is that it is often irrational to follow rules, at least when better results can be obtained by ignoring the rule in a particular case. I am arguing that participation in these practices is rational: that they best promote our interests. The rule-worship objection challenges the idea that it is rational to participate in these practices (i.e., to follow

48

Pragmatism and Normativity

the rules constituting these practices). Smart famously criticized rule utilitarianism as follows: Suppose that there is a rule R and that in 99% of cases the best possible results are obtained by acting in accordance with R. Then clearly R is a useful rule of thumb; if we have no time or are not impartial enough to assess the consequences of an action it is an extremely good bet that the thing to do is to act in accordance with R. But is it not monstrous to suppose that if we have worked out the consequences and if we have perfect faith in the impartiality of our calculations, and if we know that in this instance to break R will have better results than to keep it, we should nevertheless obey the rule. Is it not to erect R into a sort of idol if we keep it when breaking it will prevent, say, some avoidable misery? Is not this a form of superstitious rule-worship (easily explicable psychologically) and not the rational thought of a philosopher?2 Thus, it might be thought that the practice conception of rules fails, defeated by Smart’s rule-worship objection. This would be an extremely unfortunate result; indeed, it might even be part of a reductio of pragmatism. Consider the consequences of the collapse of the practice conception of rules in a pragmatist system. If this were to happen, pragmatic considerations would be applied directly to individual acts and beliefs. As noted in the previous chapter, epistemology, for example, would no longer be able to serve its purpose. If pragmatic reasons (and not just epistemic reasons) were allowed to justify individual beliefs, then the link between justiﬁcation and truth would be severed, and our knowledge-seeking practices would cease to be truth-oriented. This would hamstring our pursuit of knowledge—we could no longer take justiﬁed beliefs as likely to be true, we could never be sure that our background assumptions and theories were supported by evidence rather than merely pragmatic reasons, and so forth. It would be impossible to pursue any truth- or knowledge-seeking practice in any meaningful way, if we did not keep pragmatic reasons out of our epistemic discourse. Also, as noted previously, morality would suffer a similar fate. For example, if I promised to do X, then (if pragmatic considerations are applied directly to actions rather than practices) whether I ought to do X is determined wholly by pragmatic considerations, and the fact that I promised to do X has no particular relevance. You can in no way depend upon me to carry through with my promise. Thus, there can

Pragmatism and Rationality

49

be no reliance on promises or contracts, if the pragmatic justiﬁcation is allowed to trickle down to the level of individual actions. If rule pragmatism is not viable, and “act pragmatism” (as it might be called) has absurd consequences, then things do not look promising for pragmatism. However, I will argue that Smart’s objection rests on an over-simpliﬁed version of rationality, one that takes individuals and actions as basic. Once we see that in many cases the basic elements of rationality are teams and strategies we can defeat Smart’s objection. In responding to Smart’s objection, we can further elaborate on the idea that it is rational to follow moral and epistemic rules. That is, we can see that instrumental rationality counsels following the rules constituting our moral and epistemic practices. Let us turn now to a demonstration of this important point.

Fundamental vs derivative rationality The rule-worship charge is that if our ultimate goal is the satisfaction of our interests, then it is irrational to perform an action that ill-serves our interests (even if this action is part of an interest-serving strategy) if there is an available action that would better serve our interests. It is this charge that the pragmatist must defeat. Human endeavor can be examined at several different levels of detail. We may examine such endeavor at the level of strategy, or at the level of action (which is, in fact, several different levels; an action can fall under several more or less general descriptions). Actions can be broken down further into sub-actions or even individual bodily movements or parts of movements. The question we must answer is, “If we are to examine the rationality of human endeavor, what level of generality or detail is proper for our investigation?” That this is an appropriate question to ask can be seen by the following example: Suppose I am a maintenance engineer at a hospital where the power has gone out. The best thing for me to do (indeed, the thing I must do) is go to the basement and switch on the generator. Now, in this situation, is it rational for me to move my right foot 2 inches to the north? The question is virtually without sense. But if this movement is the ﬁrst part of my ﬁrst step toward the basement stairs, then the movement makes sense. It makes sense as part of my overall activity, which is going to the basement to turn on the generator. Examined by itself, this movement of my foot cannot be seen as rational, nor can it be seen as irrational; when viewed in isolation, the rationality of this action cannot be assessed. Rather, it is only rational in the larger context of my

50

Pragmatism and Normativity

efforts aimed at restoring power to the hospital. In other words, moving my right foot 2 inches to the north might be rational (although this claim has, to my ear, an awkwardness to it), but its rationality is wholly derived from the rationality of my more general action: going to the basement to turn on the generator. We can therefore distinguish between derivative rationality and what we might call fundamental rationality. Moving my foot 2 inches to the north is derivatively rational. If that was all I were to do (i.e., if I stopped after moving my foot 2 inches), this action would not be rational; the action is only rational because it is part of the overall action of going to the basement to turn on the generator. This more general action has what one might call fundamental rationality—it is the rational action, in the circumstances, and all sub-actions and sub-movements (such as moving my foot 2 inches to the north) inherit or derive their rationality from the rationality of this more general rational action. Thus, ﬁnding the right level of description is fundamental to rationality. This is no more surprising than the claim that we cannot understand biological speciation in terms of quarks; it is merely a recognition that if a subject matter is comprehensible at one level of description, it is not necessarily more comprehensible (or comprehensible at all) at a more detailed level of description. And the problem is not merely epistemological—it is not merely one of comprehensibility. Arguably, there simply are no facts about biological speciation that can be speciﬁed at the level of quarks. Similarly, there simply will not be many facts about fundamental rationality if we focus on movements at the level of ‘moving one’s foot 2 inches to the north.’ As I have argued above, this action taken by itself simply is not rational (and perhaps not even subject to rational evaluation). Its rationality can only be understood derivatively, as parasitic on the fundamental rationality of a more general action, such as going to the basement to turn on the generator. So what is the correct level of description for the study of rationality? The answer is that it depends on the particular circumstances of action. We can say, though, that the rule-worship objection rests on the dogma that the level of the individual action is always the correct level on which to understand rationality. This is a dogma equivalent to the positivist dogma that the level of physics is the only correct level at which science can be understood. Just as the positivists were mistaken (chemistry, biology, and astronomy are real sciences even if they are not reducible to physics), Smart and company are equally mistaken in assuming that rationality must be understood exclusively at the level of individual actions. It is to this argument that we will now turn.

Pragmatism and Rationality

51

Actions and strategies The fundamental unit of rationality is often the individual action. If I am standing before the generator switch in a hospital that has lost its supply of electricity, the rational thing to do is to ﬂip the switch. It seems clear that this action is fundamentally rational—it need not possess merely derivative rationality by virtue of being part of some larger action.3 But there are cases where the smallest unit of fundamental rationality is not the individual action, but the strategy. We can demonstrate this with a simple example: Suppose I am on a diet. If I break the diet and have a scoop of ice cream, I will get a great deal of pleasure from this, and the amount of weight I would gain (a fraction of an ounce, no doubt) would be utterly unnoticeable to me. So the act of eating a scoop of ice cream has overall positive consequences. But if every day I thus weigh the consequences of eating a scoop of ice cream and decide to do so, then in no time I will be overweight, with all of the unhappy consequences that this will bring. So it seems clear that if my goal is my long-term happiness, I cannot merely look at individual actions: no individual instance of eating a scoop of ice cream has any long-term consequences (good or bad); it has only positive short-term consequences. Instead, I must think in terms of strategies. Failing to do so will result in my long-term unhappiness. Hence, rationality—in this case, the achievement of my long-term happiness and good health— demands that I think at the level of strategies, not individual actions. Treating the action as the unit of fundamental rationality ultimately leads to the frustration of my long-term goals here, and thus is, in general, self-defeating. We must instead realize that, in this case, my goals can only be met by focusing on the level of the strategy. Thus, in this case, the strategy is the proper level to focus on, if we want to best achieve our goals. This is a case of means–end rationality. Failure to achieve the desired end owing to poor action selection signals a failure of rationality. But if I treat individual actions as the bearers of fundamental rationality, then I will fail to achieve my desired end, and I will fail because I selected the actions I did. This signals a failure of rationality. And so we must reject the idea that in this case, the individual action is the fundamental unit of rationality. I will only succeed in achieving my end if I pursue the overall strategy. On each occasion, passing up the ice cream is the rational thing to do—but this rationality is derivative; it is parasitic on the rationality of the overall strategy, just as the rationality of moving my foot 2 inches to the north was parasitic on the rationality of going to the

52

Pragmatism and Normativity

basement and turning on the hospital generator. In both cases, examining the issue at the wrong level of detail leads to the wrong answer. If I examine moving my foot 2 inches to the north, I can only conclude that this is not a rational response to a power failure. But it is rational in the larger context of restoring power to the hospital. Similarly, refusing a scoop of ice cream on a particular occasion is not rational if viewed in isolation. But when viewed in the context of my goal to lose weight, it is rational—but its rationality derives from the rationality of the overall strategy of refusing ice cream (or perhaps having ice cream only once per week, or whatever the optimal strategy turns out to be, balancing the pleasure of eating ice cream against my desire to lose weight). Thus, if what we care about is achieving our interests, and rationality is deﬁned instrumentally in terms of achieving our interests, then it is irrational to reason at the level of actions rather than at the level of strategies, for doing so will fail to promote our interests. Rationality demands thinking at the level of strategies. The act utilitarian might object as follows:4 Suppose I have eaten 100 ice cream cones over 100 days, and as a consequence have gained 10 pounds at the end of these 100 days. As a result of my weight gain, I am unhappy. Let us suppose that although I gained 100 utiles from eating the ice cream cones, this increased weight gain has caused me −110 utiles. Thus, there is a net loss of 10 utiles. My anti–act utilitarian argument above goes as follows: ‘Each day I gained 1 utile from eating the ice cream cone. I was unable to notice any incremental weight gain, and so did not feel any suffering. Thus, on any given day, there was an act-utilitarian reason for eating an ice cream cone. But at the end of 100 days, I actually suffered a net loss of 10 utiles. This shows that if I really care about promoting my happiness, I ought not be an act utilitarian.’ The act utilitarian will respond that the problem is with my accounting, and not with act utilitarianism itself.5 If I do, in fact, get −110 utiles from eating all of these cones, and if this loss of utility really is the result of eating all of these ice cream cones, then the sensible thing to do is to divide up that negative utility and apply it to each cone. Thus, I may gain 1 utile of pleasure from eating the cone, but I also lose 1.1 utiles from the unnoticeable (but nevertheless very real) incremental weight gain caused by eating this cone. I may not notice the weight gain, but the act utilitarian is, of course, interested in long-term utility. Thus, the act-utilitarian thing to do is to have an ice cream cone only occasionally. This act-utilitarian response will not work. It assumes that a gain of a fraction of an ounce in weight is in itself a bad thing. The act-utilitarian

Pragmatism and Rationality

53

assumption is that (if one is within a certain weight range) any gain of weight (even ¼ ounce) must count as a bad thing, and any loss of weight (even ¼ ounce) must count as a good thing. However, a person’s weight naturally ﬂuctuates by ounces or even pounds. Should we treat this natural ﬂuctuation as a series of good and bad events? The loss and subsequent regaining, over the course of a 24-hour period, of ¼ of an ounce strikes me as a matter of indifference. It does not strike me as a good event followed by a bad event. Slight changes in a person’s weight simply do not seem to be good or bad events; they are indifferent. A gain of ¼ ounce does not affect a person’s health, or mobility, or ﬂourishing in any substantive way. So it is not obvious why the gain of a fraction of an ounce of weight deserves to be counted as a loss of 1.1 utile. Of course, the act utilitarian can respond, ‘Granted, the gain of ¼ ounce is not by itself signiﬁcant. But in the original example, we were talking about the gain of several pounds over 100 days. This results in signiﬁcant disutility, and so it makes sense to assign a fraction (1/100) of the total disutility to each ice cream cone consumed.’ This response may be correct, but notice that this response is out of step with the basic atomistic assumptions of act utilitarianism. The response concedes that one ice cream cone by itself would not create any disutility, but that considered in aggregate, the consumption of 100 cones creates a great deal of disutility, and this disutility can reasonably be divided among the individual actions. Notice, however, that the act utilitarian must think ﬁrst in terms of groups or sets of actions; the disutility of the set of actions is prior to the assigning of disutility to the individual actions in the set. That is, the disutility of the individual cone derives from the fact that the person eats 100 cones, not just one cone. Thus, the disutility of the set of actions is prior to the disutility of the individual action. If the individual action existed in isolation—if the person ate only one cone, not 100—then there would be no disutility from this individual action (or so I have been arguing). So the act utilitarian is already being forced to think like a rule utilitarian—she is being forced to think about the consequences of a set of actions (dare I say plans or strategies?) before she can calculate the utility of any individual action constituting this set. The utility (or disutility) of the action derives from the utility (or disutility) of the set of actions. But of course, this is the line I have been pushing all along—the rationality of an individual action may, in many cases, be derivative from the rationality of a strategy or a larger set of actions. So the above act-utilitarian response actually makes a crucial concession that the rationality of individual actions is not necessarily fundamental and that

54

Pragmatism and Normativity

it can instead be parasitic on the rationality of a long-term plan or set of actions. Eating one ice cream cone creates no disutility. But if I plan (or foresee) that this ice cream cone is the ﬁrst of 100, to be consumed over 100 days, then I must treat this action as having a net utility of −0. 1. The rationality of the action derives from the rationality of the plan or strategy. Let us continue with the thread of our discussion. When talking about justifying a normative practice, a number of problems can be solved by moving from a picture where individual acts are justiﬁed by appeal to our interests to a picture where strategies are so justiﬁed and acts are seen as rational or justiﬁed insofar as they relate to these strategies. Consider some hoary examples from the utilitarian literature: water conservation during a drought and taking a shortcut across the grass in a well-traveled area. In both cases, ill effects will result only if a certain number, n, of people fail to conserve water or take the shortcut across the grass. Now, it is a well-known objection against act-consequentialism that each person ought to reason as follows: ‘My violation of the water conservation order (or my taking a shortcut across the grass) will have no appreciable effect on whether the city suffers a water shortage (or the grass dies), but will cause me a measurable amount of utility. Therefore, it is utility-maximizing for me to violate the water conservation order (or walk across the grass).’ Of course, as it is rational for each person to think this way, then if everyone is an act consequentialist, all will violate the water conservation order (or take the shortcut across the grass), and hence the city will suffer a water shortage (or the grass will die). Thus (it is argued) the best consequences can only be achieved if people do not behave like act utilitarians, but instead follow a rule: ‘Conserve water’ or ‘Do not walk on the grass.’ Act-utilitarian responses to this objection occasionally involve using mixed strategies. Smart, for example, says that an act utilitarian ought to ﬁnd out how many people can violate the water conservation order and then employ a mixed strategy to determine whether or not she ought to comply or violate. Thus, suppose that 10% of the city’s residents can violate the water conservation order without any serious negative consequences resulting. Smart recommends, then, that the act utilitarian use a random procedure to determine whether or not she should violate, a procedure that will with a probability of 0.9 recommend that she should not violate the water conservation order. Similarly, if n people (out of a total number of m people who will have the choice of walking across the grass) can cross the grass without causing it to die, then the act utilitarian might suggest that people employ a mixed strategy with

Pragmatism and Rationality

55

a probability of n/m of allowing each person to walk across the grass; or the utilitarian might suggest that the n oldest or most mobility-impaired people be allowed to take the shortcut. There are two comments to be made about these act-utilitarian replies. The ﬁrst, and most obvious, reply is that these responses are not, in fact, act-utilitarian responses. All of them counsel following a strategy. Consider Smart’s recommendation that we employ a mixed strategy. This is not an act-utilitarian move. For, an individual act utilitarian will reason as follows: ‘My defecting from this mixed strategy and simply deciding outright not to follow the water conservation order will have no appreciable effects on the city’s attempt to conserve water. However, defecting has a higher expected utility than following the mixed strategy. If my obeying the conservation order will create 0 units of utility whereas my violating it will create 1 unit of utility, then following the mixed strategy has an expected utility of 0.1, whereas violating has an expected utility of 1. Therefore, I should violate.’ Following a mixed strategy is just that: it is following a strategy. As such, it is a concession that utilitarianism can only get the job done if strategies are employed; as we just saw, a strict act utilitarian will not follow the strategy. The second comment that needs to be made about these act-utilitarian responses is as follows: a moral system must be applicable. That is, its elements must be learnable, and it must be possible for real people to follow and apply these elements. As Daniel Hunter writes, “There are cognitive constraints on the acceptance of rules. The rules, for example, cannot be too complicated, else people simply couldn’t remember them.”6 Rawls makes a similar point, writing, [P]rinciples are to be universal in application . . . Thus I assume that each can understand these principles and use them in his deliberations. This imposes an upper bound of sorts on how complex they can be, and on the kinds and number of distinctions they draw.7 However, it is sheer fantasy to suppose that the act-utilitarian ‘solutions’ suggested above are applicable. How is the average person to know how many people can violate the water conservation order without serious consequences resulting? How is the average person to know how many people can cross the grass before the grass dies? Can the average person really do a good job of calculating probabilities (including conditional probabilities, which would be required in more complex cases) and employing a suitable mixed strategy? And utilizing these strategies often

56

Pragmatism and Normativity

assumes that all people will, on short notice, be able to come up with suitable strategies (such as allowing the n most inﬁrm people to cross the grass). Again, this is fantasy. It is not a livable moral system. Thus, we can see that in a rule-utilitarian system, strategies produce the best consequences; so a system that treats strategies as fundamentally rational will produce better consequences than a system that treats actions as the bearers of fundamental rationality in all cases. Consequentialists ought then to use considerations of utility to justify strategies, not individual actions. Let us consider another argument for the conclusion that rationality need not be understood in terms of individual actions, an argument advanced by Richard Dean.8 The dogma that rationality must be understood in terms of individual actions is equivalent to embracing the theory of straightforward maximization (or ‘s-max,’ as Dean abbreviates it). According to s-max, “making a rational choice is simply a matter of choosing the action that provides you with the maximum expected utility.”9 Dean writes: Constrained maximizers usually seek to maximize their own expected utility, but in speciﬁc sorts of cases are willing to forego a maximizing action in favor of a cooperative one. In a PD [prisoner’s dilemma], constrained maximizers will act cooperatively provided they think others will cooperate as well. For the sake of engaging in mutually beneﬁcial projects with others, they will forego some opportunities to maximize their utility.10 David Gauthier famously defends the rational theory of constrained maximization.11 Gauthier argues that constrained maximizers will achieve more utility in the long run, since cooperative opportunities will be open to them that would not be open to the straightforward maximizer. The problem of pursuing strategies, as I have outlined here, is similar to Gauthier’s problem of cooperation: both Gauthier and I are trying to argue that it is rational to perform a non-maximizing action, if that action is part of a policy or strategy that will maximize utility in the long run. How can we show that it is rational to perform a non-maximizing action? Dean argues that the key to solving this puzzle is to distinguish between the aim and policy components of a theory.12 The aim component of a theory is essentially a theory of value. It “deﬁnes the objects that have value for agents.”13 The policy component is a theory of choice; it “gives agents advice on what actions to choose.”14 Now, the

Pragmatism and Rationality

57

standard theory of rationality says that a rational agent’s aim should be maximization of her expected utility. However, the standard theory then hastily concludes that if this is the agent’s aim, then each action should be aimed directly at this end. In other words, there is a (perhaps conceptual) link between the aim and policy components of a theory: if the aim is X, then every action ought to be aimed at directly promoting X. But why should we make this hasty leap from aim to policy? Dean writes, “Rather than assuming that this aim must be accompanied by a policy of directly seeking the aim, however, Gauthier maintains that the correct policy is whatever policy best fulﬁlls the aim.”15 And of course, this makes perfect sense. Given a theory of value, which deﬁnes the objects worthy of pursuit, it is rational to pursue whatever policy best promotes this aim, and irrational to adopt a policy which you know will fail to achieve this aim as well as another available policy would. This seems obvious. But it leads to the interesting conclusion that if maximization of personal utility is one’s aim, then the policy of straightforward maximization is not the rational one to pursue; rather, a policy of designing and executing strategies will often be dictated by the policy element of the theory. Similarly, if our aim is to best promote our interests, a policy of choosing actions in isolation will not be the best one; rather, we should often pursue strategies. Even though these strategies will often recommend actions that are contrary to our interests, they are rational insofar as they are the best available policy for promoting our interests. The assumption that straightforward maximization must always be rational is a dogma, and cannot withstand critical scrutiny. We should often reason in terms of strategies and rules, not in terms of individual actions. Thus, we must construe our pragmatism as rule pragmatism. If we construe it as act pragmatism, then our pragmatically justiﬁed practices will cease to achieve their ends. Consider our previous observations about epistemology: as I argued above, if pragmatic reasons (and not just epistemic reasons) were allowed to justify individual beliefs, then the link between justiﬁcation and truth would be severed, and our knowledge-seeking practices would cease to be truth-oriented. This would hamstring our pursuit of knowledge—we could no longer take justiﬁed beliefs as likely to be true, we could never be sure that our background assumptions and theories were supported by evidence rather than merely pragmatic reasons, and so forth. It would be impossible to pursue any truth- or knowledge-seeking practice in any meaningful way, if we did not keep pragmatic reasons out of our epistemic discourse. And

58

Pragmatism and Normativity

so we must treat strategies as fundamental, and adhere to our practice conception of rules.

Individual strategies and team strategies We have seen that in many cases, the strategy (not the individual action) is the basic unit of fundamental rationality. Thus, the rule-worship objection goes astray in accusing someone of irrationality if she insists on following a strategy even when the strategy dictates sub-optimal actions. However, if we look at the above examples, we note that the problem facing the act utilitarian is not merely one of failing to follow a strategy. The problem is also one of cooperation. For, the problems that occur if one thinks in terms of acts rather than strategies recur if one thinks individualistically instead of cooperatively. Consider the matter this way. Suppose I, as a pragmatist, am convinced by the above considerations that I should always think in terms of strategies, rather than in terms of individual actions. Does this mean that I will therefore always cooperate (i.e., conserve water, not walk on the grass, etc.)? Not necessarily. I might reason as follows: ‘A strategy of non-cooperation will best serve my interests. For, my non-cooperation will not result in signiﬁcant negative consequences, but will in fact beneﬁt me. Thus, the overall consequences are maximized.’ Thus, the pragmatist might follow a strategy of free riding: she recognizes that her failure to cooperate is insigniﬁcant in large-scale cooperative endeavors (such as water conservation and grass management), and so the best individual strategy is to free ride. Of course, if everyone reasons this way, then all follow a strategy of free riding, and we are deprived of the obvious beneﬁts of cooperation. What, then, is to be done? The solution is to recognize that not only is the strategy often the smallest unit of fundamental rationality, but also rationality is often irreducibly cooperative. The idea that there is a notion of team rationality, not reducible to individual rationality, is advanced by Margaret Gilbert,16 Robert Sugden,17 and others. As with the ice cream example above, where we showed that it is often irrational to think in terms of actions rather than strategies, we can show that rationality is often cooperative if we can give an example where it is clear that rationality can only be construed cooperatively, not individually. Sugden presents such an example, which he calls the Footballers’ Problem. Suppose A and B are two attacking players in a football team. A has the ball, but a defender is converging on him. B has more space, so A wants

Pragmatism and Rationality

59

Table 3.1 Decision matrix for Footballers’ Problem Player B left Player A

left right

right

10,10

0,0

0,0

11,11

to pass the ball to B. There are two directions in which B could run so as to intercept a pass from A: call these left and right. Correspondingly, there are two points on the ﬁeld, left and right, to which A could pass the ball to be picked up by B. There is no time for communication, or for one player to wait to see what the other does: each must simultaneously choose left or right. Suppose that the move to the right puts B in a slightly better position. Say, the probability that the pass will result in a goal is 10% if both choose left and 11% if both choose right. If one chooses right and the other left, the probability is zero. What should each player do?18 The Footballers’ Problem can be presented in a traditional decision matrix as shown in Table 3.1 above. What should the two players do? “The answer seems obvious: each should choose right,” writes Sugden. “But paradoxically,” he continues, “this obvious answer cannot be generated by the theory of individual rationality, as used in game theory.”19 Sugden explains as follows: Clearly, if A expects B to choose ‘right’, he should choose ‘right’ too. But equally, if A expects B to choose ‘left’, he should choose ‘left’ too. According to expected utility theory, A should choose ‘right’ if he judges the probability that B will play ‘right’ to be greater than 10/21; and he should choose ‘left’ if he judges that probability to be less than 10/21. So what is rational for A to do depends on what B can be expected to do. In situations like this, game theory invokes the assumption that the rationality of the players is common knowledge between them. Thus, in order to form a rational belief about what B will do, A has to take account of the fact that B is himself rational, and so will choose whatever is rational from his point of view. But B’s problem is exactly symmetrical with A’s: it is rational for B to choose ‘right’ if A can be expected to play ‘right’ with a probability of 10/21 or more, but rational for B to choose ‘left’ if A can be expected to

60

Pragmatism and Normativity

play ‘right’ with a probability of less than 10/21. We have entered an inﬁnite regress: what it is rational for a player in a situation like A’s to do depends on what it is rational for a player in a situation like A’s to do.20 The solution to the problem is for A and B to think not about what is rational for each of them to do, but about what is rational for the team to do. And it is clearly rational for the team (meaning A and B) to go right. Again, though, this solution cannot be reached from the perspective of individual rationality; it can only be a product of team rationality. And so we have an example where rationality is irreducibly cooperative: it is a question of what we ought to do, and cannot be reduced to separate questions about what I ought to do and what you ought to do.21 One might argue that the Footballers’ problem is misleading, as both the players have the same goal. But cooperative rationality is often called for even if the players have different goals. Consider the venerable Prisoner’s Dilemma. I will list the outcomes for each player in order of preference in Table 3.2 below. The puzzle of the Prisoner’s Dilemma is that it is rational for each person to defect, though if each person defects, then each is worse-off than he or she would have been had they cooperated. Notice, however, that this understanding of the Prisoner’s Dilemma essentially involves individual rationality. If we ask, for each player, what it is rational for that player to do, the correct answer is ‘defect.’ But this leads to the paradoxical result that each player is not very well off, achieving only his/her third preference (out of four). But if we instead ask what it is rational for the players (taking as a collective) to do, then the answer is ‘cooperate’: if the players cooperate, then both will achieve their second preference. But we can only achieve this happy result if we ask what it is rational for them to do; if we insist on asking what it is rational for each player, considered individually, to do, then the players will continue to

Table 3.2 Decision matrix for Prisoner’s Dilemma Player B

Player A

cooperate

defect

cooperate

2,2

1,4

defect

4,1

3,3

Pragmatism and Rationality

61

achieve a sub-optimal result. Individual rationality is no replacement for cooperative rationality. I have been arguing that if a particular style or method of means– end reasoning systematically fails to achieve the desired end, whereas another style of reasoning would achieve that end, then it is irrational to follow the ﬁrst style or method. And as in the Prisoner’s Dilemma individual reasoning leads to the next-to-last (third) preference, whereas cooperative reasoning leads to the second-best (second) preference, individual reasoning fails the minimal test of means–end reasoning: it systematically fails to achieve the desired goal, whereas another style of reasoning reaches this goal. Thus, the Prisoner’s Dilemma again shows that (a) cooperative rationality is not reducible to individual rationality, and (b) cooperative rationality is often superior to individual rationality. Thus, rationality must often be understood cooperatively. Again, if what we care about is achieving our interests, and rationality is deﬁned instrumentally in terms of achieving our interests, then it is often irrational to reason individually rather than cooperatively, for doing so will fail to promote our interests. Rationality demands cooperative strategies. Act consequentialists have made numerous attempts to sneak cooperation into their account, while clinging to a model of act-based individualistic rationality (as opposed to strategy-based team rationality). Brandt, for example, has argued that defection (e.g., breaking a water conservation order) is irrational for the following reason: Suppose it would maximize utility for you to break the water conservation order. If it is rational for you to do so, then it is rational for all others to do so as well. But as all others doing the same would result in overwhelmingly negative consequences, it is not rational for you to break the water conservation order. However, Brandt is trying to sneak in team rationality through the back door. If we insist on individual rationality, then we can only ask what it is rational for you to do. And if your breaking the conservation order would not cause others to do the same, then it is utility-maximizing for you to break the order, and irrational for you not to do so. But if we move to a model of team rationality, we can see that it is rational for us to obey the water conservation order, and hence to restrict our water usage. Many will be suspicious of the idea of cooperative rationality. This suspicion can be brought out more clearly by focusing on differences between the Footballer’s Problem and the ordinary Prisoner’s Dilemma. In the Footballer’s Problem, both players have the same goal—namely, to score a goal. Thus, cooperation best serves the goal of each player. It is

62

Pragmatism and Normativity

rational for player A, for example, to cooperate, because by cooperating he maximizes his personal utility. However, the key feature of the Prisoner’s Dilemma is that cooperation does not maximize one’s personal utility—defection does. This is based on the assumption that the player’s strategies are independent— that is, the strategy chosen by the other players does not depend on the strategy you choose. In this case, the critic of cooperative rationality will argue that cooperation is irrational, as it fails to maximize one’s personal utility. The ﬁrst thing to say in response to this charge is that this objection again assumes that rationality can only be understood individually. But that is the very dogma we are rejecting here. It may be true that my cooperation does not maximize my personal beneﬁt—but the point is that our cooperation maximizes our beneﬁt. By cooperating, we realize our shared goals, and so it is rational for us to cooperate. To assume that all questions of rationality can be answered in terms of how it is rational for me, considered individually, to do X, is to beg the question against the current account. The Footballer’s Problem, though not a typical Prisoner’s Dilemma, has already shown that questions of what it is rational to do cannot always be given an individualistic answer. Let us approach the matter from another direction. An incautious thinker might conﬂate satisfaction of my preferences with maximization of my well-being. But this conﬂation assumes that all of my preferences or interests are self-regarding. That is to say, this conﬂation tacitly assumes the truth of psychological egoism, the idea that all of my preferences and desires are directed toward producing the maximum possible beneﬁt for myself. If this were true, then the realization of my goals would be independent of the realization of your goals. If both my interests and your interests are self-regarding, then realization of my interests is at best independent of realization of your interests, and at worst conﬂicts with the realization of your interests (as when scarce resources are at stake). Even if a game theorist does not make this egoistic assumption, the Prisoner’s Dilemma is still geared toward individualistic rationality: it lists my utilities and your utilities, but it lists them separately, and cannot convey the degree to which they overlap or reﬂect shared goals and preferences. However, this assumption regarding the independence of our interests cannot be made with respect to other-regarding interests. If we reject the assumption of psychological egoism, we clear the way to a richer understanding of rationality. It is true that we have self-regarding interests. Indeed, our interests are often self-regarding. But humans are social

Pragmatism and Rationality

63

animals and as such have other-regarding interests. Thus, consider A and B who are members of the same community. A wants A to ﬂourish, of course, but he also wants the other members of the community to ﬂourish. B also wants B to ﬂourish, but he too wants the other members of the community to ﬂourish. So if we focus separately on A’s preference satisfaction and B’s preference satisfaction (as traditional game theory asks us to do), we ignore the fact that A and B share the same interest—namely, the interest in the ﬂourishing of their community. As the members of the community share this interest, and as coordination among the members of the community will greatly increase their success in promoting this interest, it is rational for the members of the community to coordinate in order to promote this shared interest. Thus, egoistic assumptions might lead us to treat our preferences as separate and distinct. But if we reject this assumption, we can see how the members of the community essentially share certain interests. Given the existence of community-wide interests, it is easier to see the rationality of a community-wide cooperative strategy. The existence of shared interests, then, is one way to argue that cooperative rationality is a legitimate notion. We will return to this idea shortly. But shared interests are not the only realm in which cooperative rationality is a useful idea. A cooperative moral strategy will also advance my self-regarding interests. Gauthier writes that a moral system has several attributes: First, “Morality is a system of principles such that it is advantageous for everyone if everyone accepts and acts on it.”22 Second, “acting on the system of principles requires that some persons perform disadvantageous acts.”23 Third, “each person must gain more from the disadvantageous acts performed by others than he loses from the disadvantageous acts performed by himself.”24 Fourth, “each person will do better [if a moral system is adopted] than if no system is adopted.”25 In essence, then, although morality requires that I sometimes act contrary to my self-interest, I beneﬁt from the existence of a cooperative moral arrangement, and am better off than I would be in the absence of such an arrangement. As this is true for everyone in the community, it follows that each of us is worse off if there is no such cooperative strategy in place. Thus, our selfinterest is frustrated if the community fails to adopt such a cooperative strategy. As failure to achieve one’s ends through adoption of a bad strategy (or failure to adopt a particular strategy) signals a failure of means–end rationality, it follows that it is not rational for a community to refuse to adopt such a strategy. It essentially frustrates each person’s ends.

64

Pragmatism and Normativity

Notice, however, that only the supporter of cooperative rationality can explain why such a failure is irrational. The supporter of individual rationality admits that each person is worse off if he thinks individualistically, but is forced to chalk this up as a paradox of rationality. Her view does not make room for an explanation of why failure to cooperate is irrational. That is to say, although the choices made by the members of the community thereby preclude these members from achieving their desired goals (maximization of advantage), the supporter of individual rationality cannot explain why such failure signals a failure of rationality—for the simple reason that for this supporter, no failure of rationality has occurred, even though the community members’ choices prevented them from achieving their desired ends. If we want to achieve our desired goal—that each of us is as well-off as possible—then it is rational for us to follow a cooperative moral strategy. The argument here is in some ways related to the argument, noted earlier, advanced by Brandt. You may recall that Brandt advanced the following argument against defection from morality. Suppose it would maximize utility for you if you defect. If it is rational for you to do so, then it is rational for all others to do so as well. But as all others doing the same would result in overwhelmingly negative consequences, it is therefore not rational for you to defect. What prevented Brandt’s argument from being sound is that Brandt was working with a notion of individual rationality. If we ask what it is that is rational for me to do, and consider that my actions will have only minimal (or no) effect on the actions chosen by others, then we can only conclude that it is rational for me to defect. But if we allow the notion of cooperative rationality, then the fact that all of us doing X would harm each of us does provide an argument for us to not do X. But this argument can only establish that it is rational for us to refrain from X—that is, it essentially depends on the notion of cooperative rationality. Again, if we conﬁne ourselves to the question of what it is that is rational for me to do, and if my defection would not cause others to do the same, we can only conclude that I ought to defect, and it would be irrational for me to cooperate. If we combine (a) the existence of other-regarding interests with (b) the notion that only by following a cooperative moral strategy can we best advance our self-regarding interests, we can see that the notion of cooperative rationality not only makes sense, but is a powerful and viable substitute for individual rationality in many contexts. In other words, a rejection of psychological egoism, coupled with a recognition that rationality is often both strategic and cooperative, is essential to understanding institutions such as morality.

Pragmatism and Rationality

65

I will make one ﬁnal comment. As noted above (in the section “Actions and Strategies”) Gauthier’s argument is that cooperation is often rational. However, I will not pursue Gauthier’s argument here. The reason is as follows: Gauthier is trying to prove from the standpoint of individual rationality that it is rational for the individual to cooperate. Thus, Gauthier is still taking the notion of individual rationality to be basic. But what I am trying to establish here is that there is another notion of rationality, cooperative rationality, which is separate from and not reducible to individual rationality. A demonstration that it is individually rational to cooperate with others does not show this. It is not an unwelcome addition to the theory—of course, if we want to show that it is rational to be moral, then it is helpful to be able to show that it is both individually and cooperatively rational to be moral, but the goal of the current section is to show that the notion of cooperative rationality is viable in the ﬁrst place. Showing that it is individually rational to cooperate does not show this, but merely relies on the pre-existing notion of individual rationality. However, I will note that to the extent that Gauthier and others have demonstrated that cooperation is individually rational, they have strengthened my case that it is rational to be moral. Over-determination never hurts anything. To conclude this section, the general pragmatist strategy is to show that it is rational to be moral. Morality consists of a set of strategies aimed at promoting our interests. Discussions of whether it is rational to be moral in general focus on whether it is always prudent to be moral, where ‘prudence’ is treated as synonymous with ‘self-interest.’ But we must recognize that in addition to self-regarding interests, we also have other-regarding (i.e., altruistic) interests. I will have much more to say about this in the next chapter, when I discuss the pragmatist account of morality. But for the time being, let us note that the upshot of this section is that following morality displays means–end rationality, where the end in question is the promotion of our interests.

Collective preference and joint intentions What grounds collective action? What makes it rational to participate in a cooperative strategy? I think that in acknowledging the viability and legitimacy of cooperative rationality, we should also recognize the existence of something like group preferences, joint commitment and/or joint intentions. Other philosophers have recognized the existence of such shared states. Wilfrid Sellars writes about “weintentions”;26 and David Velleman explains how to share an intention.27

66

Pragmatism and Normativity

Following Margaret Gilbert’s well-known work, I wish to argue that cooperative strategies are based on collective preferences/interests and joint commitment.28 Gilbert, like Sellars29 and others before her, has recognized a common form of reasoning carried on by humans: We aim to achieve X. I can best help achieve X by doing A. Therefore, I ought to do A. Thus, an individual action imperative is derived from the fact that a group of which I am a member has a particular aim or goal. Thus, our footballer might reason as follows: We aim to score a goal and beat the other team. I can best help achieve this by moving right. Therefore, I ought to move right. On Gilbert’s account, the ﬁrst sentence in this practical syllogism expresses a collective preference. As Gilbert writes, “I do not see how one can consistently accept the validity of this form of argument while maintaining that a collective preference does not provide group members with a sufﬁcient reason for action.”30 People who are committed to group action possess what Gilbert calls a joint commitment. Joint commitment is not reducible to personal commitments; there are structural dissimilarities between the two. For example, as Gilbert notes, one may unilaterally rescind a personal commitment (say, a commitment to take a walk by oneself), but one cannot unilaterally rescind a joint commitment (such as a commitment to win the football game or score a goal).31 When a group has a collective preference, they can form a joint commitment to satisfy this preference or commitment. Gilbert writes, “A joint commitment is a commitment to do something as a body . . . I say that those who are jointly committed to do something as a body constitute a plural subject.”32 To illustrate the connections between various concepts, we can again consider the example of water conservation during a drought. Now having access to adequate clean water serves important human interests (a concept I will explore at length in Chapter 5), such as our interests in health and life. As these interests are constitutive of human ﬂourishing (again, a topic for Chapter 5), it is rational for us to entertain a collective preference that water be conserved. Perhaps spontaneously, or perhaps as part of a publicity campaign to publicize the water shortage, we are made aware of the importance of water conservation

Pragmatism and Rationality

67

and we then form such a preference, which becomes a joint commitment to follow a cooperative strategy of water conservation. This joint commitment is manifested by individual acts of water conservation— washing one’s car and watering one’s lawn less frequently, turning off the tap while brushing one’s teeth, and so forth. Thus, we see the connection between interests, collective preferences, joint commitment, and the actions that fall under a cooperative strategy for promoting the aforementioned interests and preferences. Gilbert argues that persons belonging to a collective have an obligation to act on this joint commitment. At ﬁrst glance, this view might seem to threaten circularity if incorporated into the present pragmatist account. After all, we are trying to explain the genesis of moral obligation by appeal to collective rationality and collective preference; and it would seem that Gilbert’s account reverses the order of explanation. Gilbert, for her own part, denies that the obligation associated with joint commitment is moral obligation. She writes, “If you like, the obligations in question here are not moral obligations. They may be referred to as obligations of joint commitment to avoid misunderstanding.”33 But perhaps it would be helpful if we avoid the word ‘obligation’ altogether, as the word has unavoidable moral connotations. Nevertheless, violations of prudential rationality can merit reproach. For example, I may reproach myself for taking an extra helping of dessert. Or, to pick an example entirely devoid of any moral implications, I may reproach myself for postponing work on a project until shortly before the deadline. Similarly, I may reproach a student for failure to submit a rough draft of a paper for my review (even when this is not required for the course), on the grounds that even if the student is not morally or institutionally obligated to do so, it is foolish for the student not to take advantage of a mechanism whereby she might improve her course performance. These are not examples where a moral obligation is being violated; they are examples where one is being reproached for acting imprudently rather than unethically or immorally. By extension, we may reproach a member of our group for a failure of cooperative prudence—that is, for failing to participate in an optimal cooperative strategy. When discussing the footballers’ example, Gilbert writes of Art and Ben (as she names the two footballers) for whom it is cooperatively rational that they go right, Now suppose that at t1 Art moves left. We should expect Ben and indeed the other team members to be surprised. We would expect their reactions to go beyond mere surprise. Given that Art’s moving

68

Pragmatism and Normativity

left is not obviously explicable by some mitigating circumstance, his teammates will feel aggrieved . . . Whether or not rebukes are uttered by his teammates, Art will understand that they have the standing to rebuke him . . . [T]hose rebukes that are uttered may be backed up with such remarks as ‘We’re trying to win this game, not lose it!’ and ‘How are we going to win if you do things like that?’34 Thus, just as I may reproach myself for a violation of prudential rationality, we may reproach one of us for a violation of cooperative rationality. In neither case need there be a moral tinge to the reproach. But the reproach is appropriate, merited. But of course, morality just is, for the most part, our set of cooperative strategies for promoting our interests. So once a moral practice is established, our rebuke aimed at someone who violates an optimal cooperative strategy has a moral ﬂavor. For, I have been arguing that what it is for a person to have a moral obligation is for there to be a particular cooperative strategy, aimed at promoting a group’s interests, which it is rational for a group to follow, and which dictates a certain action for the person in question. But showing how someone can merit non-moral reproach for failure to participate in a rational cooperative strategy lends credence to the idea that morality can rest on collective preferences and joint commitment; it makes it easy to see how a moral practice could arise out of cooperative prudential rationality. The question naturally arises about the voluntariness of this cooperation. I have argued that it is rational to participate in cooperative moral strategies, but obviously people can choose to behave irrationally. The question arises as to how cooperative agreements can be binding on people who might not have consented to them in the ﬁrst place. By discussing how people might choose to opt out of a system of cooperation, we can see how this system can be binding on people even if they do not have any realistic alternatives to cooperation with society’s moral strategies. There are two possibilities for opting out. One can either entirely opt out of the system of cooperative strategies, or selectively opt out of portions of it. The one who chooses to entirely opt out of the system of cooperative rationality chooses not to be one of us. Couldn’t one reject the idea of being one of us? That is, speaking analogically, couldn’t one refuse to play on the team? Philosophers write often of the amoralist, the person who sees no reason to subject himself to any moral constraints, one who refuses to abide by our cooperative strategy. Could we demonstrate to the amoralist that he has reason to be moral?

Pragmatism and Rationality

69

There are at least three comments to make about this. First, I have argued in this chapter that morality is rational, in the sense that it is a cooperative strategy aimed at promoting our interests. Thus, assuming one has the same interests as we do, one does have a reason to be moral. I will argue in the next chapter that people in general have the same interests. So, it seems likely that the vast majority of people will have a reason to be moral—it is rational for them to do so, even if they do not acknowledge that it is rational. (Obviously, one can fail to recognize that it is in one’s interest to do something.) The second comment is this: philosophers write breezily of the amoralist, the person who feels no moral constraints. But it is easy to be overly glib about such a possibility. A person who genuinely rejected all moral standards—who refused to play along with the team strategy— would be more of a monster than a human. Careful thought reveals that someone who really did not feel any compulsion to tell the truth, preserve human life, avoid cruelty, and so forth, would be a monster, and the best strategy for us to follow would be to put such a creature where it could not do any harm—namely, a prison or a mental facility. This is not to show that the amoralist does, in fact, have a reason to be moral. This is merely to ask the reader to focus on what a genuine amoralist must be like, how rare such a creature must be, and how arguing with such a creature is the wrong strategy to pursue. The ﬁnal comment to make is this: the community need not rely on people’s recognition that following a cooperative strategy for mutual beneﬁt is in itself a good thing. The community can also introduce sanctions to help enforce such cooperation. Thus, an amoralist might not see why the mere fact that something is morally right gives him a reason to act; nevertheless, the fact that he may be punished if he fails to act (or does the wrong thing) might provide him with a reason that is sufﬁcient, by his own lights, to motivate him to act. As Kurt Baier notes, because morality is a system whereby we beneﬁt more from others’ adherence than we lose from our own adherence,35 it is rational for the community to adopt an enforcement mechanism that adds further incentives to moral behavior. But are these coercive measures against the amoralist permissible? I believe that they are. Consider an analogy, which is adopted from the conscription case discussed by Sugden36 and Gilbert.37 Imagine a group of individuals who are kidnapped and dropped in the middle of the desert together. By banding together, they can trek out of the desert to civilization and survive. Suppose that for various reasons (say, the distribution of skills among the individuals), none of these individuals

70

Pragmatism and Normativity

can achieve this goal alone and unaided. The group has an overarching joint aim or goal (to reach civilization), and various subsidiary group intentions that are aimed at achieving this goal (e.g., to march 10 miles before sunrise). Now how should the group treat a person who refused to be part of the group and did nothing to promote the group’s goals? (A genuine amoralist would, of course, often take actions that served to thwart the group’s aims, although thwarting the group’s aims would not necessarily be the point of or reason for the amoralist’s action.) Clearly, the group would not be amiss in sanctioning the individual—refusing to aid him (if he is a free rider, doing nothing to promote the group’s goal), and even forcibly ejecting the individual if he takes actions that tend to thwart the goal of the group. Thus, even though the individual does not really have a choice as to whether to join the group or not—it is either join the group and adopt their intentions, or perish in the desert—the group violates no right of his in enforcing its group intentions on him. An individual facing the choice of either joining society and sharing in its group intentions or being an amoralist and hence an outcast from society (assuming, unrealistically, that a person who could make a choice between amoralism and moral obedience were psychologically possible) is in a similar position, and similar comments apply to him. The second possibility for opting out is that one could partially opt out of our system of cooperative strategies. That is, one could agree to abide by certain of these cooperative strategies, but not by others. The point to notice about this case is that the dissenter has already placed himself within the moral system. He has already become one of us; he merely wishes to abide by only certain of our constraints. But if he is one of us, then he is by his own lights in the space of moral reasons, and he owes us an explanation for why he wishes to deviate from a particular cooperative strategy. He might argue that this cooperative strategy is sub-optimal, or that it conﬂicts with another more important cooperative strategy; but as one of us, reasoning collectively, he must argue for why his non-participation is legitimate. And of course, we do this all the time in applied moral theory—we argue why certain dominant moral rules (i.e., certain dominant cooperative strategies) do not deserve to be followed, or should be replaced by other strategies, and so on. This kind of dissent from the dominant cooperative strategy is a familiar element in moral argument, and essential to moral progress. For example, if people had not dissented from the dominant strategy of treating women and minorities as second-class citizens by disenfranchising them and in other ways limiting their rights, then that particular sub-optimal strategy would not have fallen and been replaced by a better one, one

Pragmatism and Rationality

71

that better served our interests. (Although such examples illustrate that though in some cases moral progress is made by ﬁnding better cooperative strategies, in others it is made by recognizing that a certain group is, in fact, one of us.) To consider the example of our desert trekkers, a person might dissent from a particular cooperative strategy—say, the strategy of hiking only during the night when it is cool—but he clearly owes the group an explanation of why he thinks this strategy ought not be adopted. Obviously, a larger society can be much more tolerant of individuals’ moral dissent than can our group in the desert, whose very survival depends upon everyone’s cooperation. But the point is, as illustrated by the story of the desert trekkers, that when an individual’s non-cooperation threatens the legitimate interests of others, then sanctions are appropriate and the group members are entitled to administer these sanctions. Of course, as a result of this, people will often be wrongly sanctioned when a group is following a sub-optimal strategy and the dissenters are promoting a superior one, one which ought to be adopted as a moral rule in society. History abounds with examples of people who were punished for being right, because of being thought wrong. But this is a feature of human fallibility, and I am not aware of any moral theory which can eliminate this problem. However, it is important to note that while joint action requires sharing a strategy or intention (it requires joint commitment), it does not require sharing an ultimate end (although sharing of ends can make it easier to formulate a cooperative strategy). We can illustrate this by discussing the three sorts of cases that arise in connection with cooperative rationality: shared ends, parallel ends, and idiosyncratic (unshared and non-parallel) ends. Suppose that a group shares an end—say, reducing the emission of greenhouse gases. That is, everyone in this group wants to promote a common end. In such cases, where there is a single end to be promoted by all, it is relatively easier to formulate a cooperative strategy. People can then share the intention of realizing this strategy (whose elements might include, say, using public transportation, raising the thermostat in the summer and lowering it in the winter to save fuel, using energyefﬁcient light bulbs, and so forth). It can be said that we aim to reduce greenhouse gas emissions by taking the above steps. A second case is where a group has what I shall call parallel ends. In this case, each person desires a particular end (usually for himself or herself). To give an example, each person has an interest in health and physical well-being, but what each person primarily wants to promote

72

Pragmatism and Normativity

is his or her own health. (For the sake of simplicity, I am ignoring altruistic concern for the health of others.) Thus, each person has a parallel end—Smith aims to promote her own health, Jones aims to promote his own, and so do Black, Brown, and the other members of the group. In this case, the existence of parallel ends also simpliﬁes the task of formulating a cooperative strategy which can serve as the basis for joint intention and collective action. The group can intend to create and sustain a system that promotes the health of each person in the group. Consider an example from Gilbert, that of the mushroom pickers. As Gilbert describes the group, “every now and then members of our population go on solitary mushroom picking expeditions in ‘their’ wood. When one is picking mushrooms with his head bent down, he sometimes inadvertently collides with another, who is doing the same.”38 Obviously, this is painful for everyone, and they wish to formulate a strategy to minimize such collisions. Now in this case, the members of the group have parallel aims—each member of the group wants to avoid ramming his or her own head into that of another group member. Thus, although each member of the group would use an identical sentence to express his or her goal—‘I wish not to ram my head into that of another group member’—the indexical elements in the sentence mean that the members of the group do not share the same goal. They have, instead, parallel goals. Again, the existence of parallel goals makes it simpler to formulate a cooperative strategy, because in a sense everyone wants the same thing. Thus, to use the rule Gilbert proposes, “after a mushroom picker has picked one dozen mushrooms, he must look up and make sure no one else is coming towards him.”39 This cooperative strategy serves everyone’s end, and its formulation is, again, simpliﬁed by the fact that the members of this group have parallel (not shared) ends. The ﬁnal case is when people have idiosyncratic (unshared and non-parallel) ends. This is commonplace in society, where people seek fulﬁllment, meaning, and entertainment through a wide variety of activities: Some knit; some dance; some seek communion with nature; others seek communion with God; and so forth. Each of us will recognize that it is rational to agree to a system in which we are allowed, to the extent feasible, free reign to pursue satisfaction of these unshared interests, free reign to pursue satisfaction of our interests in ways that are unique to us. Thus, it is rational to agree to a strategy allowing us a certain space in which individuals may pursue their idiosyncratic projects and desires. This latter element is made rational by the fact that we all have such unshared aims, and it is more rational for us to adopt a cooperative strategy of toleration than is it to simply forego satisfaction of these myriad

Pragmatism and Rationality

73

idiosyncratic interests. This is how idiosyncratic aims are dealt with in the theory of cooperative rationality. So we have shown how cooperative strategies are based on joint commitments to promote particular ends, whether these ends are shared, parallel, or idiosyncratic. But what happens when there is a conﬂict between what is individually rational and a particular cooperative strategy? For, while I have argued that cooperative rationality is indispensable, I have not argued that it is the only viable form of rationality. That would be too strong of a claim; surely there are cases where rationality can be construed in individualistic (rather than cooperative) terms. This problem is not identical to the familiar problem of the conﬂict between morality and self-interest (because, as we have seen, cooperative strategies are often formulated to advance one’s self-interest), but it resides in the same neighborhood. Although this answer might disappoint the reader, it is unlikely that anything too speciﬁc can be said about how, in general, one resolves conﬂicts between individual rationality and group rationality. To be sure, there have been attempts to resolve this conﬂict. For example, Raimo Tuomela presents a highly complex and formalized method for weighing individual preferences against group preferences within a game-theoretical framework.40 Though such discussions might be of theoretical interest, they are of little practical value to the moral agent who is trying to decide between personal aim and group preference. A consistent theme throughout the present work (and one on which we have already touched in this chapter) is that our moral and epistemic tools must be useable; and overly-formalized decision procedures, which make unreasonable demands on agents, cannot reasonably be said to constrain the actions of these agents—or to offer these agents with useful advice. (We will return to this theme in Chapters 7 and 8, where we will make similar arguments against epistemological theories such as Bayesianism which evaluate an agent’s belief according to a standard which the average agent is completely incapable of applying.) Thus, the very weakness of Tuomela’s account is that it aims to provide a deﬁnite method for resolving conﬂicts between individual and group rationality. There are several things to be said in response to this worry about potential conﬂicts between individual and cooperative rationality. For one, the most rational cooperative strategy is one that will minimize conﬂicts between group and individual rationality. One way in which our common-sense morality does this is that it presents itself usually as a side-constraint41 on our action. That is, in general one is allowed to pursue one’s own interests, and apply individual rationality (be it

74

Pragmatism and Normativity

atomistic or strategic) to develop one’s own life-plan and seek fulﬁllment in this way. Morality serves in general not as a guide as to what to do, but as a guide as to what not to do. Stated differently, in general we are allowed to pursue our own interests as we like, as long as we do not immorally violate the interests of another. So in an important sense, our everyday morality represents a rather weak constraint on action which already favors individual rationality (and self-interest) over cooperative rationality. By minimizing the intrusiveness of morality, our ordinary moral cooperative strategies minimize conﬂicts between cooperative and individual rationality. Another way of putting this point is in Kantian terms. We have negative and positive duties, and perfect and imperfect duties. Negative duties are duties to refrain from some action (such as lying), and positive duties are duties to perform some action (such as charitable giving). Perfect duties are those that we must always practice; imperfect ones are those that must be practiced only from time to time. Intuitively, our perfect duties are in general the negative ones—the ones that exhort us to refrain, at all times, from lying, unjustly harming others, and so on. These duties serve as constraints on our pursuit of our self-interest. They tell us what the permissible ways to pursue our self-interest are, and what are not. The positive duties, demanding as they do positive action on our part,42 are duties that one can discharge by acting sporadically (say, by occasionally giving to charity). Thus, the only constraints that are binding on us at all times are negative in nature, and rather than committing us to positive action, merely channel our pursuit of self-interest into ways that are not unjustly harmful to the interests of others. Thoreau sums up this view quite well when he writes, It is not a man’s duty, as a matter of course, to devote himself to the eradication of any, even the most enormous wrong; he may still properly have other concerns to engage him . . . If I devote myself to other pursuits and contemplations, I must ﬁrst see, at least, that I do not pursue them sitting upon another man’s shoulders.43 However, there will be cases where individual and cooperative rationality do conﬂict with each other. What ought one to do in such cases? Again, it is unlikely that anything much speciﬁc can be said about this matter. As Gilbert writes, One can sometimes validly excuse oneself from violating a commitment, joint or not. What counts as a valid excuse in such a case, as in

Pragmatism and Rationality

75

any other, will ultimately be a matter of judgment, that is, a matter of assessing the way in which reasons of various kinds play off one against the other.44 Just as contemporary particularists and anti-theorists have denied that one can give a formal decision procedure in ethics (one that will take as input a set of circumstances and will produce as output the obligatory action), it is unlikely that a formal decision procedure can be given to resolve all conﬂicts between individual and cooperative rationality. Though this result is no doubt disappointing, it should not be surprising in light of recent trends in metaethics and the theory of rationality. One ﬁnal comment deserves to be made about the relation between individual and cooperative rationality. Although we have argued above that cooperative strategies rest on joint commitments, which are not reducible to personal or individual commitments, we do have certain other-regarding interests which will enhance cooperation. In Chapter 5, I will argue that among our interests is the interest in afﬁliation. Humans are social animals, and seek afﬁliation with other humans (and possibly other species, too). This interest manifests itself also in sympathy toward and liking for others, and in a desire to see others (especially those close to us) achieve elements of human ﬂourishing. Thus, one of the interests we have, and therefore one of the goals we have reason to promote, is the welfare and ﬂourishing of others. Again, this should not be taken to suggest that we wish to ground cooperative rationality in individual rationality; cooperative rationality is irreducibly a group activity. It merely notes that our afﬁliative interests will increase our inclination to participate in group strategies. And indeed, research in game theory has revealed a variety of conditions under which cooperative, mutually beneﬁcial strategies evolve and become stable, even among parties who are ostensibly enemies.45

Further objections against rule-based systems The pragmatist system I am advancing here is teleological in the sense that the rules are aimed at promoting some good—namely, they are aimed at promoting our interests. As such, the theory is vulnerable to the objections that are raised against rule utilitarianism, the other well-known rule-based teleological system. Let us examine some of the better-known objections to see if they carry any weight.

76

Pragmatism and Normativity

The ‘simple collapse’ argument One objection against rule utilitarianism is that it collapses into act utilitarianism. A similar charge might be leveled against rule pragmatism. Suppose you have a practice consisting of a set of rules. Consider a rule, R1: ‘Whenever in circumstances C, do action A.’ However, the objection continues, when in C, the agent is actually faced with a choice between two different rules: R1: Whenever in circumstances C, do action A, and R2: Whenever in circumstances C, do action A, or whatever action will produce the best results. Clearly (the objection goes) the second rule will produce the best overall results. Ergo, the second rule is the one we should follow, and rule pragmatism collapses into act pragmatism. This argument will not work. After all, the rule pragmatism counsels adopting rules that will produce the best results in the long run. But the adoption of rules like R2 will not, in fact, produce the best results in the long run. As we saw above, rules like R2 lead inexorably into free-rider problems, Prisoner’s Dilemmas, and so forth. So, applying rules like R2 might maximize short-term beneﬁt, but in the long run, rules like R1 will produce better consequences. Consider the point this way: I have argued above that the basic unit of fundamental rationality is often the strategy, not the action. But that does not mean that any strategy will do; the strategy must be wellchosen. Thus, for example, if my dieting strategy involves having a scoop of ice cream every day, it is unlikely that my diet will succeed. It will not achieve its long-term goal. Similarly, if the strategy chosen by the rule pragmatist is R2, then the strategy will fail to achieve its longterm goals. The strategy will be brought low by all of the problems that are faced by act pragmatism, the same problems that forced us to concede in the ﬁrst place that strategy is often the basic unit of fundamental rationality. Thus, the rational strategy is R1, not R2; it is more rational to select R1 than to select R2, and R2 can therefore not be said to be the choice that produces the best results. Rule pragmatism is not ‘real’ pragmatism Other critics of utilitarianism have charged that act utilitarianism is a ‘purer’ form of utilitarianism, and that rule utilitarianism abandons what Sanford S. Levy46 calls utilitarianism’s ‘teleological motivation,’ the

Pragmatism and Rationality

77

idea that one should begin with an idea of the good, and derive one’s moral theory entirely from this idea of the good. Philippa Foot puts the objection as follows: What is it, let us now ask, that is so compelling about consequentialism? It is, I think, the rather simple thought that it can never be right to prefer a worse state of affairs to a better. It is this thought that haunts us and, incidentally, this thought that makes the move to rule utilitarianism an unsatisfactory answer to the problem of reconciling utilitarianism with common moral opinion. For surely it will be irrational, we feel, to obey even the most useful rule if in a particular instance we clearly see that such obedience will not have the best result.47 By recommending sub-optimal actions, it is thought, rule utilitarianism is “unmotivated by [the teleological motivation] or even inconsistent with it.”48 One can see how this objection might be extended to rule pragmatism. The objection is basically the rule-worship objection in another guise, and the response is the same. The objection claims that rule pragmatism is not true to pragmatism’s ‘teleological motivation’ because it often requires sub-optimal actions. But again, this objection insists on examining rationality on the level of actions, not strategies. If we can see that the rule-pragmatist strategy creates better results than the act-pragmatist one, we can see that it is in fact rule pragmatism which is truer to the teleological motivation. The rule speciﬁcation problem Another standard objection against rule utilitarianism is that it often counsels foolish action when others will not cooperate. Peter Railton, for example, writes: I suspect that rule-consequentialism is untenable . . . for it would recommend acts that . . . accord with the best set of rules even when these rules are not in fact generally accepted, and when as a result these acts would have devastatingly bad consequences. “Let the rules with the greatest acceptance utility be followed, though the heavens fall!” is no more plausible than “Fiat justitia, ruat coelum!"—and a good bit less ringing.49

78

Pragmatism and Normativity

Let me illustrate this point with an example. Suppose a car has run out of gas and needs to be pushed to the ﬁlling station. If three people are required to push the car and if three people are present, then rule utilitarianism dictates that the utility-maximizing rule in this situation would be, ‘Available and capable persons should help push the car to the gas station’, or something similar. But suppose that two of the three people refuse to help push. Nevertheless (the objection goes) the third person should push on the car, because this action is dictated by the rule that, if followed, would maximize utility in this situation. Thus, even though one person cannot by himself move the car, and it is utterly irrational for the third person to stand behind the car and push on it, rule utilitarianism nevertheless dictates that the lone person should push the car. Rule pragmatism might face a similar objection: if we are to follow the rules that would, if generally followed, best serve our interests, then I should follow this rule even if no one else does, thereby wasting my effort. This objection only has force because it assumes that the rules in a rule pragmatist system will be maximally speciﬁc. Thus, the rule pragmatist system will (goes the assumption) consist of rules such as the following: R3: If a car has run out of gas, and a ﬁlling station is near enough so that the car can be pushed to the ﬁlling station, and there are enough able-bodied persons available to push the car to the station, then you should push the car to the station. But once the rule is stated, we can immediately see the absurdity of assuming that a rule pragmatist system will consist of rules that are this speciﬁc. As I noted above, a moral system must be livable. And a system that consists of rules as speciﬁc as R3 would be far too complex to learn. A more likely rule would look something like the following: R4: If someone is in need of assistance, and you can render assistance without undue cost to yourself, then you ought to do so. This rule does not specify all the various circumstances in which one might need assistance (If a car is out of gas; if a person has a snakebite; if a person is unable to pass a test without tutoring assistance; etc.); that would result in an unlearnable profusion of rules. Nor does R4 specify how one might render help (Push the car, etc.); that too would introduce unnecessary complexity. Rather, R4 assumes that an intelligent moral agent can recognize when someone needs assistance, and crucially, that

Pragmatism and Rationality

79

this intelligent moral person can tailor a response appropriate to the circumstances. For example, let us again consider the car that is out of gas. An intelligent moral agent following R4 will decide that the owner of the car needs help. After trying (and failing) to enlist the help of the two bystanders in pushing the car to the ﬁlling station, the moral agent would then cast about for other ways of rendering assistance (say, walking to the ﬁlling station, borrowing a gas can, and buying enough gas to allow the car to be driven to the ﬁlling station). I do not think anyone seriously believes that a system of morality would consist of rules as speciﬁc as R3. R4 is a much more likely candidate. How speciﬁc should the rules be? I think that there will be a trade-off between (a) the applicability and learnability of the rules and (b) the interest-promoting nature of the rules. A more speciﬁc set of rules will better foresee exceptions; be better tailored to various, slightly different situations; and so on. But again, an overly-speciﬁc set of rules will not be learnable. And so I think a rule consequentialist will counsel us to look for an equilibrium point that represents a compromise between (a) and (b). Of course, it would be nearly impossible to sit down and construct a set of rules that satisﬁed this compromise. However, I think we should recognize rule formulation as an ongoing process of rule modiﬁcation, keeping both (a) and (b) in mind. As such, when I speak of an equilibrium point between (a) and (b), I am not trying to give a set of truth conditions for pragmatist rules, or some ideal observer theory of the correctness of rules. Rather, I am suggesting a heuristic device that would help in the ongoing project of formulating and modifying our moral code. In general, the response to the question, ‘How speciﬁc should a rule be?’ is: ‘Whatever level of speciﬁcity best serves our interests.’ If a rule is overly speciﬁc, people will have trouble learning it, will misapply it, and inefﬁciency will be introduced into the moral practice. But if the rule is not speciﬁc enough, then it will not display an appropriate sensitivity to various circumstances, and will hence not serve our interests well overall. For example, a rule that is more general and lumps together circumstances of sorts A, B, C, and D might serve our interests less well than a more speciﬁc rule that lumps together A, B, and C, but treats D as a separate circumstance, governed by a different rule. Again, though, questions of appropriate speciﬁcity must be settled ultimately by appeal to our interests, and to a consideration of the level of speciﬁcity that best serves our interests. At any rate, as I will argue in Chapter 4, the complexity of morality precludes sitting down in our armchairs and devising a set of adequate moral rules. Moral rules must be tested by living them,

80

Pragmatism and Normativity

and adopting, testing, revising, and discarding moral rules is an ongoing process. In this process, there will of course be a trade-off between speciﬁcity and learnability, but this trade-off must be negotiated over time. Thus, we need not specify in advance a particular balance between these two considerations; such a balance will be worked out in practice, as it must. Having said all this, I will concede that on occasion rule pragmatism might counsel fruitless actions. If there is an action (such as recycling) that requires a certain threshold of participation before producing a net beneﬁt, a person might be obligated by rule pragmatism to participate even though the total number of people participating is below this threshold. I am not sure, however, that this is a fatal objection to rule pragmatism. In fact, it seems intuitively plausible that even if not enough people recycle, I should continue to recycle in the hope that participation by a sub-optimal number of people might eventually grow into participation by enough people. If all participants quit recycling, then there is no hope for the future of recycling; but if the current participants stick to it, then the numbers might eventually grow large enough so that the beneﬁts of recycling are realized. Of course, rule pragmatism would have to eventually allow nonparticipation, if a cooperative endeavor is obviously going to fail. But the point at which non-participation is allowed will vary according to the circumstances. Indeed, even in the face of massive non-participation, following a moral strategy can still produce desirable results: one family hiding Jews from Nazis can (obviously) save several lives, even if few (or no) other families are willing to take similar steps. In other cases, however, (such as the case of recycling) thousands must participate before any beneﬁt is produced. So the point of allowable non-participation will vary; but non-participation will eventually be allowed. This seems, however, like a less controversial point, and does not generate obvious difﬁculties for rule pragmatism (other than the familiar difﬁculty that it will be a matter of judgment as to when the point of allowable non-participation has been reached). Thus, we can conclude that this objection seems insufﬁcient to defeat rule-based teleological morality. However, this objection calls to mind a similar one. What if a community is engaged in a cooperative strategy which, although it promotes our interests, is sub-optimal? Should an individual who is aware of the optimal strategy (and who knows the strategy is optimal) follow the community, or follow the optimal strategy, even if others don’t? The answer would depend on a number of factors, such as how suboptimal the current strategy is. But even if a strategy is sub-optimal, if it

Pragmatism and Rationality

81

is an interest-promoting strategy, it might often be rational to follow it. Let us consider a couple of examples. First, let us consider the practice of writing letters of recommendation for job candidates in academia. The current strategy is that recommenders write honest but inﬂated recommendations. A more rational strategy would be for recommenders to write honest and non-inﬂated letters. This is more rational because it would allow hiring departments to better assess the abilities of job candidates. However, it is not obviously rational for an individual to follow this latter strategy, even if it would be optimal as a cooperative strategy, as doing so would unduly harm the job prospects of the candidates one is recommending. Consider a second example. A Marxist once remarked to me that it was unfortunate that workers had to alienate their labor by working in factories, on assembly lines, but given the capitalist organization of US society, Marxists such as himself had a moral obligation to work to ensure that workers were able to obtain such jobs and to ﬁnd such jobs for workers who were unemployed. This example can be understood as follows: the Marxist believed that a Marxist economic arrangement would best serve our interests. However, the Marxist lives in a capitalist society—one which (by his lights) follows a sub-optimal economic strategy. Now, this Marxist thinks that in an ideal society, workers would not be alienated from their labor, and would not be forced to work, say, on assembly lines. However, given that society is organized around capitalist principles, not Marxist ones, what should a Marxist try to do for workers? He ought to help them get jobs. Even though by doing so he is encouraging participation in a cooperative strategy he does not endorse, nevertheless it might be the most rational course for him to pursue. He might agitate for a re-ordering of his country’s economic structure, but given that it is ordered in a particular (non-optimal, he believes) way, there are certain respects in which it is rational for him to participate in this non-optimal strategy. Thus, if one’s society is following a non-optimal strategy, it might nevertheless be rational for one to go along with this strategy, rather than follow the optimal one. Can we give general rules for when one should follow the non-optimal strategy? Perhaps not, but we can at least outline some relevant considerations: How sub-optimal is the strategy followed by society? Would an individual generate great pragmatic beneﬁt by violating this strategy? What would be the repercussions (in terms of sanctions, etc.) to the individual for violating the community’s strategy? The question is not entirely dissimilar to the question of when one ought to violate a moral rule, such as the rule against lying. A variety

82

Pragmatism and Normativity

of considerations are relevant: How serious is the lie? Do you have a relationship of particular trust with the person you are deceiving? What is the pragmatic beneﬁt of lying, and what is the beneﬁt of telling the truth? By asking these questions, we can uncover when it is permissible to lie: perhaps in telling Aunt Jane that her Bundt cake is tasty; deﬁnitely in telling the Nazis there are no Jews in this house; not necessarily in telling your father that you did not chop down the cherry tree. It is not obvious that deciding whether to violate a moral rule is itself a rule-governed activity; it is one in which practical wisdom plays a role. Similarly, in deciding when to violate and when to adhere to a non-optimal strategy, one must use practical wisdom to weigh a variety of different considerations. The discussion of when one is justiﬁed in violating a moral rule (i.e., deviating from an optimal cooperative strategy) raises a potential objection against the current account. Let us turn to this possible problem.

Rules and exceptions One form the rule-worship objection might take is of the idea that rules should admit of exceptions. After all, even the strictest rule utilitarian could hardly say that one ought to follow a rule with the highest acceptance utility even if doing so in a particular case would cause the destruction of all life on earth. If one rigidly follows a set of rules without regard to circumstances, and does not allow exceptions to the rule, then isn’t one engaging in a form of rule worship? But how can a rule have exceptions and still be a rule? This question arises out of the (mistaken) notion that there are precisely two sorts of rules: generalizations and exceptionless principles. The ﬁrst sort of rule, which Rawls calls the ‘summary conception of rules’,50 is merely a rule of thumb, without separate explanatory or justiﬁcatory force. Thus, an act utilitarian might endorse the rule, ‘It is (generally) wrong to lie,’ while acknowledging that this rule just means that as a matter of statistics, lying usually produces less utility than truth telling. Now, such a rule can be justiﬁcatory in an epistemic sense: Suppose I am trying to decide whether to tell the truth, but I am unable to perform the utilitarian calculation in a particular instance. The rule might justify the belief, ‘I ought to tell the truth in this case.’ But it justiﬁes this belief via a statistical inference: truth-telling usually produces more utility than lying, and so in this case, (probably) I ought

Pragmatism and Rationality

83

to tell the truth. But if an act utilitarian is able to determine that in this case, lying would actually produce the most utility, then she obviously ought to disregard the rule, ‘It is (generally) wrong to lie,’ on the grounds that in this particular case, it is correct (i.e., utility-maximizing) to lie. If the act utilitarian knows that lying would maximize utility in this case, then the rule ‘It is (generally) wrong to lie’ carries no justiﬁcatory force at all. Thus, this ﬁrst type of rule allows for exceptions—but only because it is a statistical generalization. And because it is a statistical generalization, it carries no independent justiﬁcatory force. In the case of act utilitarian rules of thumb, for example, the principle of utility retains all the justiﬁcatory force—the rule of thumb carries none, or is merely useful faute de mieux, in the absence of a utility calculation in a particular case. The second conception of rules holds that rules must hold without exception; otherwise, they are not real rules. This is in many ways the traditional view expressed, for example, by the dictum ‘Tell the truth, though the heavens fall!’ The rule-worship accusation against any teleological system is particularly acute if the system is forced to adopt exceptionless rules. Rule with exceptions can permit the sort of ﬂexibility that would allow one to deviate from a moral rule in exceptional cases. But neither of the above conception of rules will work for the pragmatist. Rules of thumb presuppose act-based rationality; such rules merely summarize the rationality of a number of individual actions. (‘In 98% of cases, it is utility-maximizing to tell the truth; therefore, as a rule of thumb, you ought to tell the truth.’) The whole point of moving to strategies is that it allows sub-optimal actions as part of an optimal strategy. A rule of thumb presupposes that actions are evaluated individually, and that, generally, the action prescribed by the rule will produce the best results. On the other hand, exceptionless rules do not give us the ﬂexibility to deviate from the rule even in exceptional cases, thereby making the rule-worship objection more acute. What is needed, then, is a conception of rules according to which rules have genuine justiﬁcatory force—that is, they are not merely rules of thumb or ‘summary rules’—but which allow of exceptions. Can such a conception of rules be formulated? It can, and it has been. Let us turn to a discussion of (a) why some philosophers have been driven to the conclusion that the rules governing reasoning, both practical and theoretical, admit of exceptions; and (b) how philosophers have developed a viable conception of such rules.

84

Pragmatism and Normativity

Non-monotonic inference and defeasible principles Deductive reasoning is monotonic. That is, adding more to the premise side of an entailment can never get you less on the conclusion side. That is, again, if P entails R, then P and Q must also entail R. Some philosophers, however, have defended the idea that practical and theoretical reasoning are by and large non-monotonic: that is, if one can infer R from P, it does not follow that one can infer R from P and Q. Here are two examples of allegedly non-monotonic inferences: The match has been struck. Therefore, the match will light. It is raining. Therefore, I should open my umbrella. The ﬁrst is an example of theoretical reasoning and the second, of practical reasoning. Such inferences are thought to be non-monotonic, because the inference can be defeated by the addition of more premises: The match is wet. I want to do a Gene Kelly ‘Singing in the Rain’ imitation. One way of saying that such inferences are non-monotonic is to state that they are licensed by rules or principles of inference that have exceptions. Thus, ‘If it is raining, then I ought to open my umbrella’ is a legitimate principle, but one that admits of exceptions. The same goes, on the theoretical side, for principles like, ‘If I strike this match, it will light.’ However, many philosophers believe that all good reasoning— practical and theoretical—must be deductively valid and monotonic. Such philosophers (whom we will call ‘formalists’) will hold that the above inferences are in fact enthymemes—that is, they possess suppressed premises which are contradicted by the addition of a defeating premise. Thus, the inferences really take the following form: The match has been struck. (The match is dry and well-made) Therefore, the match will light. It is raining. (I want to remain dry) Therefore, I will open my umbrella.

Pragmatism and Rationality

85

The additional premises (‘The match is wet’, ‘I want to do a Gene Kelly “Singing in the Rain” imitation’) contradict the second premise, thereby rendering the inference unsound. The existence of the suppressed premise is, according to the formalist, often indicated by an expression like ‘ceteris paribus’ or ‘all things being equal.’ Thus If the match is lit, then, ceteris paribus, it will light. If it is raining, then, all things being equal, you should open your umbrella. Which interpretation of theoretical and practical reasoning should we prefer? Should we interpret practical and theoretical inferences as nonmonotonic, or as enthymematic? Robert Brandom argues that we should take the former route. With regard to such ‘ceteris paribus’ clauses, he writes: I do not want to claim that invoking such clauses . . . is incoherent or silly. But we must be careful how we understand the expressive role they play. For they cannot (I want to say, “in principle”) be cashed out; their content cannot be made explicit in the form of a series of additional premises. They are not shorthand for something we could say if we took the time or the trouble. The problem is not just that we would need an inﬁnite list of the conditions being ruled out—though that is true. It is that the membership of such a list would be indefinite: we do not know how to specify in advance what belongs on the list. If we try to solve this problem by a general characterization, we get something equivalent to: “ceteris paribus, q follows from p” means that “q follows from p unless there is some inﬁrming or interfering condition.” But this is just to say that q follows from p except in the cases where for some reason it doesn’t. I would contend that ceteris paribus clauses should be understood as explicitly marking the nonmonotonicity of an inference, rather than as a deus ex machina that magically removes its nonmonotonicity.51 Thus, many inferences are such that they can be switched back and forth, from entailing q to entailing ∼q, by the constant addition of new premises. Consider Brandom’s example: 1. If I strike this dry, well-made match, then it will light. (p → q) 2. If p and the match is in a very strong electromagnetic ﬁeld, then it will not light. (p & r → ∼q)

86

Pragmatism and Normativity

3. If p and r and the match is in a Faraday cage, then it will light. (p & r & s → q) 4. If p and r and s and the room is evacuated of oxygen, then it will not light. (p & r & s & t → ∼q)52 And so it goes. There is no way of specifying the suppressed premises. Further, we do not need to appeal to these suppressed premises in order to explain our ability to infer conclusions from premises. This point was well-argued by Lewis Carroll in his famous article ‘What the Tortoise Said to Achilles.’ Recall that in this article, Achilles is trying to explain inference to the Tortoise. The Tortoise presents the following argument: (A) Things that are equal to the same are equal to each other. (B) The two sides of this Triangle are things that are equal to the same. (Z) The two sides of this Triangle are equal to each other. Of course, Z follows from A and B. But the Tortoise wants to know why, if he accepts A and B, he must also accept Z. Achilles explains that Z follows from A and B. The Tortoise dutifully adds another premise: (C) If A and B are true, Z must be true. But when Achilles says that if the Tortoise accepts A and B and C, then he also must accept Z, the Tortoise balks: “And why must I?” “Because it follows logically from them. If A and B and C are true, Z must be true. You don’t dispute that, I imagine?” “If A and B and C are true, Z must be true,” the Tortoise thoughtfully repeated. “That’s another Hypothetical, isn’t it?”53 And so the Tortoise adds another premise: (D) If A and B and C are true, Z must be true. and another (E) If A and B and C and D are true, then Z must be true. and so on ad inﬁnitum. Carroll’s point is that if you cannot already infer, then adding a major premise will not help. Thus, not only can we not

Pragmatism and Rationality

87

state the suppressed premises in the above examples of practical and theoretical reasoning, but it is not obvious why we have to: our ability to make such inferences is not explained by appeal to the existence of such suppressed premises. Rather, our ability to understand and follow the major premise presupposes a prior ability to infer in the ﬁrst place. If we abandon the idea that practical and theoretical reasoning are guided by exceptionless principles, then what is the alternative? Some philosophers, rejecting such principles, have gone to the opposite extreme, rejecting the idea that principles should play any role in reasoning. On the side of practical reasoning, this view is held by the radical moral particularists. Such particularists hold that principles are at best useless and at worst actually harmful in moral reasoning. Radical particularists hold that non-natural properties can ‘switch valence’—that is, they can be a bad-making property in one context, and a goodmaking property in another. It is not merely that a non-moral feature (such as the fact that x causes her pain) always carries the same moral signiﬁcance, but is sometimes outweighed by other competing moral considerations. It is rather that the same non-moral feature can have one moral signiﬁcance in one case and the opposite (or no) moral signiﬁcance in another case. Consider the example of pleasure: in one case, the fact that taking your cousins to the circus causes them pleasure contributes to the moral goodness of your action. In another case, the pleasure a sadist derives from torturing an unwilling victim is precisely part of what makes the sadist’s action so wrong; it is not, writes Dancy, the moral ‘silver lining’ of the situation.54 Particularism widens the pragmatist’s choices. We are no longer forced to choose between rules of thumb and exceptionless principles; we now have a third option: particularism. But the particularist view has its own problems. As Brad Hooker argues, radical particularism makes it impossible for us to coordinate our actions with others. Consider Hooker’s example: suppose Patty the particularist promises that if you help her get her crops in this month, she will help you get your crops in next month. Hooker writes, Assume that the only thing that could possibly make her to keep her promise [to help you get your crops in] is her moral outlook . . . As a particularist, Patty thinks that there are no considerations that always retain their moral polarity. She thinks a consideration (such as the fact that she promised to do something) might be a reason for keeping her side of the deal in one situation, but a reason against keeping it in another situation. So, will she think that having made a promise

88

Pragmatism and Normativity

to you gives her any reason to do what she promised? Not necessarily. As a particularist, she can and might attach no weight whatsoever to the promise when the time comes to keep it. And, as a particularist, she can’t point to any general considerations that mark off the situations in which a promise would be morally binding from the situations in which it wouldn’t . . . If Patty would really live by such beliefs, how much could you trust her?55 Other philosophers worry that if radical particularism were true, then morality would be unlearnable. Principles can be learned, but if there is nothing general at all shared between different contexts and different moral situations, how could we possibly learn to apply moral predicates correctly?56 Further, the particularist has a hard time explaining why certain features (that something was painful, or a promise, or cruel) are relevant in their own right, whereas certain other features (that it involved broccoli, or her shoelaces were white, or it occurred in the month of June) are never signiﬁcant in their own right, but only derivatively, by virtue of their connection to the former sorts of features.57 For these reasons and others, radical particularism does not seem a promising route. Thus, the solution is not to ﬂee from exceptionless principles all the way into the arms of the radical particularist. The solution, as I have noted previously, is to attempt to discover some conception of principles according to which the principles are genuinely justiﬁcation-conferring, but allow of exceptions. Such a conception of defeasible rules has been developed by Mark Lance and Margaret Little. According to Lance and Little, a defeasible generalization is “a kind of generalization that is both genuinely explanatory and ineliminably exception-laden.”58 Lance and Little reject the idea that pain, for example, is only accidentally and contingently connected to moral badness. However, they agree with the particularist that such non-moral properties can ‘switch valence.’ Thus, they write, Pain is always bad-making—well, except when it’s constitutive of athletic challenge; intentionally telling a falsehood is prima facie wrong—well, not when done to Nazi guards, to whom the truth is not owed, or when playing the game Diplomacy, where it’s the point of the contest. Pleasure always counts in favor of a situation—well, except when it’s the sadist’s delight in her victim’s agony, where her pleasure is precisely part of what is wrong with the situation.59

Pragmatism and Rationality

89

On the face of it, these two requirements—that (say) pain is necessarily connected to moral badness, but can sometimes switch valence—are incompatible. But the point of Lance and Little’s account of defeasible generalizations is that it allows them to sketch such a necessary connection while making sense of the cases in which the valence changes. Let us examine this account in more detail. Generalizations like ‘Fish eggs turn into ﬁsh’ and ‘Causing pain usually contributes to the immorality of an action’ are true, other things being equal. ‘Other things being equal’ does not mean that these generalizations are true in the majority of cases, for this need not be true. For example, only a small minority of ﬁsh eggs survive to become ﬁsh. However, there is also a sense in which these generalizations are necessarily true. That is to say, such generalizations “in some way are meant to capture the nature of the object in question.”60 Thus, though few ﬁsh eggs ultimately turn into ﬁsh, there is a crucial privileging of cases in which they do. One classiﬁes the ﬁsh egg as being the kind of biological organism it is, by reference to its “standard” or “normal” development. There are of course an inﬁnite number of trajectories that ﬁsh eggs could take, from developing into ﬁsh, to being enucleated with sheep DNA and becoming a sheep, to breaking down into nutrients for a turtle, to being irradiated and turning into a strange and dysfunctional pile of ﬂesh. Nonetheless, we elevate one such trajectory as a “natural” one, viz., one that does not call for explanation (at least at this level of theory); and in this sense, we circumscribe some developments as expressions of an organism’s “nature.”61 Thus, such generalizations are necessary—not in the sense of being exceptionless, but in the sense of being particularly revealing of the nature of some object (e.g., a ﬁsh egg) or some property (e.g., that of causing pain). Here are Lance and Little: Sometimes . . . when we issue a generalization to the effect that something has a certain feature, what we really want to say is not that such a connection always, or even usually, holds, but that the conditions in which it does hold are particularly revealing of that item’s nature, or of the broader part of reality in which the item is known. We might put it by saying that we’re asserting what happens in “normal” conditions, except that the notion of ‘normalcy’ is so freighted with misleading connotations. Better put, then, we are taking as privileged,

90

Pragmatism and Normativity

in one way or another, cases in which the item has the feature speciﬁed. Such generalizations tell us about the nature of something, not by eliminating exceptions to the connection, but by maintaining and demarcating their status as exceptions.62 As these generalizations hold only relative to certain privileged conditions,63 the conditions in those cases where the generalizations do not hold must in some sense be deviant. Thus, lying is bad-making, in general, but lying to the Nazis is morally permissible. However, the very moral permissibility of lying in this case can only be explained by the fact that the situation is deviant and morally defective: the Nazis are looking to round up innocent victims to be murdered. This account of the necessity of defeasible generalizations makes sense of the cases in which these generalizations do not hold. Indeed, in many cases, it is only because of the truth of the generalization that we can make sense of deviant cases. As Lance and Little write, “In athletics for instance, it is only because pain is paradigmatically something to be avoided that the notion of physical challenge has the meaning, and the status of constitutive good, that it does. It is only because pain is normally bad-making, then, that we can understand its good-making instances.”64 Thus, again, the account of defeasible generalizations they give explains how such generalizations are necessary for allowing us to make sense of the cases in which these generalization do not apply. This case also illustrates the fact that ‘deviant’ does not necessarily mean ‘defective.’65 The Nazi case is one where the circumstances are morally defective; the pain-as-constitutive-of-athletic-challenge case is merely deviant, but not thereby defective. Thus, an account can be given of defeasible rules—rules that are genuinely justiﬁcation-conferring, but which admit of exceptions. To the extent that such a conception of rules exists, it further mitigates the rule-worship objection. The existence of other prior accounts of rules with exceptions (e.g., Ross’s prima facie duties) gives further support for the idea that the pragmatist need not admit that rule pragmatism counsels adoption of a set of rigid, exceptionless rules that must be obeyed (or worshiped!) in all circumstances.

Concluding remarks A traditional question in ethics has been, ‘Why be moral?’ Generally, this question has been understood as, ‘Why should I be moral?’ The answer to this question is, ‘Because you are one of us, and it is rational

Pragmatism and Rationality

91

for us to be moral.’ Similarly, if the footballer asks why he should go right, the only possible answer is, ‘Because you are part of the team, and it is rational for the team to go right.’ Many puzzles in morality have been created by the insistence that rationality is act-based and individualistic. But I have argued here that rationality is in many cases strategy-based and team-oriented. Morality is precisely one of these cases: it is a case where the desired beneﬁt can only be achieved by strategy-based team rationality. Ergo, this is the model of rationality that is rational to apply in the moral realm; applying another model (namely an act-based individual morality) would fail to achieve the desired result, and is therefore irrational—a failure to apply the correct rational model to a problem of means–end reasoning (namely, the problem of how best to advance our interests). An assumption sometimes made by philosophers is that if our selfinterest conﬂicts with our moral obligations, then it is rational to pursue our self-interest to the detriment of morality. I hope that this chapter has undermined this assumption. First, this assumption rests on the dogma of individual rationality. To assume that in such a conﬂict self-interest must win out is to ignore the force of cooperative rationality, and the rational requirements it makes on us. Second, as morality is a rational system aimed at promoting all our interests (selfregarding and other-regarding), it will often be rational to satisfy the requirements of morality, even if this means sacriﬁcing some measure of self-interest. And in any case, as we have other-regarding (as well as self-regarding) interests, a comprehensive account of rationality based in interest satisfaction will have to acknowledge that it might often be rational to satisfy our other-regarding interests to the detriment of our self-regarding interests.

4 Pragmatism, Normativity, and Relativism

So far, I have conceded (in Chapter 2) that normative utterances do not describe an independent moral reality. Instead, morality (and epistemology) consists of a set of cooperative strategies for promoting our mutual interests. I have also asserted, though, an aversion to the relativism that is typically associated with pragmatism. But aren’t these two points at odds with each other? If there is no independent normative reality, then can morality and epistemology be non-relative? I want to argue in this chapter that although normative utterances do not describe an independent normative reality, we are not forced into a relativistic picture, or some picture where normative truth consists merely in whatever a particular community practices. The argument of this chapter will proceed as follows: ﬁrst, we will argue (initially) that it is plausible to explain moral and epistemic truth in terms of excellence of reasons: a proposition p is true if we can give a vindicatory explanation of it, that is, an explanation that demonstrates that there is nothing else to think but p. Next, we will combine this conclusion with the idea that the types of reasons supporting moral and epistemic rules are timeless: their justiﬁcatory force does not rely on their time of utterance, but instead can apply ‘retroactively’ to times before their utterance (and can have force even if never uttered at all). From these two conclusions, it will follow that the truth of normative rules is not relative to a particular time. We will then turn brieﬂy to the issue of cross-cultural relativism.

Moral truth We have conceded that moral and epistemic utterances do not describe an independent normative reality. How, then, can a normative claim be objectively true? In other words, what makes a perceived normative 92

Pragmatism, Normativity, and Relativism

93

requirement correct? Let us begin by discussing how a normative claim can be true; later, we can discuss how it can be objectively (i.e., nonrelatively) true. In replying to Cheryl Misak’s attempt to rescue truth by blurring the fact–value distinction (a typical pragmatist move, as I noted in Chapter 2), David Wiggins writes: If someone constantly emphasizes certain features of the empirical and faults other subject matters for lacking them, then either we can try to show that, on a proper understanding, those other subject matters do not lack these features or else we can show that truthdirectedness does not in itself require them . . . Misak’s response to those who fault the judgments of morality for lacking reference to experience conforms to the ﬁrst of these patterns.1 But Wiggins suggests that pursuing the second pattern might be more fruitful. And indeed, we have already committed ourselves to the second pattern. In Chapter 2, we conceded a fact–value distinction, and suggested that what it takes for a moral or epistemic sentence to be true is different from what it takes for a purely descriptive sentence to be true. However, we can give an account of what makes normative utterances true, which shows that ‘truth-directedness does not in itself require’ the features that make descriptive discourse truth-apt. As I will argue, Wiggins’s2 and McDowell’s3 discussions of moral truth lend themselves well to an account of the truth of moral utterances. Wiggins and McDowell have suggested that we can make sense of moral truth in terms of excellence of reasons. To support this picture, Wiggins draws an analogy with mathematics. He writes that such reasoning could explain the great degree of convergence displayed by people on the claim that 7 + 5 = 12. This belief (that 7 + 5 = 12) “resembles an ordinary empirical belief in being uncontroversially true . . . but resembles a moral belief in not being empirical.”4 According to Wiggins, the best explanation for why people believe this is that it can be shown by the calculating rules to be true. “There is nothing else to think but that 7 + 5 = 12.”5 Thus, even if we cannot posit mathematical facts or entities that stand in a causal relation to us (and are thereby capable of satisfying Harman’s explanatory requirement),6 we can make sense of mathematical truths in terms of excellence of reasons. The reasons for thinking that 7 + 5 = 12 are so compelling that there is nothing else to think. The best reasons support the conclusion that 7 + 5 = 12.

94

Pragmatism and Normativity

Wiggins and McDowell seem to think that their respective accounts of truth in terms of excellence of reasons can be extended to every area of discourse, from moral to scientiﬁc. Wiggins writes: A subject matter is objective or relates to objective reality if and only if there are questions about it that admit of answers that are substantially true. It is sufﬁcient for some judgment that p to be substantially true that one could come to know that p. One can come to know that p only if one can come to believe that p precisely because p. And one comes to believe p precisely because p only if the best full explanation of one’s coming to believe that p requires the giver of the explanation to adduce in his explanation the very fact that p. What follows from this is that his explanation will conform to the following schema: for this, that and the other reasons (here the explainer speciﬁes these), there is really nothing else to think but that p; so, given the circumstances and given the subject’s cognitive capacities and opportunities and given his access to what leaves nothing else to think but that p, no wonder he believes that p. Let us call such an explanation of a belief a vindicatory explanation of that belief.7 Wiggins draws parallels between his and Peirce’s respective views of truth. Peirce writes: Now, there are some people, among whom I must suppose that my reader is to be found, who, when they see that any belief of theirs is determined by any circumstances extraneous to the facts, will from that moment not merely admit in words that that belief is doubtful, but will experience a real doubt of it, so that it ceases to be a belief.8 Peirce continues, “To satisfy our doubts, therefore, it is necessary that a method should be found by which our beliefs may be caused by nothing human, but by some external permanency.”9 Now, it is easy to see how one’s belief that p can be caused by circumstances not extraneous to p, if we are talking about an ordinary theory of a perceptually caused belief. My belief that the chair is gray is caused, in the ﬁnal analysis, in part by the fact that the chair is gray. But what of non-causal knowledge, such as mathematical, or, more to the point of this chapter, moral or epistemic knowledge? We have already conceded, in Chapter 2, that moral and epistemic facts are not causally efﬁcacious; so our account of normative facts—and our epistemology of normative facts—must respect this

Pragmatism, Normativity, and Relativism

95

recognition. That is not accomplished by requiring that a belief that p be caused by circumstances not extraneous to the fact that p. However, Wiggins notes, “In the course of one of his rereading of ‘Fixation’, Peirce made an annotation against [these] words . . . Peirce’s annotation requires the word ‘caused’ to be replaced by the word ‘determined’.”10 In Wiggins’s view, this was owing to Peirce’s recognition that the earlier causal requirement on knowledge made it difﬁcult to explain certain types of knowledge, such as mathematical and moral knowledge. Given this new gloss on the permissible etiology of a belief, Wiggins writes that one should read Peirce as saying “that any opinion or belief [inquirers] arrive at to the effect that p should be determined by circumstances that are not extraneous to the fact that p.”11 Here is Wiggins: That [perceptual case] is the familiar case. But now suppose that the initiator of belief was not perception but some “elaborative process of thought,”12 one leading into a gradual accumulation of reasons that culminated at the moment of secondness in the thinker’s ﬁnding nothing else to think but that ( . . . ). Here the thing which brought the thinker to the point of conviction was not just any causal effectiveness. Still less was the thinker’s ﬁnding that there was nothing else to think but that ( . . . ) the outcome of some reality’s/Real’s causally effective agency. Rather, the reason why the thinker was unable to ﬁnd anything else to think but that ( . . . ) was that there is nothing else to think. If there is nothing else to think, no wonder the thinker thought that!13 Thus, if (to give Wiggins’s example) Peter believes that 7 + 5 = 12, and he believes this on the basis of the rules of calculating, then “Peter’s belief that 7 + 5 = 12 is determined (as Peirce requires) by a circumstance not extraneous to the fact that 7 + 5 = 12.”14 The view that we can make sense of moral truth in terms of excellence of reasons seems quite plausible. (For ease of phrasing, I will conﬁne my discussion now to morality, although the lessons apply equally to epistemology.) Consider an analogy from chess. A certain arrangement, A, of pieces on the board constitutes checkmate. But there is nothing intrinsic to these pieces of wood on the board that should, in this particular arrangement, constitute a threat to a particular piece of wood. Rather, A constitutes checkmate by virtue of the rules of chess. Crudely put, the rules make it the case that A constitutes checkmate (indeed, the term ‘checkmate’ has no meaning in itself in abstraction from the rules of

96

Pragmatism and Normativity

chess). Morality is in an important sense analogous to games like chess. There is nothing intrinsic to a particular behavior, B, by a featherless biped that makes this behavior evil; rather, it is evil by virtue of the rules of morality. Now, there are those (i.e., particularists) who deny that morality can be codiﬁed into a set of rules, so the phrase ‘rules of morality’ will rub these people the wrong way. Let me, then, use the term ‘reasons’ more broadly, to apply to non-codiﬁable reasons or codiﬁed rules. This is not an abuse of the term ‘reason’ because rules can serve as reasons: for the particularist, judgments are backed by reasons (Why was that action cruel? Because it caused her gratuitous pain). These reasons play the same role for the particularist that rules play for the philosopher who thinks morality can be codiﬁed (Why was that action cruel? Because causing gratuitous pain is cruel). That is to say, both reasons and rules justify moral judgments. More important, though, just as the rules of chess make A a checkmate, the reasons (read: rules or reasons) of morality make B cruel; there is nothing intrinsic to B that makes it cruel. It is only so by virtue of our moral practice, and the rules that constitute this practice. (Of course, we noted in the last chapter that radical particularism is not a terribly plausible view; and so it is likely that rules must play an important role in morality.) There are reasons to think that the account of truth offered by Wiggins and McDowell is not entirely satisfactory as it stands. First, both seem to think that what this vindicatory explanation must be explaining is convergence in a particular belief. But I do not think that realism requires the possibility of rational convergence. For example, few (except for a few veriﬁcationists) would deny that there is a fact of the matter regarding whether the number of neutrinos in the universe is even or odd. There is, however, not even the slightest possibility of rational convergence on this matter. Consider another example, the venerable ‘brain in a vat.’ The brain (which we shall call ‘X’) and I will never converge on the claim, ‘X is a brain in a vat.’ Even if I were somehow able periodically to hook myself up to the machine and enter the brain’s extended hallucination as a character and argue extensively with the brain, the brain would remain unconvinced, and with considerable justiﬁcation. It would regard me as I would regard some crank who kept approaching me on the street and arguing that I was a brain in a vat. Second, nor can we rescue this theory as a general account of truth by weakening it to read that what needs to be explained by the vindicatory explanation is not convergence on the belief, but instead individual instances of true belief: my (true) belief that X is a brain in a vat is best explained by the fact that X is, in fact, a brain in a vat. The weakened

Pragmatism, Normativity, and Relativism

97

version would hold that if we can explain how I came to hold p in a way that requires p to be true, then we can be realists about p. But even this weakened version (that an individual might come to believe p precisely because p) is too strong. This fact is best illustrated by the above neutrino example: it might be impossible to come to have a belief about the number of neutrinos in the universe in a way best explained by this being the correct number. Yet there is an objective fact of the matter as to whether the number of neutrinos in the universe is even or odd. Thus, the criterion for realism that Wiggins and McDowell suggest cannot work in all cases.15 However, I only wish to draw a single insight from their approach: we can make sense of moral truth in terms of excellence of reasons. Of course, this qualiﬁcation also rules out unknowable moral truths, which is an acceptable restriction to place on our account of moral truth. As I argued in the previous chapter, moral rules must be learnable; this requirement rules out any moral facts that are in principle unknowable. Such a fact would be pragmatically useless as a guide for promoting our mutual interest, and so morality would contain no such unknowable facts (although, as we will see in this chapter, there can be unknown moral facts). So what reasons determine the truth of a moral utterance, on this pragmatist account? We must keep in mind the two-level account of morality defended in Chapters 2 and 3. Pragmatic reasons justify moral rules; moral rules justify individual actions or claims about individual actions. So a general rule (e.g., ‘Cruelty is wrong’) is justiﬁed by pragmatic reasons; we appeal directly to our various interests to justify rules such as this. Cruelty is wrong because it is in our mutual interest to forbid cruelty. An individual moral claim (e.g., ‘It was wrong to make fun of that person’) is justiﬁed by appeal to the general rule. Setting the cat on ﬁre was wrong because it was cruel. Thus we can see the reasons that undergird various moral claims, general or particular. However, it seems that this account of moral truth leads to a problem. In mathematics, reasons support a unique truth. That is, given the rules of mathematics, there is no other possible solution one could get to the problem ‘7 + 5’ except 12. Reason determines a unique answer to each mathematical question.16 But is the same true of morality? Do pragmatic reasons determine a unique set of cooperative strategies for promoting our mutual interests? On the contrary, it seems as though our interests may well under-determine moral truth. That is, there might well be multiple, incompatible rational strategies for promoting our mutual ends. In part, this is because of the very general terms in which interests must be stated. But also, even if these interests were identiﬁed in a very

98

Pragmatism and Normativity

speciﬁc manner, it still would not follow that there is a unique best set of cooperative strategies for promoting them. I will not give an account of human interests until the next chapter, but let me make a few comments. Any account of the basic ends of humans—whether these are the goods of the natural-law tradition, or the capabilities of Nussbaum and Sen—will state these ends or goods in very general terms. John Finnis, for example, lists the following basic human goods: life, knowledge, play, aesthetic experience, sociability (friendship), practical reasonableness, and religion.17 Nussbaum lists out basic human capabilities in equally general terms: 1. Life: Being able to live to the end of a human life of normal length; not dying prematurely, or before one’s life is so reduced as to be not worth living. 2. Bodily Health: Being able to have good health, including reproductive health; to be adequately nourished; to have adequate shelter. ... 9. Play: Being able to laugh, to play, to enjoy recreational activities.18 The list of interests I give will be similarly general.19 Not only are they general, but there are many ways of achieving many of these goods or interests. Thus, the good of knowledge might be achieved by attending continuing-education classes, or reading a magazine devoted to a particular hobby or some other topic of interest, or in any number of other ways. Similar comments apply to the other interests and goods. It seems clear, then, that such general lists of interests and goods as these will underdetermine morality. Given the general nature of our interests, it seems that there will not be a unique set of cooperative strategies that will successfully promote our interests. We should be careful not to overstate the degree of underdetermination. Although our interests underdetermine morality to some extent, the underdetermination is not radical. Clearly, a system that enslaved one-third of the population would not be a satisfactory system for promoting our mutual interests, nor would be a system that ignored the interests of half of the society (say, the interests of women). The mere fact that morality aims to promote our mutual interests supports a principle of equal consideration; and, as I will argue in the next chapter, appeal to our interests and their various manifestations supports other principles such as respect for autonomy. So let us be careful not to overstate the degree of underdetermination that exists in morality.

Pragmatism, Normativity, and Relativism

99

Nevertheless, if we wish to say that morality is determined by our interests, then we must conclude that there is a degree to which morality is underdetermined. This conclusion seems to threaten us with relativism. It seems as though a non-relativistic pragmatism would require the conclusion that our interests determine a unique cooperative strategy, a requirement that seems unlikely to be fulﬁlled. We will discuss this underdetermination worry in this chapter, and return to it in the next chapter. In the meantime, let us begin to examine how we can give a non-relativistic account of normative ascriptions. In what follows, I will often talk about moral truth. This is done for the sake of simplicity of phrasing. The lessons to be learned from the following should be easily transferable to other sorts of normative discourse (in particular, epistemology, which is, of course, the other focus of this book).

Underdetermination, complexity, and social practice I have already conceded that our interests underdetermine a unique optimal cooperative strategy in the case of morality. This would seem to raise the threat of relativism. However, by endorsing a social-practice account of the justiﬁcation of moral norms, we can alleviate this relativistic worry. Let us examine the details of such an account, and see how it can be pressed into the service of a non-relativistic pragmatism. First we should make some brief remarks about social practices themselves. I think we all have a rough, intuitive idea of what a social practice is. It is, for example, what anthropologists posit to explain certain regularities in the behavior of a community. Here is one example: “[C]onsider the practice in basketball of not stepping outside the sidelines when one is in possession of the ball. It is clearly useful as part of a systematic account of the ‘dance of the basketball players’ to take them to be committed to following such a rule, even though they do not always follow it, nor does any penalty for non-compliance universally follow.”20 We attribute commitments such as these to communities whose behavior we seek to explain. It is the fact that a bit of behavior (staying inside the lines, stopping at stop signs) is caught up in such a web of commitments that makes possible its classiﬁcation as an action (and its performer as an agent), as opposed to mere behavior.21 In interpreting a community in terms of social practices, one takes the members of the community to be committed to and bound by certain appropriatenesses. The normativity is only shallow, however, as positing such practices does not involve committing oneself to following

100

Pragmatism and Normativity

them.22 Again, though, as a member of a social practice, one is prima facie committed to the norms embodied in this practice, unless one can demonstrate (through what Brandom calls ‘the game of giving and asking for reasons’) that the norm should be eliminated or replaced. As I argued in the previous chapter, if one has already acknowledged that he is ‘one of us,’ then he is by his own lights in the space of moral reasons, and he owes us an explanation for why he wishes to deviate from a particular cooperative strategy, if he does. Otherwise, he is bound by it. A social-practice account has certain features that make it well-suited to the task at hand. By contrasting this account with other theories of justiﬁcation, we can better understand the social-practice theory and its advantages. A social-practice account recognizes the fact that when one is trained into a set of beliefs (as a child, or as a novice scientist or philosopher, etc.), one is given a starting point that is (from the perspective of the child or novice scientist or philosopher) entirely arbitrary.23 Now, a coherentist would say that the elements of this system are justiﬁed to the extent that they cohere with the overall belief system and the degree to which the system itself is coherent. But such a view ignores the complexity of our belief system: we are trained into a system whose coherence we cannot evaluate all at once. The coherence of a belief system cannot simply be seen by casting an appraising eye over the whole of one’s belief system; it is sheer fantasy to suppose that we have such mastery over our belief system. Coherence can only be achieved slowly and piecemeal. Indeed, incoherencies in one’s system can only be discovered slowly and piecemeal. Thus, the agent is forced to take her belief system as being in order as it is, and revise it as incoherencies become manifest. This fact should not encourage complacency, nor should it encourage excessive conservatism. It should compel us to acknowledge, however, that we must accept an arbitrary starting place, and then use that starting place as a platform for revising our theory, always using (subject to their revision) the tools that are initially given to us by this starting place. Thus, the social-practice account is characterized by two main features, conservatism and revision: the initial elements of the belief system must be accepted as prima facie in order, subject to their revision in light of apparent inconsistencies, explanatory deﬁciencies, and so forth. We will return to this account of justiﬁcation in Chapters 7 and 8, when we discuss the pragmatist account of epistemology; there I will argue for (among other things) the legitimacy and unavoidability of this brand of conservatism. Thus, we might say that morality displays justiﬁcatory dependence on our social practices in that the social practice forms the

Pragmatism, Normativity, and Relativism

101

starting point for our moral inquiry, and confers prima facie justiﬁcation on certain basic moral claims. A social-practice account of justiﬁcation works best for the account of morality and epistemology we are presenting here. There are two features of morality and epistemology that make this so: the ﬁrst is the fact that our interests underdetermine a unique cooperative strategy. That is, presumably multiple incompatible cooperative arrangements might satisfy our interests equally well.24 This is true in epistemology as well as in morality: different institutions and norms might equally well serve our interests in truth, explanation, and instrumental control over nature. The second feature is what we might call the complexity of morality. This is the same problem that faces coherentism in epistemology (about which I will say more in Chapter 7): just as we cannot simply see the coherence (or lack thereof) of our belief system, we cannot simply read off the entire moral code from our list of interests. That is, morality is too complex to simply be derived in its entirety from our interests as one would derive theorems from a set of axioms. Not only is morality too complex, but the derivation of a cooperative strategy cannot be done in one’s armchair. The choice of a cooperative strategy has real-world consequences (and hence, unforeseen, unintended, and unpredictable consequences). Thus, these strategies must be lived; only this way can problems with these strategies be discovered. To consider an example, some might argue that, in principle, communism seems like a satisfactory principle of justice; but that in practice it does not work (or only works in very artiﬁcial and circumscribed circumstances, such as kibbutzim, with their strict interview and admissions processes, etc.). Thus, though a strategy might in the abstract seem like it will satisfy our interests in a satisfactory manner, only by living the strategy can we fully realize the consequences of implementing the strategy. Again, this indicates that we cannot simply derive the correct set of moral strategies from our interests. The selection of strategies must be an ongoing process, as strategies are attempted, reﬁned, and (in some cases) ultimately rejected.25 Similar comments apply to epistemology: choice of epistemic methods is bound up with the progress of science, and which epistemic methods are legitimate (and which ones are not) cannot be decided from one’s armchair, but must instead be discovered over time. Correct epistemic methods must be uncovered as inquiry progresses; only investigation will tell us which methods are epistemically sound and which ones are not. Epistemic methods evolve over time, and the process of discovery of optimal methods of inquiry is one that takes place over

102

Pragmatism and Normativity

generations, with later investigators building on the knowledge of earlier ones. Consider, for example, measures taken to mitigate the effects of observer bias in clinical trials. Although the double-blind randomized trial is now taken for granted, it is a fairly recent innovation in clinical medicine, as a study of the history of blind medical assessment reveals.26 For a long time, doctors were content to rely on anecdotal evidence regarding the efﬁcacy of various remedies. James Lind’s experiments in 1747 on the use of lemon juice to prevent scurvy are recognized as one of the ﬁrst trials to use a control group of patients receiving an alternative treatment. But the ﬁrst double-blind randomized trial—where patients are randomly assigned to an experimental and a control group, and neither physician nor patient knows to which group the patient is assigned—did not take place until 1948, when Austin Bradford Hill used this technique to test the efﬁcacy of streptomycin in the treatment of tuberculosis. Thus, though such epistemically sound techniques may seem obvious to us, they are only obvious because they have become such an accepted part of experimental practice. It took centuries for the epistemic methods of clinical medicine to develop to their current advanced state. Thus, we must view the discovery of norms—the moral norms that best serve our interests, and the epistemic norms that best serve our epistemic ends (including knowledge, which is a fundamental human interest)—as a process which is extended over time (indeed, over centuries). These two features—underdetermination and complexity—support the adoption of a social-practice account of justiﬁcation in morality and epistemology. Let us begin by looking at the issue of complexity. (For the sake of simplicity, I may revert to discussing only morality; but again, these lessons apply equally well to epistemology.) As I argued above, the whole of morality cannot simply be derived at once from our interests. Thus, a moral system must be developed over time. We enter this development process mid-stream: we enter a society where the moral system has been under development for some time. We take this starting point and move on from there, revising the moral system generation after generation. Owing to the complexity of morality, we must develop a moral system piecemeal, trying various moral rules, seeing how they play out in practice, and rejecting those rules that do not serve our interests. Consider an example: in the United States, it was at one time considered (ceteris paribus) immoral for a woman to hold a full-time job, as by doing so she took the job away from a man, whose obligation it was to provide for his family. The thought was that a woman should instead get married and have her husband support her. Very few people

Pragmatism, Normativity, and Relativism

103

now would endorse such a moral rule; we have realized that women’s interests deserve equal consideration and a woman may wish to seek fulﬁllment through a career outside of the home. We have further (almost) realized that the man in the family need not be the primary breadwinner. (I foresee a time when it is accepted that a family need not consist of a man and a woman, even. In time, the previous sentence may even come to appear quaint.) In this way, moral norms are accepted; but we can see that this acceptance is provisional: every moral rule is subject to revision, as we work out which rules work effectively to promote our interests, and which rules do not work and should be replaced. The connection with a social-practice account of justiﬁcation should be clear: the complexity of morality forces us to accept provisionally a moral code, which is the product of generations of moral inquiry, and take this moral code as a starting point to be revised and reﬁned. Only over time can we hope to develop a moral system that is fully responsive to the entire range of human interests. The moral code as it stands is a step on the way toward this goal: it is the starting point for our process of revision. Thus, we can see the social-practice elements of conservatism and revision as a solution to the problem of complexity: the current moral code must be accepted as prima facie in order (this is the conservative element), as a starting point; but it is also viewed as the starting point for revision and reﬁnement. In this way the social-practice account of justiﬁcation provides a solution to the problem of moral complexity. This social-practice account also helps with the problem of underdetermination. Although a complete solution must wait until later in the chapter, we can begin here. As indicated, our interests underdetermine a unique optimal set of moral rules. But the social-practice account counsels us to treat the moral code currently embodied in our social practice as the starting point for future deliberations. On the conservatism that forms a part of the social-practice account of justiﬁcation, we must treat the social practice as in order, as is, subject to future emendations of this practice. Thus, the social practice helps determine morality by setting a starting point—a starting point which may in some sense be arbitrary, but which provides us with grounds for preferring one strategy for satisfying our interests over another. Consider the matter this way. When we consider our interests in isolation, alone, our interests do not give us a reason for preferring strategy A over strategy B, if both of these strategies would equally well promote our interests. But we do not have to consider our interests in isolation— we must consider our interests in the context of our current social practice. But decisions have already been made as to which direction

104

Pragmatism and Normativity

we ought to take in promoting our interests, and these decisions are embodied in the moral code that forms part of our social practice. Thus, although strategies A and B might be equally good, considered in the abstract, our social practice might already be moving in the direction of A, which gives prima facie justiﬁcation for carrying on in this way and continuing to endorse A as the correct strategy for promoting our interests. Further, our strategies might be to a large degree interdependent. That is, strategy A (currently embodied in our social practice) might be interdependent on other strategies that promote our interests. Hence, we could not simply switch to strategy B without doing damage to other strategies that we are pursuing to promote our interests, and which display interdependence with strategy A. Again, then, there is some reason for favoring the strategy currently in place. Thus the conservatism which forms part of the social-practice account of justiﬁcation helps solve the underdetermination problem. Let us sum up our progress. The problems of underdetermination and complexity lead us to a social-practice account of justiﬁcation. This account is characterized by conservatism: the elements of social practice are prima facie justiﬁed by virtue of belonging to the practice. The burden of proof is thus on those individuals who wish to revise the practice. Thus, a moral norm is justiﬁed if it is part of the current practice, unless one can demonstrate (using, subject to their revision, the tools made available by the practice) that another norm is better-justiﬁed, and would serve our interests better than the existing norm.27 Thus, the socialpractice account involves fallibilism as an essential element, too: the social practice provides prima facie justiﬁcation for the moral elements in it, but these elements are always subject to revision. We can always decide that we were wrong in endorsing a particular moral principle, and that there is a better cooperative strategy for promoting our interests. Moral norms implicit in our social practice are cooperative strategies that have evolved over time. We need not say that we made a conscious decision to follow these particular cooperative strategies, any more than we need to say that we consciously decided to use particular nouns (‘boat,’ ‘tree,’ etc.) to refer to particular objects. Rather, cooperation has evolved over time. Now, I have argued above that we must take the current cooperative strategies (the current state of the evolution of our social practice) as our starting point; but the pragmatist should not be committed to thinking that merely because a particular cooperative strategy has evolved over time, it is the best one possible. (For example, compare the various attempts in history to simplify English spelling: clearly, these attempts were motivated by the thought that the

Pragmatism, Normativity, and Relativism

105

English spelling system that had evolved over time was cumbersome and inefﬁcient.) Thus, though the pragmatist takes the current stage in the evolution of our moral development as a starting place for moral reasoning, current practice is always subject to revision, as we have seen. We will talk more about the revision of our social practice as the chapter progresses. One might worry about our conclusion that morality has some dependence (both epistemic and practical) on the social practice. After all, if the connection is too tight, then we are threatened by the very relativism this pragmatist account aimed to avoid: if moral truth just amounts to whatever the social practice endorses, then morality is relative. Fortunately, we need not concede that the connection between morality and social practice is tight enough to result in this sort of relativism. We will see that the revisability of social practice and what I will call the ‘timelessness’ of reasons prevent the connection from being too tight.

Moral truth and the temporal logic of reasons We have seen that on a social-practice account, moral reasons ultimately have their ground in our interests, but underdetermination and complexity force us to take as our starting place in moral reasoning the current state of the social practice (i.e., morality has justiﬁcatory dependence on our social practice). Further, we have seen how Wiggins and McDowell explain moral truth in terms of excellence of reasons. If we combine these two conclusions (our view of the nature of moral reasons, and the idea that excellence of reasons constitutes moral truth), the resulting view of normative discourse has a distinct advantage: it allows us to reconcile two theses that are, on their face, difﬁcult to reconcile. The ﬁrst claim is that morality and epistemology are creations of humans. The second claim is that the truth of a normative claim (such as ‘That was immoral’) can often be entirely independent of anything people say or believe about this claim. Let us see how, given this account of normative discourse, we can reconcile these two claims. Social dependence (i.e., dependence on a social practice) can be strong or weak. If ‘which performance is correct’ is merely a matter of ‘which performance most people take to be correct,’ then that correctness is strongly socially dependent. Consider, for example, the town in Virginia named Buena Vista. Contrary to what anyone acquainted with the Spanish language might think, the name of this town is pronounced ‘Byoona Vissta.’ What is more, this is the correct way to pronounce

106

Pragmatism and Normativity

the name of this town. It would be wrong to pronounce the name of this town as a Spanish-speaker would. Or, to take a more intuitive case, consider the state of Illinois. According to my dictionary, this is a word of French origin. But it would be incorrect to pronounce ‘Illinois’ as a French person would (‘Ill-een-wah,’ or some such). In both cases, the pronunciation of the word in question is correct by virtue of what people take the correct pronunciation to be. If Virginians pronounced ‘Buena Vista’ differently from how they do now, or if Americans pronounced ‘Illinois’ differently from how they do now, then that pronunciation would be the correct one. In short, if most people in the relevant community take a certain pronunciation to be correct, then that pronunciation is correct.28 There are performances whose correctness does not depend on what most people take to be correct. The truth of some propositions is not a question of what most people believe about these propositions. The paradigmatic cases of such propositions are scientiﬁc ones. For example, water would be identical to H2 O even if no one believed that it was. Whether water is identical to H2 O is a socially independent fact. However, it is difﬁcult to understand how social dependence can be anything other than strong social dependence. I wish to argue, though, that morality and epistemology are only weakly socially dependent. Their dependence on the social practice is justiﬁcatory only, and this connection does not support a claim of strong social dependence. That is, both are socially dependent but most, or even all, of us could be wrong about the truth of a given moral or epistemological claim. For ease of phrasing, I will discuss this weak social dependence in terms of morality, but it will be easy to apply our conclusions to epistemology and other sorts of normative discourse as well. An important fact to note, if we are to head off worries of relativism, is that though rules are conventional, they are not arbitrary. (This is especially clear in morality, where the moral rules are justiﬁed by our interests, and in epistemology, where the rules are justiﬁed by the goals of inquiry.) There are good and bad rules, and we might think of compelling reasons to change some of the rules of the game. For example, the National Basketball Association used to allow goaltending. But it became too easy for players such as Wilt Chamberlin to stand under the basket and block shots, so it was decided that goaltending should be disallowed.29 It was determined that the ends of the game would be better served by disallowing goaltending than by allowing it. Of course, certain elements of basketball practice are arbitrary: the decision to have as the point of the game the throwing of a ball through

Pragmatism, Normativity, and Relativism

107

a horizontal hoop was arbitrary. But within the context given by a certain set of arbitrary rules, there are better and worse ways to structure the game. Indeed, one way of looking at the process of creating rules is as a process of codifying reasons. If we make the basket narrower than two basketball-widths, the game will be too hard, but it will be too easy if we make it wider, so we will make the basket to have a width of two basketballs. Also, if we allow people to carry the ball, that will favor the offense too much, so we will require them to dribble. If we allow charging, that will make the game too rough and the incidence of injury too high, so we will disallow that. The examples can be multiplied indeﬁnitely; given the minimal framework of arbitrary rules that provide initial structure to the game, the rules of basketball have rationales behind them. The same is true of morality: the decision to drive on the right side of the road (in the United States) was arbitrary, but within the framework given by such arbitrary rules, it is clearly immoral to drive on the wrong side of the road, as such activity needlessly endangers others. Thus, certain aspects of social practice are arbitrary, but they form the framework in which objective rules can operate. Indeed, it might be instructive (even if not strictly accurate) to view morality as an objective function from such arbitrary rules and contingent features of society and the makeup of agents to a set of prescribed and proscribed behaviors. The function is determined by our interests—our interest in not dying provides the connection between the custom of driving on the right side of the road and the moral rule against driving on the left side of the road. Notice, though, that even these arbitrary features can be criticized on moral grounds. As we will see below, we ought to regard no feature of our practice as de jure unrevisable. There is an important disanalogy between morality and games such as basketball. In basketball, all of Wilt Chamberlin’s goaltending was permissible before the rule revision. Since the revision, though, it is a foul. But it would be unfortunate if we had to regard morality in this way. This would have the result that slavery was permissible in the United States before the Civil War, and impermissible afterward (think of the Civil War as a big rule revision). But we need not draw this unfortunate conclusion: even if moral truth depends on reasons, this does not directly entail that moral truth is historically local. To make my case that basketball and morality are disanalogous in this respect, I must ﬁrst discuss a feature of reasons: reasons display a sort of ‘timelessness.’ Let us look at how the practice of reason-giving works. Suppose we are trying to ﬁgure out whether to do A (stay home) or whether to do B (go to the beach), and we think we ought to do B. Assume there is a

108

Pragmatism and Normativity

decisive reason R (the weather service is saying that a massive storm is moving into the area within the hour) that supports performing action A. R can be a particularist reason, such as ‘That caused her gratuitous pain,’ or a rule such as ‘In battle, you should never leave your ﬂank unprotected.’ If someone utters R, thereby convincing us that we really ought to do A, then we are not justiﬁed in thinking, ‘B was the best course of action until his utterance made A the best course of action.’ Rather, we are justiﬁed in thinking that A was best all along; it merely took his presentation of R to make us realize it. Suppose nobody thinks of R at the time, and we go on to do B. Even centuries later, someone might truly utter, ‘They ought to have done A,’ and present R to support this conclusion. (It is crucial to distinguish between whether a reason obtained at time t and whether agents at time t can be blamed for failure to recognize this reason. In other words, we must keep in mind the distinction between ‘morally wrong’ and ‘blameworthy.’) So we can now say both that Galileo’s observations provided good reason to think the earth orbits the sun (even though few at the time regarded his observations as such); and we can also give good reasons for thinking that the Church’s treatment of Galileo was unjustiﬁed. In both cases, we are applying reasons ‘retroactively,’ scientiﬁc reasons in the one case, and moral and epistemic reasons in the other. But in an important sense, we are not really applying them retroactively; rather, we are presenting reasons that obtained at the time of Galileo, but which were not widely recognized. As Mark Lance writes, “A proposed rule has normative authority because we can show it—within a practice governed by norms of rational argumentation—to be a good one to impose. It has normative authority over us—and already had it over our past uses . . . because we already had allegiance to a practice within which, as we now know by having done it, it is possible to defend such a rule.”30 Thus, even if at an earlier time a particular moral rule was underdetermined, we know now that it is possible to defend such a rule within the practice, because the pragmatic reasons we appeal to in justifying a given moral principle applied even before they were appealed to in justifying the moral principle in question. So reasons are in an important sense timeless; their truth and their normative binding-ness do not depend on the time of their utterance. If at time t it is true that ‘S ought to do A at time t,’ then at time t + 100 years it is true that ‘S ought to have done A at time t.’ Indeed, we still have a decisive reason to do A even if no one ever thinks of R. B would have been a bad course of action even if no one ever thought of R. Reasons, then, can be decisive even if they remain un-uttered. I will discuss in a moment how

Pragmatism, Normativity, and Relativism

109

this conclusion regarding the timelessness of reasons is consistent with my above comments regarding the temporal relativity of the rules of basketball as displayed by the earlier goaltending example. As I argued above, morality has some dependence (namely, justiﬁcatory dependence) on our social practice. But moral truth also has autonomy from social practice in that decisive reasons remain decisive even when they go un-uttered or un-thought. People who do not think of the best reason, and go on to perform an inferior action, are mistaken, and the action they pursue is inferior to the one recommended by the un-uttered decisive reason. This ‘timelessness of reasons’ will allow us to conclude, in a moment, that moral truths are weakly socially dependent. Their truth is not merely a matter of what most people take their truth to be, although morality is a socially instituted practice. In this instance, socially dependent does not entail relative or historicist. For, as we have seen, reasons have an autonomy from our practice of reasongiving, and it is this autonomy which makes the social dependence of reasons weak rather than strong. We now have two important results. First, moral rules are licensed by pragmatic reasons. Second, reasons display a sort of timelessness: a reason can be decisive before it is uttered, or hours, days, or centuries after the decision is made, or even if never uttered (or even thought of) at all. When we combine these two conclusions we get an interesting result. Although morality is a creation of humans, it need not be conceived of as local. Even in societies where slavery is accepted, slavery is still wrong; there are still decisive reasons to abolish slavery (even if no one, not even the slaves themselves, think of these reasons). There might now be a decisive reason for or against legally permitting abortion; there might be a decisive reason why we should or should not allow gay marriages; or why we should all be vegetarians. These reasons might exist, even though we do not know it. Thus, moral truth, though possessing some important ties to our social practice, need not be thought of as relative or local. This is because reasons mediate between moral truth and social practice, and keep the connection between the two weak enough to avoid the relativism found in areas of practice (such as word pronunciation) that are more strongly socially dependent. So even if the social practice provides us with a starting point for justiﬁcation, this starting point is only provisional, and if reasons support a rule revision, then this revision is justiﬁed—even if this fact is not currently recognized. We must return to the disanalogy between morality and games such as basketball. Recall that basketball is relative in a particular way: before the

110

Pragmatism and Normativity

introduction of the Wilt Chamberlin rules goaltending was permissible, and afterwards it became a foul. We should be reluctant to regard morality as relative in the same way. But basketball is relative in this way because our reasons need to be reiﬁed to some extent to make the game of basketball work. Considerations of fairness recommend that the rules be standardized for all games, and the best way to do this is to reify the rules. Thus, basketball has an ofﬁcial rulebook, and new rules only become effective upon their ofﬁcial adoption; they are not applied retroactively. Rules in basketball display temporal relativity, owing to this reiﬁcation. I should note that even games such as basketball do not absolutely need to be relativistic in the way described above. Imagine a group of children playing basketball on the playground. Player A has the ball, and he says to player B of the opposing team, ‘Uh-oh; here comes your mom, and she looks really mad,’ when in fact player B’s mother is nowhere in the area. While B is distracted, A drives to the basket and scores. We can well imagine B saying, ‘That wasn’t fair;’ and we can further imagine our group of playground particularists agreeing that it wasn’t fair, and demanding a ‘do-over.’ Player A could argue that there is no rule against such deceit; but considerations of fair play compel the other players to decide against him. Or we can imagine the players saying at some later date, ‘We shouldn’t have allowed that play to stand.’ Thus, games such as basketball need not be reiﬁed as they are. The important lesson to be gleaned from this is that the decision whether or not to reify the rules of a game, and therefore make it historically relativistic, is itself a decision made on the basis of reasons. In the National Basketball Association, virtually all would agree that fairness demands that all play by the same set of rules, and this is best assured by reifying the rules, and modifying them as reason demands. On the playground, though, the same considerations of fairness could allow for punishment or re-play even when no explicitly reiﬁed rule forbids the action in question. Similarly, we could argue that we ought not play the moral game relativistically, because we want people to do what they have the best moral reason to do; we do not want to absolve of responsibility people living in evil societies (such as Nazi Germany, or the antebellum American South) for the actions they perform, such as genocide and slave-holding.31 We do not want people to be able to use as an excuse, ‘Everyone around me was doing it, too.’ We do not want people to be forced (morally) to do what everyone else is doing, because what everyone else is doing is often bad. Another good reason not to reify moral reasons is that few would submit to a single reifying

Pragmatism, Normativity, and Relativism

111

authority, a problem that does not arise in the formation of new sports leagues.32 Notice that once Wilt Chamberlin became adept at goaltending, but before goaltending was disallowed, the fact that goaltending was too easy for players like Chamberlin was still a good reason to prohibit goaltending. However, because the game being played was a reiﬁed one, this reason was not in itself a sufﬁcient reason for referees to penalize the practice. The rule had to be explicitly reiﬁed ﬁrst. However, in a non-reiﬁed game, the fact that a decisive reason obtains for prohibiting an activity is (or can be) a sufﬁcient reason for punishing that activity. We thus have some important conclusions. First, moral claims are licensed by the giving of reasons. Second, reasons display a sort of timelessness. Third, morality is a non-reiﬁed game. These three conclusions together entail a fourth: it may sometimes be permissible to punish or condemn activities which have not been explicitly prohibited by current moral practice, or which are even condoned by current moral practice. So the mere fact that a game is rule- or reason-constituted does not entail that the game must be relativistic. It is true that morality is underdetermined by our interests. But this should not make us despair of coming to moral agreement in those areas that are underdetermined. Using the practice as our starting place, it may eventually be possible to justify a moral claim in areas where currently our interests do not yet determine a unique answer. Morality is about coming to cooperative arrangements; and the fact that our interests do not determine a unique correct cooperative arrangement does not rule out a future state where such a cooperative arrangement evolves as part of the social practice, and hence acquires (prima facie) justiﬁcation. As I said at the beginning of this section, I spoke only of moral reasons for ease of phrasing, but the conclusions reached in this section apply equally to other sorts of reasons as well: in all cases—moral, epistemic, semantic, and so forth—reasons provide a buffer between social practice and the truth of normative judgments. Thus, a social-practice account of normative judgment can avoid strong social dependence and even historical localism. I will note, however, that epistemology is in an important sense relativistic. We can concede that it was rational to believe 5000 years ago that the earth was ﬂat. It is obviously not rational to believe this now. And so epistemology does display a sort of temporal relativity. It is always an option to structure a practice relativistically; but it makes more sense to structure epistemology this way than it does

112

Pragmatism and Normativity

to structure morality this way. Again, questions of how to structure the practice are pragmatic questions, to be decided by pragmatic reasons. The question of epistemic justiﬁcation and relativism is a complex one, but we do not yet have the tools to construct a complete account of the temporal logic of epistemic reasons. These tools will be developed in Chapter 7, where we will then discuss the temporal relativism of epistemic justiﬁcation. Returning to the example of morality, I indicated that the revisability of our social practice stands as an additional bulwark against relativism. I have argued that owing to the complexity and underdetermination of morality, the social practice must form an important element of the justiﬁcation of moral claims. But these same features (particularly the complexity) of morality indicate that we should always be cautious about claiming that we have attained a system that perfectly satisﬁes our interests. That is, we should always be prepared to revise our social practice (and the moral claims embodied in it) should compelling reasons arise for doing so. In Chapters 7 and 8, which develop a pragmatist epistemology, we will see how crucial revision is to rationality; but let us turn brieﬂy to this issue now. We have seen that conservatism is an element of the social-practice account of justiﬁcation: certain fundamental claims, embodied in our social practice, must be accepted conservatively. But this conservatism should not be allowed to become dogmatism. We must ever be willing to revise any element of the practice, should a compelling reason for doing so arise. Indeed, we need regard nothing as in-principle unrevisable. Some claims may be de facto unrevisable. For example, it is difﬁcult to imagine what could possibly license revision of the rules of arithmetic. The important lesson, though, is that nothing is de jure unrevisable—not a community’s moral claims; not its epistemic standards; and not even its standards for revising these epistemic standards. As Quine points out in ‘Two Dogmas of Empiricism,’ Any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system. Even a statement very close to the periphery can be held true in the fact of recalcitrant experience by pleading a hallucination or by amending certain statements of the kind called logical laws. Conversely, by the same token, no statement is immune to revision. Revision even of the logical law of the excluded middle has been proposed as a means of

Pragmatism, Normativity, and Relativism

113

simplifying quantum mechanics; and what difference is there in principle between such a shift and the shift whereby Kepler superseded Ptolemy, or Einstein Newton, or Darwin Aristotle?33 Thus, nothing need be treated as de jure unrevisable. Indeed, there is good reason to think that we ought to treat nothing as de jure unrevisable. As Lance and O’Leary-Hawthorne write, “A practice could in effect adopt the positivist proposal of treating a whole bunch of claims as de jure unchallengeable (if the challenge is to their truth). But we would not be tempted to adopt such a practice ourselves. Such a practice seems to encourage—even be constitutive of—dogmatism, preclude dialogue, induce cognitive sterility, and all at no obvious gain.”34 I will argue in Chapter 7 that revisability is a crucial element of justiﬁcation, and that a practice that does not permit revision soon ceases to be justiﬁed. Convinced as we may be that certain moral rules promote our interests better than any available alternatives do, we must be cognizant that many others in history have been equally convinced, but nevertheless wrong.

Cross-cultural relativism So far, this chapter has been devoted to a discussion of what one might call historical relativism, the claim that moral truth varies according to era within a particular culture. But discussions of relativism frequently concern what might be called cross-cultural relativism: the claim that moral truth varies according to culture. Let us brieﬂy discuss this latter type of relativism. Naturally, we will not always be able to convince the members of another culture that their moral way is wrong. What follows from this? There is a difference between S’s being wrong and our being able to convince S that she is wrong. One advantage possessed by the pragmatist account is that it ties morality to interests, which (as I will argue in the next chapter) are universal features of human nature. Thus, we have a cross-cultural basis for evaluating various moral codes. If a particular moral code (say, one that forbids women to leave the house unaccompanied) inhibits the satisfaction of certain fundamental human interests, then this moral code is inferior, whether or not we are able to convince the practitioners of the code of this fact. Convincing people that their moral code is mistaken may be particularly difﬁcult if this code is justiﬁed by appeal to religion. Indeed,

114

Pragmatism and Normativity

one might hold that the interests identiﬁed in the next chapter are of trivial importance compared with religion, and that one should be willing to forego satisfaction of these interests for the sake of religious purity. Though it is certainly true that many abominable practices (such as female genital mutilation) are justiﬁed by appeal to religion, no relativistic conclusions follow from this. First, we do not (and should not) allow religion to be used as an excuse for immorality. If a cult were to appeal to religion to justify their practice of infant sacriﬁce, we would hardly regard such an appeal as convincing. Part of the reason for this is that, as Martha Nussbaum argues, the term ‘religion’ has a moral connotation. Nussbaum writes: We may and do . . . judge that any cult or so-called religion that diverges too far from the shared moral understanding that is embodied in the core of the political conception does not deserve the honoriﬁc name of religion. Thus U.S. law has persistently refused to give religious status to Satanist cults and other related groups . . . Even when a group clearly counts as a religion, we sometimes judge that it forfeits its claim to state deference when it goes outside of certain moral understandings.35 Nussbaum here is writing primarily about legal protections for religious practice, but the lessons can be extended to moral practice: a religion that is involved in substantial suppression of the satisfaction of our fundamental human interests may forfeit the consideration we normally allow to religions and therefore may legitimately be condemned. Let us not forget too that, at bottom, disagreement about religion is factual disagreement. Factual disagreement often leads to moral dispute—Would allowing euthanasia lead society to descend a slippery slope? Would condoning homosexuality lead to the breakdown of the traditional nuclear family?—but these are not the sorts of disputes that are taken to be symptomatic of relativism. These are at bottom factual, not moral, disputes. Two comments need to be made at this point, to avoid misunderstanding. First, if we decide that another country’s moral practice is inferior to our own, nothing immediately follows from this. In other words, it does not follow that we should intervene in the other culture to alter their practice. Perhaps such an intervention would be counterproductive; perhaps the other culture’s immoral practice is not serious enough to warrant intervention. Or the problem could be epistemic— we cannot be sufﬁciently sure that the other culture’s practice is morally

Pragmatism, Normativity, and Relativism

115

inferior to our own, and hence condemnation (much less intervention) cannot be justiﬁed. Second, the claim that we can criticize the moral standards of other cultures should not be taken as an endorsement of chauvinism. If another culture has a different moral practice from our own, it does not follow automatically that our practice is the correct one. Rather, in cases of cross-cultural dispute, we should examine which of the two practices better serves our interests. Thus, someone from the United States should not condemn the Dutch merely because they allow euthanasia; rather, she should examine the practice of euthanasia in light of our various interests and decide whether the Dutch practice of allowing euthanasia is superior or inferior to the US condemnation of it. When we run into a culture with a moral code incompatible with ours, it should be borne in mind that their code might contain elements superior to our own. The social-practice account of justiﬁcation given here emphasizes revision, and the impulse for revision can come from outside of one’s own culture as well as from within it. Granted, the social practice-account of justiﬁcation here also endorses conservatism. But the conservatism is not dogmatic. That is, it does not say that the current practice must be maintained come what may. It merely says that the elements of the practice are prima facie justiﬁed, subject to their revision. Confrontation with another culture, which embraces a different set of moral norms, can result in the types of moral arguments that lead to the revision of the prima facie justiﬁed elements of one’s own practice. Again, the social-practice account endorses conservatism; but it also embraces revision as a necessary element of rationality. There is a ﬁnal version of cross-cultural relativism, which might be true. We have conceded earlier that our interests underdetermine morality. Thus, different cultures might have developed different cooperative strategies that are equally well-justiﬁed from a pragmatist perspective. Would this be a worrisome result? It does not seem that it would be. If another society’s cultural norms are different from ours, but serve their interests well, then it is not clear what the source of this worry is. Suppose, for example, that in one culture elderly parents are cared for by putting them in a nursing home or retirement community, whereas in another culture, elderly parents move in with and are taken care of by their children. If both of these practices serve the needs of elderly parents (and their children) equally well, then this is not a worrisome cross-cultural difference.

116

Pragmatism and Normativity

But when people worry about differing (but equally justiﬁed) cultural practices, they are often worried about cultural practices that we would ﬁnd immoral. What if such a practice were part of another culture’s pragmatically justiﬁed moral practice? In this case, there are two possibilities. First, the other culture’s practice might not serve people’s interest as well as our own cultural practice does. If this is the case, then we may legitimately criticize the other culture’s moral norms. Thus, the ﬁrst possibility does not describe a genuine case of relativism. The second possibility is that the other culture’s practice might serve people’s interests just as well as our own does. If this is the case, then the fact that we feel uncomfortable with the other culture’s norms is no more important than the fact that we might ﬁnd another culture’s culinary practices disgusting. If another culture’s culinary practices meet their dietary needs, then the fact that we ﬁnd them uncomfortable is no valid criticism of them. Similarly, if the other culture’s moral norms adequately serve human interests, then the example is similar to the above example of caring for one’s elderly parents: it is not a worrisome case of relativism. It is mere cultural diversity, which is not the same as relativism.

Relativism of circumstances Although we have seen that the version of pragmatism defended here is not relativistic in any pernicious sense, there is one sense in which morality is local. Morality is a cooperative strategy for promoting our interests; but which strategy best promotes our interests will depend on a group’s environment and circumstances. Thus, as a group’s circumstances change, so might the optimal moral strategy. To give a (possibly controversial) example, chastity might well have been a genuine virtue before the advent of reliable and readily available birth control methods; but the current availability of such methods makes chastity optional rather than obligatory. It is not clear whether this is a worrisome form of relativism. Indeed, philosophers often appeal to difference of circumstances so that they might explain different societies’ differing norms without appealing to relativism. To take a hackneyed example, relativists sometimes appeal to the Eskimos’ alleged practice of putting their elderly on an ice ﬂoe to die as evidence that moral relativism is true. Opponents of relativism will say that this practice is not evidence of relativism, and is explained not by a genuine difference of moral principles, but by the radically different circumstances in which Eskimos live. As Eskimos live a marginal

Pragmatism, Normativity, and Relativism

117

existence in a difﬁcult environment, having to support non-productive members can put an entire community at risk, and so it is necessary to allow such members to die. James Rachels, discussing the Eskimo practice of infanticide, writes: Among the Eskimos, infanticide does not signal a fundamentally different attitude toward children. Instead, it is a recognition that drastic measures are sometimes needed to ensure the family’s survival . . . The Eskimos’ values are not all that different from our values. It is only that life forces upon them choices that we do not have to make.36 To conclude, then, the adoption of different cooperative strategies based on differing circumstances between cultures does not seem to represent a worrisome form of relativism.

Summary and conclusion The argument of this chapter has been complex, so let us retrace our steps brieﬂy. I began by arguing that the issue of moral and epistemic truth can be illuminated by Wiggins’s and McDowell’s idea of truth as consisting of excellence of reasons: a proposition p is true if we can give a vindicatory explanation of it, that is, an explanation that demonstrates that there is nothing else to think but p. What reasons support moral and epistemic claims? On the pragmatist account under discussion, moral rules are justiﬁed by our interests; epistemic rules are, similarly, justiﬁed by their success in promoting our epistemic ends (such as truth and explanation, and knowledge, which is also a fundamental human interest). However, this does not give us a straightforward set of truth conditions for moral and epistemic rules. The problems of underdetermination and complexity forced us to adopt a social-practice account of justiﬁcation, whereby the elements of the social practice are taken as prima facie justiﬁed, subject to their revision. However, even if a particular moral claim might be justiﬁed by the social practice at a particular time, that does not mean that the claim is true (i.e., that it is supported by the best reasons, by a vindicatory explanation). If we combine the idea that moral truth consists of excellence of reasons with the idea of the timelessness of reasons, we get the result that the truth of moral rules is not relative to a particular time. Some skepticism about social-practice accounts of normativity no doubt stems from what Rorty calls “the religious need to have human

118

Pragmatism and Normativity

projects underwritten by a nonhuman authority.”37 This way of thinking underlies our intuition that science is the paradigm of objectivity; science is objective because it is made true by something non-human. I suggest that this intuition should be construed as follows: robust objectivity requires that the truth of a type of discourse be independent of what we take it to be. This is why we think science is the paradigm of objectivity: there is a world out there that (in some sense) makes scientiﬁc truth independent of what we take it to be. But morality and epistemology have this same virtue: moral and epistemic truth is also independent of what we take it to be. In this respect, then, morality and epistemology are as objective as science. And I suspect this is the most interesting kind of objectivity, in any case.

Part II Pragmatism and Morality

This page intentionally left blank

5 Pragmatism, Interests, and Morality

The conclusion of Chapter 3 is that morality is justiﬁed directly by instrumental rationality, and that it is rational to be moral. Questions of whether it is rational to be moral have often been cast in terms of the relation between morality and prudence, where ‘prudence’ is taken to be synonymous with ‘self-interest’. But we have many interests, and not all of them are self-directed. Some of our interests are other-directed (i.e., altruistic). Thus, if rationality consists of promoting our interests, then a strategy to promote all of our interests—self-directed and otherdirected—will be rational. As noted in Chapter 2, we can continue to say that morality is prudentially rational, as long as we always keep ﬁrmly in mind that we are not using ‘prudential’ in an egoistic sense, but instead to refer to the satisfaction of all of our interests—both self- and otherregarding; indeed I will use the word ‘prudential’ in this way throughout this chapter. And, of course, when we speak of morality as being rational, we must keep in mind the notion of strategic cooperative rationality defended in Chapter 3. Pragmatism is not rational egoism. Although the last chapter argued that it is rational to be moral, the discussion of the last two chapters might also have convinced the reader that the account given here is similar to utilitarianism, or at least has some points of contact. The current chapter should dispel this impression. It is true that, in some sense, the pragmatist theory presented in this work is teleological: morality and epistemology exist to serve our interests, and so are goal-oriented. But this is not to say that there is some quantity (e.g., utility) that can be measured and which ought to be maximized. Indeed, I think the account given here has at least one signiﬁcant advantage over utilitarianism. Utilitarianism is presented as a ﬁrstorder ethical theory. What utilitarianism does not provide is a suitable 121

122

Pragmatism and Morality

meta-ethic. That is, utilitarianism gives us an account of which acts are right (those that best promote utility), but it does not give us a good explanation of why we should maximize utility. Indeed, one might see the ﬁrst-order utilitarian theory as in fact undermining any utilitarian meta-ethic. The best-known version of utilitarianism equates utility with preference satisfaction, and maximal preference satisfaction is the goal. One problem with this approach is that it does not really ground morality in preferences. Preferences are used as a measure of utility (and by extension, a measure of whether a moral system is the best one), but preference does not ground utilitarianism in the ﬁrst place. Let me be more clear: in order to be truly grounded in preferences, utilitarianism would have to be supported by our preferences. That is, we (or most of us) would have to prefer a utilitarian system to any other system. But this is not the case—most people’s instinctive distrust of utilitarianism indicates that maximization of utility is not a preference we possess. So, though utilitarianism counsels the maximization of something, it is not something we have a preference for maximizing. And hence utilitarianism (or at least the preference version of it) is in an important sense self-refuting. Even if you reject the idea that utilitarianism is self-refuting, it is still clear that utilitarianism has failed to ground morality in our preferences—our set of preferences does not justify the claim that one ought to maximize overall utility. (Of course, one might argue that it follows from the nature of rationality that one ought to maximize preference satisfaction. Even if this is true, it only follows that one ought to maximize the satisfaction of one’s own preferences; it does not follow that one ought to maximize overall preference satisfaction.) Thus, though utilitarianism might offer us a ﬁrst-order moral theory (i.e., a theory of which actions or rules are morally correct), it does not justify the utilitarian principle. The principle of utility must be given a separate, external justiﬁcation. Pragmatism, on the other hand, offers a coherent meta-ethic and a ﬁrst-order theory. At the level of meta-ethics, pragmatism claims that moral rules are rational strategies for the mutual promotion of our interests. Thus, pragmatism tries to demonstrate that it is rational to be moral. But the same considerations that ground morality—that is, the demonstration that morality is rational—also provide us with a ﬁrst-order moral theory: the correct moral rules are those that satisfy our interests, those that qualify as rational strategies for the mutual promotion of our interests. So not only does pragmatism provide us with both a meta-ethic and a ﬁrst-order moral theory, but these two elements are of a piece. Thus, the pragmatist account is complete and

Pragmatism, Interests, and Morality

123

coherent in a strong way. This is a powerful advantage that pragmatism has over utilitarianism.

Interests What is it that justiﬁes our continued participation in moral and epistemic practice? We have talked loosely, over the course of this book, of morality and epistemology as satisfying our interests. But what are these interests? Two natural candidates for the role of interests are (a) preferences and (b) desires. Identifying interests with preferences would allow us to tie our account here ﬁrmly to the literature on game theory, thus enabling us to give some theoretical weight to our discussion, in Chapter 3, of the rationality of morality (and epistemology). The other option is to identify interests with desires. This move would dovetail nicely with the so-called Humean theory of normative reasons, according to which all justifying reasons have their ground in desire. By identifying interests with desires, we can then take advantage of this Humean theory of normative reasons to explain how our interests serve to justify our moral practice. Our moral practice, on this account, is a cooperative strategy that serves to further our mutual desires. Thus, it would be theoretically neat if we could argue that desires are what that justify morality. Unfortunately, the idea that desires give us reasons for action has been subjected to a great deal of destructive criticism lately. Much of this criticism is to be found in Scanlon’s What We Owe to Each Other, although Jean Hampton is another trenchant critic of the Humean theory of normative reasons.1 Scanlon has several arguments against the idea that desires are a source of justiﬁcatory reasons. I will brieﬂy examine three of them. First, according to Scanlon, reasons and desires have different structures. Reasons interact with each other in a way that desires do not. When desires conﬂict with each other, the fact that one desire outweighs, or is stronger than, the competing desire does not cause the competing desire to cease to be felt. For example, if I have a desire to eat ice cream, but an even stronger desire to stick to my diet, then I will stick to my diet; but that will not make my desire for eating ice cream go away. Reasons, on the other hand, are capable of interacting in a more complex way than this. In some (but not all) cases, a reason can completely silence a competing reason, and make it not a reason at all. To use Scanlon’s example, suppose I am playing tennis. Suppose also that after taking into account all relevant considerations, I decide I ought to

124

Pragmatism and Morality

play to win (as opposed to, say, letting my opponent win). This decision makes some reasons relevant to my choices and other reasons irrelevant. For example, if I am at match point, “the fact that a certain shot represents the best strategy will count as sufﬁcient reason for making it” (p. 51). The fact that my making this shot will cause my opponent distress and unhappiness is entirely irrelevant to whether or not I should make this shot (given that I have decided I ought to play to win). That is, the reason in favor of making this shot entirely silences this particular competing reason (my opponent’s unhappiness). This is not to say that my opponent’s (un)happiness is a matter of complete indifference to me; it is merely to say that it is not a relevant consideration to be taken into account when deciding whether to make this shot. Thus, reasons interact in more complex ways than do desires; desires are therefore not reasons. The second of Scanlon’s objections against the Humean theory of justiﬁcatory reasons is that the mere fact that we have a desire does not provide us with any reason at all to perform the desired action. Scanlon cites Warren Quinn’s example of the man who has a desire to turn on every radio he encounters. The man in question does not see anything good about doing so: he does not want to hear the news, or listen to music; he merely has an impulse to turn on all radios he sees. According to Quinn and Scanlon, this desire does not provide the man with any reason to turn on these radios. Such a desire does not justify the action in question because it, in Quinn’s words, leaves out “precisely that element of desire that does the rationalizing,”2 that is, the element that justiﬁes the action in question: namely, the judgment that the desired action or object is in some way good. Thus, a desire by itself does not justify action; rather, the justiﬁcation is done by the evaluative element of the desire: the judgment that the object or action desired is good. (And, indeed, Scanlon argues that rather than providing a reason for action, a desire is more often the result of the judgment that a particular object or course of action is good.) Scanlon’s third objection is related to the second one. Suppose I have a desire to buy a computer. As the second objection argued, this desire by itself provides me with no reason to buy a computer. But suppose I judge that I do, in fact, need a new computer. I now have a reason to buy the computer—but the reason does not stem from my desire. Scanlon writes: Now suppose that I endorse the judgment to which the desire involves a tendency, and take myself to have good reason to buy a

Pragmatism, Interests, and Morality

125

new machine. Even in this case, the reason that I have for buying a computer is not that it will satisfy my desire, but rather that I will enjoy having it, or that it will help me with my work, impress my friends and colleagues, or bring some other supposed beneﬁt.3 Thus, it is not the desire for a new computer that provides me with a reason to buy a new computer. Rather, it is (say) the fact that the new computers have features that my old computer does not have, and which I would enjoy or ﬁnd useful, etc. Roughly stated, it is not the desire that provides a reason for action; rather, it is the object of the desire (its features, its role in my life and my projects, etc.) which is the source of such reasons. How serious are these objections against the Humean theory of normative reasons? It seems as though a sophisticated version of the theory could overcome the ﬁrst two objections. An unsophisticated Humean might say that all desires are reasons, or give reasons. A more sophisticated Humean might merely say that all reasons derive from desires. Thus, if I have a set of desires, I do not have a reason to act on all of them; but all of my reasons for action ultimately derive from this set of desires. My reasons are not to be identiﬁed with these desires; rather, my reasons are the strategies, etc. that would best enable me to fulﬁll these desires, or to fulﬁll the greatest number of them, or the strongest among them, etc. Thus, the sophisticated Humean need not concede Scanlon’s ﬁrst objection, which simply equates reasons with desires. Nor does the sophisticated Humean need to concede the second objection, which claims that every desire gives rise to a reason. The third objection, however, presents a more formidable obstacle to the account under discussion. According to this objection, desires simply are not the right sorts of things to act as a source of justifying reasons. If I have a set of desires, it might be rational for me to act to satisfy these desires, but my actions would not be rational because they contribute to the satisfaction of my desires. No; the action would be rational if it satisﬁed ends which I have deemed worthy of pursuit. Scanlon’s ﬁnal objection is a powerful one. Although complicating the connection between desires and reasons may have helped answer the ﬁrst two objections, this move does not seem as though it will provide us any way of escaping Scanlon’s third objection. And so it seems we must, unfortunately, abandon our attempt to identify interests with basic desires. It would seem that similar comments would apply to an attempt to ground morality in preferences. Would the mere fact that I have a

126

Pragmatism and Morality

preference for turning on all radios in my vicinity give me a reason for doing so? And if I have a preference for buying a new computer, is it the preference that gives me a reason for action, or the computer’s qualities? Thus, it seems as though the objections to grounding morality in desires may also apply to grounding morality in preferences. With what will we replace our fallen account? Scanlon’s ﬁnal objection seems to provide us with some direction. Remember, according to Scanlon, it is not the desire for an object that gives us a reason for action, rather, it is the fact that the object in question has some value for us, that it will help promote our valuable ends. Thus, Scanlon’s objection encourages us to focus not on the desires, but rather on the ends that are the objects of our desires. These, insofar as they are genuinely worthy of pursuit, provide us with our reasons for action. What ends do we have in common? What ends are valuable for all of us? After all, we saw that one advantage of the interests-as-basic-desires account was that it gave us something shared on which to ground morality. Do we have ends in common which can serve the same purpose? In fact, we do: each of us has a reason to pursue the ends that constitute human ﬂourishing.

Interests and ﬂourishing Two of the objections against the interests-as-desires account offered above push us in a particular direction. One objection had it that the presence of a desire does not give us a reason for action unless we endorse or consider valuable the goal aimed at by the desire. Another objection (related to this ﬁrst one) is that if I have a desire to do X, and also a reason to do X, the latter exists not by virtue of the former, but instead by virtue of the features of X that make X worthy of pursuing or achieving. Both of these objections locate the source of reasons not in the desire, but in the end aimed at by the desire. This suggests that we should seek reasons for action not in desires, but in the ends that are themselves desirable, worthy of being desired. What ends do we have reason to pursue? Plausibly, we have reason to pursue the ends that constitute a good life. What ends are these? Derek Parﬁt has identiﬁed three theories regarding what constitutes human ﬂourishing.4 First, there are hedonistic theories according to which what makes the person experience the most happiness (where happiness is understood as a subjective state) constitutes ﬂourishing. Next are the desire fulﬁllment theories, according to which ﬂourishing is constituted

Pragmatism, Interests, and Morality

127

by (naturally) fulﬁllment of one’s desires. Finally, there are objective-list theories, according to which ﬂourishing consists of achieving ends that are objectively valuable for humans. Richard Arneson has argued forcefully in support of the third type of theory, the objective-list theory.5 A large part of his argument consists of demonstrating the inadequacies of the other two types of theory. Let us examine some of these arguments. Arguments against hedonism as a theory of human welfare are well-known. One of the best-known is Nozick’s ‘experience machine’ argument.6 This argument is aimed against any theory that, like hedonism, attempts to identify the good with some type of experience (such as an experience of pleasure or happiness, as hedonism would have it). In this example, Nozick asks you to imagine that you have the opportunity to spend the rest of your life in an ‘experience machine,’ which would feed you a whole range of artiﬁcial but pleasant and rewarding experiences. If you thought that pleasure were the only good, you would not hesitate to spend the rest of your life in the machine. However, Nozick argues that intuitively, you ought not spend the rest of your life in such a machine. Nozick’s reason is that we care about actually having real accomplishments, friendships, etc., and that our experience of such events is not the only—and perhaps not even the most important—source of their value. Desire fulﬁllment views fare little better. There are several objections against the idea that ﬂourishing or welfare is constituted by desire fulﬁllment. One of the strongest objections, raised by Arneson (against desire fulﬁllment accounts of welfare) and also by Martha Nussbaum (against preference satisfaction accounts of welfare), is the problem of adaptive preferences or desires. If one is lacking in opportunities, often one’s desires and preferences will become accordingly modest. We should regard this lack of opportunity as an obstacle to ﬂourishing; but the desire or preference fulﬁllment theorist will say that a person who has modest hopes (based on modest prospects) is ﬂourishing to the maximal degree, as long as these hopes are satisﬁed. Arneson writes: Being blind, I don’t form ambitions that require eyesight, and being impoverished, I don’t form ambitions that require wealth to have a reasonable prospect of success, and being unintelligent, I don’t form ambitions that would strain my limited brain power, and lacking social connections, I don’t form ambitions that can be achieved only with the help of powerful allies.7

128

Pragmatism and Morality

A desire satisfaction theorist would be forced to say that this person has an excellent chance of leading a remarkably ﬂourishing life, as his desires are so modest. But intuitively, this is incorrect. Nussbaum notes that according to a 1944 poll, 45.6% of widowers in India rated their health as ill or indifferent, but only 2.5% of widows did so. This result is despite the fact that widows were, as a group, in manifestly worse health than the widowers were. However, as they had come to expect poor health, they came to accept their poor health as satisfactory. The problem of adaptive preferences presents a serious difﬁculty for utilitarian accounts of welfare. As Amartya Sen writes, “Quiet acceptance of deprivation and bad fate affects the scale of dissatisfaction generated, and the utilitarian calculus gives sanctity to that distortion.”8 The failure of hedonistic and desire-fulﬁllment theories leaves the objective-list theory as the sole candidate for a theory of ﬂourishing or welfare. What ends constitute human ﬂourishing? While lists of human goods differ in detail, they are in broad outline very similar. The following seem like good candidates for essential elements of human ﬂourishing: 1. Life: Life and self-preservation, avoidance of injury and death, are all human goods. All things being equal (lack of ill health, etc.), it is better to live longer, rather than have short lives. 2. Health: It is a human good to remain healthy, and have one’s body intact. Sickness and injury are to be avoided, illnesses are misfortunes, and so forth. 3. Control and autonomy: Humans should possess at least some control over their own actions and environment. Even though humans readily submit to authority, they are quick to resent excessive control over their actions and their environment. We should be able, within limits, to make our own decisions, pursue our own projects, and so forth. 4. Pleasure and fun: Pleasurable and otherwise enjoyable experiences are valuable, whether these experiences are sexual, gustatory, or involve the enjoyment of play. 5. Afﬁliation: Humans are social animals, and seek afﬁliation with other humans (and possibly other species, too). This interest manifests itself also in sympathy toward and liking for others, and in a desire to see others (especially those close to us) achieve elements of human ﬂourishing. 6. Knowledge: Humans are inquisitive creatures, who seek to understand their environment.

Pragmatism, Interests, and Morality

129

This list of interests bears a resemblance (and is inspired by) Martha Nussbaum’s conception of human functionings.9 Nussbaum is concerned to present an account of human ﬂourishing, an account that is truly universal. She argues that there are certain functionings that enhance human life (and in many cases constitute human life). These functionings include things like life, bodily health and integrity, practical reason, afﬁliation, and so forth.10 A human life lacking one or more of these functionings is deﬁcient in an important way. Nussbaum writes: The intuitive idea behind the approach is twofold: ﬁrst, that certain functions are particularly central in human life, in the sense that their presence or absence is typically understood to be a mark of the presence or absence of human life; and second—this is what Marx found in Aristotle—that there is something that it is to do these functions in a truly human way, not a merely animal way. We judge, frequently, that a life has been so impoverished that it is not worthy of the dignity of the human being, that it is a life in which one goes on living, but more or less like an animal, unable to develop and exercise one’s human powers.11 Such an account can be used as an account of human interests, and these functionings can be identiﬁed with interests: humans have an interest in these functionings, in the sense that any human life devoid of certain of these functionings is to that degree impoverished. There are several notable features of Nussbaum’s list of functionings. First, it is universal. It “represents the result of years of cross-cultural discussion.”12 It was developed with reference not merely to values that Westerners hold, but in discussion with both Westerners and nonWesterners. Nussbaum writes that the list’s “very close resemblance to other similar lists worked out independently in parts of the world as divergent as Finland and Sri Lanka gives some reason for optimism about consensus.”13 Second, though the list is intended to be universal, it is resolutely not intended to be neutral. The elements of the list are morally infused, and have moral import. Nussbaum writes: The basic intuition from which [this] approach begins, in the political arena, is that certain human abilities exert a moral claim that they should be developed . . . Not all actual human abilities exert a moral claim, only the ones that have been evaluated as valuable from an

130

Pragmatism and Morality

ethical viewpoint. (The capacity for cruelty, for example, does not ﬁgure on the list.) Thus the argument begins from ethical premises.14 Third, Nussbaum’s list (like the interests approach developed in this chapter) acknowledges that these functionings may be realized in different ways by different people (and among different cultures): “part of the idea of the list is its multiple realizability: its members can be more concretely speciﬁed in accordance with local beliefs and circumstances. It is thus designed to leave room for a reasonable pluralism in speciﬁcation.”15 Finally, the list is not grounded in some tendentious metaphysical conception of the human being (such as, for example, an Aristotelian teleological conception of biology). Nussbaum writes that her approach is “not one that relies on a particular metaphysical or teleological view . . . Thus the argument begins from ethical premises and derives ethical conclusions from these alone, not from any further metaphysical premises.”16 The basic interests we identiﬁed above can be said, in a very general way, to deﬁne human ﬂourishing. Of course, human nature is enormously ﬂexible; this is why basic interests must be stated in such general terms. Let us introduce a distinction between basic interests and derivative interests. This distinction will help us capture this ﬂexibility inherent in human nature; it will also allow us to connect more ﬁrmly these interests (which are prescriptive in nature) with the actual interests or ends recognized and pursued by individual agents. Derivative interests are speciﬁc, detailed manifestations of our basic interests, and reﬂect the ﬂexibility of human nature. For example, afﬁliation is a basic human interest. But many factors, including acculturation, determine how this basic interest is translated into speciﬁc derivative interests. In some people and some cultures, this basic interest will manifest itself in treating marriage as a legitimate end, or in an interest in going to bars and nightclubs, etc. In others, this basic interest will manifest itself in an interest in joining a group centering around a particular activity (such as a book club or a recreational sports team). In yet other cases, this basic interest in afﬁliation will manifest in an interest in owning pets. Obviously, a derivative interest need not be a manifestation of a single basic interest; it might reﬂect a number of basic interests. A person who decides to join a recreational basketball team may well be manifesting basic interests in afﬁliation, health, and in pleasure and fun. Most of the interests people pursue are derivative interests. So, for example, the content of one’s motivation in checking out a book from

Pragmatism, Interests, and Morality

131

the library is not ‘to increase my knowledge;’ rather, it is ‘to ﬁnd out more about the life of Admiral Nelson.’ The content of one’s motivation in putting antibiotic cream on a cut is not ‘to preserve my health;’ rather, it is ‘to avoid infection.’ However, these derivative interests are instances of basic interests, and by promoting these derivative interests, one is in fact promoting the basic interests under which they fall. Basic and derivative interests serve different roles in our moral theory. It is much easier to develop a set of cooperative strategies that will help promote our interests if we all have similar interests. If we did not have similar interests—if person A had interest a, and B had b, and C had c, and so on—it would be difﬁcult to conceive of a cooperative strategy that would help secure our respective interests—a for A, b for B, c for C, and so on. There would be a complex set of rules, dictating a certain behavior toward A (aimed at helping her achieve a), whereas A would have to display distinct behaviors toward B and C, to help them achieve their respective (and different) ends. If we had radically different interests, then our cooperative strategy would have to resemble a barter, rather than a simple cooperative strategy. That is, if I have interest a, but you have interest b, then the mutually beneﬁcial strategy for us would be for you to help me promote a in turn for my helping you promote b. Thus, if we had radically different interests, a mutually fruitful interaction would require that each person ﬁrst identify the interest possessed by the others, and only then decide how to proceed. Each agent in such a cooperative system would have to possess an enormous amount of information in order for the cooperative system to work. But if each person has the same interests, then a cooperative strategy will allow us, as a general rule, to exhibit the same behaviors toward everyone: we do not have to inquire what interests a person has before deciding which set of rules to follow (Is your interest a? Then I should do x. Or is it b? Then I should do y. etc.). For moral purposes, each person is (prima facie) interchangeable. Each person has the same interests, and so we ought to treat each person the same. For example, the rule, ‘Do not kill person P’ is the rule we should follow toward people who have an interest in life. But if everyone has this interest, then we can observe this as a general rule to apply to everyone. Thus, the fact that we all share basic interests makes it much easier to develop a set of cooperative strategies to promote our interests. Of course, we will have different obligations toward some people, based on special relationships we have toward those people. But it will be easier to keep track of special obligations to a selected few individuals, rather than having to keep track of

132

Pragmatism and Morality

distinct obligations for each person we might encounter. The former is possible; the latter is not. Of course, we need to be cautious when we say that each person has the same interest. When I say that A and B both desire health, it is primarily the case that A desires A’s health and B desires B’s health. But the fact that everyone shares the same interest—the interest in promoting his or her own health—makes it easier to construct a cooperative strategy that satisﬁes each person’s interests. It would be much more difﬁcult if we did not have similar interests (such as the interest in health). Many, perhaps most, of our legitimate interests are instantiations of these basic interests. Often, it is rational to act to promote a basic interest as a means of promoting the derivative interests which fall under that basic interest. Thus, suppose I have a variety of derivative interests which are all health-related. I might simultaneously promote several of these interests by acting in a general way to promote my overall health—say, by joining a health club, or getting regular physicals from a physician. This action will serve to promote my various derivative interests, perhaps more effectively than merely addressing them one by one will. Thus, by getting regular physicals I can monitor my cholesterol (and take physician-recommended steps to lower it, if necessary), get dietary tips, discover various health ailments before they become serious, and so forth. Acting at the level of the basic interest is often more rational than addressing my derivative interests piecemeal, as it can allow me to fulﬁll multiple interests in a strategic fashion. Let me give another example of how acting at the level of the basic interest can be more rational than acting to promote each individual derivative interest that falls under this basic interest. Suppose I aim to acquire knowledge about horticulture. I could join the library, at a cost of $35, or I could by a book on horticulture for $20. In this case, it would seem more rational to buy the book. Suppose I also aim to acquire knowledge about the feats of Admiral Nelson. A book on Nelson costs only $25, again cheaper than a library membership. But if I satisfy these goals piecemeal, I end up spending $45, more than a library membership would have cost me. If instead, I acted on my general interest in acquiring knowledge—by joining the library, thereby opening multiple avenues for the satisfaction of this interest—I would be able to satisfy my interests more efﬁciently. So it is not merely rational (or sometimes not even rational) to act to promote individual derivative interests; it is sometimes rational to act to promote a more general basic interest. Often, by organizing our derivative interests under basic interests, and acting to promote those basic interests, we can achieve our ends more

Pragmatism, Interests, and Morality

133

efﬁciently than we can if we act piecemeal to promote each individual derivative interest. We can conclude, then, that it is often rational to promote our basic interests. As I noted above, focusing on basic interests allows us to treat moral agents interchangeably, and thereby allows us to learn a vastly more simple set of moral rules than would be possible if agents did not share interests, or if we tried to develop a cooperative moral strategy that focused purely on unshared interests. Now we have seen that it can be rational to promote basic interests, as well as (sometimes instead of) the derivative interests, which fall under these basic interests. Thus, we can conclude that it is often rational for us to act to promote our basic interests, both purely for ourselves and as part of a cooperative strategy. Let us see how this approach works to justify moral rules. Consider our interest in autonomy and control over our actions. This interest might be exempliﬁed in a variety of ways—I might want the freedom to pursue my dream of writing the great American novel, whereas someone else merely wants to walk the streets free and unmolested. But these derivative interests all fall under the interest of autonomy. Our moral rules cannot be as speciﬁc as, ‘If X wants to write the great American novel, then X should be allowed to do so;’ or ‘If X wants to walk the streets free and unmolested, then X ought to be allowed to do so.’ This would result in an unlearnable profusion of rules; and as we noted in Chapter 3, moral rules must be learnable. Rather, a more general rule is called for, one that encompasses the derivative interests that fall under the general interest in autonomy: ‘If X wants to do A, and his doing A does not violate the legitimate interests of others, then ceteris paribus, X should be allowed to do A.’ Thus, by abstracting away from our idiosyncratic derivative interests, we can establish moral rules at the level of general interests, rules that protect our derivative interests—that is, rules that protect the idiosyncratic ways in which each of us realize the various human interests. Thus, our derivative interests are in general instantiations of our basic interests: we satisfy our basic interests by satisfying the derivative interests that instantiate them. And so in participating in a cooperative strategy, we are precisely seeking to make room for the satisfaction of our basic interests, as these basic interests are instantiated in our concrete derivative interests. This conclusion helps us see more clearly why understanding basic interests as categories under which derivative interests fall makes these basic desires well-suited for grounding morality. As I noted above, it is easier to develop a cooperative strategy among people

134

Pragmatism and Morality

who have similar interests. As basic interests are shared, this makes it easier for us to develop a cooperative strategy that serves each of us well. Let us examine the developing picture. We have seen that we can justify morality by appeal to our interests. As noted in Chapter 3, our afﬁliative interests will not just give us an incentive to promote others’ interests (as well as our own), but also increase our incentive to form joint commitments with others and thereby work with others cooperatively, if a mutually fruitful arrangement can be made. Again, this should not be taken to suggest that we wish to ground cooperative rationality in individual rationality; cooperative rationality is irreducibly a group activity. It is merely to note that our afﬁliative interests will increase our inclination to participate in group strategies. And indeed, research in game theory has revealed a variety of conditions under which cooperative, mutually beneﬁcial strategies evolve and become stable, even among parties who are ostensibly enemies.17 As we noted in Chapter 3, means–end reasoning involves choosing the right action or strategy for the accomplishment of the end. But we further noted that sometimes the correct strategy is a cooperative strategy. That is, what is rational must be understood cooperatively, not individualistically. I further argued that our interests are not best satisﬁed by an action-based, individualistic notion of rationality. Rather, we must think of rationality in terms of cooperative strategies. And so the rational promotion of our interests dictates that we should execute a cooperative strategy for the mutual promotion of ends. That is, we should cooperatively follow a set of strategies (a practice, consisting of a set of strategic rules) to promote our interests. The practice (i.e., set of strategies) aimed at promoting our interest in knowledge is the practice of epistemology. (We will talk about the structure of a pragmatist epistemology in Chapter 7.) The most general practice (general in the sense of being aimed toward the promotion of all of our interests) is morality. Moral norms are justiﬁed by appeal to our interests. Our interests in life and health underwrite the rule against harming and killing. Our interest in knowledge supports a prohibition of lying. Our interest in control over ourselves and our environment supports rules protecting autonomy and freedom of choice. And so interests can give rise to moral rules, where the latter are cooperative strategies designed to promote our mutual advantage. This approach takes the mystery out of goodness. Talk about good, bad, right, and wrong is merely a special type of legislative or expressive discourse introduced as a cooperative strategy, which strategy is justiﬁed rationally by appeal to our interests. So moral goodness is not

Pragmatism, Interests, and Morality

135

a mysterious entity; it is merely an artifact of cooperative rationality. There is nothing queer in that. So, for example, if we say ‘Mary is a good person,’ we are saying something like ‘Mary reliably participates in optimiﬁc cooperative strategies, and is motivated to do so by taking part in joint commitment,’ and we are endorsing this behavior, or expressing allegiance to the set of norms followed by Mary. In this way, moral normativity is reduced to instrumental rationality, and is no more (or less) queer than the latter. There is some superﬁcial similarity between this approach and the natural-law approach. Similar to the pragmatist approach, the naturallaw theory starts with an objective list of basic human goods, and develops moral principles based on this list. But natural law takes an overly-formalist approach that in many cases impedes our interests rather than promoting them. Like pragmatism, natural-law theory identiﬁes certain fundamental human goods. We must act to promote these goods, and must never act with the intention of impeding or harming these goods. However, owing to its overly-formalistic nature, the natural-law theory often proves harmful to these interests. For example, the refusal to countenance birth control (as it impedes the good of reproduction) might make a family unable to afford education for its (many) children, thereby impeding a number of interests, such as the interest in knowledge and autonomy (as education provides options). Overpopulation might also serve to impede interests such as health (from pollution owing to overcrowding, etc.). Thus, the pragmatist approach here better observes the spirit of interest promotion than does a natural-law theory.

Commonality and diversity The fact that morality consists of cooperative strategies necessarily shifts the focus to what is shared among us. Common ends are appropriate for justifying cooperative strategies. Naturally, what we share are our interests. And so naturally morality will consist primarily of a set of cooperative strategies aimed at promoting our mutual interests. But one might experience a certain dissatisfaction with the strategy of attempting to ground morality on interests alone. After all, if the goal is to show that it is rational to be moral, then shouldn’t we focus on other states whose satisfaction it is rational to seek? Can we focus only on interests? In addition to basic interests, we have derivative interests. Among these derivative interests, we have both shared and non-shared interests. We have already discussed the role that basic interests play in the

136

Pragmatism and Morality

justiﬁcation of morality. Let us now look at unshared and shared derivative interests to see what role they might play, or what obstacles they might place, in the path of justifying morality by appeal to cooperative rationality. We noted above that our interests can manifest in a variety of ways. And so, even though we share basic interests, the way in which these interests are instantiated will exhibit some degree of idiosyncrasy. This diversity of derivative interests might make one despair of grounding morality in rationality. ‘After all,’ one might reason, ‘if morality is at base grounded in pragmatic (and primarily prudential) reasons, and if different people have different desires, then we must inevitably wind up with some version of subjectivism, as what is prudentially justiﬁed will differ among different people.’ There are at least two things that need to be said in response to this: First, not every prudential strategy is a moral strategy. Thus, if I have an idiosyncratic goal of tasting every variety of coffee at Starbucks, a strategy aimed at promoting this goal will not be a moral strategy. That is, the principle that I follow to satisfy this goal will not be one of the principles or rules constituting a moral practice. Recall that only our cooperative prudential strategies can be considered moral rules. The second comment that needs to be made is that we should appreciate the constructive role that our diversity of interests plays in our moral practice. It might be appropriate to say that while we share basic interests, we also share in that each of us possesses our own idiosyncratic derivative interests, whose fulﬁllment we also seek. Put differently, we share in that each of us pursues the interests in ways that are not shared, that are idiosyncratic. Instead of standing in the way of a cooperative moral strategy, this diversity will support a certain liberal element in our moral theory. After all, each of us will recognize that it is rational to agree to a system in which we are allowed, to the extent feasible, free reign to pursue satisfaction of these unshared interests, free reign to pursue satisfaction of our interests in ways that are unique to us. There will no doubt be some tension between this liberal element and the element aimed at mutual satisfaction of our shared interests, but such a tension is neither surprising nor alarming; rational satisfaction of our interests will obviously require certain trade-offs. My interest in losing weight conﬂicts with my interest in eating ice cream daily. Conﬂict among interests is a constant feature of prudential rationality. Thus, morality will consist of a set of strategies aimed not only at promoting our mutual interests but also at allowing us a certain space in which individuals may pursue their idiosyncratic projects and desires.

Pragmatism, Interests, and Morality

137

This latter element is made rational by the fact that we all have such unshared interests, and it is more rational for us to adopt a liberal element in our morality than it is to simply forego satisfaction of these myriad idiosyncratic interests. Calling this element of pragmatist morality ‘moral liberalism’ is perhaps a bit misleading, inviting (as it does) comparisons with political liberalism. Although the two are similar in that their shared catchphrase is ‘tolerance,’ they are different in various ways. While political liberalism is supposedly neutral between conceptions of the good (and this is supposed to be political liberalism’s Achilles heel), moral liberalism is obviously not neutral. A moral code is certainly not going to be neutral among all practices. Certain practices, which go against our mutual interests, will be condemned. Moral liberalism is not the same thing as moral subjectivism, or excessive moral permissiveness. I am merely arguing that given the fact that each of us manifests our common interests in idiosyncratic ways, it is rational to allow each of us room to pursue our own idiosyncratic derivative interests, so long as our doing so does not offend (too greatly) against our overall cooperative moral strategies. One way of looking at what I am saying is as a rational defense of the category ‘permissible acts.’ Of course, every moral system recognizes three categories of actions: obligatory, forbidden, and permissible/optional (with the possible exception of strict utilitarianism, which does not obviously recognize obligatory and permissible as constituting separate categories of actions). It is clear how the pragmatist account here justiﬁes the ﬁrst two categories. What is seldom discussed is how an ethical system can support the third category, that of the permissible. I am arguing that the diversity of our derivative interests (that is, the idiosyncratic ways in which our shared interests are instantiated in individual agents) makes it rational for our moral system to possess a category ‘permissible,’ and that this category should extend over the various ways in which each of us engages in his or her own projects in pursuit of the satisfaction of our basic interests, as these interests manifest themselves in his or her derivative interests. Thus, what we have in common justiﬁes our labeling of certain acts as ‘obligatory’ and ‘forbidden’; the respects in which we differ justify the category ‘permissible’. It is a strength of pragmatism that it can make rational sense of this central element of our ordinary moral practice. Of course, tolerance can be taken too far. After all, what if a person has an idiosyncratic desire that is strongly opposed to our cooperative moral strategy? Is it rational for this person to pursue this desire? As we will discuss later in the chapter, as we want to see our cooperative

138

Pragmatism and Morality

strategy succeed, it is in our interest to adopt a practice of sanctions, thereby punishing people who offend against the rules of our moral practice. Thus, though pragmatism recognizes the importance of tolerance and the category of permissibility, obviously tolerance should not be taken too far. We may decide, to a certain extent, to attempt (via sanctions) to make it irrational for a person to pursue an idiosyncratic interest that is opposed to or destructive of our cooperative moral strategy. A balance must be struck, then, between promotion of our interests and undue interference in the personal projects of others. Many compromises have been suggested; a common one is that sanctions should only be applied to prevent someone from harming someone else.18 Thus, actions that are solely self-regarding fall in the sphere of permissible actions, and should not be condemned. Though all such views face certain difﬁculties, these are not difﬁculties that are faced by pragmatism in particular; rather, they are difﬁculties faced by any theory which seeks to carve out a sphere of non-interference where we may conduct our personal affairs unmolested. Thus, the pragmatist is under no special obligation to resolve the question of when private conduct may be sanctioned by society at large, and I will hence not pursue the question here.

Shared non-basic ends It is perhaps unlikely that a vast majority of people could come to share the same non-basic interest, goal, or purpose. But it is at least theoretically possible that we could come to share a non-basic end; that is, we could come to have an end that, though shared among us all, is neither a basic nor a derivative interest. There seems to be no reason why, as morality consists of a set of cooperative rational strategies, a shared end might not justify a moral principle. What does our theory say about such a possibility? Naturally, our verdict will depend on the nature of the shared end. The end might be something more or less innocuous, such as an interest in preventing the extinction of endangered species. If we shared such an end, then rationality would endorse a cooperative strategy to advance such an end. In this way, a shared non-basic end could contribute to the set of cooperative strategies constituting our moral code. But this possibility will make some nervous. ‘After all,’ they will say, ‘isn’t it theoretically possible that we could all be instilled with a wicked end, such as the goal of (say) practicing female genital mutilation? And would that mean, then, that it is morally permissible to practice FGM?’

Pragmatism, Interests, and Morality

139

Although such a possibility seems remote, it cannot be entirely ruled out. But notice one thing: we have already conceded that our ends are not neutral. That is, we can judge an end as worthy or unworthy of pursuit. The mere fact that we are somehow made to share this end does not mean that this end is acceptable or should be the basis for a cooperative strategy. But, one might ask, on what basis can we condemn such a shared end? After all, if morality is supposed to be supported by rationality, then how can we morally condemn this shared end in a non-questionbegging way? Isn’t this end itself the basis for a rationally justiﬁed strategy (namely, a strategy of promoting FGM)? The answer to this challenge is as follows: it could never be rational to adopt an end that is radically opposed to our interests. By doing so, we would be introducing an unacceptable level of rational dissonance into our cognitive apparatus. Though it might be possible to instill in the overwhelming majority of people the goal of promoting FGM, this practice is contrary to several of our interests—our interest in health, in pleasure, and so forth. Thus, the goal of promoting FGM should be suppressed. Consider the matter this way: we are operating with a means–end conception of rationality. Given this fact, it can never be rational to adopt an end that directly conﬂicts with our interests, as this would automatically mean that one or both of these (our interest or the conﬂicting non-basic end) must be frustrated. If we possessed such a conﬂicting non-basic end, it would be rational to get rid of it or the interest with which it conﬂicts. As the latter is more fundamental and important than the former, it is rational to eliminate the goal of promoting FGM. Granted, we cannot simply choose to eliminate unwanted goals, if people genuinely desire them. But if a practice is condemned and forbidden, then the desire for that practice may diminish in time, as the desire to acquire slaves or deny women the vote has diminished with time. The point of condemning a desire is not just that we wish to limit expression of the desire; we also hope that in time, the prevalence of the desire may diminish in society.

The threat of circular justiﬁcation One potentially serious problem arises for this account. It seems, initially, that whatever we identify these interests with must satisfy an important requirement: namely, these interests cannot themselves be

140

Pragmatism and Morality

‘morally loaded,’ for that would introduce an unwelcome circularity into our account of the justiﬁcation of morality. One apparent problem with a pragmatist approach to morality is that all the candidates for grounding morality (desires, preferences, primary goods, human capabilities/functionings) already seem to have moral content, and so presuppose the existence of the very morality they are supposed to justify. Doubts about the possibility of founding a moral (or political) theory on ends that are themselves morally neutral are familiar. Consider the well-known example of political liberalism. Much has been written about the problem of “liberal toleration, the idea that liberal theory must remain neutral among conceptions of the good.”19 The point of liberalism is that it does not choose among conceptions of the good, that it is neutral among conceptions of the good: individuals in a liberal society are thereby allowed to pursue their own conceptions of the good. The problem, as Henry Richardson writes, is that “a defense of liberal neutrality that is itself fundamentally neutral is . . . impossible.”20 Toleration is itself a substantive and controvertible good, not merely a value-neutral attitude one can adopt toward various conceptions of the good. Rawls faced a similar problem with his account of primary goods. These goods were supposed to be in an important sense neutral: that is, goods anyone would accept. But Rawls was gradually forced to concede that these goods embody a controversial, normatively loaded conception of the good, and therefore cannot provide a neutral foundation for a political theory. Thus, one might question whether the pragmatist can give a morally neutral account of interests, which can then serve as a foundation for morality. But it seems that initially, if morality is to have some sort of grounding (and not merely be unanalyzable and basic, as the intuitionists think), morality’s grounds must themselves be non-moral. But can this requirement be satisﬁed? For example, our desires and preferences (popular starting places for teleological conceptions of morality) are shaped by moral reasons, and so are already ‘morally infused.’ Some preferences and desires are moral, and others immoral. At the very least, they are shaped by moral norms: my preference for making money in a legal way might reﬂect my moral conviction that laws ought to be obeyed. My desire to write a ‘thank you’ note to my aunt for her nice gift reﬂects my conviction that gratitude is a virtue. Examples can be multiplied indeﬁnitely. As for interests, we have conceded that they are not morally neutral in character. There are at least two different ways that one could respond to the idea that our interests are morally loaded. The ﬁrst is to tell a story

Pragmatism, Interests, and Morality

141

(a myth, really) demonstrating how morality could arise out of nonmoral ends, and could evolve to the stage where ends become morally infused. The second is to grab the bull by the horns and argue that circular justiﬁcation is permissible in practical reasoning. Let us explore these two options.

Justiﬁcation-conferring myths Rebecca Kukla writes that “[m]ythical legitimations of authority, from the social contract to the original position, have been time-honored tools for philosophers.”21 In telling a myth, one tells how our current practice could have evolved via a rational process, and by demonstrating this process, one confers legitimacy on our current practice. For example, in telling how people can move from a State of Nature into civilization, Hobbes is not giving a history of how moral norms came about. Rather, he is offering a justiﬁcation for such norms, and the vehicle for justiﬁcation is the myth of the State of Nature. One way of defusing the threat of circularity is to tell such a myth. This myth would explain how our current moral practice—with its morally infused ends—could have arisen out of a practice of prudential reasoning with no moral presuppositions whatsoever. If we were to tell such a myth about the origins of morality, we would perhaps begin with a community of simple reasoners. The members of this community recognize certain ends to be worth pursuing: food, sex, shelter, health, continued survival, and so forth. Although they perhaps do not have a sophisticated evaluative language, they might express the value of certain ends practically: successful hunters are respected and envied, as well as being sought-after mates; whereas community members who pursue other ends, judged not to be valuable (say, sketching likenesses of animals on the walls of caves) are sneered at as time wasters.22 There need be no moral reproach implicit in this condemnation, any more than there is moral reproach implicit when a person says that someone else is stupid. (We might think it is rude to call someone stupid, but let us say that such rules, which are partly rules of etiquette and partly rules of morality, have not yet evolved in this hypothetical community.) Suppose that at ﬁrst this community’s prudential reasoning is simple: they use act-based means–end reasoning: I want X, Y is a means to X, and so I shall do Y. They do not employ the style of reasoning that I argued in Chapter 3 is so important: that is, they do not reason strategically, nor do they reason cooperatively. They do not have a moral code, or recognize morally infused interests or ends. Their interactions with

142

Pragmatism and Morality

each other are governed by desires for promoting ones self-interest (primarily in the form of avoiding retaliation) and promoting the interests of those for whom one feels affection. Imagine that into this community is born a prudential genius Jones. Jones, like everyone else, pursues these simple ends listed above. But Jones realizes that sometimes the best way to achieve such an end is indirectly, through pursuing actions that do not directly aim at this end. For example, Jones knows that more experienced hunters are better hunters, in part because they are more accurate spear-throwers. Thus, foreseeing that he will want food not only in the near future but again and again, Jones realizes that one way of ensuring that this end will be achieved in the future is to perform actions now that are not directly aimed at food-gathering: namely, practicing throwing his spear when not hunting, so that he will be a more accurate thrower when on the hunt. Thus, strategic thinking is born: throwing his spear at a knot in the tree does not directly satisfy any need of Jones’s; but Jones knows that if he assiduously practices his spear throwing, then when he does go on a hunt, he will have a better chance of success. Thus, Jones is the ﬁrst strategic thinker. His rationality is not purely act-based, but strategy-based. (His reasoning, though, is still individualistic, and not cooperative.) At ﬁrst, we may suppose that the other members of the community laugh at Jones, and make fun of his practice of ‘hunting for trees.’ But soon, Jones becomes one of the most successful hunters in the community. Other members, noting his success, begin to emulate his practice of practicing, covertly at ﬁrst, but more openly as the practice becomes more accepted. Other strategies of Jones are picked up by other members of the community, and soon, community members begin to think of their own strategies for indirectly (but more effectively) pursuing the ends they value. At this stage, we may suppose that there are no moral rules in the community. One individual might avoid harming another merely because he fears the latter’s revenge, or out of sympathy or some other bit of fellow-feeling, but the members of the community do not follow moral constraints per se. Let us suppose that Jones, our genius, makes another brilliant prudential leap. He has already realized that one’s ends can be more effectively achieved through following a strategy, rather than by reasoning always in terms of actions. But suppose Jones now realizes that a strategy is something that can be pursued by more than one person, and that these two people might ﬁnd following such a strategy mutually beneﬁcial. So, for example, while Jones realizes that it might

Pragmatism, Interests, and Morality

143

be rational for Jones to steal Smith’s spear (and thus save himself the trouble of making a new spear), Jones also realizes that if Smith follows the same reasoning, Smith might steal Jones’s, or even Green’s (of whom Jones is fond, and for whom Jones wishes only good things) spear. Jones realizes that he and Smith (and Green) could, however, follow a strategy together: neither will take any of the other’s possessions. And this is a strategy, because not only will they refrain from taking each other’s possessions now, but they will continue to refrain from doing so; on no occasion will one take a possession from another. Perhaps Jones suggests to the members of this community that they adopt a certain number of such strategies: say, that no member of the community will kill or harm another member, or take something another member of the community has made or killed, etc. The members of the community, much like the residents of Hobbes’s State of Nature, recognize that following such rules will help them pursue their ends. If one does not have to worry about one’s spear being stolen, one does not have to constantly guard it, or always carry it around. Before, the theft of one’s hunting implements meant one could not hunt until he had made new implements; and one who could not hunt could not attract a mate, etc. And of course, not eating does not contribute to health or continued survival, nor does constantly having to defend oneself from other members of one’s community. So the community members see that the ends they regard as valuable are promoted by the cooperative strategies that Jones is suggesting. Thus, moral reasoning is born, via joint commitments to respecting individual’s possessions and so forth. In time, the list of ends becomes modiﬁed. Previously, social domination had been a valued end; but as the community members become more convinced of the value of cooperation, social domination comes to be discounted, even looked upon slightly askance—even as somewhat immoral, as attempts at domination disrupt cooperation. New ends emerge from these cooperative strategies: honesty, for example, emerges as a valued end as our community members recognize that a cooperative strategy of truth telling will greatly enhance the exchange of information. Other ends such as generosity, loyalty, and so forth emerge in a similar fashion. Thus, some pre-existing ends acquire a moral value in that they become the objects of joint commitment, and cooperative strategies aimed at promoting these ends and new morally loaded ends arise. Over the years, a complex revisionary process occurs, whereby the value of various ends, and the cooperative strategies used to promote those ends, undergo various revisions in pursuit of mutual accommodation.

144

Pragmatism and Morality

The latest stage of our myth is, of course, society as it exists today, with its many morally loaded ends and complex cooperative strategies. But even this latest stage is not the ﬁnal stage. As noted earlier, the process of moral inquiry is ongoing, and we are ever experimenting with different ways of promoting our ends, and accommodating competing ends with each other. Through telling such a myth we can perhaps explain how justifying morality by appeal to interests is not question-begging. By explaining how our current moral practice—with its morally infused ends—could have arisen out of a practice of prudential reasoning with no moral presuppositions whatsoever, we might thereby justify our current practice, just as Hobbes justiﬁes morality with his myth of the State of Nature, and Rawls (to a lesser degree) uses the device of myth (agents in the Original Position) to justify the constraints of justice. However, such myth-telling is not the only way to respond to the charge of circularity. There is a more direct way: we can admit that moral justiﬁcation is circular, but argue that this constitutes no objection to morality. Let us see how such an argument would proceed.

Practical rationality and circular justiﬁcation It is generally conceded that in theoretical justiﬁcation, circular reasoning is illegitimate.23 But perhaps it is unwise to move too quickly to generalize this result to all sorts of reasoning; doing so might obscure important differences between practical and theoretical justiﬁcation. And indeed, it appears that some sorts of circularity are permissible in prudential reasoning. For example, a goal may serve to justify a practice (i.e., a set of rules) even if this goal is in large part itself justiﬁed by this very practice. Let us illustrate this point with an example. Consider in football (aka soccer) the desire that goals be scored. Though many purists are capable of enjoying the beauty of a low-scoring or scoreless game, many people (both players and viewers) enjoy the excitement of a higher-scoring game. Now, the practitioners of the game24 may decide that the aim that goals be scored is being frustrated by certain of the game’s rules (e.g., by an overly restrictive offside rule). Thus, these practitioners may therefore justify relaxation of the offside rule and the consequent production of a higher-scoring game by appeal to the aim of scoring goals. Now notice the features of this example: the aim of scoring goals is justiﬁed in part by the rules of the game, as the rules of the game specify that the game is to be won by scoring goals. However, this very aim is used to justify

Pragmatism, Interests, and Morality

145

a rule within the game. Thus, an aim justiﬁed in part by the rules of a practice can be used to justify rules within that practice. Let us examine this conclusion further. It may be that the sort of circularity allowed in practical reasoning is similar to a type of circularity that is allowed in theoretical reasoning. Let us begin by asking, ‘What are the grounds or ultimate source of (a) the justiﬁcation of our moral rules in morality and (b) the justiﬁcation of our theories in theoretical reasoning?’ In the former case, the answer defended here is our interests; ultimately, morality is justiﬁed by appeal to our interests. In the latter case, the answer seems to be observation. Though a good anti-foundationalist ought to say that observation cannot by itself justify theoretical claims,25 there is nevertheless a point in talking about observation as the ultimate source, the sine qua non, of theoretical justiﬁcation. However, do we require, in the theoretical case, that the source of justiﬁcation for our theories be itself neutral with respect to the theories it justiﬁes? Or do we allow theoretically-infused observations to justify the very theories that infuse them? Clearly, the latter is the case. Although philosophers long dreamed of a theory-neutral observation language, this dream has long been given up as impossible, and perhaps even undesirable.26 And it is not merely the case that our observations are inﬂuenced or infused by the theories they justify; rather, our observations are in part justiﬁed by these very theories! Consider, as an example, the phenomenon of St. Elmo’s ﬁre, which is a glowing region of atmospheric electricity that appears on pointed objects (church steeples, airplane wings or propellers, etc.) during thunderstorms. Now it is a commonplace that observation is theory-laden: observational judgments you make depend on the theories you hold. When I observe St. Elmo’s ﬁre, I judge that I am observing atmospheric electricity; and furthermore, I am justiﬁed in making this observational judgment. I would not be justiﬁed in judging that I am observing a ghost or some other supernatural phenomenon. What justiﬁes the former (but not the latter) judgment? The fact that I am justiﬁed in believing current theories of electricity, but I am not justiﬁed in believing theories about the existence of ghosts and other supernatural entities. And so we see that the justiﬁcatory relation between theory and observation is a two-way street: observation is the ultimate source of justiﬁcation for our theories; but these observations are themselves justiﬁed in part by the very theories that depend on observation for their justiﬁcation. Our epistemic ground is itself infused and justiﬁed by the theories that it grounds.

146

Pragmatism and Morality

We can now note that the fact that has worried us—the fact that our interests, which justify morality, are morally infused—is precisely the same phenomenon on the practical side, as it appears on the theoretical side. And as on the theoretical side this fact is taken to be puzzling, but not fatal to the very idea of theoretical justiﬁcation, it is not clear whether we should regard it as any more threatening on the practical side. So what if our interests are ‘theory-laden?’ This merely goes to show that in practical reasoning, as in theoretical reasoning, justiﬁcation between theory and its ground is a two-way street. In both cases, there is a ‘raw material’ that is shaped by our theoretical commitments. On the theoretical side, we have our existing perceptual mechanisms, but our judgmental capacities are shaped and altered by our theoretical commitments so that the very observations we make depend on the theories we hold. On the practical side, we have our basic human desires, needs, and inclinations, and these are to some extent shaped by our moral judgments about what is morally valuable. In either case, the pre-existing machinery—of desires, of observation and judgment—is shaped and formed by the theoretical commitments that this machinery serves to justify. Practical and theoretical reasoning are no different in this respect. Thus, a certain sort of circularity is permitted in prudential rationality as well as in theoretical rationality. Therefore, there is no in-principle reason why a morally loaded interest might not serve to help justify the practice of morality. A hypothetical goal of donating to charity is caused and justiﬁed by moral requirements of beneﬁcence. But that is no reason why recognition of this end cannot be used to justify a moral rule (which will be a rule of beneﬁcence) requiring people to give a certain percentage of their income to charitable causes. Because we have recognized that charity or beneﬁcence is a desirable end, it is rational for us to adopt (or to continue following) a rule requiring such giving, and it would be irrational to adopt a rule saying that no such giving was required, or a Nietzsche-esque rule stating that charitable giving was in fact immoral. And so the above discussion seems to clearly illustrate that morally loaded interests can justify moral principles.

Moral rules and exceptions We have admitted, in Chapter 3, that to completely evade the ruleworship objection, the pragmatist must adopt the notion of defeasible rules, rules that allow of exceptions. But as I stated in the introduction to the current chapter, we wish to emphasize the degree to which the

Pragmatism, Interests, and Morality

147

current account differs from utilitarianism. One might worry, though, that the admission of defeasible rules pushes the account given here too close to utilitarianism. Consider, for example, one of the moral rules discussed in the Chapter 3: R4: If someone is in need of assistance, and you can render assistance without undue cost to yourself, then you ought to do so. What, precisely, counts as an undue cost to yourself? Do we mean, ‘Unless the cost to you is greater than the beneﬁt conferred on the person you are helping?’ If so, then this directly seems like utilitarianism: we should act to maximize net beneﬁt.27 But such a conclusion is not licensed by the pragmatist foundations we have laid in this chapter. Given the foundations we have laid, morality can impose only weak constraints on us; it certainly cannot require us to treat our own interests as entirely interchangeable, one for one, with those of someone else. Let me explain. I have argued above that we have associative interests, and some interest in seeing the interests of others satisﬁed. However, we have explicitly connected the notion of an interest to the agent’s ﬂourishing, and this means that my promotion of interests will primarily involve promotion of my self-regarding interests. After all, my death in general harms my ﬂourishing more than someone else’s death harms my ﬂourishing. My good health contributes to my ﬂourishing more than someone else’s good health contributes to my ﬂourishing. We do have other-regarding (i.e., altruistic) interests, and we also share group preferences and joint commitments to promote our mutual interests, but we must recognize that our self-regarding interests predominate. (As I noted in Chapter 3, morality consists primarily of a set of negative constraints, and hence imposes much weaker constraints on action than it does if all our perfect duties were positive duties.) Given this fact, it will hardly be rational for us to adhere to a strategy that dictates that we promote others’ ﬂourishing at signiﬁcant cost to ourselves. In particular, it will be irrational to follow a strategy that demands that we be willing to promote others’ interests up to the point where the cost to us is equal to the beneﬁt conferred upon those whose interests we are promoting. And as the point of pragmatism is to show how being moral is rational (i.e., how it serves our interests), pragmatism cannot support a utilitarian conception of morality. Thus, recalling R4 above, what counts as an ‘undue cost’ to me will be small, relative to the beneﬁt I am considering conferring on another

148

Pragmatism and Morality

person. If we brieﬂy entertain the fantasy that interest satisfaction can be quantiﬁed and compared, if helping Smith push his car to the gas station will confer upon Smith 10 units of interest satisfaction, I might only be obligated to help him if the cost to me is (say) 2 units or less of interest satisfaction. Granted, by helping him I will be satisfying an other-regarding interest (and so helping others should not be viewed as imposing only costs upon oneself in terms of interest satisfaction) and at the same time acting on a collective preference for mutual aid, but as our self-regarding interests predominate, we should acknowledge that, on balance, helping others usually imposes costs upon ourselves, in terms of the (dis)satisfaction of our self-regarding interests. The pragmatist insistence that morality should not in general be too demanding accords with the actual structure of our moral practice. I argued in Chapter 3 that in general, people are allowed to pursue their self-interest in whatever manner they choose, and morality primarily consists of a negative constraint: people must not pursue their self-interest in a way that unjustly harms the interests of others. Thus, I may in general conduct my life in a way that serves my interests— I may get a job that is satisfying, pursue my own hobbies and pleasures, and so forth—and morality merely constrains my manner of pursuing these personal beneﬁts. It forbids me to get a job through lying or sabotaging the efforts of others; it forbids me to pursue gratiﬁcations that harm others without their consent; and so forth. When we think about how morality constrains our lives, we must recognize that most of our life-decisions primarily involve considerations of what will beneﬁt us: which school will provide me the best education, and maximize my professional opportunities? Which job will provide the most satisfaction? Morality primarily provides negative constraints; within these constraints, we are in general permitted to pursue our self-interest in an unfettered fashion. Of course, not all moral constraints are negative. Many moral constraints are positive: we have, for example, obligations of beneﬁcence. But the intuitive Kantian way of understanding positive and negative duties supports the idea that negative duties are stronger than positive ones. Arguably, our only perfect duties (i.e., the only duties we must constantly fulﬁll) are our negative duties.28 If this is true, then our positive duties are imperfect—we must fulﬁll them from time to time, but they are occasional constraints. And so, pragmatism—by grounding morality in a notion of ﬂourishing that is primarily self-regarding—recognizes that moral constraints are weak, and in general merely constrain our pursuit of self-interest.

Pragmatism, Interests, and Morality

149

They in no way discourage our pursuit of self-interest, or demand that we consistently sacriﬁce our self-interest in favor of the interests of others. That would not be rational, if rationality is deﬁned instrumentally, in terms of interest satisfaction; for, most of our interests are self-regarding. It makes sense, then, that if we wish to justify morality in terms of instrumental rationality, then moral constraints must play a minor role relative to the role played by the pursuit of self-interest. Flourishing is not an entirely self-regarding notion, but it is mostly self-regarding. Of course, we must not lose sight of the fact that morality is a cooperative strategy aimed at promoting not only our other-regarding interests, but also our self-regarding interests. If it were not for our other-regarding interests, we could still form collective preferences and joint commitments aimed at promoting our mutual self-interest. Indeed, I argued in Chapter 3 that not doing so would be self-defeating, and hence irrational. Thus, we should not overstate the opposition between morality and self-interest. Nevertheless, the point is to show that morality is rational; and we wish to understand rationality in terms of serving our interests. So, if morality requires excessively large sacriﬁces of selfinterest, then it is no longer clear how it is supposed to be rational for us to participate in this cooperative strategy. Thus, morality cannot demand that we make enormous sacriﬁces for the sake of morality. This conclusion helps further demonstrate the fruitfulness of the pragmatist approach developed here. One advantage of this conclusion is that our meta-ethical conclusion can help us solve various puzzles in applied ethics. For example, many who read Peter Singer’s writings on famine and afﬂuence29 object to his conclusion that people in afﬂuent nations have a moral obligation to donate until any further donation would make them worse off than the people they are trying to help. However, it is difﬁcult to ﬁnd any ﬂaw with his argument. With the analysis we have given of the source of moral obligation, we can see the ﬂaw: as we are trying to establish that moral constraints are rational constraints, the recognition that the goal of ﬂourishing directs us in large degree to promote our self-interest shows that it is simply not always rational to make enormous sacriﬁces in one’s personal welfare for the sake of others. For similar reasons, we can see why a long-standing objection against consequentialism—that consequentialism is too demanding—simply does not arise against the account offered here. According to this objection, if given a choice between (say) buying himself a $100 pair of shoes and giving $100 to charity, the thoroughgoing consequentialist should

150

Pragmatism and Morality

choose the latter option (as it will likely produce the best overall consequences), and should always make a similar choice in like cases. But again, this objection assumes that moral considerations always trump other sorts of rational considerations. The meta-ethical underpinnings of the pragmatist theory offered here deny this, and so cut off the ‘demandingness’ objection at the knees. There is at least one problematic type of case, however. There are circumstances where abiding by a cooperative moral agreement will impose substantial costs upon an individual. For example, suppose someone has an opportunity to embezzle $2 million and retire to Bermuda. Can we really say, with a straight face, that it would be irrational for him to do so? This seems to be a case where consideration of an individual’s selfinterest seems to substantially outweigh any interest he or she might have in abiding by a cooperative strategy; it seems in these cases that it is rational for the person to perform the immoral action. Does this undermine our claim that morality is justiﬁed by rationality? The pragmatist must recognize that an individual’s joint commitments are often a weak reed upon which to lean. And this is why we cannot always rely on appeal to collective preferences and joint commitments to justify continued participation in cooperative moral strategies. We understand that it is in our interest (in the interest of each of us individually and as a group) if moral constraints are in general observed. As Gauthier says, ‘Morality is a system of principles such that it is advantageous for everyone if everyone accepts and acts on it.’30 And thus, it is rational for us to put certain additional strategies in place to ensure that it is rational for individuals to obey morality. In particular, it is rational to put in place strategies of punishment. This helps ensure the survival of our mutually advantageous moral strategy by making it irrational to violate this strategy in cases where (absent punishment) it would not be rational to abide by the cooperative strategy. In other words, there are cases where, if there were no such thing as punishment, it would be rational to violate morality, because considerations of self-interest are sufﬁcient to outweigh the rationality of abiding by the cooperative moral agreement. As we know that such cases arise, we must give the individual contemplating moral violation additional considerations that count in favor of abiding by the cooperative strategy. These additional considerations take the form of punishments that are inﬂicted upon violators. In this way, our other-regarding interests are supplemented by self-regarding interests, and (hopefully) these

Pragmatism, Interests, and Morality

151

additional considerations make it rational to abide by the cooperative moral agreement.

Conclusion In this chapter, we have examined the interests which underlie and justify our moral (and epistemic) practice. We have seen that although hedonistic and desire fulﬁllment theories of welfare do not appear successful, we can use an objective-list theory of welfare to delineate our interests. We have further seen the connection between basic interests and derivative ones. Finally, we have seen that although one might worry that justifying morality through morally-loaded interests may be circular, there are at least two strategies for rendering this circularity non-vicious. We have mentioned that it might be necessary to apply sanctions to people in order to sustain our cooperative moral strategy. Sanctioning agents presupposes a practice of holding people responsible for their actions. Pragmatism has the advantage of being able not only to justify and illuminate our moral practice, but also to justify our practice of applying sanctions and ascribing freedom and responsibility to agents and their actions. Let us turn to a pragmatist account of freedom and responsibility in the next chapter.

6 Pragmatism, Freedom, and Responsibility

An account of the justiﬁcation of morality presupposes a defense of the existence of free will and the viability of our practice of holding people morally responsible for their actions.1 Most meta-ethicists are content merely to justify morality, and leave the task of defending a viable account of free will to the metaphysicists. Of course, there are exceptions, the most famous being (of course) Aristotle. Aristotle recognized that morality (or at least an account of virtuous behavior, as arguably Aristotle was not providing an account of morality per se) and free will go hand in hand, and so supplemented his account of the virtues with his famous discussion of voluntariness and involuntariness in Book Three of the Nicomachean Ethics. The virtue of the pragmatist approach is not merely that it can provide an account of free will. The pragmatist approach can present a uniﬁed account of morality and free will, providing a consistent theoretical underpinning for both types of discourse (and for epistemology, as well). This helps demonstrate the fruitfulness of the pragmatist approach to normative discourse. What will a pragmatist justiﬁcation of free will look like? Obviously, one might argue that we should believe in free will because it is pragmatically useful to do so. This approach has been defended by John Fowler Corbin,2 Saul Smilansky,3 and others. While conceding that determinism is true (and that incompatibilism might well be true), this approach maintains that we are nevertheless pragmatically justiﬁed in believing that we are free and responsible, because the consequences of not believing so are presumed to be quite bad. It should be obvious by now that there is no room for this type of justiﬁcation in the version of pragmatism defended in this book. One of the primary theses of this book is that if pragmatism is to be viable, 152

Pragmatism, Freedom, and Responsibility 153

it must move away from this conﬂation of practical and theoretical justiﬁcation in the direction of a more sophisticated account that makes room for theoretical objectivity. I wish to concede that determinism is true; and I also wish to give a pragmatist justiﬁcation of our practice of labeling some actions as free, others as unfree, and holding people morally responsible for their actions. However, we should not rush to embrace a type of pragmatism (such as the ones described in the previous paragraph) that does not allow for a robust distinction between practical and theoretical justiﬁcation. The goal of a pragmatist justiﬁcation of free-will discourse should be to show the purpose that such a discourse serves and to justify the discourse—by appeal to this purpose—without conﬂating practical and theoretical justiﬁcation. The emergent conception of free will must be theoretically respectable—it must withstand philosophical scrutiny. Again, in giving a pragmatist account of free-will discourse, we must ask, ‘What role do free-will ascriptions play in our language? What interests do these ascriptions serve?’ As it turns out, free-will discourse serves our interests indirectly. Free will serves our interests by serving as an adjunct to our moral discourse: free-will judgments help enforce our cooperative moral strategy by identifying those actions that ought to be sanctioned or rewarded, thereby reinforcing moral behavior. A pragmatist examination of the free-will debate leads to another important conclusion: that people are responsible for their actions is a necessary conclusion, and cannot be avoided merely by citing incompatibilist arguments. Responsibility is inescapable. This conclusion will emerge as follows: the criterion that a pragmatist uses to decide which actions are free, and which ones are not, turns out to be the same criterion a hard determinist would use to decide which actions should be subject to sanctions or rewards, and which actions should not. We will thus be able to argue that by applying pragmatist reasoning, hard determinists will end up being compatibilists, subscribing to the pragmatist version of compatibilism. That is, hard determinism will end up identical to a pragmatist conception of compatibilism. Hence, pragmatism allows us to see that from every perspective in the free-will debate—libertarian, compatibilist, and hard determinist—humans are free. There is a powerful rhetorical reason for pursuing this strategy. One might worry that though it might be pragmatically useful to treat people as free and hold them responsible for their actions, doing so for pragmatic reasons is not fair or just, especially given the truth of determinism. But the strategy I am pursuing in this chapter is designed to alleviate this worry. We will see that even if you reject the fairness or

154

Pragmatism and Morality

justice of pragmatically justiﬁed punishment and responsibility ascriptions, and resort to a version of hard determinism, you will still end up punishing and holding responsible in a way that mirrors the pragmatistcompatibilist account. Thus, the pragmatist can vindicate free-will and responsibility ascriptions by demonstrating that they are inescapable: no matter which position you endorse in the free-will debate (libertarianism, compatibilism, or hard determinism), you will end up holding people responsible for their actions and treating them as free. What shape should our pragmatist defense of free will take? Obviously, one might defend a libertarian approach. This, however, does not seem promising. Libertarianism is ailing. Once the preferred account of free will, it has now fallen into considerable disfavor. Most philosophers are suspicious of libertarian accounts of agent causation and unmoved movers, and are convinced that the opposite of determinism is not freedom, but indeterminism. I think that libertarianism’s current low state is entirely justiﬁed. I will not argue for this conclusion, as it is a conclusion that has been adequately argued elsewhere. This leaves only hard determinism and compatibilism on the ﬁeld. Although hard determinism is scarcely more popular than libertarianism, many philosophers seem to reject it not because of its philosophical implausibility, but because they fear the consequences of its being true. Metaphysicians’ inability to come up with a satisfactory compatibilist account is thus a source of worry, and might lead one to fear that hard determinism is true. Ideally, then, a pragmatist discussion of free will will accomplish two things: First, it will give us some deﬁnition or criterion for ‘free action.’ Second, it will vindicate compatibilism (e.g., by defending it against incompatibilist arguments). This chapter aims to do both. First, we will argue that the actions a compatibilist will punish or reward tend to ﬁt a certain criterion: they tend to have a particular feature. This criterion will not amount to a deﬁnition of ‘free action,’ but is instead a defeasible criterion licensing the following defeasible inference: if an action falls under this criterion, then, prima facie, the action is free. We will show how this criterion dovetails with pragmatist considerations concerning freedom and morality – how it is a good criterion for a pragmatist conception of free will. Second, we will turn to the task of showing that pragmatism can offer a defense of compatibilism. We will use pragmatist tools to defeat hard determinism. (As noted above, I am not going to discuss libertarianism, preferring instead to assume that determinism is true.) We will accomplish this task in a novel manner: I will show that hard determinists will end up punishing or rewarding action (and inaction), and the actions

Pragmatism, Freedom, and Responsibility 155

they punish and reward will also tend to fall under the pragmatist freewill criterion. This will happen because a hard determinist community will reason about punishment and reward in an essentially pragmatist manner, and therefore will end up endorsing the same conclusions as the pragmatist compatibilist. Thus, we will see that the hard determinist endorses the following thesis: Sanctions thesis (S): Some human actions are sanctionable. Also we will see that what determinists call ‘sanctionable actions’ and what compatibilists call ‘free actions’ have the same extension, that is, they pick out the same set of actions. After (hopefully) accomplishing these two things, I will respond to some objections. At this point, it will emerge that not only will the same actions be regarded as sanctionable by the hard determinist and free by the compatibilist (so that the two theories are equivalent at the practical level), but we ought to translate ‘sanctionable’ as ‘free’. Thus, the hard determinists end up endorsing the compatibilist thesis (some human actions are sanctionable/free). Thus, hard determinism is quite literally a form of compatibilism. Therefore, assuming that libertarianism is false, compatibilism is the only remaining account of free will, and must be the one we endorse. If successful, this will again show that pragmatism is a fruitful way of approaching traditional philosophical problems and dilemmas.

Free action Let me begin by expressing some doubt regarding metaphysicians’ attempts to give a one-size-ﬁts-all deﬁnition of such concepts as free will and personal identity. First, it is not clear whether our concepts have sufﬁciently sharp boundaries to admit of such explicit deﬁnition. Consider Wittgenstein’s point that there is nothing had in common by all instances of our application of the term ‘game.’4 The terms we use generally cannot be deﬁned in a way that captures all correct applications of the term. Second, it is far from clear that we need to give such explicit deﬁnitions to our terms. It might ease application, but it is not clear whether the beneﬁts of this outweigh the disadvantages of having to use the resulting Procrustean concepts, with their awkward, contrary-topractice contours and counterintuitive results. Thus, I do not think we need to worry if we cannot give a set of explicit rules for the application of terms such as ‘free action’ (e.g., ‘An action A is free if and only if . . . .’).

156

Pragmatism and Morality

A related lesson is gaining prominent attention in ethics, where particularists of various stripes argue against the existence of exceptionless moral principles.5 We should thus expect that we will not be able to give explicit and exhaustive deﬁnitions for most of our concepts. Instead, we may only be able to give criteria, or rules for the application of a concept, which will license defeasible inferences of the form ‘Ceteris paribus, if action A has feature F, then action A is free.’6 In other words, there might be times when correct application of a term violates the explicit rule we have provided for the application of this term (much as the duck-billed platypus violates the rule that the term ‘mammal’ is applied only to animals that give birth to live young). With this in mind, I now turn my attention to providing a criterion for the term ‘free action.’ Consider the case of a child drowning in a shallow pool. Person A walks by and, deciding he does not want to ruin his slacks, declines to wade into the pool and rescue the child. We condemn his inaction as evil. Consider, though, person B, who is chained to a nearby tree, and hence is helpless to rescue the child. No one would blame person B for his inaction. What is the difference in the two scenarios that causes us to condemn person A and not person B? It cannot be that person A could have done otherwise, whereas person B could not have. Given the set of physical laws P and the circumstances C immediately prior to A’s inaction, there is only one behavior that could have been exhibited by person A: inaction. (This is actually a matter of debate, but for rhetorical reasons, I wish to concede this point to the incompatibilist, for I believe that we can still give a successful pragmatist defense of free will even after making this concession regarding determinism and necessitation.) In other words, there are no possible worlds in which, given the same circumstances and physical laws, person A would have done otherwise than he actually did. So A could not have done otherwise, any more than B could have done otherwise. What, then, is the difference? The difference seems to be that A’s behavior is the sort of behavior that could be altered through education and application of sanctions (or threat of sanctions). In other words, if we could somehow get A to grasp the suffering of the child’s family, or if we could offer him praise for rescuing the child, or punishment for failing to rescue the child, then his psychology might be altered in a way such as to cause him to rescue any future drowning children he might encounter (or the mere threat of punishment might induce him to rescue the present child). However, no amount of education, praise, or punishment could alter the behavior of the man chained to the tree. In other words, ‘Failure to rescue a child

Pragmatism, Freedom, and Responsibility 157

by reason of being chained to a tree’ is not the sort of behavior that can be altered by education or (threat of) sanctions; ‘Failure to rescue a child by reason of not wanting to ruin your slacks’ is a behavior that can be reshaped through appropriate means. Or consider the case of a weak-willed man who often at parties drinks a third glass of wine when his limit is two. His behavior is the result of an internal struggle between his moral desires and his desire to drink more alcohol. We could change his behavior by any number of means. We could show him statistics about how alcohol impairs one’s ability to drive, or show him grisly car accident photos. Or we could ﬁne him for driving under the inﬂuence. Or we could, as a group, ostracize him until he stops drinking and driving. Any of these courses of action has, as a possible outcome, alteration of the man’s future behavior. His is the type of behavior that can be reshaped in this way. However, punishing the man chained to the tree for not rescuing the child will not affect the chances of his rescuing a child when the next time chained to a tree in full sight of a drowning child. His is not the type of behavior that is amenable to such alteration. Thus, we hold the incontinent man responsible for his actions, but not the chained man. Of course, the fact that a certain type of action or inaction might be altered by sanctions and rewards does not mean that that action should be punished; it just means that the person is responsible for this action, good or evil. Consider person C, a wheelchair-bound quadriplegic, who observes the child drowning and does not enter the water. Perhaps some extreme form of social sanctioning could alter the behavior of person C, causing her to enter the water. But any such attempt to rescue the child would lead only to the death of C as well, and would beneﬁt the child in no way. So if we do label C’s action as free, then we ought to praise her inaction, for entering the water would surely have been foolhardy. Similarly, a healthy person who entered the water and saved the child is to be praised: although her action is the sort that is susceptible to ‘engineering,’ it is also the sort of action that we want to encourage. If this same person were being remote-controlled by a mad scientist, though, we would not praise her, because praising her would have no effect on her behavior in future cases of being under control by a mad scientist. We might praise the scientist, though, for directing this person to rescue the child (although we might instead punish the scientist for placing people under remote control; we will return to the question of the value of freedom later in the chapter). It thus appears that free actions share a common feature: a behavior is free if and only if it is of the type of behavior that is amenable to

158

Pragmatism and Morality

alteration by reward, punishment, retraining, etc. Application of moral praise and blame is part of this ‘moral engineering:’ it reinforces desirable behavior and provides a disincentive for undesirable behavior. This criterion seems to neatly capture the contours of the concept of freedom (and hence responsibility). Considerations distinct from this criterion might justify us in applying the term ‘free’ in a case where the criterion is violated, or in withholding application of the term in a case where the criterion is satisﬁed. But this is the way it goes with defeasible criteria and entailments. This criterion meshes nicely with pragmatism. In our justiﬁcation of morality, I argued that morality is justiﬁed as a set of cooperative strategies for promoting our ends. Free-will discourse supports these strategies: by identifying which actions are amenable to alteration through punishments and rewards, free will serves the end of causing behavior to conform to our moral norms. The pragmatist recognizes that morality serves our interests, and that our moral practice can in turn be advanced successfully by instituting a practice of praise and blame, punishment and reward. As is the case with morality, pragmatism can offer both a justiﬁcation for a practice and guidance on the content of that practice. Pragmatism gives us a justiﬁcation for the practice of free will-ascriptions: it serves as a valuable adjunct to our moral practice, which is itself justiﬁed by appeal to our interests. Further, pragmatism can give us guidance on content by justifying our criterion for free action: it recognizes that as the success of morality requires conformity with our cooperative moral strategy, it makes sense to punish and reward all and only actions that belong to the type of action that is amenable to alteration or reinforcement through such punishment and reward. Why do we say that the pragmatist will punish the type of action that is amenable to alteration by punishment? It is acknowledged that in individual cases, threat of punishment might not cause a person to refrain from performing an immoral action. But the goal of enforcing communitywide compliance with our cooperative moral strategies requires that we punish these actions anyhow, so long as they belong to the type of action whose incidence can be reduced through sanctions (or, in the case of good actions, belong to the type of action whose incidence can be increased by reward). Consider the matter this way: if a person performs an immoral action, despite the threat of punishment, then there is a sense in which this person’s behavior was not amenable to alteration by threat of sanctions.

Pragmatism, Freedom, and Responsibility 159

But if we allow this fact to prevent us from administering sanctions, then no one would ever be sanctioned, and any bad behavior would by deﬁnition be behavior that was not amenable to alteration by the threat of sanctions. Thus, sanctions can never be applied, and there would not be the practice of sanctioning bad behavior. But the whole point of the pragmatist conception of responsibility is that sanctions are threatened (and applied) to increase conformity with morality. And so instead of focusing on individual actions, and on whether this token action is amenable to alteration by sanctioning, we must ask whether the token action falls under an action type that is susceptible to alteration by sanctions or threat of sanctions. So, for example, if someone robs a grocery store while of sound mind, obviously the threat of sanctions was not capable of altering this person’s behavior (as it did not alter the person’s behavior); but we must sanction the person anyhow, because the type of behavior displayed (robbing a grocery store while of sound mind) is the type of behavior whose incidence can be reduced by the application of sanctions to violators; and even if punishing this particular robber does not deter other criminals, punishing all tokens of this type of behavior is part of a strategy of behavior correction which has, as its goal and outcome, the reduction of anti-social activity. Thus, focusing on action types rather than tokens is a consequence of following strategies rather than calculating the pragmatic consequences of individual actions. The same considerations apply to good actions, the type of actions we want to encourage. Suppose Smith performs a good action. It may be the case that Smith would have performed this action even if she would not have been praised or rewarded for her performance. Thus, the promise or reward may have had no effect on her behavior, just as the threat of punishment did not have an effect on the behavior of our grocery store robber. But if her behavior is the type of behavior we want to encourage, and (crucially) if it is the type of behavior whose incidence is increased by praise and rewards, then we should apply positive sanctions to Smith. Thus, if we want to alter a community’s behavior, we must punish actions that fall under a certain type and reward actions that fall under a certain other type. If punishment is to have the desired effect, then potential offenders must know that any behavior of type N will open them to sanctions. If sanctions are unevenly applied (say, only 50% of actions of type N are regarded as punishable), then this will alter the rational calculations of a person contemplating perpetrating an offense of type N. So again, we must follow a strategy of punishing all actions that fall under a certain type (and rewarding all actions that fall under a certain other type). The focus on token actions is self-defeating,

160

Pragmatism and Morality

in the same way that focusing on the rationality of individual actions (without regard for strategic considerations) is self-defeating. This, in outline, is the pragmatist conception of freedom. As I said in the introduction to the chapter, however, the pragmatist is in a position to give a unique vindication of compatibilism: the pragmatist can show that hard determinism turns out to be equivalent to the pragmatist conception of compatibilism. Again, this is because hard determinist community will reason about punishment and reward in an essentially pragmatist manner. Hard determinism is a form of compatibilism; and hence, free will is philosophically inescapable.

Hard determinism and freedom On their face, compatibilism and hard determinism are very different. Of course, compatibilists and hard determinists agree that the thesis of determinism is true: Determinist thesis (D): Every event (including human actions) has a cause, and the chain of causes leading to any given action by an agent extends back in time to some point before the agent was born.7 Eventually, this chain of causes becomes a question for cosmologists, not for philosophers of free will. But this is where the similarity between compatibilism and hard determinism is alleged to end. The two theories disagree about the truth of a crucial thesis, the compatibilist thesis: Compatibilist thesis (C): Determinism is compatible with free will and moral responsibility. Compatibilists, who endorse both C and the thesis of determinism, also endorse the freedom thesis: Freedom thesis (F): Some human actions are free. Hard determinists, on the other hand, endorse D, but reject C, and consequently must also reject F. They realize that humans have a ﬂourishing practice of labeling acts as free and unfree, and granting and withholding ascriptions of moral responsibility, but they think this practice is unjustiﬁed.

Pragmatism, Freedom, and Responsibility 161

However, I believe we can show that hard determinism actually turns out to be a form of compatibilism. We will demonstrate this by showing that hard determinist community will reason about punishment and reward in an essentially pragmatist manner. Let us look at how a community of hard determinists might apply praise and blame in a way that follows the contours of our pragmatist criterion for freedom. In order to defend my thesis, I will tell a fable about a community with a particularly active and committed intellectual life. Let us imagine a community of philosophically minded citizens who are not afraid to bite the bullet in their quest to live life according to their philosophical convictions. We can further imagine that this community, embracing both determinism and incompatibilism, decides to jettison the term ‘free will’ and quit applying moral praise and blame. Suppose further that the worst fears of the foes of hard determinism are realized: these citizens, freed of moral responsibility and without threat of punishment to keep them in line, cease to observe the constraints of conventional morality and begin to murder, steal, rape, and pillage at will. The town elders, seeing that this is a disaster, which is reducing the quality of life for everyone, decide that even if free will and moral responsibility do not exist there needs to be a certain amount of social engineering to keep society well-ordered and smooth-running. They thus decide to punish actions which hinder the well-ordering of society (murder, theft, etc.) and reward actions which promote order in society (honesty, benevolence, etc.). Immediately, we can see the afﬁnity this reasoning has with the sort of pragmatism outlined in this book: a practice (in this case, a practice of punishing and rewarding) is instituted for the reason of promoting a particular end. This practice works reasonably well; the incidence of lying, murder, theft, and so on decrease dramatically. There are, however, a few cases in which this strategy does not work. Punishing people who fail to rescue drowning children because they are chained to a tree has led to no decrease in the incidence of inaction in similar circumstances. The elders meet again and decide that such behavior simply is not the sort of behavior that is amenable to alteration by punishment. Further, as punishing these people is costly, and yields no beneﬁts to society (as this punishment does not alter future behavior of people chained to trees), it is decided that such behavior should no longer be punished. Thus, the elders decree the following: a behavior is sanctionable if and only if it belongs to type of behavior that is amenable to alteration by reward, punishment, retraining, etc. The elders are pleased with this decree, because it has some pleasing results. Whereas before they were rewarding people who under remote control from mad scientists rescued

162

Pragmatism and Morality

drowning children, they are no longer obliged to reward such people. One of the elders points out that such rewards were an unnecessary expense, as they had no effect on whether the person in question performed future rescues while under remote control. Rewarding the mad scientists, however, does seem to have a positive effect on the instance of remote-controlled life-saving (although the scientists are later punished when it is discovered that, in general, the people of the community do not want to be remote-controlled—this is a result that later turns out to be important).8 Thus, the hard determinists see the value of socializing people to certain sorts of behavior. They further realize that the application of praise and blame is a much more effective method of social control than a Hobbesian leviathan. So they encourage each other to resume labeling acts as right and wrong, punishing those who commit wrong acts and rewarding those who commit right acts. Again, notice that the reasoning being applied here by the hard determinist is essentially pragmatist in nature. The pragmatist wants to promote certain strategies because conformity with these strategies promotes our interests. The hard determinists also conclude that the members of the community have certain interests to promote, and these interests are best promoted by instituting a system of punishments and rewards. But this is the same reasoning as was applied by the pragmatist—our interests are best served by instituting a practice of punishment and reward, and in general, the actions that should be punished or rewarded are those actions of a type amenable to alternation by such punishment or reward. And again, for reasons outlined above, for a practice of sanctions to be effective, it must prescribe and proscribe types or classes of actions. The community’s policy must be, in this sense, strategic. To approach the issue from a slightly different direction, in the harddeterminist community an action must be sanctionable before sanctions are warranted, so only those actions that ﬁt the given criterion for sanctionability are subjected to praise or blame. But this will have the result that our hard determinists will become compatibilists again. Consider the similarity between the hard-determinist thesis and the compatibilist thesis. The compatibilist thesis is the following: A behavior is free if and only if it is of the type of behavior that is amenable to alteration by (threat of) reward, punishment, retraining, etc.

Pragmatism, Freedom, and Responsibility 163

The hard-determinist thesis is this: A behavior is sanctionable if and only if it is of the type of behavior that is amenable to alteration by (threat of) reward, punishment, retraining, etc. Eventually, then, the behavior of our hard determinists becomes quite similar to that of our compatibilists. If we were to ﬁnd ourselves among these hard determinists, and had to write a translation manual, it is not beyond the realm of possibility that we would translate the term ‘sanctionable’ as ‘free’ (although I will return to this point anon).9 Hard determinism is beginning to look a lot less like a brave new world, and more like business as usual. The reason for this seems again to be the pragmatist character of the hard determinists’ reasoning about punishment and reward. The thesis generalized It seems likely that hard determinism and compatibilism will coincide regardless of what theory of compatibilism you endorse. That is, the account of free will that is most rational for the compatibilist to endorse will likely be the one that is best for the hard determinists to adopt. What makes this likely is that what is at stake in the free-will debate is responsibility, typically moral responsibility. Acts that are free are acts for which we are responsible; and we are not responsible for unfree acts.10 If you are responsible for an action, then you are the one praised or blamed for this action. Praise and blame, in turn, serve to reinforce behavior we value and discourage behavior we do not value. But this is precisely why the hard determinist is distributing praise and blame: to encourage certain behaviors and discourage others. So it appears that the hard determinist and the compatibilist will end up sanctioning the same behavior. Consider this: the compatibilist will want to punish those free acts that are wrong, or which exhibit traits that we do not value, and reward those that are right, or which exhibit traits we value. The hard determinist will want to select those acts for which we are responsible, and either reward or punish such acts, depending on the hard determinist’s system of values. As free acts and acts for which we are responsible are the same, the compatibilist and the hard determinist will end up sanctioning the same behaviors. (This assumes, of course, that the hard-determinist community has the same interests and values as the compatibilist community. We will return to this point later.)

164

Pragmatism and Morality

To state things another way, if the compatibilist is a pragmatist about free-will talk, then the compatibilist and the hard determinist will both be approaching the free-will debate in exactly the same manner: both will recognize that there are certain legitimate ends (such as causing behavior to conform to certain moral norms, which are themselves justiﬁed pragmatically by appeal to peaceful and fruitful coexistence, etc.) that can be advanced successfully by instituting a practice of praise and blame, punishment and reward. There is no reason in particular to think that the practice of praise and blame that serves these legitimate ends best in the compatibilist community would be different from the one that serves these ends best in the hard-determinist community (although we will discuss this point further); after all, these are the same ends being served by a pragmatically justiﬁed practice of free-will ascriptions. Consider the matter this way: for a pragmatist, varieties of normative discourse are structured so as to serve our interests. Moral norms serve our interest in peaceful and fruitful coexistence; epistemic norms serve our interest in truth, explanation, instrumental control over the environment, and so forth. There is no reason why the hard-determinist community should have different interests from ours—they, too, want to live together peacefully and fruitfully; they, too, have an interest in understanding, explaining, and controlling the world around them; they, too, have an interest (arising out of sympathy) in seeing their friends and neighbors ﬂourish. They will even have an interest in liberty: though they might reject the metaphysical notion of freedom, they will not want to be deprived of political and social liberty, which is the right to pursue their interests and preferences more or less free from coercion by others.11 If they have the same interests as ours, then we have no reason to suppose that the varieties of pragmatically justiﬁed discourse (normative, epistemic, freedom/sanctionability, etc.) will differ between the two communities. Both communities will participate in the same pragmatically justiﬁed (but not necessarily utilitarian)12 practices. A compelling argument for the thesis under discussion is this: the compatibilist wants to hold us responsible for voluntary actions. Voluntary actions are, roughly, those that are chosen by us in the absence of coercion, etc. But actions that are the subject of choice are precisely those actions that can be encouraged or discouraged through the application of sanctions or rewards—that is, if we choose (and our choice is determined), then we can be given external incentives to choose differently. That is the outcome of punishment and reward. Thus, these voluntary actions are precisely those actions that are candidates for

Pragmatism, Freedom, and Responsibility 165

sanctioning in the hard-determinist community. At the same time, the compatibilist will not punish actions that are not under the agent’s voluntary control—say, acts that are coerced or performed under the inﬂuence of hypnosis. But these same involuntary actions are just the actions that are not amenable to correction via sanctioning, so the hard determinist will not sanction these actions. The end result is that the hard determinist and the compatibilist will end up sanctioning the same set of actions. Thus, it seems likely that a successful compatibilist account will have contours similar to the contours endorsed by the community of hard determinists. So it seems that a true theory of which acts we are responsible for (i.e., which acts are free) will serve equally well as a compatibilist account of freedom and responsibility and a hard-determinist account of freedom and responsibility. Thus, the result I am defending does not depend particularly on the account of free will I sketched at the beginning of the chapter. Actions that are chosen voluntarily (according to the compatibilist’s criterion) are precisely those actions that are subject to external incentivizing in the form of sanctions, either positive or negative. If an action is voluntarily chosen, then it is the type of action that can be encouraged or discouraged by sanctions: sanctions can encourage to choose differently the next time (or to choose the same way again, if the action in question was a desirable one). And so ‘voluntary actions’ (as spoken by the compatibilist) will pick out the same set of actions as ‘sanctionable actions’ (as spoken by the hard determinist). Let us approach the matter from yet another direction: the point of punishment, from both the pragmatist and hard-determinist perspectives, is to give people an additional reason to adhere to the set of behaviors we ﬁnd mutually beneﬁcial. The set of behaviors it is rational for us to want the members of our community to subscribe to are (as I argued in Chapter 3) the cooperative strategies that aim at promoting our interests. There is no reason to suppose that fundamental interests would diverge between a hard-determinist and a compatibilist society.13 Thus, the rational cooperative strategy (which is, in essence, the moral practice) will be the same in both communities. Further, both communities want to inﬂuence people’s behavior to encourage them to behave in accordance with this cooperative strategy. Both the compatibilist and the hard determinist recognize that it only makes sense to sanction certain actions. But which actions? The hard determinist and the pragmatist approach this issue by asking the same question: which actions should we sanction if our goal is to encourage compliance with our rational strategy? I have argued above what I think the answer to

166

Pragmatism and Morality

this question is; and we have seen that there is no reason to suppose the hard determinist and pragmatist will answer this question differently. Why should they? Their goal is the same: to encourage participation in the cooperative strategy which is aimed at promoting our interests. We can see, then, that the sanctioning strategy is the same for both the hard determinist and the pragmatist. Of course, a critic will not be satisﬁed with these few brief comments. She will have two related sets of objections: First, the hard determinist’s practice will diverge from the pragmatist’s; second, a pragmatically justiﬁed practice of free will will be unsatisfactory—for example, it will allow punishment of the innocent, if this serves a useful purpose. It is important for the argument of this chapter to see that these two sets of objections as related. I have argued that the hard determinist and the pragmatist will approach the question of free will in the same way: both will regard the practice of making free-will ascriptions (and of punishing and rewarding in accordance with these ascriptions) as justiﬁed by pragmatic reasons; both groups will adopt such a practice because the practice serves a useful end. The objection against the hard determinists is that as they are adopting sanctioning practices for merely pragmatic reasons, the resulting practice will not conform to our common-sense conception of the practice of holding people responsible for their actions. For example, the hard determinist might punish innocent people, if doing so is useful, even though our common-sense practice of ascribing responsibility, punishing, etc., forbids this. Notice, though, that this is the same line of thinking that motivates the objection to a pragmatist conception of free will and responsibility: if the practice is pragmatically justiﬁed, will it have the same shape as our current, common-sense practice of holding people responsible for their actions, punishing and rewarding them, and so on? So responding to objections against the pragmatically justiﬁed hard-determinist practice of sanctioning certain actions is a fortiori responding to these same objections against a pragmatically justiﬁed compatibilist practice of punishing and rewarding certain actions. Let us see what sort of objections might be raised against these practices.

Objections We will now turn to several objections against the idea that a practice of free will and moral responsibility can be justiﬁed pragmatically. I will argue that these objections do not disprove my general thesis, and that

Pragmatism, Freedom, and Responsibility 167

a pragmatically justiﬁed practice of free-will ascriptions (whether it is advanced by the hard determinist or the pragmatist compatibilist) will not differ from a standard compatibilist account of free will. Responding to these objections will give me further opportunity to demonstrate that the hard-determinist and compatibilist communities will behave the same. It will also emerge that ‘sanctionable’ and ‘free’ mean the same thing; this will allow us to conclude that hard determinism genuinely is a form of compatibilism. For ease of phrasing, I will refer throughout the rest of the chapter to anyone who wants to justify a practice of free-will or sanctionability ascriptions pragmatically as a ‘pragmatist,’ whether this person is a hard determinist or a pragmatist compatibilist. As we are comparing a pragmatically justiﬁed practice of free-will (or sanctionability) ascriptions to the views a typical non-pragmatist compatibilist might hold, I will contrast pragmatism (the free-will theory held by the pragmatist) with what I will call ‘standard compatibilism.’ This term should merely be taken to refer to a standard non-pragmatist compatibilist theory.

Punishing the innocent? A persistent worry about justifying punishment pragmatically is that it seems to follow that if it would serve some social purpose to ‘punish’ a person who had committed no crime, then this would be permissible, according to the pragmatist. It seems clear, though, that one ought not punish an innocent person if he or she has committed no a crime, even if punishing the person so might serve some societal purpose. However, it is not clear whether the pragmatist must allow punishment of the innocent, if doing so would serve a pragmatic purpose. As we saw in Chapter 2, to say that one justiﬁes a practice (such as the practice of retributive punishment) pragmatically is not to say that one can appeal to these pragmatic considerations to justify individual actions. The reasons that justify the practice are not allowed to ‘trickle down’ to the level of individual acts; the practice ceases to serve its purpose if that is allowed. This is a fact that was seen clearly by Rawls in his article ‘Two Concepts of Rules.’ In this article, Rawls is trying to argue that rule utilitarianism is superior to act utilitarianism. With the latter, utilitarian considerations justify individual actions. However, as allowing utilitarian considerations to justify individual actions leads to free-rider problems and the like, act utilitarianism fails to maximize utility. With rule utilitarianism, however, utilitarian considerations justify a practice (consisting of a set of moral rules), and individual actions are justiﬁed

168

Pragmatism and Morality

by appeal to these rules, and not by appeal to utilitarian considerations. As Rawls writes, There are obvious utilitarian advantages in having a practice which denies to the promisor, as a defense, any general appeal to the utilitarian principle in accordance with which the practice itself may be justiﬁed. There is nothing contradictory, or surprising, in this: utilitarian (or aesthetic) reasons might properly be given in arguing that the game of chess, or baseball, is satisfactory just as it is, or in arguing that it should be changed in various respects, but a player in a game cannot properly appeal to such considerations as reasons for his making one move rather than another. It is a mistake to think that if the practice is justiﬁed on utilitarian grounds then the promisor must have complete liberty to use utilitarian arguments to decide whether or not to keep his promise. The practice forbids this general defense; and it is a purpose of the practice to do this. Therefore what the above arguments [against utilitarianism] presuppose—the idea that if the utilitarian view is accepted then the promisor is bound if, and only if, the application of the utilitarian principle to his own case shows that keeping it is best on the whole—is false. The promisor is bound because he promised: weighing the case on its merits is not open to him.14 So even if we acknowledge that pragmatic considerations justify a practice, we are not forced to admit that these same considerations can be appealed to in justifying individual acts within the practice, any more than we have to concede that a chess or basketball player can appeal to pragmatic or utilitarian considerations to justify a particular play. Thus, the objection that pragmatically justifying the practice of punishment will lead to the punishment of innocents fails to recognize the difference between the pragmatic justiﬁcation of a practice and the pragmatic justiﬁcation of an individual action. The former is permitted, but the latter is forbidden, for allowing the reasons that justiﬁed the practice to seep into the practice and justify individual actions causes the practice to collapse, rendering it incapable of serving its purpose.15 Retributive punishment One alleged difference between the pragmatist and the standard compatibilist is that though the standard compatibilist might endorse purely retributive (as opposed to corrective) punishment, the pragmatist will not employ this form of punishment. After all, the pragmatist punishes

Pragmatism, Freedom, and Responsibility 169

to alter future behavior, or to present a deterrent to other would-be miscreants, but purely retributive punishment seems an unnecessary expense for the community of pragmatists. The pragmatist might endorse non-corrective punishment to stand as an example to others, and thereby act as a deterrent, but many standard compatibilists will say that evil deserves punishment, even if no one ever knows about the punishment (which would, of course, preclude the possibility of this punishment serving as a deterring example). Evil actions, they will say, simply deserve to be paid back. There are two possible responses to this attempt to distinguish between pragmatism and standard compatibilism. First, one might claim that failure to endorse the practice of purely retributive punishment hardly puts one outside of the standard compatibilist community. Many claim that retributive punishment panders to the lowest of human impulses, the desire for revenge, and that it has no place in civilized society. Punishment should be purely corrective, these people will argue; non-corrective punishment is pointless inﬂiction of suffering. Indeed, we often seem willing to give a second chance to those who have turned their lives around, or who seem genuinely remorseful for what they have done. It seems, then, that our commitment to retributive punishment is not as strong as one might think. The second (and, I think, more plausible) response is this: the practice of retributive punishment will be necessary in a pragmatist community. Suppose there is no purely retributive punishment. Then people know they can get away with single, unrepeatable infractions, such as killing your only relative (who happens to be wealthy) to get his inheritance. No corrective punishment is required, as you have no more relatives, and recidivism is already ruled out. But a society that practices purely retributive punishment is also a society in which people know they cannot get away with even these unrepeatable infractions. Thus, retributive punishment serves a role in the well-ordering of society, and hence can be practiced by the pragmatist.16 Indeed, this is a point worth dwelling on. For a practice of punishment to serve the goals of the pragmatist community, the members of the community must associate guilt with punishment: unless you know that punishment follows guilt, the punishment serves as no disincentive to incur guilt. This is why we emphasized so strongly action types as meriting sanction: knowing that a type of action is sanctionable serves as a disincentive to perform a token of that type. But a practice of purely retributive punishment makes the connection between guilt and punishment a necessary one, and would hence well serve the pragmatists’

170

Pragmatism and Morality

goal of deterring wrong behavior. It is clearly possible, then, for standard compatibilists and pragmatists to be on the same wavelength regarding retributive punishment. So non-corrective punishment cannot be the wedge that pries the two theories apart. One might argue, though, that punishments doled out in accordance with this practice are not really retributive punishments, as the purpose of such punishments is (essentially) to deter wrong-doers. This is a serious objection, but one that can be met ultimately. Recall our earlier discussion of practices. Rawls emphasized that the considerations that justify the practice are not allowed to ‘inﬁltrate’ the practice and justify individual acts within the practice. Individual acts are essentially justiﬁed by the rules of the practice. As Rawls puts the point It is a mistake to think that if the practice [of, e.g., promise keeping] is justiﬁed on utilitarian grounds then the promisor must have complete liberty to use utilitarian arguments to decide whether or not to keep his promise. The practice forbids this general defense; and it is a purpose of the practice to do this.17 Thus, when the hard determinist punishes a person retributively, she must punish this person merely because the person deserves to be punished—merely because the rules of the practice demand it. No act of retributive punishment is justiﬁed by considerations of deterrence or social utility; allowing such considerations to seep into the practice causes the practice to collapse, rendering it incapable of serving its purpose. And so, for the hard determinist, retributive punishments genuinely are retributive. Is freedom intrinsically valuable? Many of us think that freedom is intrinsically valuable. We should be allowed to be free, even if this freedom leads to bad results. Indeed, a standard response to the problem of evil is to say that a world of people with free will (who hence perform some evil acts) is better than a world of people with no free will (but who are constrained to perform only good acts). Will the pragmatist place intrinsic value on freedom? Can freedom be considered intrinsically valuable, if free-will discourse is pragmatically justiﬁed? It is not clear, though, that endorsing standard compatibilism requires that we place intrinsic value on freedom. First, there is nothing in the

Pragmatism, Freedom, and Responsibility 171

deﬁnition of standard compatibilism that says autonomy is intrinsically valuable. Second, it is not clear that we do place intrinsic value on freedom. To take a related example, one of the classic defenses of political liberty, John Stuart Mill’s On Liberty, argues that we ought to allow liberty for the beneﬁts this liberty brings (societies that progress more quickly than do conformist societies, the contribution of new truths to our belief system, the strengthening of old truths by their clash with false views, etc.). Nowhere does he defend the claim that liberty is intrinsically valuable. It seems likely that the standard compatibilist can pursue a similar line regarding metaphysical freedom. For these two reasons, I believe, philosophers need not necessarily think that freedom is intrinsically valuable, for them to be counted as compatibilists. Thus, if our pragmatists place only instrumental value on freedom, this does not preclude them from being genuine compatibilists. And I think that the pragmatist can endorse the type of defense of freedom that Mill gives for liberty. A free society is a better and more pleasant place to live; we have an interest in living free from coercive authority. Thus, though the way of thinking that leads philosophers to embrace pragmatism might lead them to reject placing intrinsic value on freedom, this rejection is not in itself incompatible with genuine compatibilism. In any case, I suspect members of the pragmatist community would place some intrinsic value on freedom. Let us discuss this in relation to our hard-determinist community: think about what it would mean for there to be no sanctionable (i.e., ‘free’) behavior in the community, that is, no behavior amenable to alteration by punishment or reward. That would require that some authority was exercising such massive control over people’s psychology that these people’s behavior would cease to be alterable by the application of sanctions and rewards. I doubt that the members of the hard-determinist community would ﬁnd this idea very appealing; they would not want (any more than we would) to cede such control or authority to anyone. They would like and value sanctionability. Similar comments apply to the pragmatist compatibilist’s views toward freedom. Hence, this worry about the value of freedom/sanctionability seems not to be justiﬁed.

Draconian sanctions? Another objection that might be raised against my thesis is the claim that the sanctioning behavior of the pragmatist community would actually be quite different from that of the standard compatibilist

172

Pragmatism and Morality

community. Pragmatists would be less averse to forms of control (brainwashing, etc.) that standard compatibilists ﬁnd abhorrent. Thus, pragmatism is not really just compatibilism by a different name. I think that this objection also has a response. Remember why, for example, the hard determinists re-instituted freedom talk and responsibility talk. Order was breaking down, and the members of the community did not like this. They re-instituted social controls because they liked order, for order brings security and allows us to pursue the things we like and value. In short, the hard-determinist (and pragmatist compatibilist) community will discourage behavior that they dislike ordisvalue (or which produces things they dislike or disvalue), and encourage behavior that they like or value (or which produces things they like or value). But is this different from what we do now? We encourage moral behavior because we value it; we encourage hard work and industry because these are traits we value. Our distribution of praise and blame, punishment and reward, is based on our values. This is no different from the pattern used by the pragmatists: their distribution of sanctions reﬂects their values. So, assuming they have the same interests, the two communities (pragmatist and standard compatibilist) will encourage the same sorts of behaviors. And why wouldn’t these communities have the same interests? Why should hard determinists cease to value life and health, satisfaction of curiosity and instrumental control over nature, afﬁliation with others, and the other things we identiﬁed as constitutive of human ﬂourishing and the good life? As I argued above, even the interest in autonomy and control can be understood within a hard-determinist framework. This will be so not just because of the above-mentioned reasons regarding the intrinsic value of sanctionable behavior, but also because the myriad idiosyncratic derivative interests that people have require autonomy in order to be satisﬁed. Now, the behaviors that our hard determinists and compatibilists will seek to encourage or discourage, depending on their attitude toward these behaviors, will include sanctioning behaviors. In other words, we sanction the methods others use to sanction if we think that these people are going about it the wrong way. We may disapprove of pornography, but think that ﬁrebombing a pornographer’s house is an inappropriate punishment. We might disapprove of theft, but think that allowing Clockwork Orange-esque behavioral modiﬁcation to be practiced on such individuals gives too much power to too few individuals,

Pragmatism, Freedom, and Responsibility 173

or makes us fear the government; or perhaps our sense of sympathy and fellow-feeling condemn it. (The reader should be reminded of our discussion of the value of sanctionability at the end of the previous section.) And this is why the pragmatist community will not necessarily be a 1984-ish police state, with mind control and brainwashing. The community sanctions behavior to produce behavior it is willing to tolerate, and if the community is unwilling to tolerate draconian punishments (and clearly, we do not tolerate these sorts of punishments), then this sort of sanctioning behavior will not be tolerated. The sanctioning behavior of the pragmatist community will be determined by their interests. And (as I argued earlier) as the pragmatists’ interests will be substantially the same as those possessed by a standard-compatibilist community (up to and including interests in liberty and aversion to coercive authority; recall our interest in control and the intrinsic value of sanctionability), the members of the pragmatist community will not permit such objectionable punishments.

Freedom and the mentally ill There might be a difference in the behavior of our pragmatists and our standard compatibilists regarding the treatment of the mentally ill. Consider the criminal behavior of one who is mentally ill but whose illness might be treatable through medication or therapy. This criminal’s behavior ﬁts the criterion for ‘sanctionable behavior,’ but standard compatibilists would hesitate to call the action ‘free.’ Indeed, juries often are given the option of ruling a defendant not guilty by reason of insanity. Such people are thought to be not responsible for their actions. In short, they are not free, although (for example) the hard determinist might call their actions sanctionable. What this means is that the hard determinist (and other pragmatists) will be sanctioning actions the standard compatibilist regards as unfree; and this might be thought a salient difference between the two positions. An important thing to note is that even though we do not call such people free, we sanction them all the same. We commit them (typically against their will) to mental hospitals, and force them to undergo therapy with the goal of altering their behavior (or, if that is not possible, then protecting the public from them). The difference seems to be that with not-guilty-because-insane violators, our treatment is aimed at rehabilitation, not retribution, whereas our treatment of violators we consider sane is aimed both at punishment and improvement.

174

Pragmatism and Morality

Notice, though, that if the standard compatibilists did away with retributive punishment, and aimed all treatment of violators at rehabilitation, then their behavior would be brought into line with that of the pragmatists who make no distinction between the sane and the criminally insane, in terms of labeling actions as sanctionable. Both the standard compatibilists and the pragmatists would then focus solely on rehabilitation. If the standard compatibilists did away with retributive punishment, such an action would probably be accompanied (or even caused) by the realization that criminals are poorly socialized, or often suffering from mental or physiological disorders, in much the same way that the criminally insane are. If this were the case, then the rationale behind calling the actions of one group free while calling those of the other group unfree would be removed, and the standard compatibilists might begin regarding the actions of both groups in the same way. Remember, though, that the pragmatist need not give up on retributive punishment. As we discussed above, the same considerations that might cause the compatibilists to employ retributive punishment could also persuade the hard determinists to do the same. But the important thing to remember is that the pragmatists’ and the standard compatibilists’ treatment of the mentally ill are not signiﬁcantly different from each other. Both treat the mentally ill with an eye to rehabilitation, or, if rehabilitation proves futile, then both will at least ensure that the person in question is prevented from doing any harm. Thus, this cannot be the source of any important differences between the pragmatist and the standard compatibilist. As an addendum, I should note that similar comments apply to the two communities’ treatment of children. A three-year-old child is in general regarded (in a standard compatibilist community) as not morally responsible for her actions. Thus, one might object that as the behavior of this child might be affected by aversive therapy, the pragmatist might well punish the child, whereas the compatibilist would not. Two things are to be noted at this point: ﬁrst, we do punish three-year-old children. We give them ‘time outs,’ or send them off to bed early, or revoke certain privileges, etc. Second, there is no reason why the pragmatist community must use punishment as its only method for discouraging bad behavior. The pragmatist has the same concerns as the standard compatibilist with regard to such a child: a desire not to cause the child to suffer, when causing such suffering is not necessary to morally educate the child; concern for the child’s psyche and future development; and so forth. So the pragmatist will often choose education over punishment.

Pragmatism, Freedom, and Responsibility 175

Thus, the pragmatist and the standard compatibilist need not differ in their treatment of children.

Is punishment permissible in a hard-determinist community? It has occasionally been suggested that punishment is unjust, if incompatibilism and determinism are both true. Thus, our hard determinists ought not punish people for stealing, killing, etc., because these people are not really free (and hence are not morally responsible). It is not clear whether this objection is coherent, though, for the critic is caught on the horns of a dilemma. On the one horn, if we may with justiﬁcation say that the hard determinists ought not punish people, then we are admitting that the hard determinists are subject to moral evaluation and moral constraint. But if this is the case, then they are justiﬁed in punishing people, for the very reason that they are subject to moral evaluation and moral constraint. So the objection defeats itself the moment it is spoken. On the other horn, if the hard determinists are not subject to moral constraints, then they can hardly be morally constrained not to punish people according to their criterion for ‘sanctionability.’ So either way, the hard determinists may punish people without being subject to moral criticism. This result is not merely a sleight of hand resulting from judging the hard determinist from our own compatibilist standpoint. This result still obtains if we imagine the hard determinists contemplating this dilemma. Suppose a hard determinist judges that given the truth of determinism people cannot be held responsible for their actions, and therefore she ought not sanction people for their actions. But judging that she ought to do or refrain from an action, the hard determinist is judging that she and her community are, in fact, subject to moral constraints and judgments of moral responsibility. Therefore, a hard determinist cannot consistently make this judgment. Alternately, if she judges that members of her community are not subject to moral constraints, then she is not morally constrained to avoid punishing people who engage in behavior harmful to the community, and thus may permissibly engage in the sanctioning of behavior along the lines indicated above. So in neither case is the hard determinist community behaving unjustly by sanctioning people. In any case, as I will argue in the next section, the hard determinists are free and morally responsible in precisely the same sense as the compatibilists are. We will see that ‘sanctionable’ means the same thing as ‘free,’ and ‘moral responsibility’ as spoken by a hard determinist

176

Pragmatism and Morality

means the same thing as ‘moral responsibility’ as spoken by a compatibilist. Therefore, there should be nothing morally objectionable about punishment in a hard determinist community.

Are we free or are we ‘free’? There is one ﬁnal objection to be met. This objection attacks the idea that hard determinism really is equivalent to compatibilism (of the pragmatist or standard variety). Suppose I have shown that at the practical level, compatibilism and hard determinism are identical, that the two communities behave identically. One might nevertheless wonder whether it follows that hard determinism is a form of compatibilism; that is, one might wonder whether the word ‘sanctionable’ as spoken by a hard determinist means the same thing as ‘free’ spoken by a compatibilist. An example will illustrate this worry:18 suppose Smith believes that morality consists of following God’s commands (call this ‘God worship’), whereas Jones holds that morality consists of following the patterns of nature (call this ‘nature worship’). Suppose further that what Smith believes God to have commanded is exactly what Jones takes the patterns of nature to be. Thus, their behavior is the same. But it clearly does not follow that nature worship is a form of God worship. Similarly, one might claim that though the hard determinists behave the same way as do the compatibilists, hard determinism is still not a form of compatibilism. Let me begin by noting that on many accounts of meaning, the question ‘The two communities’ behavior is identical, but do their words mean the same thing?’ is deeply confused, and reveals a misunderstanding of what meaning-talk is all about. For these philosophers, then, the ‘sanctionable’ and ‘free’ have the same meaning, as the two words are used identically in the respective communities. Nevertheless, I think this objection is worthy of closer examination than this. We will examine some key terms, and it will emerge that these terms mean the same thing when spoken by the compatibilist and the hard determinist. So, while ‘nature worship’ and ‘God worship’ might mean different things, we will see that ‘sanctionable’ and ‘free’ have the same meaning. Other key terms (‘retributive punishment’ and ‘moral responsibility,’ in particular) also mean the same thing when spoken by the hard determinist and the compatibilist. Let us consider these terms one by one. Sanctionable vs free: The worry here is that ‘sanctionable’ as spoken by a hard determinist does not have the same robustness as has the word ‘free’ spoken by the compatibilist. I think this is not true. I think

Pragmatism, Freedom, and Responsibility 177

the truly robust sense of freedom (freedom from all compulsion, external and internal; ability to do otherwise; etc.) was lost along with libertarianism. What remains for the compatibilist is a more modest conception of freedom: a free act just is an action that is a candidate for reward and punishment, etc. Not only does this deﬁnition pick out the actions that are free, but also it is not clear how the phrase ‘free action’ could mean anything more than what I have used here to characterize the actions it picks out. But all of this is precisely what characterizes the hard determinist’s notion of freedom, as well. For the hard determinist, a free action is just one that is a candidate for reward and punishment, etc. I can see no compelling rationale to refuse to translate ‘sanctionable’ as spoken by a hard determinist as ‘free’ in the mouths of the compatibilists. However, the hard determinists and compatibilists might have auxiliary commitments that will cause ‘sanctionable’ and ‘free’ to diverge in meaning. Let us examine some other important terms connected to the notion of freedom/sanctionability and see whether these words have different meanings for the two communities. Moral responsibility: Sets of pragmatically justiﬁed norms can be classiﬁed according to the interests that they serve. Thus, our epistemic norms are those that serve our interest in truth and explanation. Semantic norms are those that serve the interest in communication and forging linguistic bonds between disparate communities. By contrast, our moral norms are those that serve our interest in peaceful and fruitful coexistence. So far, we have encountered no reason to suppose the interests of the hard-determinist and compatibilist communities will diverge. And so each community will have a set of norms characterizable as moral norms, norms aimed at promoting peaceful and fruitful coexistence. Assuming similarity of interests between the two communities (and I have argued that interests are universal), these norms will themselves be similar between the two communities. So the norms in question will be moral norms in both communities; the only question remaining to be answered is, ‘What does it mean to be responsible vis-à-vis this set of norms?’ Well, according to a standard deﬁnition, “moral responsibility” refers to “the fulﬁllment of the criteria for deserving blame or praise (punishment or reward) for a morally signiﬁcant act or omission.”19 But of course, ‘sanctionable’ and ‘free’ are both just shorthand for ‘deserving blame or praise.’ And so ‘morally responsible’ will mean, for both communities, ‘free/sanctionable with respect to norms of morality,’ which

178

Pragmatism and Morality

norms (as we have seen) do not differ between the two communities.20 Thus, the phrase ‘moral responsibility’ appears to have the same meaning in both communities. Retributive punishment: One might argue, though, that punishments doled out in accordance with this practice are not really retributive punishments, as the purpose of such punishments is (essentially) to deter wrong-doers. However, as we saw above (in the section ‘Retributive Punishment’), when the hard determinist punishes a person retributively, she must punish this person merely because the person deserves to be punished—merely because the rules of the practice demand it. No act of retributive punishment is justiﬁed by considerations of deterrence or social utility; allowing such considerations to seep into the practice causes the practice to collapse, rendering it incapable of serving its purpose. And so, for the hard determinist, retributive punishments genuinely are retributive. To conclude this section, we have seen that at the end of the day, the hard determinists and the compatibilists have the same linguistic commitments. Terms such as ‘moral responsibility’ and ‘retributive punishment’ have the same meaning in the two communities, and ‘sanctionable’ has the same meaning as ‘free.’ Thus, hard determinism and compatibilism are not like God worship and nature worship; hard determinism really is a form of compatibilism.

Conclusion In this chapter, we have reached a number of important conclusions. First, we have argued that the pragmatist can develop a viable conception of free will and responsibility. Second, we showed that the pragmatist can show how the practice of making free-will ascriptions is pragmatically justiﬁed. Third, and perhaps most importantly, pragmatism has shown us that free will is, in an important sense, necessary: there is no theory under which free will does not exist. Those determinists worried about the justice or fairness of holding people responsible for their actions must realize that holding people responsible and treating them free is inescapable: the libertarian and compatibilist both endorse the existence of free will and responsibility; but crucially, so does the hard determinist. Of course, whereas the compatibilist will answer the question, ‘Are some actions free?’ in the afﬁrmative, the hard determinist will give an afﬁrmative response to the question ‘Are some actions sanctionable?’

Pragmatism, Freedom, and Responsibility 179

As I have argued, as the behaviors of the two communities will coincide, and, as I have also argued, as we should translate ‘sanctionable’ as ‘free,’ the hard determinist will, it seems, be a variety of compatibilist. Certainly, your average hard determinist would deny this. But such a blanket denial is insufﬁcient: to deny that the two positions are the same, the hard determinist must be able to point out salient differences between his position and that of the compatibilist. I have argued that such a difference cannot be found, even if we look at such likely candidates as retributive punishment and the manner of sanctions allowed by the hard determinist community. It is possible that a hard determinist or a pragmatist may end up being an atypical compatibilist, and endorsing views that few standard compatibilists would be comfortable holding. But if you identify a difference between hard determinism and traditional compatibilism, you must ﬁrst (before deciding that hard determinism is not a form of compatibilism) ask yourself, ‘Is this difference really sufﬁcient to make the hard determinist no longer a compatibilist? Is this difference really sufﬁcient to push the hard determinist out of the compatibilist camp?’ I have argued, though, that even when differences might arise between the two views, these differences do not sufﬁce to push the hard determinist out of the compatibilist camp. I must conclude, then, that hard determinism is a form of compatibilism; and that pragmatism has demonstrated the necessity of free will.

Postscript on epistemic responsibility In this chapter, we have been discussing freedom and responsibility as it relates to morality. In Part III, we will move to a discussion of epistemic (as opposed to moral) norms. Does the account of moral responsibility outlined here tell us anything about epistemic responsibility (i.e., about our responsibility to abide by epistemic norms)? This depends on whether beliefs are under our voluntary control; that is, it depends on whether doxastic voluntarism is true. What if beliefs were under our voluntary control? If belief, like action, were voluntary, then could we assimilate our account of epistemic responsibility to the above account of moral responsibility? It is not entirely clear that we can. Consider the pragmatic point of sanctions, on the account given in this chapter. The point of sanctions is to alter behavior by introducing an additional pragmatic reason for conforming to certain norms. This is well and good, when the norms in question are practical norms. But what about the norms in question when they

180

Pragmatism and Morality

are theoretical norms? If doxastic voluntarism is true, can we apply sanctions to people with the goal of making them conform to proper epistemic norms? It is not clear whether this is legitimate. After all, by applying sanctions to people, you would be giving them a practical reason to hold a particular belief. This would seem to be a conﬂation of practical and theoretical norms. After all, it might be practically rational to follow a moral norm because of threat of sanctions; but it can hardly be theoretically rational to hold a particular belief because of threat of sanctions. And the beliefs are properly subject to theoretical, not practical, justiﬁcation. I am not entirely certain what force this argument has. It is true that ideally, one should hold one’s beliefs for theoretical (rather than practical) reasons. But it is equally clear that in the moral case, one should do the right thing for moral reasons, not merely because of fear of sanctions. (On the two-level account defended here, moral rules are justiﬁed by prudential reasons, to be sure; but individual actions are justiﬁed by moral rules, not prudential reasons. Thus, individual actions should be performed for moral, not prudential, reasons.) Sanctions are applied in those cases where the proper (i.e., moral) reasons are insufﬁcient, and we must give the agent an additional reason for conforming. Similarly in the epistemic case, sanctions would only be applied when the proper (i.e., theoretical) reasons are insufﬁcient, and we must give the agent an additional reason for conforming to proper epistemic norms. Thus, in both cases (moral and epistemic) we must recognize that the sanctioned agent is conforming for basically the wrong reasons: in these cases, the agent is conforming not because conformity is (morally or epistemically) correct, but instead because of the threat of sanctions. Even if the agent is conforming for the wrong reasons, society recognizes the value of enforcing compliance, even if certain people must be compelled to comply for reasons that are, in a sense, the wrong reasons (i.e., not moral or epistemic reasons). Thus, it is perhaps the case that the account of moral responsibility given in this chapter could be extended to the epistemic realm, if doxastic voluntarism is true. However, I will not pursue this issue further. The insurmountable barrier to extending our account of moral responsibility into the realm of epistemic norms is that our actions are voluntary, whereas our beliefs are not. That is, we choose our actions, but we do not choose our beliefs: Doxastic voluntarism is false. One of the most inﬂuential criticisms of doxastic voluntarism is found in William Alston’s ‘Internalism and Externalism in Epistemology’.21 . Alston writes:

Pragmatism, Freedom, and Responsibility 181

When I see a truck coming down the street, I am hardly at liberty either to believe that a truck is coming down the street or to refrain from that belief. Even if there are special cases, such as moral or religious beliefs, where we do have pinpoint voluntary control (and even this may be doubted), it is clear that for the most part we lack such powers.22 Alston’s view seems exactly right. Indeed, although some hold the belief (as Alston notes) that religious beliefs are under our voluntary control, it seems that this belief is more likely to be held for ideological rather than philosophical reasons, and that involuntariness is a feature of all beliefs, without exception. How does this bear on the notion of being held responsible to epistemic norms? The ﬁrst thing to note is that some epistemic norms govern action, not belief. Thus, a clinician wishing to discover whether one drug works better than another in the treatment of a disease has an epistemic obligation to perform a double-blind study, and not merely base his results on anecdotal evidence, and so forth. To give another example, if a person holds certain beliefs dogmatically, and reads only materials that he knows support his position, we might criticize the person—not necessarily for his beliefs, but for his willful failure to expose his belief to divergent opinions. Thus, many epistemic norms relate to actions one can perform, and actions are the sort of things subject to control via sanctions. Failure to abide by action-related epistemic norms may, therefore, subject one to sanctions. This is particularly clear when the action in question has both a moral and an epistemic component: one’s failure to perform an epistemically adequate study to determine the safety of a particular medicine is subject not merely to epistemic criticism, but also potentially moral blame. Clearly, moral sanctions are in order in such a case. However, many (perhaps most) epistemic norms concern not action, but belief. Many philosophers have argued that as beliefs are not under our voluntary control (i.e., as doxastic voluntarism is false), we cannot be held responsible to epistemic norms. Alston writes: If I cannot believe or refrain from believing that p at will, then it is futile to discuss whether I am permitted to believe that p at t or whether I would be irresponsible in choosing to believe that p at t. And it seems that we just don’t have any such control.23 It cannot be said that I ought to believe that p, if I cannot control whether I believe that p. There is a clear connection here to the thesis

182

Pragmatism and Morality

that ‘ought’ implies ‘can.’ I ought to believe that p (or ∼p) only if I can believe that p (or ∼p). In the pragmatist terms we have used in this chapter, sanctioning Smith (positively or negatively) will not have any effect on whether or not Smith believes that the Earth is ﬂat. It might make him pretend to believe that the Earth is round, but epistemic norms deal with what a person believes, not with what he acts like he believes. If sanctions do not alter what a person believes, then failure to conform to epistemic norms (at least those norms related to belief) is not sanctionable. Thus, while sanctioning a person for a moral failing can successfully make a person act in conformity with a moral norm, sanctioning a person for an epistemic failing cannot make a person believe in conformity with an epistemic norm. This is how action and belief diverge in respect of voluntariness. However, even if a person is not necessarily responsible for his or her epistemic performance, it does not follow that we cannot evaluate this performance as good or bad. Thus, a person living in a modern, technologically advanced society, who nevertheless believes that the Earth is ﬂat, fails to conform to relevant epistemic norms. Given the information available to him, he ought to believe that the Earth is round, and his failure to do so means he has performed badly, from an epistemic standpoint, even if he cannot be held responsible for this failure. Evaluation does not imply an ascription of responsibility: I can say my computer, or car, or kitchen knife is performing badly, even though (obviously) it is not responsible for its bad performance. What is the practical import of such a judgment? That is, if we judge that someone has performed epistemically badly, does it make a difference? After all, we have already conceded that irrational belief is not sanctionable. And indeed, people are in general not punished for holding irrational beliefs.24 However, even though holding irrational beliefs is not a basis for sanctions, it is often the basis for exclusion. What I mean is this: if Smith holds irrational beliefs, he may as a result be excluded from certain groups or activities, even if these beliefs are not under his voluntary control. Let me give some examples. Holding certain beliefs is a necessary condition for membership in certain professions, or admission to the ranks of recognized experts. If one holds, in spite of all the evidence, that electrons have a positive charge, she will be excluded from among the ranks of professional physicists. That is, she will not be hired as an academic physicist, or as a researcher in a laboratory; her articles (if they embody a commitment to the idea that electrons have a positive charge) will not appear in recognized professional journals. She will not be recognized by

Pragmatism, Freedom, and Responsibility 183

membership in those professional organizations (such as the National Academy of Sciences) that recognize professional standing. In short, she will be excluded from the profession. In other contexts, one with irrational beliefs might be excluded from discussions where such beliefs are relevant. Thus, by possessing such irrational beliefs, one loses the power to participate in and inﬂuence the course of such discussions. Such a discussion might be, for example, on a discussion board on the internet, or a discussion among peers or colleagues, or so on. The opinions of one with irrational beliefs will not be sought (or will be discounted), his views not canvassed, and so forth. Thus, recognition that one fails to live up to certain epistemic norms results (loosely speaking) in exclusion from the community of cognizers, or from some subset of this community. One might be excluded in the sense of not being allowed to participate in the discussion (as one who is excluded from the community of professional physicists); or one may be excluded in that although one is allowed to participate, her contributions are ignored or dismissed because of a failure to satisfy the relevant epistemic norms. The latter type of exclusion is most common in settings (such as an informal conversation or an internet discussion board) where there are few (or no) formal procedures, or no ofﬁcial body responsible for vetting a discussant’s qualiﬁcations and excluding those who are unqualiﬁed. Of course, irrational people are not always excluded from public discourse, any more than immoral people are always punished. (Indeed, immoral people often ascend to positions of power, and irrational people often succeed in exerting epistemic control over groups of people in the population.) The system functions imperfectly; I am only describing the purpose of the system. The system of moral praise and blame (and its adjunct, the system of freedom ascriptions) is supposed to function to apply sanctions to encourage certain behaviors and discourage others; the system of epistemic ascriptions is supposed to function in part to exclude certain people from ‘having a say,’ from the privilege of inﬂuencing the opinions of rational cognizers. The fact that these systems do not function perfectly is merely an unfortunate fact, not evidence that I have misidentiﬁed their correct function. Excluding people who fail to adhere to important epistemic norms may seem unfair; it may seem as though people with irrational beliefs are being sanctioned, even though their beliefs are not under their voluntary control. But the exclusion of people who fail to abide by important epistemic norms merely reﬂects the fact that membership in certain communities requires that one satisfy certain standards.

184

Pragmatism and Morality

Consider a non-epistemic example: if Jones has terrible hand–eye coordination, he will not be admitted to the community of professional athletes. By excluding him, we are not saying that he is somehow responsible for his bad hand–eye coordination; we are merely saying that he is not qualiﬁed to join the ranks of professional athletes. This exclusion may cause Jones great suffering (perhaps he has always dreamed of being a famous athlete), but the division of labor in society requires a certain degree of sorting according to aptitude. Those who do not have the aptitude for professional sports are excluded, just as those who lack the aptitude to be carpenters, physicists, airplane pilots, or professional drivers are excluded from these professions. It is not necessarily a person’s fault that she cannot qualify for this profession or that one; but achieving the goals of a given profession or activity requires that certain standards be upheld. Thus, we can see that in society, exclusion is often practiced on the basis that the person excluded fails to meet certain standards, even if she has no control (and hence no blame or responsibility) over whether she meets these standards or not. It is so in epistemology, too: failure to meet certain epistemic standards will result in certain types of exclusions. This exclusion revolves around the notion of testimony, and the epistemic role it plays. Testimony is a source of prima facie justiﬁcation.25 That is, if S tells you that p, you are prima facie justiﬁed in believing p. This justiﬁcation is, of course, prima facie: it can be defeated if you possess good evidence of p’s falsity, or evidence of S’s unreliability, etc. This testimony, then, is not justiﬁcation-conferring. But admission to the ranks of those whose testimony provides prima facie justiﬁcation requires that one meet certain epistemic standards (or, conversely, failure to meet these standards is grounds for exclusion).26 If one fails to meet these standards, then she is excluded, and her testimony is not justiﬁcation-conferring. Again, this exclusion can be partial: even if Smith is not a reliable reporter in questions of physics, she might still be relied upon to tell you the correct time. Thus, we can evaluate epistemic performance (and exclude poor performers), even if doxastic voluntarism is false. It is no different from, say, evaluating a person’s eyesight and excluding her from certain professions (e.g., pilot) based on the results. Exclusion is part of the self-policing mechanism in an epistemic community. Let us consider a real-life example, dealing with exclusion from a particular epistemic community, namely, the community of professional scientists: it has been established beyond a reasonable doubt that the HIV virus causes AIDS. And yet for a time, Thabo Mbeki, the president of South Africa, persisted on questioning the link. He placed on

Pragmatism, Freedom, and Responsibility 185

his AIDS Advisory Panel people such as Sam Mhlongo, who questions whether the HIV virus even exists. Such dissenters blame AIDS not on the HIV virus, but instead on other factors such as poor diet and drug addiction. Such dissent comes in the face of overwhelming evidence to the contrary. Predictably, AIDS dissenters decry their exclusion from mainstream scientiﬁc journals as a form of censorship. The legitimacy of this practice of exclusion is eloquently argued in a Nature editorial criticizing Mbeki’s ﬂirtation with AIDS dissenters: No one who has been impressed by your success in what many claimed impossible—the peaceful transition to democracy in South Africa following the long struggle against the iniquities of apartheid—will reject the argument that there are times when the voice of those challenging the dissenting order must be heard . . . But this does not mean that all dissidents and ‘heretics’ can claim equal legitimacy merely on the basis of their persecution; democratically endorsed procedures exist through which their ideas can be put to the test, and viable heresies separated from those that, after close scrutiny, deserve to be placed aside . . . Science has developed . . . the peer-review system, [which] is little more than a way of speeding up the process of sorting out those ideas which have a greater chance than others of surviving intellectual scrutiny and testing through experimentation. One hypothesis that has survived this process is the idea that there is a direct, causal relationship between infection with the human immunodeﬁciency virus (HIV) and the onset of AIDS . . . Our columns have been—and remain—open to anyone offering evidence to the contrary, but on one simple condition: that their evidence passes the same rigorous tests of scientiﬁc robustness that are applied to any scientiﬁc paper that we receive. So far this has not happened. Those who have experienced rejection may choose to castigate this as ‘censorship’, but the vast majority of authors of the scientiﬁc papers that we reject on technical grounds accept the process as valid and necessary for the health of science.27 Thus, the exclusion of those who fail to meet adequate epistemic standards (i.e., to prove that their position is adequately supported by evidence) helps promote the ‘health of science,’ as well as reserving scarce resources (e.g., journal pages and conference slots) for those scientists who have met the epistemic burden placed on them by their epistemic peers. This passage from Nature helps demonstrate not only that epistemic exclusion does in fact occur among self-policing

186

Pragmatism and Morality

epistemic communities; it also argues that such exclusion is legitimate and even necessary, and not an objectionable form of censorship. To tie together our notion of testimony with the above description of how epistemic communities (like the scientiﬁc community) police themselves, we are allowed to treat scientists as experts and rely on their testimony in part (though not entirely) because the scientiﬁc community polices itself and excludes members who violate important epistemic norms. If a community, like a scientiﬁc community, ceases to police itself, ceases to exclude those who violate epistemic norms—if, for example, scientiﬁc consensus were allowed to be swayed by researchers who had demonstrable conﬂicts of interest (e.g., climate researchers who are funded entirely by the coal and oil industries)—then we might no longer be entitled to accept the testimony of scientists as legitimate expert testimony. That is, such testimony is no longer prima facie reliable. Thus, self-policing via exclusion is necessary (but again, not sufﬁcient!) for the health of epistemic communities. This is particularly true of communities of experts, and their status as experts depends in large part on their belonging to such a self-policing community. Exclusion should not be regarded as sanctioning. Sanctioning is intended to alter behavior. If you want to alter a person’s beliefs, you argue with or instruct the person; exclusion does not serve to alter a person’s beliefs, and is therefore not a version of sanctioning. Further, a person with irrational beliefs is only excluded, and no further punishment is forthcoming. The point of applying epistemic labels (unjustiﬁed, irrational, unsupported by the available evidence, strongly conﬁrmed, etc.) to beliefs is to indicate which beliefs ought to be held and which ones rejected. As an adjunct to this, epistemic communities will then exclude agents who hold rejected beliefs, and admit those who hold beliefs taken by the relevant community to be well-supported. It is in this way that epistemic communities police themselves, and encourage investigation into what are considered to be worthwhile theories and beliefs, while discouraging beliefs that are taken to be unsupported or false. What standards should be used to determine good or bad epistemic performance? We turn to this question in Section III of the book, where we will discuss a pragmatist epistemology, and the types of epistemic evaluation that comprise this pragmatist epistemology.

Part III Pragmatism and Epistemology

This page intentionally left blank

7 Pragmatism and Epistemology

As I noted in Chapter 2, the terms ‘morality’ and ‘epistemology’ can each be construed in two different ways: broadly and narrowly.1 Epistemology, construed broadly, it is our actual epistemic practice. That is, it is the practice common in every ﬁeld, from chemistry to philosophy, of evaluating beliefs as justiﬁed or unjustiﬁed, scientiﬁc methods as rational or irrational, etc., and of evaluating and revising not only those beliefs, but the standards of evaluation themselves. Construed more narrowly, it is explicit theorizing about epistemology in the broad sense. This theorizing can be the attempt to construct a set of formal conditions for the justiﬁcation of beliefs (e.g., explanatory coherence among beliefs; proper inferential relation to foundational beliefs; proper causal genesis, à la reliabilism, etc.); or it can merely be an attempt to uncover some of the features of justiﬁcation and knowledge, without any commitment to the possibility of formal theorizing. Epistemology in the narrow sense is in essence our practice of theorizing about epistemology in the broad sense, about our practice of epistemology. One can be committed to the viability of epistemology in the broad and the narrow sense, and think that we can make true theoretical claims about justiﬁcation and knowledge, without necessarily thinking that we can construct a formal theory of justiﬁcation. Similar comments can be made about the term ‘morality’: this term can be construed broadly, as our practice of praising certain acts and traits and condemning certain others; or it can be construed narrowly, as referring to the activity of theorizing about morality in the broad sense. As in epistemology, this theorizing can be an attempt to construct a theory of right and wrong (e.g., Kantianism, utilitarianism, etc.). Again, as in epistemology, one might think it is not possible to construct a formal moral theory, which gives a decision procedure for determining our obligations;2 but to doubt the 189

190

Pragmatism and Epistemology

viability of formal moral theorizing is not necessarily to doubt the legitimacy of morality in the broad sense, nor is it to think that we cannot fruitfully theorize about morality. In both the epistemic and the moral cases, a person can make moral and epistemic judgments without understanding the theoretical underpinnings of these judgments. That is, one can practice morality and epistemology in the broad sense without their having a grounding in theoretical morality and epistemology. I have tried to show that our moral practice (in the broad sense) is justiﬁed, that is, it is rational to follow moral norms. This practice is justiﬁed as a cooperative strategy for promoting our mutual interests. How, though, is our epistemic practice (understood in the broad sense) justiﬁed? How do we justify the practice of evaluating beliefs and research techniques according to canons of rationality? The answer is that doing so promotes certain of our interests. In particular, as I argued in Chapter 2, epistemic evaluation promotes our interest in truth and understanding, and in instrumental control over the world. Epistemology, like morality, is prudentially rational (where ‘prudential’ is, as I have repeatedly emphasized, not understood to mean ‘self-regarding’). In this chapter I will argue that epistemology (in the broad sense) consists of two distinct evaluative practices, each of which plays a different pragmatic role in promoting our inquiry into the world. Let us examine this pragmatist conception of epistemology.

Two epistemic perspectives The form of pragmatism we have relied on in this book denies that normative discourse is true by virtue of describing normative facts or corresponding to some independent normative reality. This feature of pragmatism gives rise to the following question: if normative vocabulary does not describe a factual reality, then what justiﬁes the varieties of normative discourse? The answer, as we saw in Chapter 2, is that these types of discourse are justiﬁed by appeal to various interests we have. Thus, moral discourse is justiﬁed not by appeal to some causally efﬁcacious moral reality, but by appeal to our interest in peaceful and fruitful coexistence. Semantic discourse is justiﬁed by our appeal to our associative interest in unifying diverse linguistic communities and simplifying communication. And epistemological discourse is justiﬁed by appeal to our interest in truth, understanding of the world, and instrumental control over nature. Now, from such a pragmatist perspective, the key to understanding justiﬁcation or knowledge is focusing on the question, ‘What role do

Pragmatism and Epistemology

191

justiﬁcation and knowledge attributions play? What purpose do they serve?’ And when we focus carefully on this question, we realize that justiﬁcation attributions play not one but two distinct roles in our language, corresponding to two related but different purposes. That this is so can be seen by considering the following sort of claim that one might make: C: Astrology as a system has been soundly refuted, but given her upbringing and community, her belief in astrology is perfectly rational. Notice that both clauses in this sentence make epistemic claims, and that there is a sense in which these claims oppose each other. Let us brieﬂy examine these two types of claims. The ﬁrst claim (the one made by the ﬁrst clause of C) addresses whether the evidence, objectively viewed, speaks for or against a certain theory or proposition: the claim is that the evidence is overwhelmingly against astrology. To speak in Lakatosian terms, astrology (as a research project) has degenerated to the point of death. The second clause of C, on the other hand, makes a claim about whether a person’s token belief in that theory or proposition is rational or irrational. More precisely, the second clause evaluates a person’s epistemic performance relative to a particular epistemic background (about which we will say much more in this chapter). These two types of claims are, respectively, externalist justiﬁcation attribution and internalist justiﬁcation attribution. The former attributes what we shall call objective justiﬁcation; it abstracts from the person’s individual performance and his/her epistemic background, and evaluates theories or propositions, or beliefs, in the abstract. The latter attributes what we shall call subjective (or personal) justiﬁcation; it evaluates the agents’ epistemic performance and beliefs, relative to a certain background. The distinction between subjective and objective justiﬁcation is captured nicely by Robert Audi’s description of the difference between personal and impersonal justiﬁcation. Audi writes, “I distinguish between personal justiﬁcation—that of a person’s belief that p—from impersonal justiﬁcation—that of the proposition that p, or of ‘the belief that p’ in the abstract.”3 Substitute ‘subjective’ and ‘objective’ for ‘personal’ and ‘impersonal,’ and you have the distinction I am interested in here. An agent can be subjectively justiﬁed without being objectively justiﬁed. That is, an agent can perform well epistemically (or, if you believe in epistemic responsibility, can be epistemically free of blame), even if

192

Pragmatism and Epistemology

an external observer can see that the belief or theory the agent holds as a result of this performance is not the best-justiﬁed one, or that it has structural ﬂaws, etc.4 That is to say, an agent can be subjectively justiﬁed without her theory being objectively justiﬁed. In C above, we are saying that she has not performed badly relative to her epistemic circumstances, but that her theory is not objectively justiﬁed: perhaps there is evidence available in her community, not widely disseminated (so she is not held accountable for failing to obtain it), that astrology is false; or perhaps the history of astrological inquiry in her community has not been characterized by sufﬁcient openness to refutation, or so on.5 Alvin Goldman comments on this case that we are inclined to say that there is a sense in which she is justiﬁed because of “the cultural plight of our believer . . . Our believer has good reasons to trust his cultural peers on many matters, and lacks decisive reasons for distrusting their conﬁdence in astrology.”6 On the other hand, there is a sense in which we want to say this person’s belief, which results from reading zodiacal signs, is not justiﬁed because the method of astrology “looks improper and inadequate.”7 Thus, in the same case, a belief can be justiﬁed in one sense, but not in another. Another way in which a person can be subjectively justiﬁed without her theory possessing objective justiﬁcation is that the evidence for the superior theory is simply not available in her culture or historical epoch. For example, Einstein’s theory of general relativity is extremely well-justiﬁed, which is to say that it is objectively justiﬁed. But an English scientist living in the early 1700s would have been subjectively justiﬁed in believing Newtonian mechanics. Indeed, it is difﬁcult to imagine a reason that would have made belief in relativistic mechanics subjectively rational in the early 1700s. Thus, although belief in relativity is (objectively) rational, the 18th-century physicist was (subjectively) justiﬁed in believing the truth of Newtonian physics, as evidence for the former theory was simply unavailable during that time period. Thus, we have the interesting result that an 18th-century physicist ought not have believed a theory (relativity) which is objectively justiﬁed. This demonstrates another way in which subjective and objective justiﬁcation can diverge. The names I have given these two types of justiﬁcation attribution anticipate the argument of the next chapter. There, I will argue that by recognizing that there are two types of justiﬁcation attribution, we can offer a solution to the internalism–externalism debate in epistemology. Adopting a pragmatist account of epistemology and justiﬁcation, I will argue that we have two sets of competing intuitions (internalist and

Pragmatism and Epistemology

193

externalist) regarding justiﬁcation because there are in fact two distinct types of justiﬁcation attribution: internal justiﬁcation attribution and external justiﬁcation attribution. Neither is a substitute for the other, but each serves a distinct and important purpose. So, dividing justiﬁcation attribution into internalist and externalist varieties has signiﬁcant advantages. First, we can show (as we will in Chapter 8) that this offers a solution to one of the more intractable problems in contemporary epistemology. Further, each type of justiﬁcation attribution serves a distinct and important purpose: internalist justiﬁcation attributions evaluate an agent’s epistemic performance and beliefs, whereas externalist justiﬁcation attributions evaluate propositions or beliefs in the abstract, and judge whether or not these beliefs should be adopted or maintained by the evaluator and his or her epistemic peers. We care about the former because we often want to know whether an agent has performed well, epistemically; we often want to know whether an agent was justiﬁed in holding a particular belief. Such evaluation can be important, for example, in moral and legal contexts, when questions of culpable ignorance arise: we may know a particular fact beyond a reasonable doubt, but we can still ask (of an agent who did not know this fact), ‘Was the agent’s ignorance of this fact epistemically excusable?’ Would an epistemically satisfactory person (or a person who turned in an epistemically satisfactory performance) have become aware of this fact, or is the agent’s ignorance the result of a poor epistemic performance? So, internalist justiﬁcation attributions play a valuable, indeed crucial, role in many contexts. Externalist justiﬁcation attributions play perhaps an even more important role. Often, we do not care about any particular agent’s epistemic performance; what we want to know is the truth. Does the evidence support the claim that the universe will keep expanding? Does the evidence support the claim that the death penalty serves as an effective deterrent? We do not want to know whether a particular agent was justiﬁed in believing that the universe would recollapse, or that the death penalty has a deterrent effect; we want to know what theory or proposition we ought to endorse. And so internalist justiﬁcation attributions are of little use here.8 Thus, it is clear that both perspectives are crucially important, and that neither can be eliminated. So each perspective, as it were, is spoken for independently; each addresses an important aspect of epistemic evaluation. (Jointly they are supported by the fact that by retaining both perspectives, we can satisfy both internalist and externalist intuitions, and are forced to abandon neither. We will pursue this line of inquiry

194

Pragmatism and Epistemology

in the next chapter.) Let us discuss each type of justiﬁcation attribution separately.

Internalist justiﬁcation attributions As I indicated above, internalist justiﬁcation attributions evaluate a person’s epistemic performance and beliefs: has the person in question behaved epistemically badly or well? To answer this question in particular cases, we must ﬁrst ask, ‘What does count as a good epistemic performance?’ Evaluations of ‘good’ and ‘bad’ are relative (a) to a particular evaluative frame of reference and (b) to what can be reasonably expected of a person. What can be reasonably expected of a person depends (i) on the circumstances surrounding a person’s particular performance, and (ii) on what sorts of things people are (or the person in question is) normally capable of. Thus, normative evaluations (including internalist epistemic evaluations) are context-sensitive in several ways. First, they are sensitive to a particular evaluative frame of reference. Second, they are sensitive to the individual circumstances surrounding a particular performance. Third, they are determined in part by what is normal, or what can be normally expected of human cognizers. Let us brieﬂy examine these considerations in order. First: our evaluation of a person’s performance depends on the background relative to which we are making this evaluation. There are different backgrounds and different human norms with respect to which a person’s epistemic evaluation can be evaluated. A person’s performance could be at the same time bad by one standard and good by another. Usually, the background relative to which an evaluation is made can be assumed. Suppose, for example, that I have a friend who plays in a recreational basketball league. If, after the game, I tell him that he played well, obviously the relative background that is determining the standard of evaluation is the background of recreational basketball. Thus, my utterance is true. But if some talent scouts for a professional basketball team are watching the game, they might say, ‘None of these guys can play well at all.’ This utterance can also be true. Although this statement appears to conﬂict with my utterance, the assumed background—that of professional basketball—sets a different standard, and so both their utterance and mine are true. Thus, at least some of our evaluative terms display a sort of contextualism: what, precisely, one is saying when one praises a performance as good will depend on one’s circumstances and an assumed set of background standards relative to which the evaluation is made.

Pragmatism and Epistemology

195

Since internalist justiﬁcation attributions are evaluating individuals’ epistemic performance, such utterances are similarly backgroundrelative. For example, we might say, ‘Georg Stahl argued brilliantly in favor of phlogiston theory.’ This sentence might be true—but it is true relative to the background of early 18th-century science. Suppose that a contemporary scientist (let us call him Smith) advanced the same arguments in favor of a theory similar to phlogiston theory. We would not say that this scientist had argued brilliantly, though he is advancing in essence the same arguments as had Stahl. Why? Because we would say, ‘Smith has ignored three centuries of research in chemistry, all the subsequent objections to phlogiston theories, etc.’ Thus, we would say that Smith argued badly—but we are evaluating Smith’s performance relative to the background of early 21st-century science. Let us consider a different example. Suppose I attend a high school debate on the subject of free will. I might say, ‘Jones did an excellent job of justifying his views.’ Nevertheless, if a professional philosopher (Smith) advanced similar arguments, I might justiﬁably argue, ‘Smith has done a poor job of arguing for the existence of free will.’ Again, the difference is that the background—high school debate team vs professional philosophical argumentation—is different in the two cases, and the latter background sets a higher standard of performance and evaluation. So again, by varying the background relative to which the person is evaluated, we come to a different conclusion regarding the quality of the person’s performance. But there is an important sense in which these different conclusions do not conﬂict with each other: each is true relative to a particular background. Thus, epistemic evaluation of individuals is not monolithic—an individual can be evaluated relative to any number of different backgrounds, which determine different standards of performance and evaluation. In most cases, the relevant background is clear; but in some cases, we will have to specify the background relative to which this type of epistemic evaluation (i.e., the internalist justiﬁcation attribution) is made. Second: evaluation of someone’s performance is sensitive to the individual circumstances surrounding a particular performance. (There is some overlap between this consideration and the previous one.) For example, if a particular runner runs a 20-minute mile, our evaluation of whether this is a good or a bad performance will be altered by, say, our knowledge that the person was running up a 20◦ incline against a 15 mile per hour headwind. In that case, we might say that the person turned in an outstanding performance, an opinion we would not

196

Pragmatism and Epistemology

render if this 20-minute mile were completed on level ground with a slight tailwind. The same sorts of considerations apply to the person described by C above. Though we would normally condemn belief in astrology by someone living in a scientiﬁcally advanced society, we recognize that if someone is raised in a community characterized by belief in astrology, and evidence of astrology’s falsehood is not available to this person, then belief in astrology is not irrational, but may be entirely rational. The principle at work here should be clear: if an agent has no access to certain facts, then her epistemic performance cannot be faulted on the grounds that she failed to take these into account. That is, our judgments as to whether one has performed well (epistemically) should not take into account facts the person could not have accessed. And so we can see that an agent’s epistemic circumstances (e.g., the epistemic community in which one lives, the evidence made available in such a community, etc.) affect our evaluation of that epistemic agent. Third: the standards for subjective justiﬁcation are determined in part by what is normal, or what can be normally expected. To take an example unrelated to epistemology, a person who lives to the age of 90 in good health and then dies is regarded as having been fortunate— but only because the average human life span is closer to 80 than 150. If the latter were the case, we would instead mourn this person’s untimely passing. Thus, how we evaluate a person’s fortune in this respect depends on the normal human lifespan. Consider a moral example. Imagine a member of (say) the French Resistance. If he is captured and beaten, we will not necessarily condemn him for revealing the identity of his comrades—but only because humans are sensitive to pain and injury. If humans were much less sensitive to pain, then we might reasonably expect our Frenchman to endure a severe beating rather than betray his comrades. Thus, what we can morally expect of a person depends on contingent factors about the human constitution, and what a normal human might be capable of enduring. Similar comments apply to epistemological performance. Whether an epistemic performance counts as good or not depends on what humans are normally capable of. Humans have certain signiﬁcant cognitive limitations, and epistemic appraisal loses its point if we evaluate as bad a performance that goes up to, but naturally not beyond, these limits. Again, we consider a baseball pitcher to have a good fastball if he can throw it 90 miles per hour; we do not criticize his fastball as slow because it is not 200 miles per hour. Humans are not capable of such

Pragmatism and Epistemology

197

performance. We consider a mathematician excellent if he can prove a long-disputed theorem during his or her lifetime; we do not judge a mathematician to be bad because he cannot produce results like the proof of Fermat’s last theorem every week. Humans are not capable of such performance. Similarly, we cannot judge a person’s epistemic performance as bad if he performs as well as a human can be reasonably expected to perform. Our epistemic evaluations must be relative to the human norm. Other authors have argued that our evaluation of epistemic agents must be relativized to agents’ actual abilities. For example, Alvin Goldman writes: Advice in matters intellectual, as in other matters, should take account of the agent’s capacities. There is no point in recommending procedures that cognizers cannot follow or prescribing results that cognizers cannot attain. As in the ethical sphere, ‘ought’ implies ‘can’.9 And so when attributing subjective justiﬁcation, we must take into account the actual epistemic abilities of agents. What can be reasonably expected of humans, epistemically speaking? Well, we cannot expect them to calculate conditional probabilities for all their beliefs and have a belief set that conforms to Bayesian axioms. Perhaps it would be ideal if they could—but it would also perhaps be ideal if humans could withstand any amount of torture without betraying their principles and ideals. What would be ideal is not relevant to what counts as good in the actual world, given our actual abilities. Can humans be expected to have a belief system that satisﬁes coherentist constraints? Probably not; again, as we saw in Chapter 4, our belief systems are massive things, and examination of whether the system is coherent or not can only proceed piecemeal. Clearly, there is a sense in which a coherent belief system is superior to an incoherent one—but if facts about the coherence of our belief system are inaccessible to us, then they are useless to our actual practice of epistemology. Only insofar as our ongoing epistemic practice reveals incoherencies in our belief system (or insofar as good epistemic practices would reveal such incoherencies) can we in any pragmatically useful sense be held accountable for these incoherencies. This point has been argued by others. Mark Lance, writing about theories (like Bayesianism and coherentism) that impose structural constraints on agents’ beliefs, argues that “the requirement of coherence is too strict, since it is no epistemic fault to violate coherence so long

198

Pragmatism and Epistemology

as the violation is not something one could reasonably be expected to notice.”10 The standard Bayesian response that he cites to this completely undermines the potential role of such theories in subjective justiﬁcation attributions. Kaplan, responding to the well-known criticism that Bayesianism represents requirements that are too strict for agents to actually fulﬁll, writes: Being a regulative ideal, Modest Probabilism [as Kaplan refers to his version of Bayesian epistemology] is not to be understood as giving you a set of marching orders which you can violate only at the cost of being classiﬁed as irrational. Modest Probabilism is, rather, to be understood as putting forth a standard by which to judge the cogency of a state of opinion—a standard according to which any violation of Modest Probabilism opens your state of opinion to legitimate criticism. To say that your state of opinion is open to legitimate criticism if it violates Modest Probabilism is not to say that you are open to legitimate criticism in such an event. You can hardly be held open to criticism for violations of Modest Probabilism . . . that are due only to your limited cognitive capacities, limited logical acumen, or limited time.11 But of course what we are looking for in subjective justiﬁcation attributions is precisely a set of standards according to which you can be legitimately criticized (or praised) for your epistemic performance. If, as Kaplan suggests, Bayesianism proposes standards which it is unrealistic to expect normal epistemic agents to be capable of meeting, then we need a different standard to judge when agents are performing well relative to the human norm—that is, relative to what humans are in fact normally capable of. To the extent that Bayesianism is offered as a regulative ideal, and an agent is not necessarily irrational for failing to adhere to Bayesian constraints, Bayesianism does not offer a complete theory of rationality, and must be supplemented with something else. However, Kaplan’s suggestion is that we should use Modest Probabilism as a tool for evaluating ‘states of opinion’ considered somewhat in the abstract. This suggests that theories like coherentism or Bayesianism might be pressed into service at the level of objective justiﬁcation. We will see in the next chapter that this suggestion turns out to be false, but let us not get ahead of ourselves. So our subjective justiﬁcation attributions will be affected by these three factors: the evaluative frame of reference, the agent’s circumstances, and what can normally be expected of people. The fact that

Pragmatism and Epistemology

199

epistemic evaluation moves along these three axes makes it difﬁcult to say anything general about what, precisely, subjective justiﬁcation consists in. However, we can draw some conclusions, both negative and positive, from our discussion so far. Let us begin with the negative conclusions. (1) We cannot require that an agent’s belief system satisfy the structural constraints set forth in contemporary epistemological theories such as coherentism, foundationalism, and so forth. Many current theories of justiﬁcation are ruled out by virtue of the fact that they expect impossible feats of intellectual prowess on the part of the epistemic agent. I have already argued that it is impossible for a human agent to calculate conditional probabilities for all her beliefs and have a belief system that conforms to Bayesian axioms. I have also claimed that it is impossible for an agent to ascertain whether her belief system satisﬁes coherentist constraints. Incoherencies in one’s belief system cannot simply be seen by casting an appraising eye over the whole of one’s belief system; nor can one reveal the coherence of one’s belief system in this manner. It is sheer fantasy to suppose that we have such mastery over our belief systems.12 Incoherencies in one’s system can only be discovered slowly and piecemeal. Indeed, it seems that any demand that one’s belief system meet certain overall formal constraints is an impossible demand to make on an agent. Consider, for example, a philosophical book. Although the author tries, of course, to make the overall account coherent, it is always possible (and indeed likely) that only once other philosophers examine the book carefully will the book’s inconsistencies and incoherencies emerge. If a philosopher cannot root out all incoherencies in something as self-contained as a book, what hope is there that an agent will be able to root out the incoherencies in her entire belief system, merely by critically examining it? This hope is, of course, vain. Whether the formal constraints are foundationalist—all beliefs in the system must be inferentially derived from foundations—or coherentist or Bayesian or whatever, it is not possible to expect a person to be able to determine whether her belief system conforms to these constraints, much less bring the system into conformity with these constraints. And so the requirement that our internalist justiﬁcation attributions reﬂect what is normally possible of epistemic agents rules out most formal epistemological systems. As Lance writes, “The reason is that the logical properties of reasonably complex systems . . . are no more transparent to the mind than are, for example, the causal pedigrees considered by reliabilists.”13 I distinguished earlier between epistemology in the broad

200

Pragmatism and Epistemology

and narrow senses. In the narrow sense, epistemology is our practice of evaluating beliefs as justiﬁed or unjustiﬁed, methods of inquiry as rational or irrational, agents as having performed epistemically well or badly, and so forth. Epistemology in the narrow sense can merely be an attempt to uncover some of the features of justiﬁcation and knowledge, without any commitment to the possibility of formal theorizing; but many philosophers engaged in philosophy in the narrow sense take their task to be the construction of a set of formal conditions for the justiﬁcation of beliefs (e.g., explanatory coherence among beliefs; proper inferential relation to foundational beliefs; proper causal genesis, à la reliabilism, etc.). It appears that human cognitive limitations eliminate such formal theories from the realm of subjective justiﬁcation. Unless one of these theories can be pressed into service at the level of objective justiﬁcation, the prospects look dim for formal theories of epistemology that impose structural constraints on our belief systems. Let us move to the second negative constraint on internalist justiﬁcation attributions. (2) Agents cannot be required to keep track of the justiﬁcations for their beliefs. It has commonly been argued that agents often forget why they hold certain beliefs; they cannot keep track of the justiﬁcations for all of the beliefs they hold. Thus, for example, I know that kangaroos are marsupials, but I do not recall where I learned this fact, or on what authority I know it. As agents cannot keep track of the justiﬁcations for their beliefs, this cannot be required of epistemic agents. (3) Agents cannot be expected to be able to justify inferentially each of their beliefs. Commonly, an epistemologist will claim that though it is not necessary for every belief to be inferentially justiﬁed, it is necessary for the agent to be able to present an inferential justiﬁcation for a given belief, if challenged.14 However, agents cannot be expected to do this. This is not merely because the agent might have forgotten the justiﬁcation for a particular belief (as in (2) above), but because in many cases competent epistemic agents simply cannot provide such a justiﬁcation. For example, Sellars has argued that for a perceptual belief to be justiﬁed, the agent in question had to be able to present an argument of the form, “Perception is reliable under standard conditions, which now prevail, and so since I am disposed to form the perceptual belief that p under these circumstances, I am justiﬁed in believing that p.”15 As Brandom argues,16 it is unlikely that all competent epistemic agents can produce an argument along these lines to justify their perceptual beliefs. The typical epistemic agent is nevertheless a reliable perceiver: her perceptual judgments are in general reliable. That is, perceptually speaking,

Pragmatism and Epistemology

201

most agents perform well, epistemically. And so we can endorse an agent’s epistemic performance even if the agent is unable to provide an inferential justiﬁcation for her perceptual beliefs. One might regard the above incapacities as forming the basis of a reductio of internalism. Clearly, subjective justiﬁcation has clear afﬁnities with traditional internalism in epistemology, and I have already indicated that I take my account of internalist justiﬁcation attributions to be a form of internalism (a claim I will make good on in the next chapter). Thus, one might regard our inability to provide inferential justiﬁcation for our beliefs, to recall the justiﬁcations for our beliefs, to evaluate the coherence of our belief system, and so forth as inabilities that cast doubt on the viability of internalism. Here is why one might regard the above incapacities as part of a reductio of internalism. First (one might argue), the incapacities described above demonstrate that the average epistemic agent is unable to fulﬁll certain reasonable epistemic obligations: for example, although it is reasonable to require an agent to know why it is reasonable to hold a particular belief, epistemic agents are unable to provide any inferential justiﬁcation for many of their beliefs. The other incapacities (inability to evaluate the coherence of one’s belief system, etc.) also demonstrate that epistemic agents are unable to fulﬁll basic epistemic obligations. Thus, the deontic conception of justiﬁcation (i.e., the conception of epistemology in which being justiﬁed consists of adhering to ones epistemic duties) must go. The pragmatist might respond that since ‘ought’ implies ‘can’, our inability to fulﬁll these obligations shows that they are not epistemic obligations in the ﬁrst place. But this reply fails to understand the force of the anti-internalist objection. The objection states that any internalist epistemology must be able to recognize such basic actions as being able to justify one’s beliefs to be epistemic obligations. But as we cannot perform these basic actions, internalism cannot recognize as obligations things which it must reasonably be able to recognize as obligations, if it is to be a viable theory of justiﬁcation. That is, the notion of being responsible for one’s beliefs includes, among other things, that one can justify one’s beliefs, that one can explain why the belief is reasonable, or likely to be true. But as normal epistemic agents are unable to do this for a wide variety of beliefs that they hold, it is not possible for epistemic agents to be responsible for their beliefs. Epistemic responsibility is not a viable notion; agents cannot therefore be held responsible for their beliefs, and internalism is false.

202

Pragmatism and Epistemology

The only problem with the above anti-internalist argument is that it assumes that to evaluate an agent’s epistemic performance is the same as evaluating how well one has fulﬁlled one’s epistemic obligations. At the end of Chapter 6, we as much as conceded that the deontic conception of justiﬁcation needed to be abandoned. But subjective justiﬁcation has nothing to do with the fulﬁllment of obligations. In evaluating whether an agent has performed badly or well, we are not asking whether the agent has fulﬁlled her epistemic duties; rather, we are asking whether the agent has performed well or badly (relative to a set of circumstances, an evaluative frame of reference, etc.). Evaluating a performance as good or bad is not the same as holding the subject of our inquiry responsible for this good or bad performance. To give an example, to say that my watch performs well is not to hold my watch somehow responsible for its good performance; in no way am I saying that the watch has fulﬁlled its chronological obligations, or that it is in any way responsible for its character. Analogously, we can evaluate the performance of epistemic agents without holding them responsible for their epistemic performance. Recall the argument from the end of Chapter 6: to hold an agent responsible is to say that the agent may be subjected to sanctions. But we do not sanction agents for poor epistemic performance; we merely exclude them. And I argued at the end of the last chapter that excluding someone is not the same as sanctioning her. But the most important rejoinder to this attempted reductio is this: recall that epistemic evaluations are always relative to a particular background, including facts about what humans are normally capable of. Thus, against a background of normal human abilities, we can evaluate certain epistemic performances as better or worse. The inability of humans to provide any epistemic justiﬁcation for most of their beliefs might mean that in some abstract sense, humans are not capable of performing very well epistemically, and therefore every human epistemic performance is in some sense badly ﬂawed. But to give an argument along these lines trying to show that no human epistemic performance is ever good from an internalist standpoint is like arguing that Carl Lewis is not a fast runner because he is so much slower than a photon. Relative to human capabilities, Carl Lewis is very fast. And relative to human capabilities, a given epistemic performance might be outstanding. A hypothetical perfect reasoning machine that can somehow represent its entire belief system instantly, detecting all inferential relations within the system and assessing the system’s coherence, might be conceivable; but the performance of such a machine does not set the standards by which human epistemic performance is to be judged. Thus,

Pragmatism and Epistemology

203

the above objection does not serve as an effective reductio of internalist justiﬁcation. If, in rendering a judgment about subjective justiﬁcation, we are merely evaluating an agent’s epistemic performance, does that mean assimilating subjective justiﬁcation to the model of reliabilism? This might seem like a natural way of construing this notion of subjective justiﬁcation; we construe it in terms of whether an agent is forming beliefs in an accurate way or not, that is to say, whether the agent used a reliable mechanism in forming her belief. Thus, it seems plausible to say that ‘performed well, epistemically’ equals ‘used a reliable belief-forming method’, and ‘performed badly, epistemically’ equals ‘used an unreliable belief-forming method.’ This may seem plausible, but it is misguided. Performing well, epistemically, is not the same as using a reliable belief-forming method. As I noted above, evaluation of one’s epistemic performance can be relativized to circumstances such that even if one uses an unreliable belief-forming mechanism, it might still be true that one performed well in the circumstances. Reliabilism cannot distinguish between someone who performed well epistemically, but whom circumstances prevented from using a reliable belief-forming mechanism, and someone who performed well epistemically, and who due to favorable circumstances succeeded in using a reliable belief-forming mechanism. As one commentator puts it, reliabilism “has difﬁculty doing justice to our intuition that a Newton and an Einstein may be roughly equal in virtue, while far apart in terms of truth.”17 Thus, subjective justiﬁcation cannot be construed in terms of reliabilism. The relativization of subjective justiﬁcation to circumstances, etc. allows us to say that a person might be subjectively justiﬁed, even if the belief-forming mechanism she used is not reliable. We have drawn several negative conclusions from the three factors affecting subjective justiﬁcation (the evaluative frame of reference, the agent’s circumstances, and what can normally be expected of people). However, some positive conclusions also follow. These require a fair amount of elaboration and defense, and it is to this task that I now turn.

Internalism, conservatism, and revision In the following pages, I hope to demonstrate two things. First, I hope to show that conservatism is an essential part of subjective justiﬁcation. The version of conservatism defended here applies only to what I will

204

Pragmatism and Epistemology

call our basic beliefs (which, for dialectical reasons, I shall only fully characterize later). For a variety of reasons to be outlined below, an epistemic agent is entitled to these beliefs even if she is not able to provide any inferential justiﬁcation for them. The second conclusion I will draw is that the truth of conservatism places a good deal of importance on the practice of belief-revision, when evaluating an agent’s subjective epistemic performance. The principle of conservatism that is defended in this chapter will have several qualiﬁcations, but it would be premature to introduce them now; let us see how they emerge dialectically as the chapter progresses. We must be sensitive, or course, to the charge that conservatism is inherently anti-epistemic, and so I will foreshadow some of the conclusions we will reach concerning conservatism when we discuss objective justiﬁcation. It will emerge that although conservatism is an essential part of subjective justiﬁcation, we may attribute objective justiﬁcation to conservatively held basic beliefs. Those who deny that conservatism is supported by epistemic considerations do so because they fail to take into account the necessarily social, diachronic and self-correcting nature of our epistemic practice. It will emerge that justiﬁcation is essentially historical in nature, and that the history an epistemic practice must have to be objective-justiﬁcation-conferring is a generations-long history. Thus, this historical requirement on objective justiﬁcation makes objective justiﬁcation social as well. But let us not get ahead of ourselves; let us begin our discussion of conservatism as it relates to subjective justiﬁcation. It has been argued that conservatism is an unavoidable part of justiﬁcation. For example, Lawrence Sklar has argued that “conservatism lies at the very basis of any possible structure for justifying beliefs at all.”18 Sklar starts with the seemingly undeniable premise that “all epistemic justiﬁcation is relative to an assumed background of believed theory.”19 From this starting point, Sklar argues as follows: [W]e must realize that all justiﬁcation is “local.” We justify the beliefs we take to be in need of justiﬁcation “one at a time,” using all the resources of our unchallenged background belief in the process. Such “local” justiﬁcations are the only justiﬁcations of which we can make sense, for all justiﬁcation requires a body of unchallenged background belief, and we never could justify our totality of beliefs “all at once.”20 Thus, all local justiﬁcation occurs in the context of a set of background beliefs that is merely accepted, and cannot be argued for. This is not a

Pragmatism and Epistemology

205

new position. In On Certainty, Wittgenstein writes, “I did not get my picture of the world by satisfying myself of its correctness; nor do I have it because I am satisﬁed of its correctness. No: it is the inherited background against which I distinguish between true and false.”21 To give an example, when I get out of bed in the morning, I do not have to consider the hypotheses that putting my feet on the ﬂoor will cause them to explode, or that turning off my alarm clock will cause demons to rain out of the heavens. Without a background theory of the world ruling out such hypotheses, uncertainty would rule, and no action would be possible. Of course, I do not explicitly consider these bizarre hypotheses, and judge them unworthy of further consideration. I merely behave as if they were false, and they never enter my mind. This is why, I think, Wittgenstein says our knowledge is grounded in a certain way of acting:22 our knowing something is in many cases not a matter of considering and rejecting rival hypotheses; instead, it is in large part a matter of acting in a way that is incompatible with belief in such rival hypotheses. We have a way of acting, a way of ignoring certain hypotheses, that guides our empirical inquiry, and shapes how we explore the world. Normally, we do not need to respond to challenges to this way of acting; I do not need to justify ignoring these rival hypotheses. Rather than standing in need of justiﬁcation, this way of behaving provides the background against which our other beliefs are subjectively justiﬁed. But our way of acting as if certain hypotheses are false forms the riverbed in which the water of our empirical knowledge ﬂows, to use Wittgenstein’s analogy. To summarize our positive conclusion regarding subjective justiﬁcation, conservatism is an unavoidable part of subjective justiﬁcation. As Sklar and Wittgenstein recognize, all inferential justiﬁcation takes place against a background of beliefs that the individual is not capable of justifying. There are certain principles we must merely accept. And so conservatism is forced upon us by at least two considerations: First, it is forced on us by our limitations as human agents, by our inability to evaluate massive chunks of theory at once. Second (and perhaps more fundamentally), conservatism is forced on us by certain unalterable facts about empirical inquiry. It is not really just a human frailty that we cannot evaluate our entire belief system at once; rather, it is a logical difﬁculty: Something must remain ﬁxed if we are to judge; something must serve as the hinge on which judgment turns. In discussing which theory we ought to endorse, it is taken for granted that certain basic propositions must be accepted—accepted as the basis for making further judgments and for evaluating various hypotheses that

206

Pragmatism and Epistemology

confront us. So conservatism is unavoidable. When we are evaluating a particular agent’s epistemic performance, we must bear in mind that it is impossible (for purely logical reasons and also for reasons related to the limitations of human cognitive abilities) for the agent to avoid conservatism altogether. One interesting consequence of the centrality of conservatism to subjective justiﬁcation is that this conservatism gives us reason to emphasize the importance of revision in subjective justiﬁcation. Consider, again, the issue of conservatism: when one is trained into a set of beliefs (as a child, or as a novice scientist or philosopher, etc.), one is given a starting point that is (from the perspective of the child or novice scientist) entirely arbitrary.23 We do not choose which theory of the world we are dogmatically trained into as children; this is beyond our control. Thus, one’s epistemic performance cannot be judged by the theory one holds initially; rather, the theory one is trained into is one of the circumstances to which epistemic evaluation is relativized. One’s epistemic performance cannot be judged by the theory one holds because, as I argued earlier, humans are not capable of evaluating their belief systems tout court. One is forced to take one’s belief system as a given, and revise it piecemeal. Thus, consideration of our epistemic capacities (and of the fact that subjective justiﬁcation attributions must be relative to what is humanly possible) supports conservatism: one’s epistemic performance cannot be faulted on the grounds that one cannot evaluate (and then accept or reject) one’s entire belief system tout court. One must accept one’s belief system as given—one must accept it conservatively—and use this system as the basis for further revisions. Thus, emphasis on conservatism supports giving revision an important role in subjective justiﬁcation attributions. As each of us is given a more or less arbitrary starting place, it seems contrary to the spirit of subjective justiﬁcation to evaluate a person’s epistemic performance based on his or her starting place. After all, we do not choose which theory of the world we are dogmatically trained into as children; this is beyond our control (and hence, we are not to be epistemically praised or blamed for holding this system). Similarly, as we will emphasize later in this chapter, an aspiring scientist being trained into the currently held scientiﬁc theory does not have the knowledge to evaluate whether this theory is correct or not; only once the scientist has been trained into the theory and given its epistemic resources can she go about trying to prove, disprove, or revise the theory. We must merely accept an arbitrary starting place (although we will see that from the perspective of objective justiﬁcation this starting place need not be regarded as entirely

Pragmatism and Epistemology

207

arbitrary), and then use that starting place as a platform for revising our theory, always using (subject to their revision) the tools that are initially given to us by this starting place. This strongly suggests that our evaluation of a person’s epistemic performance should depend not so much on her starting place (and whether she has evidence for her ‘starting theory’), but rather on how she revises her beliefs over time in light of new evidence and so forth. So it is not epistemically irresponsible or unsound to hold a belief for which one has no evidence (we cannot help but do that, as we are simply dogmatically trained into a particular practice); rather, it is epistemically unsound for an agent to revise her theory without being directed by evidence, or failing to revise the theory in the face of new evidence.24 We will return to these points in the next chapter. A critic, however, will no doubt say that this picture of justiﬁcation is a picture of dogmatism, a system that allows a background system to be held conservatively, without (inferential) justiﬁcation. The way of acting needs no justiﬁcation, indeed! There are several points to be made in response to this. First, as noted above, subjective justiﬁcation attributions must be sensitive to what is humanly possible in terms of cognition. And as I have argued, both here and in Chapter 4, conservatism is forced on us by our cognitive limitations. A person who holds certain beliefs conservatively does not go wrong, epistemically, because we can expect no better of someone with limited reasoning ability. Indeed, as conservatism is (as Sklar and Wittgenstein recognize) a logical requirement on justiﬁcation, and not merely a function of contingent human cognitive limitations, we are forced to treat the necessity of conservatism as a sort of mitigating factor in evaluating people’s epistemic performance from the subjective standpoint. Second, conservatism as deﬁned here is limited in scope. It applies only to a narrow category of beliefs, those I am characterizing as our basic beliefs. The third point, which is crucially related to the second, is that these basic beliefs are only in one sense conservatively held. We can transcend mere local justiﬁcation, and give a justiﬁcation that shows how, when viewed from the standpoint of objective justiﬁcation, these basic beliefs can be justiﬁed. I should note that as conservatism is an unavoidable part of the human epistemic condition, it must be present even in objective justiﬁcation attributions. That is, often we will attribute objective justiﬁcation to conservatively held beliefs, or to beliefs that rely for their objective justiﬁcation on their inferential relations to conservatively held beliefs. We will show, however, that when viewed from the standpoint of objective justiﬁcation, conservatively

208

Pragmatism and Epistemology

held beliefs can be considered rational, because we can give a historical vindication of them (as I will endeavor to do below). To further anticipate the course of the argument, I will argue later that knowledge requires both subjective and objective justiﬁcation; and so, for S to know that p, the conservatively held beliefs that provide the necessary background for S’s knowing that p must themselves be rational (in the objective, external sense). Thus, no one can know something on the basis of a non-rational or irrational conservatively held belief. I believe the points outlined in this paragraph, once established, will clear our pragmatist epistemology from any charges of countenancing dogmatism. Let us discuss external justiﬁcation attributions, and how certain features of objective justiﬁcation help allay worries about dogmatism.

Externalist justiﬁcation attributions As noted above, a person can be subjectively justiﬁed even if her belief or theory is not objectively justiﬁed. That is, even if we recognize that a person’s epistemic performance cannot be faulted, that does not mean that we recognize the person’s belief system is one we rationally ought to adopt. I cannot fault Newton’s epistemic performance (at least not when it comes to physics), but we now recognize that there is in fact a better-justiﬁed theory regarding the interaction of massive bodies over great distances. A person living 5000 years ago may have performed well, epistemically, in forming the belief that the earth was ﬂat. However, we now recognize that this theory is not tenable, and that the evidence available to us refutes this theory of the earth’s shape. Thus, though we might say that Newton’s beliefs were subjectively justiﬁed (in that he performed well epistemically given the relevant background of evaluation), we must say that his theory is objectively unjustiﬁed, because we know that Einstein’s theory is better-justiﬁed than Newton’s. An externalist justiﬁcation attribution is made when we judge that the best reasons support a particular proposition, p. Here we are not evaluating whether a particular agent performed well or not—an agent can perform well (i.e., be subjectively justiﬁed) even when we recognize that her theory is not one that we ought to adopt. So in attributing objective justiﬁcation, we do not evaluate a person’s performance relative to a background. We abstract from relativizing background frames, and judge what theory is best-justiﬁed or most rational without qualiﬁcation. This is not to say that we adopt a ‘view from nowhere’ or somehow slip the surly bonds of our own theoretical commitments. Let me show how such attributions work, the relation they bear to internalist justiﬁcation

Pragmatism and Epistemology

209

attributions, and the relations objective justiﬁcation attributions bear to our current theoretical commitments. At time t, we may judge ‘We ought to believe P,’ based on the fact that as far as we can tell, P is the best-justiﬁed theory. This is an externalist justiﬁcation attribution; it is about which theory we ought to believe. Of course, at time t + n, we may revise our judgment, and judge that Q (which is incompatible with P) is the best-justiﬁed theory. Does that mean we were wrong to claim that P was objectively justiﬁed? It does not necessarily mean this. However, note that in judging our past performance, we are changing perspectives from objective to subjective justiﬁcation: we are saying, in essence, ‘Q is the best-justiﬁed theory, but at time t, we did not go wrong in believing P; it was perfectly rational, given the evidence we had.’ Such retrospective attributions are one place where internalist justiﬁcation attributions have their place. Thus, objective justiﬁcation attributions are made using the theoretical tools we have at our disposal. At time t, we judged (using, of course, the background theory we had at our disposal at that time) that P was the best-justiﬁed theory without qualiﬁcation. We may go on to reject those theoretical tools, and decide that a different and incompatible theory is objectively justiﬁed. But that does not mean that we went wrong in endorsing the earlier theory. At t + n, we judged that at time t, we were correct in endorsing P—given the theoretical tools we had at our disposal at time t. So this judgment relativizes the justiﬁcation attribution to a background, and evaluates the performance of the epistemic agents within and relative to that background. It is a subjective justiﬁcation attribution. Objective justiﬁcation attributions are, in a sense, the limiting case of subjective justiﬁcation attributions. A very self-reﬂective agent who understands human fallibility, in saying that Q is objectively justiﬁed, might recognize that in a sense she is making a subjective justiﬁcation attribution—but the assumed background is our epistemic performance and our background theory. She is saying that given our theoretical background and assuming competent epistemic performance on our part, we do not go wrong in believing Q. But such a judgment is only relativized in the trivial sense that all judgments of objective correctness that we make are made using our conceptual devices and our theoretical commitments. Objective justiﬁcation attributions claim that Q is the best-justiﬁed theory (i.e., the one best-supported by the evidence)—by our lights. Subjective justiﬁcation attributions are intrinsically relativized in the three ways outlined above, and crucially, in making a subjective justiﬁcation attribution, we are in no way endorsing the belief in

210

Pragmatism and Epistemology

question, or saying that it is one that we should adopt. We are only evaluating the performance of a particular agent (or group of agents), relative to a certain background. An objective justiﬁcation attribution essentially involves endorsement of the belief or theory in question, and this represents a fundamental difference between subjective and objective justiﬁcation attributions. I said in Chapter 4 that in Chapter 7 we would develop the tools to see in precisely what respect epistemology is relative. We have seen, in the above paragraphs, the sense in which epistemological evaluation is temporally relative: in evaluating past epistemic performance, we shift to the internalist perspective, and these internalist attributions are relative to the time of the performance being evaluated. Thus, belief in P is subjectively justiﬁed relative to the background conditions of time t; It would not be subjectively justiﬁed relative to the theoretical background obtaining at time t + n. One pragmatic purpose served by externalist justiﬁcation attributions is to help maintain scientiﬁc consensus. As Kuhn argues,25 science advances much more rapidly when consensus is maintained than when it is characterized by ‘divergent’ thinking. Thus, ensuring convergence by excluding those who reject the current paradigm aids rapid scientiﬁc advancement. Consider an example: for a physicist working today, it is rational to believe that relativistic physics is correct; this theory is objectively justiﬁed. It is rational to carry out research projects using relativistic techniques, and it would be irrational to attempt to use Newtonian physics to account for (say) planetary motion.26 A physicist who incorporated unjustiﬁed techniques (such as those of Newtonian physics) would be sanctioned by partial or complete exclusion from the profession of physics. Those who are adept at applying relativistic techniques to the study of planetary motion are rewarded by inclusion. And this is not a matter of subjective justiﬁcation: that is, inclusion is not based merely on the judgment that a scientist has performed well, epistemically. A scientist raised in an insular community of Newtonians might have performed brilliantly, epistemically; but he will still be excluded from the community of physicists at large because of his use of a discredited theory. This example brings out a further detail of externalist justiﬁcation attributions. Such attributions need not be explicit; they can be discursive or practical. In other words, we can call a methodology irrational or a belief unjustiﬁed, or we can carry out our scientiﬁc practice in a way that favors that methodology or presupposes the falsity of that belief. And when our professional physicists apply relativistic techniques to the

Pragmatism and Epistemology

211

study of planetary motion, they are practically (though not discursively) endorsing these techniques. Similarly, the exclusion of the Newtonian physicists from their ranks is a form of practical evaluation; it is an action that practically judges belief in Newtonian methods to be unjustiﬁed. This practical evaluation is crucial to scientiﬁc progress. As I mentioned above (in this chapter and at more length in Chapter 2), Kuhn27 notes that if we do not practically favor a particular paradigm over all others, science does not progress rapidly. Rapid scientiﬁc progress requires practical epistemic evaluation. Hence, epistemology serves the pragmatic end of hastening scientiﬁc progress. We can see that externalist justiﬁcation attributions serve a variety of pragmatic goals: they hasten scientiﬁc progress by helping maintain consensus; they serve to endorse a particular theory for one’s epistemic peers; and so forth. In short, they are the tool by which we claim that one theory is better-justiﬁed than the rest, and ought to be the one that we endorse. Given the argument of Chapter 2, we should perhaps hesitate to talk about objective justiﬁcation as though it were some thing or property. But as we endorsed in Chapter 2 a legislative or expressive account of normative discourse, we need not think of objective justiﬁcation in this way; rather, making externalist justiﬁcation attributions is something we do. Objective justiﬁcation need not be some property (like reliability) that may or may not be accessible to us. Rather, objective justiﬁcation attributions are a tool that serves the pragmatic goals of endorsing certain theories. We may endorse a particular theory only to realize at a later date that this theory is mistaken, and another one better-justiﬁed. But this only shows that the possibility of error is ever-present (a point I will discuss in more detail anon). The important thing, however, from the pragmatist point of view, is that objective justiﬁcation is not a property (such as reliability) that may only be contingently accessible to us. As I will argue below, such potentially inaccessible properties are not pragmatically useful. Rather, externalist justiﬁcation attributions are a tool that we use for various pragmatic purposes, to endorse theories or propositions, to legislate belief for our epistemic peers. We argued in Chapter 2 that normative discourse should probably not be understood as primarily descriptive, but should instead be understood as legislative or expressive. Thus, viewing objective justiﬁcation attributions as endorsing a theory ﬁts well with the pragmatist view of normative discourse. It also ﬁts with recent contextualist trends in which philosophers examine not knowledge as an objective property or natural kind, but rather knowledge attributions.

212

Pragmatism and Epistemology

Objective justiﬁcation and the importance of revision Above, we raised the worry that conservatism is inherently dogmatic and anti-epistemic. As I noted there, this worry also arises with objective justiﬁcation attributions, as conservatism must play a part in such attributions as well. Conservatism is forced on us by both logical necessity and by our limitations as human cognizers, and so we will often attribute objective justiﬁcation to conservatively held beliefs, or to beliefs whose justiﬁcation depends on conservatively held beliefs. I indicated earlier that the response to this worry lay in our account of external justiﬁcation. We must now turn to the task of explaining why our pragmatist epistemology is not dogmatic. The refutation of this charge of dogmatism consists of two parts: First, we will show that conservatively held beliefs can be rationally vindicated, and hence shown to possess objective justiﬁcation. Second, we will see that knowledge requires not just subjective justiﬁcation, but also objective justiﬁcation; and so no agent can know a proposition P on the basis of a conservatively held belief that is not itself objectively rational or justiﬁed. The picture of justiﬁcation we have outlined—where justiﬁcation takes place against a conservatively held background of beliefs—only looks dogmatic if you look at it as a static system, with an immovable foundation. However, our background theories that tell us which hypotheses to ignore are themselves revisable. The riverbed moves over time. Consider the proposition, ‘The earth is the center of the universe.’ In the 1920s, the astronomer Edwin Hubble made an interesting observation: in every direction you look, galaxies are receding from the earth. Furthermore, the farther away a galaxy is, the faster it is receding. How did scientists explain this observation? Signiﬁcantly, nobody suggested the hypothesis that the earth is at the center of the universe; that hypothesis was not one that even merited discussion. Scientists eventually settled on the hypothesis that space itself is expanding. It was a revisionary hypothesis, but no one was willing to postulate a geocentric universe. Notice, though, that 1000 years ago the proposition ‘The earth is at the center of the universe’ was not merely a hypothesis that needed to be taken seriously; cosmological hypotheses that conﬂicted with this belief were immediately rejected. Thus, the hypothesis that the earth is at the center of the universe started out as a proposition used to test hypotheses, and ended up as a hypothesis to be discarded without serious consideration.28 It is for this reason that Wilfrid Sellars writes, Above all, the [traditional picture of knowledge] is misleading because of its static character. One seems forced to choose between the picture

Pragmatism and Epistemology

213

of an elephant which rests on a tortoise (What supports the tortoise?) and the picture of a great Hegelian serpent of knowledge with its tail in its mouth (Where does it begin?). Neither will do. For empirical knowledge, like its sophisticated extension, science, is rational, not because it has a foundation but because it is a self-correcting enterprise which can put any claim in jeopardy, though not all at once.29 As our system of empirical knowledge evolves through revision (which revision occurs through argument and the ‘tribunal of experience,’ as Quine calls it), we sometimes realize that what were groundless beliefs, in no need of inferential justiﬁcation, are false and need to be discarded. That is why our system of knowledge is rational: it is rational because no belief has been in principle immune to revision. We are objectively justiﬁed in ruling out hypotheses such as ‘Demons will fall from the heavens if I turn off my alarm clock’ because our theory of the world, which dictates that such hypotheses be ignored, has survived the tribunal of experience, and because the theory has been revisable in light of evidence that it is false. Revisability, not a foundation, is the source of objective justiﬁcation when it comes to basic beliefs.30 So judging requires that certain background beliefs merely be accepted without being argued for. Recall Wittgenstein’s quote, which was cited earlier: “I did not get my picture of the world by satisfying myself of its correctness; nor do I have it because I am satisﬁed of its correctness. No: it is the inherited background against which I distinguish between true and false.”31 But this is not dogmatism, because these background beliefs evolve as our empirical knowledge grows.32 It is through the evolutionary pressures imposed on theory by experience that old background beliefs get rejected and new ones introduced. And it is this evolution over time—this revision of background beliefs over time, to better accord with experience—that allows us (correctly) to regard these background beliefs as adequate.33 For even if we did not choose these background beliefs ourselves, they are not arbitrary: they are the product of millennia of empirical inquiry. This fact is what makes our body of empirical knowledge rational. Our critic, then, misunderstands the nature of justiﬁcation: she thinks that for a system to be justiﬁed at time t, it must be possible at time t to give an explicit justiﬁcation for every belief in the system. But this picture of justiﬁcation ignores the fact that only some beliefs are justiﬁed this way; others (namely our basic beliefs) are justiﬁed purely by their history. Thus, the critic ignores the temporal element crucial to an understanding of objective justiﬁcation: the basic beliefs in the system, the ones for which we can offer no inferential

214

Pragmatism and Epistemology

justiﬁcation, are objectively justiﬁed—and thereby ﬁt to serve a justiﬁcatory role—because they themselves have withstood the test of time and evidence, because these beliefs are the product of epistemic evolution, because the system has been allowed to evolve over time. A system that became immune to revision would before long cease to be justiﬁed. The system, then, is rational, not because of its structure at time t, but because no belief in the system has always been de jure immune to revision. Let me put this point another way: having made a judgment, you may go on to justify the judgment, but this justiﬁcation will, of course, rest on ‘assumed background of believed theory,’ as Sklar puts it. One might, if the dialectical situation requires it, go on to justify some or all of these assumed background beliefs; but of course, such a justiﬁcation will itself rely on an assumed background of theory. You might suppose that at some point, we will reach a set of beliefs that we cannot justify; there are no beliefs more basic than these that we could use to justify these ‘foundational’ beliefs. As Wittgenstein says, at some point “I have exhausted the justiﬁcation, I have reached bedrock and my spade is turned. Then I am inclined to say, ‘This is simply what I do.’”34 Does that mean these ‘bedrock’ beliefs are arational? No;35 to think so is to think that a belief is only justiﬁed if we can present an explicit inferential justiﬁcation for that belief. Some beliefs are justiﬁed that way, but our ‘foundational’ beliefs are justiﬁed for a different reason: they are justiﬁed because they are the result of a millennia-long inquiry of the world, because they are the product of epistemic evolution. It follows that if we were to declare a set of basic beliefs de jure unrevisable, then in short order these beliefs (and all that they support) would cease to be justiﬁed. The reason is that justiﬁcation for our basic beliefs relies in large part on their having faced the tribunal of inquiry and survived, and declaring a set of basic beliefs de jure unrevisable is to remove them from before this tribunal.36,37 But if we have not treated our background beliefs as de jure unrevisable, then our system of beliefs is justiﬁed. Thus, although many philosophers (including Wittgenstein) have argued that justiﬁcation requires foundations, the correct view is that justiﬁcation requires revisability. It is only because of their historical revisability that our so-called foundations (the basic propositions) are justiﬁed. Because of the fallibilism implicit in this epistemology, with its emphasis on revision, such beliefs are not foundational in the sense of immutable, of course; so it would be inaccurate to label the epistemology being offered here foundational. One might object that some basic beliefs, such as logical beliefs, have not been revisable, but are nevertheless objectively justiﬁed. There are

Pragmatism and Epistemology

215

two ways of thinking about this issue. One way (which I take to be Wittgenstein’s view) is that such beliefs have not been revisable, and therefore are not objectively justiﬁed—but these beliefs form the framework of our language, the framework within which justiﬁcation takes place. The other way of thinking about this issue is more Quinean: such beliefs may never be revised, but what makes them objectively justiﬁed is that they have been open to revision. If a belief (such as a belief in the law of the excluded middle) has been open to revision for centuries, but so far no compelling reason has risen to revise it, then the belief has earned its epistemic status by virtue of having survived this centuries-long test. Thus, objective justiﬁcation does not necessarily require that the belief is the product of revision; objective justiﬁcation requires only that all of our basic beliefs must have been subject to revision as a condition of their being justiﬁed. As Mark Lance and John O’Leary-Hawthorne put the point, Treating a whole bunch of claims as de jure unchallengeable . . . seems . . . constitutive of dogmatism . . . Meanwhile, the recognition that some claims may turn out to be de facto unchallengeable (and even necessarily so) runs no similar cognitive risks.38 It should be clear that this chapter adopts the latter Quinean approach. The reason it is not dogmatic to judge by our framework propositions is that these framework propositions (not just laws of logic, but more malleable framework propositions, such as the belief that no human has ever set foot on Mars) have not been immune to revision. Despite the defense offered above, one might nevertheless think that the present account suffers from a kind of dogmatism. The objection goes as follows: You say our background beliefs are justiﬁed because they derive from a history of revision. But any history will start with arbitrarily chosen beliefs. Our current beliefs are determined by an arbitrary initial choice, and there is no reason to prefer one choice to another. So our current background beliefs are arbitrary, and holding them is dogmatic.39 This objection overstates the inﬂuence of our starting place on our current theory. It is plausible to think that our current theory depends more on the revisionary pressures our theories have encountered over the years than on our starting place in the distant past. Consider

216

Pragmatism and Epistemology

an analogy from evolution. Richard Dawkins writes that “eyes have evolved no fewer than forty times, and probably more than sixty times, independently in various parts of the animal kingdom.”40 So similar environmental inputs led various lineages to develop the same feature. Now it may be pointed out that all of these different lineages have a common descent, but it is not their common descent that explains their common evolution of the eye. Rather, similar environmental and evolutionary pressures caused the eye to evolve again and again; common descent plays little or no role in the explanation of this. Plausibly, the same applies to the evolution of our belief systems over time. Consider the situation of different communities characterized by different theoretical starting points. As long as they experience epistemic evolution characterized by progress (which we will endeavor to deﬁne in the following few paragraphs), the fact that these communities have different starting points will be less important than the fact that they live in the same world and are hence subject to the same causal inﬂuences and pressures on their scientiﬁc theories. Just as external pressures proved more important than starting place in the case of evolution, so will external epistemic pressures prove more important than starting place in the evolution of our theories. So far, we have used an evolutionary analogy to discuss the type of change our theories undergo over the generations. If researchers allow themselves to be guided by goals such as a desire to reach the truth, it seems likely that epistemic evolution will result in progress. As Karl Popper writes, What characterizes the empirical method is its manner of exposing to falsiﬁcation, in every conceivable way, the system to be tested. Its aim is not to save the lives of untenable systems but, on the contrary, to select the one which is by comparison the ﬁttest, by exposing them all to the ﬁercest struggle for survival.41 In a locus classicus of evolutionary epistemology, Donald Campbell writes that “a blind-variation-and-selective-retention process is fundamental to all inductive achievements, to all genuine increases in knowledge, to all increases in ﬁt of system to environment.”42 It is important to note that the change that confers objective justiﬁcation on basic beliefs must be a certain sort of change. We have talked about the importance of epistemic evolution in the objective justiﬁcation of our background beliefs. But presumably not all change is positive, and hence not all change is justiﬁcation-conferring. What

Pragmatism and Epistemology

217

is needed is that we revise our background theories in a way that ﬁts experience. That is, what is needed is not just change, but progress. But of course, not all researchers are guided by such pure motives, and one might worry, then, that not all theory change represent progress toward better ﬁt between theory and environment. Thus, it is necessary to have some tool in hand for distinguishing progress from mere change (or even regress). Such a tool has been developed by Philip Kitcher in The Advancement of Science. Kitcher notes that progress is not one-dimensional; rather, there are different types of progress, which Kitcher identiﬁes as follows. First is practical progress, which is an increase in our ability to control the world. Then there are varieties of cognitive progress. First is conceptual progress, which Kitcher deﬁnes as follows: Conceptual progress is made when we adjust the boundaries of our categories to conform to kinds and when we are able to provide more adequate speciﬁcations of our referents. Striking examples come from the history of all sciences: ‘planet,’ ‘electrical attraction,’ ‘molecule,’ ‘acid,’ ‘gene,’ ‘homology,’ ‘Down’s syndrome,’ are all terms for which faulty modes of reference have been improved.43 Another type of cognitive progress is explanatory progress. Kitcher writes that Explanatory progress consists in improving our view of the dependencies of phenomena. Scientists typically recognize some phenomena as prior, others as dependent. For example, ever since Dalton, chemists have regarded molecular arrangements and rearrangements as prior to the macroscopic phenomena of chemical reactions, and, since the 1960s, geologists have viewed interactions among plates as prior to facts about mountain building and earthquakes.44 Given these speciﬁc varieties of progress, Kitcher deﬁnes progressive change in a practice as follows: . . . let us say that the sequence of practices P1 , . . . , Pn is broadly progressive just in case for every pair of adjacent members there is a component of practice with respect to which the change from the earlier to the later is progressive and the change from P1 to Pn is progressive with respect to every component of practice.45

218

Pragmatism and Epistemology

Kitcher argues that science in general does progress, and outlines in Chapter 6 the conditions (which he thinks in fact normally obtain) under which science advances. Thus, using tools of the sort supplied by Kitcher, we can deﬁne the sort of progressive change that confers justiﬁcation on our background theories. To say more about the rational progress of science would require a lengthy diversion into the history and philosophy of science, so we will have to rest content with these few suggestions. Before moving on to draw some intermediate conclusions from the above discussion, let me pause to make some comments about the epistemic values that are at work here. In keeping with the evolutionary analogy, the primary notion we are working with is one of ﬁt: beliefs that do not ﬁt with their environment are discarded and replaced with those that do ﬁt. But a creature’s environment has two elements: First, there are the physical, non-living elements of the environment. Second, there are other organisms. Both of these environmental features create pressures on a particular creature; and to survive, a creature must exhibit ﬁtness with respect to both elements of its environment. The analogy can be applied, with only a little strain, to the evolutionary view of epistemology under discussion here: evolutionary forces work through conﬂicts with inputs from the physical world as well as with other beliefs (‘creatures’). Conﬂict with either environmental feature can cause a belief to ‘die out’ and be replaced by a ﬁtter specimen. And thus, we see that the emphasis on ﬁtness makes coherence with other beliefs and coherence with empirical inputs to be important considerations. There are two reasons, however, for regarding environmental inputs as, in an important sense, prior to coherence with other beliefs. First, incoherence is most often introduced into a belief system through observation: observation introduces a new belief into the system which creates incoherence, and thus a ‘struggle for survival’ among the beliefs present in the system. The exponential growth of knowledge over the past centuries has mostly been the growth of empirical knowledge— this is the primary source of new inputs into the system, and so will understandably be the source of most of the inconsistencies. Ideally, the inconsistency will then be reconciled by the familiar considerations of simplicity and so forth. The second reason for the emphasis on observation and experience relates to the issue discussed above: namely, some beliefs are not revised, and are yet justiﬁed. As I indicated above, such beliefs are objectively justiﬁed because they have survived so long without needing revision. But, for their survival to be epistemically meaningful, they must have

Pragmatism and Epistemology

219

survived constant testing. As we noted above, the most torrential source of new information in our system is observation; it is observation that will be placing the most evolutionary pressures on a given belief. If this ﬂow of new information, and the exponential increase in knowledge that we have become accustomed to, does not dislodge a particular belief, then this is a powerful argument that the belief is true: it has survived such a long and severe test at the hands of ever-increasing empirical data. We are now in a position to draw several conclusions about objective justiﬁcation: (1) As we already noted, conservatism is an essential feature of objective justiﬁcation. This conservatism is not anti-epistemic, because of the diachronic nature of objective justiﬁcation.46 We are led to the following, interesting conclusion: Our basic beliefs’ objective justiﬁcation is essentially diachronic, and their objective justiﬁcation requires revisability. If these requirements are not met—if a system of basic beliefs is not revisable—then our conservatism becomes mere dogmatism. Our system of background beliefs becomes objectively unjustiﬁed (as evolutionary forces are not allowed to work on them); we are no longer objectively justiﬁed in ruling out remote hypotheses (as this practice depends on having a justiﬁed background theory, and the objective justiﬁcation of this background theory can only be understood in terms of revisability and diachronic evolution). Objective justiﬁcation has an essentially diachronic element. We will come back to this point in a few pages, to clear up any potential misunderstandings that might arise. (2) The type of conservatism that is defended here is very different from the version of conservatism discussed by most commentators that says, roughly, that a belief acquires some degree of justiﬁcation merely by being believed.47 But this simple version of the principle is almost certainly false. Rather, the principle of conservatism applies only to certain propositions, which we will call basic beliefs. Further, which propositions ﬁt this category is determined by the diachronic progress of our particular social practice. Finally, such beliefs are justiﬁed by their history, rather than by other sorts of reasons we might adduce in support of them. (We will discuss this last point in the following paragraphs.) Let us further explain and deﬁne the notion of basic beliefs and the Version of conservatism defended here. (a) Basic beliefs are propositions that are so basic that we ﬁnd it difﬁcult, if not impossible, to provide any sort of inferential justiﬁcation

220

Pragmatism and Epistemology

for them. Propositions that might ﬁt into this category include ‘Most humans have two hands;’ ‘The sky is usually blue;’ ‘The earth is not at the center of the universe;’ ‘Demons do not interfere with scientiﬁc experiments;’ and so on. If asked to provide an inferential justiﬁcation for such a belief, we might ﬁnd ourselves unsure of what to say. In many cases, we simply do not know how to go about inferentially justifying these beliefs; they simply seem too basic. Another type of case is illustrated by one of Wittgenstein’s famous examples: “My not having been on the moon is as sure a thing for me as any grounds I could give for it.”48 I could provide an inferential justiﬁcation for the claim, ‘I have two hands’ or ‘I have never been on the moon,’ but it is pointless to do so, as the considerations I would cite as support are not more basic than the claim they are supposed to support. Indeed (and this is Wittgenstein’s point), such an inferential justiﬁcation would actually fail to be a justiﬁcation at all, as my not having been on the moon is already as justiﬁed—or even more justiﬁed—than any proposition I could attempt to use as evidence for my belief. Thus, our inability to justify basic beliefs is not always (or not necessarily) a result of human cognitive limitations; it may be logically impossible to justify them. It may be possible for a proposition or set of propositions to entail (strictly or defeasibly) a basic belief, but it may be logically impossible for this relation to count as an evidential or justiﬁcatory relation. (b) These basic beliefs are objectively justiﬁed by the diachronic progress of our process of inquiry. Even though the average person cannot provide an inferential justiﬁcation for such beliefs, these beliefs are (as I argued earlier) objectively justiﬁed by the fact that they are the product of centuries of epistemic evolution, and have so far withstood the challenges that are part of the progress of knowledge and the exponential increase in knowledge and information that accompanies such progress. Thus, the difference between a basic belief and an inferential belief is that a basic belief needs no justiﬁcation beyond its history, whereas an inferentially justiﬁed belief is one that is ultimately justiﬁed on the basis of some basic beliefs.49 (c) It is important to emphasize the sense in which what I am calling conservatism here really is a version of conservatism. From one perspective, it is not conservatism at all: the basic propositions are objectively justiﬁed not merely because they are believed by us, but also because they are the product of a millennia-long process of empirical inquiry. That is, they are justiﬁed by their history. So there is something justifying these basic propositions, something beyond the mere fact that they are believed by us. Thus, from an external view,

Pragmatism and Epistemology

221

taking our theory of the world as our object of study, conservatism is not conservatism at all. However, from the perspective of individual agents, the basic propositions are conservatively held. Most agents are not capable of telling a story about what justiﬁes these basic propositions; they hold them (and are entitled to hold them) for reasons that are conservative in nature. These basic propositions represent a starting point from which the individual can reason.50 Thus, conservatism really is conservatism—at least from the perspective of the individual. And so subjective justiﬁcation attributions can be made to beliefs that are genuinely conservatively held, even though objective justiﬁcation requires that a basic belief have the proper history. (d) A fourth feature of basic beliefs follows from other features of basic beliefs, namely, that basic beliefs are objectively justiﬁed, and yet we generally cannot provide any justiﬁcation (inferential or noninferential) for them, beyond pointing to our history of inquiry (and individual agents will often be unable to do even this). It follows from this that when a basic belief is challenged, the burden of proof is on the challenger to defend his or her position. The holder of the basic belief cannot be expected to defend his belief (as part of the deﬁnition of ‘basic belief’ is that such a defense is in general not possible), and yet the belief is justiﬁed all the same; and so a challenger to this belief must herself shoulder the burden of proof. If the challenger cannot give us evidence suggesting that the proposition in question is false, then we are justiﬁed in continuing to believe that the proposition is true. Indeed, it virtually follows from any deﬁnition of conservatism that the burden of proof is on the challenger: the holder of the basic belief is prima facie entitled to this belief, and so it is up to the challenger to demonstrate that this entitlement does not hold. (e) To say that basic beliefs are prima facie justiﬁed in no way entails that they are immune from revision. Consider an example of a basic belief offered by Wittgenstein: “No one has ever been on the moon.” Wittgenstein writes:

What we believe depends on what we learn. We all believe that it isn’t possible to get to the moon; but there might be people who believe that that is possible and that it sometimes happens. We say: these people do not know a lot that we know. And, let them be never so sure of their belief—they are wrong and we know it. If we compare our system of knowledge with theirs then theirs is evidently the poorer one by far.51

222

Pragmatism and Epistemology

And yet someone who now denied that humans had ever set foot on the moon would be dismissed as ignorant or crazy. So basic propositions are revisable. With the basic proposition that no one has ever been on the moon, the process of revision was speedy, as the event was televised. But with other basic beliefs (such as the belief that the earth is at the center of the universe), conservatism will often entail that their revision will be a lengthy (and sometimes painful) process. And indeed, as I argued above, such basic beliefs must be revisable; it is their revisability that is ultimately the source of their rationality. (3) Our ﬁnal intermediate conclusion is that objective justiﬁcation is social in character in that it presupposes a community with a history of inquiry, with the results of this inquiry (and the evolving epistemic background governing this inquiry) passed down through generations of inquirers. Remember the course of our epistemic argument for the justiﬁcation of basic beliefs: our so-called basic beliefs are justiﬁed by the generations-long history of inquiry that produced them. The argument is this: certain beliefs (such as a belief that demons do not interfere with scientiﬁc experiments) are objectively justiﬁed because they have endured generations of inquiry, practiced by countless researchers, and they have survived this history without refutation. That our basic beliefs survived so long—without becoming encumbered by ad hoc epicycles, open to revision and refutation but not having succumbed—is what lends epistemic weight to our background theory. Thus, objective justiﬁcation is social in that local justiﬁcation requires that our basic beliefs are objectively justiﬁed, and these basic beliefs have their objective justiﬁcation in the history of inquiry that has produced them. As this history spreads over generations and countless inquirers, there is a social dimension to objective justiﬁcation. I will return to this point in a moment, when addressing objections, but the key thing to note is that this history is crucial to the objective justiﬁcation of our basic beliefs; a community lacking such a history would have theories built on sand, lacking objective justiﬁcation. Before moving on, I want to clear up one potential misunderstanding. I am only attempting to establish that revision and diachronicity are crucial to the objective justiﬁcation of basic beliefs—those beliefs that are conservatively held. As I noted earlier in the chapter, some beliefs are inferentially justiﬁed, and others seem to defy such inferential justiﬁcation. It is the latter beliefs whose objective justiﬁcation essentially relies on their place in our history of inquiry, whose objective

Pragmatism and Epistemology

223

justiﬁcation is essentially diachronic and essentially relies on openness to revision. The former are inferentially justiﬁed, and so their justiﬁcation need not rely on the conservatism defended above. Thus, when explaining how an ordinary claim (such as ‘Pelé led Brazil to three World Cup victories’) is objectively justiﬁed, we will appeal not to conservatism, diachronicity and revision, but instead to memory, authority, or some other recognized justiﬁer.52 The objective justiﬁcation of such claims might ultimately trace back to the objective justiﬁcation of our basic beliefs, and so the objective justiﬁcation of ordinary-belief claims might ultimately rely on revision, etc. But this would only show that a claim such as ‘Pelé led Brazil to three World Cup victories’ indirectly relies on revision and diachronicity (in the same way that a foundationalist would say that an inferentially justiﬁed belief indirectly relies on some foundational belief); it is not directly objectively justiﬁed by such considerations. Let us now turn our attention to a preliminary objection. One might object as follows: ‘An appeal to revisability and the long history of human inquiry cannot be part of an argument vindicating those theories, because the appeal makes use of those theories. Therefore, the argument is circular.’53 This objection fails, because I am not trying to justify those theories constituting the history of inquiry. I am trying to justify the current theory of the world, held by an epistemic community. That theory is objectively justiﬁed because the previous theories led, through a more or less objective process of inquiry, to the current theory. The current theory represents the culmination of centuries of inquiry, and its authority rests on the process that led up to it. The current theory is not justiﬁed by itself; it is justiﬁed by the history that preceded it.

Some interactions between subjective and objective justiﬁcation There seem to be, however, several classes of counterexamples to the theory of justiﬁcation developed here. By examining these alleged counterexamples, we can better understand the relation between subjective and objective justiﬁcation attributions. Consider the following: A community that springs into existence (say, Adam and Eve) presumably have justiﬁed beliefs, even though there is no history of inquiry in this community. (For future reference, we will call this community the Ruritanians.)

224

Pragmatism and Epistemology

Beliefs about goblins, etc. are not justiﬁed, even though they may be part of the fundamental assumptions in a community with a long history of inquiry. A victim of an evil demon is justiﬁed in her beliefs, even though she is part of no such history. Initially, these seem like daunting counterexamples; however, I think they can be defused. These examples each exploits one set of intuitions (internalist or externalist), and hence does not take account of the fact that there are two types of epistemic justiﬁcation. In evaluating an agent’s subjective justiﬁcation, we are evaluating the agent’s performance and beliefs. We ask whether the agent performed well, epistemically, given the evidence and epistemic resources available to her. But remember that in addition to subjective justiﬁcation, we must also recognize the relevance of objective justiﬁcation attributions. With objective justiﬁcation attributions, one takes a view that abstracts from the particular circumstances of the agent being evaluated and asks, ‘Objectively speaking, does the evidence in the agent’s possession establish, by our lights, the theory held by the agent?’ As a more externalist standard of justiﬁcation, the historical diachronicity requirement belongs to objective justiﬁcation. So there is a simple way to account for these three counterexamples. Consider the ﬁrst example, of a community that is created ex nihilo, with a ready set of beliefs about the world. The Ruritanians may be perfectly (subjectively) justiﬁed in believing their theories, but we should withhold attributing objective justiﬁcation to their theories because their background theories have not been diachronically vindicated. Thus, for example, suppose the Ruritanians hold a belief we agree with: say, the belief that relativistic physics is the correct account of the large-scale features of the universe. Now this belief, considered in the abstract, is objectively justiﬁed; it is (again, by our lights) supported by the best evidence. However, even if the Ruritanians have similar evidence for this belief, we will deny that their evidence is objective-justiﬁcationconferring, because it lacks the appropriate history. Our belief in general relativity is justiﬁed by the numerous experiments that have conﬁrmed it, and by the failure of disconﬁrming experiments—in short, by its survival, for nearly a century, in the face of a deluge of observation and empirical inquiry. The Ruritanians have no such history one can point to as objectively justifying their belief. We can see that the evidential basis they have for their beliefs is weak (although this fact may not be

Pragmatism and Epistemology

225

apparent to them), and that it therefore is insufﬁcient to objectively justify the beliefs they hold, even if by our lights those beliefs are objectively justiﬁed for us (that is, even when we judge that our evidence justiﬁes similar claims for us). This brief account gives the rough idea of the solution considered here, but there are complications, depending on how the Ruritanians came to hold their beliefs in the ﬁrst place. Let us examine the case in more detail. As I noted in the previous paragraph, it is perhaps true that the Ruritanians are subjectively justiﬁed in forming beliefs about the world, and perhaps in maintaining the beliefs they already have.54 To put things in terms of a deontic conception of subjective justiﬁcation, the Ruritanians are not epistemically to be blamed for believing as they do. But when we shift perspectives, and view the situation through the lens of objective justiﬁcation, matters are different. It depends, ﬁrst, on how the Ruritanians acquired their beliefs. Suppose Ruritania was created by a race of super-beings, who gave the new community many of the beliefs that they themselves held. In this case, the Ruritanians might well be objectively justiﬁed—but the objective justiﬁcation of their beliefs is parasitic on the objective justiﬁcation of the beliefs of the race of super-beings. Because the super-beings’ beliefs are objectively justiﬁed (because of their history of inquiry), the Ruritanians’ beliefs are also justiﬁed, as the super-beings endowed the new creatures with these same beliefs. But suppose, in the creation of Ruritania, there was no conscious or intelligent control over the beliefs of the new community. It is not clear, then, why we should attribute objective justiﬁcation to the Ruritanians. If the new beliefs of the Ruritanians were formed not via intelligent control, but by some non-intelligent process (owing to some feature of the ex nihilo formation process the Ruritanians just happened to wind up with this set of beliefs), then it is not clear at all why we should say that these beliefs are objectively justiﬁed. Again, the Ruritanians may be subjectively justiﬁed—no blame accrues to them for believing as they do—but the beliefs they have are not worthy of epistemic respect, as there is no particular reason to suppose the process by which these creatures were endowed with this particular set of beliefs would give them any true beliefs at all. Thus, if their beliefs are true, they are only accidentally true—they are not true because they have survived the long-term tribunal of experience as required by the diachronic constraint on justiﬁcation. Thus, while it is certainly understandable for the Ruritanians to hold the beliefs they do, it is extremely counterintuitive to attribute to them a strong version of justiﬁcation, especially given that one of the chief functions of the justiﬁcation condition is to rule

226

Pragmatism and Epistemology

out accidentally true beliefs from counting as knowledge. (I suppose it is open to one to argue that a community of sentient, sapient believers could instantaneously come into existence, ex nihilo, with a set of beliefs, and the process of belief-endowment, though not controlled by any intelligent force, somehow guarantees that the beliefs are true, and that the beliefs of this community are therefore objectively justiﬁed. But actually, on second thought, it is not really open to someone to argue this absurd position.) In any case, it is not clear what relevance science ﬁction examples like this one have to our actual conception of justiﬁcation. If our epistemic circumstances were radically different than they are, then perhaps a different conception of justiﬁcation would be appropriate. Perhaps, for example, the Ruritanians would at ﬁrst be entitled only to claim a very weak type of justiﬁcation (similar, say, to a notion of permission: they are permitted to believe as they do, but not justiﬁed in a strong sense). Only as the community develops, and acquires a history of inquiry, do they become entitled to claim for themselves (or we are entitled to attribute to them) a stronger notion of justiﬁcation. Thus, I am not forced to concede that the ﬁrst counterexample is a threat to the theory of justiﬁcation developed in this chapter. Let us consider the second alleged counterexample, the community in which belief in (say) goblins forms part of the fundamental assumptions of the community. Now remember that objective justiﬁcation attributions evaluate whether a belief is justiﬁed by our lights, that is, according to our best evidence. When we ask ourselves whether a particular community is justiﬁed, relative to their particular circumstances, in believing that P, then we are essentially asking a question about subjective justiﬁcation. Thus, we can only say (as we must judge by our standard) that belief in goblins is not objectively justiﬁed, because we know that goblins do not exist. So in evaluating this hypothetical community, we can only be asking a question about subjective justiﬁcation: is this community epistemically entitled to its belief in goblins? Has it performed well (relative to its epistemic circumstances) in maintaining this belief? Even if the belief is false, by our lights, the community in question may not go wrong in believing in goblins. Since objective justiﬁcation essentially involves judging by the best evidence we have, we can really only discuss objective justiﬁcation in this hypothetical community if we imagine, for a moment, that this community is ours. Suppose that our community is characterized as above: belief in goblins is a basic assumption of our community, which itself has a long history of inquiry. May we attribute (by our standards,

Pragmatism and Epistemology

227

of course) objective justiﬁcation to belief in goblins? It depends. First, we must ask about the epistemic evolution of our community. As I noted above, the history of epistemic inquiry in the community must be characterized in part, at least, by some degree of responsible revisability. Thus, if beliefs about goblins have long been de jure unrevisable in our community, then in fact these beliefs are not objectively justiﬁed (although again, the members of the community who have not reﬂected on these facts about de jure unrevisability and its consequences for epistemology may be subjectively justiﬁed in believing in goblins). And indeed, in real life there are probably issues about which it is impossible to have a justiﬁed opinion because investigation into these issues has been largely taboo. To give a (necessarily, because of the point I am illustrating!) contentious example, it is widely believed that some ethnic groups are genetically ‘smarter’ than others. However, research into this question has been limited by the taboo nature of the question itself. Suppose, for the sake of argument, that the taboo has been strong enough to prevent research into this topic from being pursued in any depth or detail. (I do not know whether it has been or not; I am merely presenting this as a hypothetical scenario.) In such a scenario, little could be known about the supposed ethnic correlates of intelligence. So even if people have long believed that some ethnic groups are smarter than others, it would be difﬁcult to see how this belief could be regarded as justiﬁed because investigation of these claims has by and large not been allowed. This is just another way of saying that we cannot have justiﬁed opinions on questions that we are not allowed to investigate, even if our opinions on the matter are of long standing. Longevity of opinion is not the only criterion for objective justiﬁcation; the beliefs in question must have repeatedly stood before the tribunal of experience. And so if our belief in goblins is treated as de jure unrevisable, then quite plausibly we cannot attribute objective justiﬁcation to this belief. But suppose the epistemic history of our community meets these diachronic constraints, and thus our community is somehow ‘entitled’ to its belief in goblins. We might then be able to say that at this particular stage in our community’s epistemic evolution, belief in goblins is objectively justiﬁed. Of course, future developments might cause our community to quit believing in goblins, but in the meantime, our belief is objectively justiﬁed—just as our epistemic community might previously have been objectively justiﬁed in believing in caloric or phlogiston. There is no reason, in principle, why a false belief cannot be rationally held; to deny this is to conﬂate truth and justiﬁcation. Just as in our community it was no doubt rational to believe in caloric or

228

Pragmatism and Epistemology

phlogiston, our community in this hypothetical example is entitled to its belief in goblins. So for this latest example, we must recognize that having a long history of inquiry is a necessary but not sufﬁcient condition for justiﬁcation: the way in which this history went, and the way in which the community went about revising its commitments to reach its current point, are relevant too. But at the risk of conﬂating justiﬁcation and truth, we must recognize that there are cases where this history can go correctly, yet still result in some false (yet justiﬁed) beliefs. Our discussion of the goblin example brings out a further aspect of the Ruritanian example. Objective justiﬁcation judges whether a theory is justiﬁed by our lights—that is, whether we ought to endorse the theory. And so we will judge the particular theories held by the Ruritanians to be objectively justiﬁed or not depending on the evidence we possess; the history (or lack thereof) of the Ruritanian community is irrelevant to that. But we can also judge, from the objective standpoint, the evidence possessed by the Ruritanians. This is why, in discussing the Ruritanians, I was careful always to discuss whether their evidence was objective-justiﬁcation-conferring, and did not talk about whether their theories were objectively justiﬁed. Thus, although the theory of relativity (believed by both us and the Ruritanians) is objectively justiﬁed—that is, it is established by our lights—our evidence for the theory is objective-justiﬁcation-conferring while theirs is not. (It is quite conceivable that we could encounter a community that had evidence that objectively justiﬁed a belief we did not possess, such as an advanced civilization that taught us many new theories and what justiﬁed them. The Ruritanians, on the other hand, have evidence that fails to objectively justify a belief which is for us objectively justiﬁed anyhow because of the evidence we possess.) By now, the response to the ﬁnal counterexample (the victim of the evil demon) should be clear. She is subjectively justiﬁed in holding the theoretical commitments she does—she has not gone wrong in thinking there is an external world, etc., given the evidence at her disposal— but we would not attribute objective justiﬁcation to these commitments because we know the evidence she has for her theories is manufactured, and hence does not in fact support her theory that she is living in a real world (as opposed to being deceived by a demon). I will have more to say about this particular counterexample in Chapter 8.

Justiﬁcation and knowledge The question naturally arises, ‘Which type of justiﬁcation is relevant to knowledge?’ As we will see, both subjective and objective justiﬁcation are necessary, but neither is sufﬁcient. This section constitutes

Pragmatism and Epistemology

229

the ﬁnal part of our answer to the charge of dogmatism. As knowledge requires objective (as well as subjective) justiﬁcation, and objective justiﬁcation requires that our basic beliefs be historically vindicated, the pragmatist epistemology rules out knowledge based on basic beliefs that are (because of a lack of historical vindication) not objectively justiﬁed. First, subjective justiﬁcation is necessary for knowledge. Consider an example of someone who has a true belief, but who has performed epistemically badly in forming this belief (i.e., who was not subjectively justiﬁed in forming this belief). Thus, suppose Jones came to believe, through assiduous reading of tea leaves, that Bill Clinton was born in 1946. The proposition believed by Jones is true and well-documented, but we hesitate to call Jones’s belief of this proposition knowledge, as Jones performed badly (epistemically) in coming to hold this belief. So it appears that the internal perspective is important to knowledge: performing epistemically badly precludes possession of knowledge. But the subjective justiﬁcation is not sufﬁcient for knowledge. Suppose that Jones belongs to a community where tea leaf reading is widely believed to be a valid epistemic method, and where evidence of tea leaf reading’s unreliability is not widely available. Given these circumstances, we might say that Jones is perfectly rational in believing the results of his readings. Thus, Jones’s true belief that Bill Clinton was born in 1946 was subjectively justiﬁed. But was it knowledge? Intuitively, it seems not to be knowledge. No, subjective justiﬁcation is not sufﬁcient. It seems it must be supplemented by objective justiﬁcation. Although we know that Jones is subjectively justiﬁed, he is not objectively justiﬁed, and this seems to preclude knowledge. Thus, it appears that to possess knowledge, one must have a true belief that is subjectively justiﬁed, and whose subjective justiﬁcation would sufﬁce for objective justiﬁcation. Thus, both types of justiﬁcation (subjective and objective) are necessary for knowledge. But neither type is sufﬁcient. We have already seen that subjective justiﬁcation is not sufﬁcient for knowledge. But neither is objective justiﬁcation. Suppose that a review of relevant documents has demonstrated conclusively that Bill Clinton was born in 1946. In other words, suppose belief that Bill Clinton was born in 1946 was objectively justiﬁed. If Jones did not possess this information, but instead held his belief on the basis of his tea leaf readings, then we still hesitate to call Jones’ true belief knowledge, even though the claim ‘Bill Clinton was born in 1946’ was objectively justiﬁed. Thus, objective justiﬁcation is not by itself sufﬁcient for knowledge. The person must possess subjective justiﬁcation which would qualify as an objective justiﬁcation for the view in question. In other words,

230

Pragmatism and Epistemology

if the agent believes P, and is subjectively justiﬁed in believing P, and the evidence on which the agent bases her belief would warrant an attribution of objective justiﬁcation to the proposition P, and P is true, then the agent knows that P. Now suppose reliable, authentic publicly available documents indicate that Bill Clinton was born in 1946. Thus, we would say that the claim that Bill Clinton was born in 1946 is objectively justiﬁed. Suppose, further, that Smith, in the course of his research, reads these documents and forms the belief, on the basis of this data, that Clinton was born in 1946. We would say that as Smith is basing his belief on such reliable sources, he has performed well epistemically; his belief is subjectively justiﬁed. And as evidence leading him to believe that Clinton was born in 1946 objectively justiﬁes the claim that Clinton was born in 1946, this claim is objectively justiﬁed in the abstract, and subjectively justiﬁed for Smith. And as, further, Smith’s belief is true, Smith knows that Bill Clinton was born in 1946. If this true belief of Smith’s were only subjectively justiﬁed, but Smith’s evidence did not pass muster to objectively justify the claim in question [say, Smith non-culpably believes in the reliability of tea leaf reading, and used this method (regarded in his community as reliable) to divine Clinton’s birth year], then we would deny that Smith knows that Bill Clinton was born in 1946. So as we can see, our intuitions track the pragmatist’s theoretical account of how justiﬁcation and knowledge attributions work. And so we can see that both types of justiﬁcation are necessary for knowledge. Neither is alone sufﬁcient. Thus, as knowledge requires objective (as well as subjective) justiﬁcation, and objective justiﬁcation requires that our basic beliefs be historically vindicated, then the pragmatist epistemology rules out knowledge based on basic beliefs which are (because of a lack of historical vindication) not objectively justiﬁed. This should put to rest charges of dogmatism against our pragmatist epistemology. We should note that if we attribute objective and subjective justiﬁcation to an agent, but later rescind our objective justiﬁcation attribution, we would also rescind the attribution of knowledge. Consider the famous eclipse observation made by Arthur Stanley Eddington in 1919 that was generally taken to conﬁrm Einstein’s theory of relativity. Suppose that at the time, an agent (call him Jones) had judged that Eddington was subjectively justiﬁed in believing his results supported Einstein’s theory—Eddington conducted his observations with great care, and therefore did not go epistemically wrong in believing Einstein’s theory on the basis of these observations. Suppose further, that Jones judged that Eddington’s evidence objectively justiﬁed belief in Einstein’s

Pragmatism and Epistemology

231

theory. On this basis, Jones would have been licensed in attributing knowledge to Eddington—knowledge that Einstein’s theory was true. Suppose, however, that after further examination of the case, Jones judged (as many have) that Eddington’s technique was too crude and the margin of error too great to objectively justify belief in Einstein’s theory. To the extent that Eddington’s belief in Einstein’s theory was based on the 1919 observations, Jones would now say that Eddington did not know that Einstein’s theory was true. Again, this result seems to agree with our intuitions on the case—if Eddington’s evidence was insufﬁcient to objectively justify belief in Einstein’s theory, we are inclined to say that Eddington could not have known Einstein’s theory to be true on the basis of this evidence, even if he was subjectively justiﬁed in believing Einstein’s theory (that is, even if we conclude that his judgment—that the evidence proved Einstein’s theory—represented an acceptable epistemic performance, given Eddington’s circumstances and the care with which he conducted his observations). This conclusion—that in order to know, a person must have subjective justiﬁcation, and this subjective justiﬁcation must sufﬁce for objective justiﬁcation (and of course the belief in question must be true)—may be of further use: it may help provide a solution to the Gettier problem. Consider Gettier’s ﬁrst case. Smith has good evidence that Jones will get a certain promotion. Smith also knows that Jones has 10 coins in his pocket. Smith performs the requisite inference, and becomes justiﬁed in believing that the person who will get the promotion has 10 coins in his pocket. This belief turns out to be true, but only because Smith (who, unbeknownst to anyone, also has 10 coins in his pocket) in fact receives the promotion. Now, Smith is justiﬁed—but in what sense? Clearly, he is subjectively justiﬁed—his epistemic performance cannot be faulted. He formed a valid inference from justiﬁed premises, and therefore obtained a justiﬁed conclusion. Yet this conclusion does not constitute knowledge. The pragmatist account allows us to explain why Smith does not have knowledge: Smith’s belief may be subjectively justiﬁed, but we, examining Smith’s case, recognize that his evidence bears the wrong relation to the conclusion, and hence cannot support a knowledge relation. Thus, from the externalist perspective, we can recognize the inadequacy of Smith’s justiﬁcation. Even if he has performed well, epistemically, and is therefore subjectively justiﬁed, we recognize the ﬂaw in Smith’s evidential chain, and recognize therefore that, from the external perspective, this evidential chain is not justiﬁcation-conferring.

232

Pragmatism and Epistemology

In general, this is the form that Gettier problems take. There is some misrelation between evidence and conclusion that prevents the agent from having knowledge, even if the agent is subjectively justiﬁed in believing the conclusion in question. But we are reluctant to ascribe knowledge in these cases precisely because from the external perspective, we see the misrelation and judge on this basis that the evidence does not confer objective justiﬁcation on the conclusion. As knowledge requires both subjective and objective justiﬁcation, the pragmatist theory gives us an explanation for why Gettier cases do not count as knowledge. We will, in Gettier cases, deny that the agent’s evidence is objectivejustiﬁcation-conferring because we can see the ﬂaw in the evidence; we can see how the evidence fails to establish the truth of the claim in question. So Gettier problems, in general, present us with cases of subjectively justiﬁed true belief that is not knowledge. This is not surprising to the pragmatist—it is not knowledge because it is only subjectively justiﬁed, and Gettier-style counterexamples are written such that by our lights the evidential chain is faulty and not knowledge-conferring (because not objective-justiﬁcation-conferring).

Conclusion Much work in epistemology has focused on what I call the internalist perspective: it focuses on what it is for a particular agent to be justiﬁed in holding a particular belief. Recent moves in the direction of virtue epistemology (of reliabilist and non-reliabilist varieties) are not helpful to the extent that they focus even more closely on the performance of particular epistemic agents (an obviously internalist project). Thus virtue theory, like traditional epistemic theorizing, can only tell half of the story, and is to that extent an inadequate epistemology. No doubt more can be said about these two perspectives. I hope, though, that I have succeeded in making one thing clear: given the importance of this internalist–externalist distinction, and given how crucial and ineliminable each perspective is, any viable epistemic theory must give an account of both types of justiﬁcation attribution, internalist and externalist. It is not clear whether all epistemic theories (reliabilist, coherentist, etc.) have the resources to make the internalist– externalist distinction I have outlined. For example, a simple teleological virtue theory (one that deﬁnes an agent’s epistemic virtue in terms of the reliability of one’s belief-forming mechanisms) “has difﬁculty doing justice to our intuition that a Newton and an Einstein may be roughly equal in virtue, while far apart in terms of truth.”55 In our terms, such

Pragmatism and Epistemology

233

theories may give an account of the externalist perspective, but they are unable to give plausible internalist justiﬁcation attributions. This should be seen as a serious problem that such theories need to overcome. We may (from the externalist perspective, of course) reject or demand the revision of any epistemological theory that lacks the resources to make this crucial distinction.

8 Pragmatism, Internalism, and Externalism

When a debate drags on interminably, and when both sides have the force of appealing intuitions and plausible arguments on their respective sides, one begins to think that perhaps both sides are latching onto some truth of the matter, and that neither position can be rejected as entirely false. The task, then, is to locate the truth in each position, and try to formulate a view that retains the insight in each position. The debate on internalism and externalism in epistemology has reached just such a point. Each side has compelling arguments, and each side can point out (quite plausibly) some important feature of epistemic evaluation that the other side fails to capture. Both internalism and externalism appeal to powerful intuitions, yet the motivating intuition of each seems to exclude the other. Some philosophers are recognizing that the time has come to try to mine both positions for their insights, and to attempt to formulate a view that retains the insights from both sides.1 Adopting a pragmatist account of epistemology and justiﬁcation, I have argued that there are in fact two distinct types of justiﬁcation attributions: internal justiﬁcation attributions and external justiﬁcation attributions. Neither is a substitute for the other, but each serves a distinct and important purpose. In this chapter I will argue that internalist and externalist justiﬁcation attributions should be seen as satisfying internalist and externalist intuitions, respectively, and that we have two sets of competing intuitions regarding justiﬁcation because there are two sorts of justiﬁcation attributions. Thus, I am arguing that adopting a pragmatist view of epistemology allows a reconciliation of internalism and externalism in epistemology. I will also argue that the traditional connection between justiﬁcation and truth has been misunderstood, and I will present a new way—that sheds light on the internalism–externalism debate—of understanding this connection. 234

Pragmatism, Internalism, and Externalism

235

The motivating insights of internalism and externalism There are two primary intuitions motivating epistemological internalism. The ﬁrst one is similar to the intuition that motivates much discussion of moral luck.2 We think that there should not be such a thing as moral luck; we think that something that is outside of the agent’s control should not affect our moral evaluation of the agent. For example, suppose I make up my mind to kill someone, and I attempt to shoot this person from a moderate distance. Now imagine two scenarios. In the ﬁrst scenario, imagine that a lucky (for my intended victim) gust of wind blows the bullet slightly off-target, so that my intended target is only wounded. In the second scenario, imagine that there is no lucky gust of wind, and the bullet is fatal. Opponents of moral luck ask the following question: why should anyone’s moral evaluation of me be different in the two cases? Even though I failed in the ﬁrst case, I failed for reasons beyond my control; my intention was the same in both cases. A similar case can be constructed in which an ill-timed gust of wind either does or does not thwart my attempt to save someone’s life. Again, we will think that my moral praiseworthiness is the same in both cases, notwithstanding the intervention of bad (but not moral!) luck in one case. The very plausible intuition in these cases is that if something is beyond my control, then it should not ﬁgure into anyone’s evaluation of my actions. A similar intuition motivates epistemological internalism. Thus, facts to which one has no access should not alter whether or not one’s beliefs are justiﬁed. The idea, roughly, is that if one has no access to certain facts, then one cannot be faulted (epistemically) for failing to take them into account. That is, our evaluation of a person’s epistemic performance should not take into account facts the person being evaluated could not have accessed. So the ﬁrst intuition motivating epistemological internalism is negative in nature: our evaluation of an agent cannot, it seems, turn on factors inaccessible to the agent. The second motivating intuition is more positive in nature: it says that if an agent is to be justiﬁed in believing a proposition, then the agent’s justifying reason for this proposition must in some sense be available to the agent. This motivating intuition is connected to one of the most powerful intuitions in epistemology: justiﬁcation is supposed to prevent accidentally true beliefs from qualifying as knowledge. Bonjour sums up internalism’s second motivating intuition as follows: Whatever plausibility attaches to externalism seems to derive from the fact that if the external relation in question genuinely obtains,

236

Pragmatism and Epistemology

then P will not go wrong in accepting the belief, and it is, in a sense, not an accident that this is so. But it remains unclear how these facts are supposed to justify P’s acceptance of B. It is clear, of course, that an external observer who knew both that P and accepted B and that there was a law-like connection between such acceptance and the truth of B would be in a position to construct an argument to justify his own acceptance of B . . . But P himself has no reason at all for thinking that B is likely to be true. From his perspective, it is an accident that the belief is true. And thus his acceptance of B is no more rational or responsible from an epistemic standpoint than would be the acceptance of a subjectively similar belief for which the external relation in question failed to obtain.3 So both of the motivating intuitions of internalism are at odds with an externalist construal of epistemology. The ﬁrst intuition denies that it is appropriate to allow facts to which an agent has no access to alter our epistemic evaluation of the agent in question. The second intuition says that if a belief whose justiﬁcation an agent cannot access (i.e., whose justiﬁcation is external to the agent) is at best accidentally true, as far as the agent can tell, and that it would be epistemically unsound for the agent to invest conﬁdence in this belief. The two intuitions are really two sides of the same coin. Both ask the question, how can an external fact, a fact to which one does not have access, legitimately alter our epistemic evaluation of an agent? Although these intuitions are often cast in terms of epistemic blameworthiness (e.g., one cannot be held responsible for facts one cannot access), they do not need to be construed in this way. As we saw in Chapter 6, many philosophers argue that owing to the failure of doxastic voluntarism, there is no viable notion of epistemic praise- or blameworthiness. There, I conceded that we cannot hold agents responsible for their beliefs, but argued that we can nevertheless evaluate agents’ epistemic performance. These two internalist intuitions can be understood in this latter sense. When we are evaluating the epistemic performance of an agent, we cannot criticize this performance on the grounds that the agent’s belief formation process did not take account of facts that were not accessible to the agent. This does not reﬂect any notion of epistemic responsibility. We simply cannot fault the performance of an agent who is not ‘plugged in’ to certain potentially relevant facts, any more than we can criticize the performance of an electrical appliance that is not plugged in to an outlet. As I argued in the previous chapter, our evaluation (positive or negative) of an epistemic agent must take

Pragmatism, Internalism, and Externalism

237

into account the context in which the performance took place. Which facts the agent has access to should be seen as part of the context of evaluation. Although these compelling intuitions speak in favor of internalism (and, at the same time, against externalism), externalism is itself motivated by two compelling intuitions, intuitions that speak against internalism. The ﬁrst motivating intuition behind externalism is that internalism leads to an inﬁnite (or circular) regress of justiﬁcation: that is, after all, if P’s belief that B is to count as justiﬁed, then (according to internalism) P must have another belief B1 (say, of the form ‘I had a perceptual experience of type T, which is typically caused by the world being as belief B represents it to be’); but then this belief B1 also needs justiﬁcation; and so forth. One way of stopping this regress is to stipulate that there are some beliefs that are justiﬁed not by other beliefs, but by the obtaining of some relation between the belief and the world (for example, by the fact that belief B was formed by a reliable belief-process).4 Bonjour describes the externalist as one who “might argue that although it is indeed necessary for a belief to be justiﬁed and a fortiori for it to be basic that a justifying argument . . . be in principle available in the situation, it is not always necessary that the person for whom the belief is basic (or anyone else) know or even justiﬁably believe that it is available.”5 Bonjour, of course, rejects this externalist position, and argues against it by appealing to internalism’s second motivating intuition, which I have discussed above.6 Now, the traditional internalist move against externalism’s ﬁrst motivating intuition is to argue that justiﬁcation is non-linear, but is instead holistic, in nature. Thus, a belief is justiﬁed by being appropriately connected (through explanatory links) to other beliefs in a coherent system. Thus, it is considerations of coherence among an agent’s beliefs that do the justifying work. This position, it is said, solves the regress problem while preserving internalism (after all, it is coherence among the agent’s beliefs that generates justiﬁcation). This solution to the regress problem leads to the second motivating intuition of externalism, the intuition that internalism does not preserve a sufﬁciently strong connection between justiﬁcation and truth. This second intuition is manifested in a number of objections against internalist epistemological theories. A traditional worry about coherentism, for example, is that there is no obvious connection between coherence and truth: maximizing the coherence of a theory is not the same as maximizing the truth in that theory. As Bonjour puts the point, “there seems to be no

238

Pragmatism and Epistemology

clear connection between the coherence of a system of beliefs and the cognitive goal of truth.”7 So, externalism is motivated by two intuitions:8 ﬁrst, internalism leads to an inﬁnite regress; and second, internalism severs the connection between justiﬁcation and truth. Taken together, our four intuitions— two for internalism and two for externalism—seem to tell us that internalism and externalism each must be and can’t be true. This is, of course, an unwelcome conclusion. I think the four intuitions we have just rehearsed are valid; each is latching on to a genuine and important feature of justiﬁcation. But doesn’t this commit us to thinking that internalism and externalism are both true, a manifest contradiction? The key lies in recognizing that internalism and externalism are not mutually exclusive accounts of justiﬁcation, but are instead two different kinds of epistemic ascription that one can make—each type of ascription with its own pragmatic signiﬁcance. As we saw in the last chapter, justiﬁcation attributions play not one but two distinct roles in our language. Internalist justiﬁcation attributions judge the epistemic performance of individual agents. Externalist justiﬁcation attributions evaluate whether a particular theory or proposition is justiﬁed. In this chapter, I hope to demonstrate that internalist justiﬁcation attributions satisfy powerful internalist intuitions, and externalist justiﬁcation attributions satisfy powerful externalist intuitions. Thus, we can show that pragmatism allows us to reconcile internalism and externalism in epistemology. Let us examine how our two perspectives satisfy the intuitions motivating internalism and externalism.

Internalist justiﬁcation attributions As I indicated above, internalist justiﬁcation attributions evaluate a person’s epistemic performance and beliefs: has the person in question behaved epistemically badly or well? Is (say) the person’s belief that p justiﬁed? As I argued in Chapter 7, our evaluation of a person’s performance always takes into account the circumstances under which the performance was executed. Recall our example from the previous chapter: if a particular runner runs a 20-minute mile, our evaluation of whether this is a good or a bad performance will be altered by, say, our knowledge that the person was running up a 20◦ incline against a 15 mile per hour headwind. In that case, we might say that the person turned in an outstanding performance, an opinion we would not

Pragmatism, Internalism, and Externalism

239

render if this 20-minute mile were completed on level ground with a slight tailwind. The same sorts of considerations apply to the person, described in the previous chapter, whose belief in astrology is rational (for her) because of the epistemic community in which she was raised. Although we would normally condemn belief in astrology by someone living in a scientiﬁcally advanced society, we recognize that if someone is raised in a community characterized by belief in astrology, and evidence of astrology’s falsehood is not available to this person, then belief in astrology is not irrational, but may be entirely rational. It should be apparent what we are getting at here: we are exploiting the internalist’s ﬁrst intuition, the intuition that facts to which one has no access should not alter whether or not one’s beliefs are justiﬁed (notice again, we are talking about belief tokens possessed by an individual, rather than propositions or beliefs in the abstract; these latter are addressed by the externalist perspective). Again, the idea is that if an agent has no access to certain facts, then her epistemic performance cannot be faulted on the grounds that she failed to take these into account. That is, our judgments as to whether one has performed well (epistemically) should not take into account facts the person could not have accessed. And so we can see that internalist justiﬁcation attributions are sensitive to and answer to the ﬁrst internalist intuition. Similar comments can be made about the internalist’s second intuition. Recall that according to this intuition, if one cannot access the justiﬁcation for a belief, then it is irrational for one to hold that belief. As Bonjour put the point, if “P himself has no reason at all for thinking that B is likely to be true,” then “[f]rom his perspective, it is an accident that the belief is true. And thus his acceptance of B is no more rational or responsible from an epistemic standpoint than would be the acceptance of a subjectively similar belief for which the external relation in question failed to obtain.”9 Our internalist justiﬁcation attributions will reﬂect this intuition: if a person holds a belief for which he has no evidence, then that person has not behaved well, epistemically; this person’s belief is not (subjectively) justiﬁed. Having explained how the internalist perspective can account for these intuitions, I should now express some reservations about how these two internalist intuitions are usually expressed. As I argued in the previous chapter, when one is trained into a set of beliefs (as a child, or as a novice scientist or philosopher, etc.), one is given a starting point that is (from the perspective of the child or novice scientist) entirely arbitrary.10 As each of us is given a more or less arbitrary starting

240

Pragmatism and Epistemology

place, it seems contrary to the spirit of internalism to evaluate a person’s epistemic performance based on his or her starting place. After all, we do not choose which theory of the world we are dogmatically trained into as children; this is beyond our control (and hence, we are not to be epistemically praised or blamed for holding this system). Similarly, an aspiring scientist being trained into the currently held scientiﬁc theory does not have the knowledge to evaluate whether this theory is correct or not; only once the scientist has been trained into the theory and given its epistemic resources can she go about trying to prove, disprove, or revise the theory. We must merely accept an arbitrary starting place, and then use that starting place as a platform for revising our theory, always using (subject to their revision) the tools that are initially given to us by this starting place. This strongly suggests that our evaluation of a person’s epistemic performance should depend not so much on the person’s starting place (and whether one has evidence for one’s ‘starting theory’), but rather on how the person revises the beliefs over time in light of new evidence and so forth. So it is not epistemically irresponsible or unsound to hold a belief for which one has no evidence (we cannot help but do that, since we are simply dogmatically trained into a particular practice); rather, it is epistemically unsound for an agent to revise her theory without being directed by evidence, or failing to revise the theory in the face of new evidence.11 This suggests a new way of casting internalism’s two intuitions. The ﬁrst intuition should be re-cast as follows: one’s epistemic performance cannot be faulted on the grounds that one failed to revise one’s theory on the basis of evidence to which one had no access; and the second intuition should be re-cast as follows: if one makes a revision and this revision is not made on the basis of evidence the agent acquires, then the revision is not rational, but irrational (or at best arational). This recasting of the second intuition says not that it is epistemically unsound to hold a belief without evidence; it says rather that it is epistemically unsound to revise one’s belief system in a way that is not itself prompted by some bit of evidence (say, a newly acquired belief, or a recognition that there is a pre-existing inconsistency within the agent’s belief system). This way of re-casting the two intuitions preserves their tie to internalist justiﬁcation attributions (by tying them to our evaluation of an agent’s epistemic performance) while better taking account of the importance of conservatism and the diachronic nature of rationality. (We will see below, however, that the original version of internalism’s ﬁrst intuition still has a place.)

Pragmatism, Internalism, and Externalism

241

Notice that this re-casting of internalism’s two intuitions brings internalism directly into line with the account we gave in Chapter 7 of internalist justiﬁcation attributions. In Chapter 7, we noted the importance of conservatism and revision to our evaluation of an agent’s epistemic performance (that is, to internalist justiﬁcation attributions). In the spirit of internalism, we cannot fault an agent for failing to revise her belief system on the basis of evidence to which she had no access; but we can fault an agent for revising her belief system without evidence. So internalist justiﬁcation attributions satisfy these two internalist intuitions. Even if one rejects the diachronic re-casting of internalism’s two intuitions, in their original form they represent plausible criteria for evaluating the epistemic performance of an agent (which is what internalist justiﬁcation attributions are supposed to do). So it seems that internalist justiﬁcation attributions answer nicely to internalism’s two intuitions. One might object that this re-casting of internalism’s intuitions makes them too weak. I concede that these requirements are weak. However, two comments are appropriate at this point. First, these intuitions do not propose to give sufﬁcient conditions for justiﬁcation. The intuitions behind internalism and externalism merely give constraints that an epistemic theory must meet; these intuitions might impose necessary conditions on justiﬁcation, but they are not the whole story. The second comment is this: our epistemic practice will simply not support very strong requirements. To give an example, the (seemingly plausible) requirement that an agent must in every case be able to give a reason for a justiﬁed belief is now almost universally rejected as being a toostrict requirement on justiﬁcation. Similarly, a room full of introductory epistemology students can almost always be convinced of the following principle: Given two hypotheses, h1 and h2 , where h1 is incompatible with h2 , if you cannot rule out h2 , then you are not justiﬁed in believing h1 . But this principle leads inexorably into skepticism.12 Despite its innocuous appearance and its initial plausibility, it imposes a requirement on justiﬁcation that is far too strict for our epistemic practice to bear. Thus, the weakness of these internalist intuitions should not count against them. Rather, it should add to their acceptability, as these intuitions do not place impossible demands on our epistemic practice. As I noted earlier in the chapter, many philosophers cast internalism’s intuitions in terms of epistemic blameworthiness. For example, Mylon

242

Pragmatism and Epistemology

Engel, Jr, who endorses a two-perspective approach different from the one presented in this chapter, writes that “we can deﬁne personal justiﬁcation [i.e., the sort of justiﬁcation that responds to internalist concerns] in terms of epistemic praiseworthiness and blameworthiness.”13 Again, I think this move is unwise. It is not clear whether there is a viable notion of epistemic praise- and blameworthiness. As we saw in Chapter 6, it seems that this notion depends on doxastic voluntarism, the extremely dubious idea that we choose our beliefs. If our beliefs are not under our voluntary control, then it is hard to see how we can be praised or blamed for holding the beliefs we do. Thus, there seems to be good reason to understand internalist justiﬁcation attributions as evaluating agent’s epistemic performances not in terms of epistemic blameworthiness. Fortunately, I noted that we need not construe internalism’s two intuitions in terms of epistemic blame- or praiseworthiness. Rather, we can construe them in terms of how well an agent has performed epistemically: we cannot criticize an agent for failing to revise her beliefs on the basis of evidence to which she had no access. We can, however, criticize the performance of an agent who revises her belief system without evidence. In neither case are we saying the agent is responsible for her beliefs or their revision. Rather, we are discussing what counts as a good epistemic performance, and what counts as a bad performance. To recall my earlier analogy, we cannot criticize the performance of an agent who is not ‘plugged into’ certain potentially relevant facts any more than we can legitimately criticize the performance of an electrical appliance that is not plugged into an outlet. There is no question of responsibility; the only question is whether the context (having access to certain facts; being plugged in to an outlet) supports criticism of the agent’s (or appliance’s) performance. So our re-casting of internalism’s two intuitions does not rely on the notion of epistemic responsibility; it only makes reference to contextual features (such as access to certain facts) that are relevant to our evaluation of an agent’s epistemic performance.

Externalist justiﬁcation attributions I have argued that internalist justiﬁcation attributions satisfy internalism’s two motivating intuitions. But this is only half the story. Let us now turn to a discussion of externalism’s two intuitions. Let us begin with externalism’s ﬁrst intuition, the intuition that only externalism can avoid an inﬁnite and vicious regress of justiﬁcation. This is a thorny intuition, and difﬁcult to satisfy. Perhaps this intuition

Pragmatism, Internalism, and Externalism

243

is also satisﬁed by the move to the perspective of externalist justiﬁcation attributions. Let us consider the case of perceptual belief. Suppose one forms a belief that an object is red. If justiﬁcation is demanded for this belief, one can justify the belief by claiming that in these circumstances, she is a reliable reporter of red objects. Of course, if she must justify this claim, then she is in danger of falling into an inﬁnite regress. Robert Brandom, in discussing Sellars’s epistemological internalism, has suggested that the regress is solved if one moves this attribution of reliability to what I am calling the externalist perspective. Here is Brandom: [The reliabilist’s] claim is that the real function of the traditional justiﬁcation condition on knowledge is to rule out accidentally true beliefs. If so, then the rationale for engaging in assessments of whether various beliefs qualify as knowledge is perfectly well-served by insisting only that candidate beliefs result from reliable belief-forming mechanisms—that is, mechanisms that are likely to lead to truths, whether or not the reporter knows that they are . . . [I]t is not obvious why Sellars should resist the reliabilist’s suggestion. Why isn’t it enough that the attributor of knowledge know that the reporter is reliable, that the attributor of knowledge endorse the inference from the reporter’s responsive disposition noninferentially to apply the concept red to the thing’s (probably) being red? Why should the reporter herself have to be able to offer the inferential justiﬁcation for her noninferential report?14 Of course, the internalist’s answer to Brandom’s rhetorical question is embodied by the internalist’s second intuition: one is irresponsible in forming a belief if one has no access to the belief’s justiﬁcation, for the belief is accidental from the person’s perspective. We should recall, though, my re-casting of internalism’s second intuition in light of our recognition of the importance of the conservative and diachronic aspects of justiﬁcation. I suggested that the second intuition should be re-cast as follows: if one revises ones belief system, then this revision must be made on the basis of evidence to count as rational. Now, when one acquires a belief that there is a red object before him, he is certainly revising his belief system. But this revision is on the basis of evidence: the agent in question is seeing the red object (or having a characteristic visual experience, etc.). And so the agent does have evidence for this new belief: namely, perceptual evidence.15 Thus, we see that the externalist’s ﬁrst intuition is best accounted for not by externalism, but (ironically) by our re-casting of the internalist’s

244

Pragmatism and Epistemology

second intuition. More accurately, the regress problem (a problem recognized by the externalist’s ﬁrst intuition) is best resolved by positing the conservatism of epistemic systems. Certain classes of belief are prima facie justiﬁed by belonging to a system of beliefs, and the onus of evidence is always on the challenger: that is to say, it is revision that requires evidence; maintaining an existing belief (of a certain type) does not normally require that one present evidence for this belief. So, for example, conﬁdence in the reliability of one’s visual perception is prima facie justiﬁed, and the burden of proof is on those who want to deny that this conﬁdence is justiﬁed: they must present reasons for thinking that visual perception is not reliable. Thus, this conﬁdence allows us to form new justiﬁed beliefs about the world, without itself standing in need of inferential justiﬁcation. However, externalism also contributes to the solution of the regress problem. As we saw in the last chapter, simply allowing basic beliefs to be conservatively justiﬁed opens us to charges of dogmatism. Our reply consisted of two parts: First, we argued that the beliefs are only conservative when viewed from the perspective of subjective justiﬁcation; from the external, objective standpoint, basic beliefs can be given a justiﬁcation based on their history. Second, we pointed out that for a belief to count as knowledge, it needs both subjective and objective justiﬁcation. So if there is a regress of justiﬁcation leading back to our basic beliefs, the external perspective explains why these stopping points can be objectively rational, and how they must be objectively rational to qualify as knowledge. (And the diachronic, revisionary and fallibilistic nature of our epistemology separates this account from traditional foundationalism, which is, of course, anathema to most pragmatists.)

Justiﬁcation and truth The externalist’s second intuition was that internalism does not preserve a sufﬁciently strong connection between justiﬁcation and truth. Externalist accounts have typically avoided this by deﬁning justiﬁcation in terms of reliable belief-forming processes. However, this advantage of externalism comes at a cost: justiﬁcation is severed from our actual practice of evaluating beliefs as justiﬁed or unjustiﬁed, rational or irrational. Let me explain this last point in the following way: According to the version of pragmatism I am employing here, the key to understanding justiﬁcation or knowledge is focusing on the question, ‘What role do justiﬁcation and knowledge attributions play in our language?’ The

Pragmatism, Internalism, and Externalism

245

obvious answer is that epistemic talk allows us to sort beliefs into those that are to be believed, and those that are not. This serves several goals: it hastens the achievement of adequate explanations of the world; it advances our understanding of the world; it advances us toward the truth; it helps us gain mastery over the world.16 But if we embrace traditional forms of externalism, then epistemic discourse cannot serve this purpose. Epistemic facts to which we have no access are useless toward achieving our epistemic aims of understanding, explanation, truth, and instrumental control over nature. But of course, for the traditional externalist, epistemic facts are quite often inaccessible—we often have no access to facts about whether our beliefs are justiﬁed or not. To claim that we must be able, in principle, to access such justiﬁers is to abandon traditional externalism in favor of some form of access internalism. Thus, if we embrace traditional forms of externalism, then epistemic discourse cannot serve the role of guiding inquiry. So pragmatism counsels against externalism; assessments of objective justiﬁcation should not be seen as external to the game of giving and asking for reasons.17 This view also distinguishes this twoperspective solution to the internalism–externalism debate from other such proposed solutions,18 which generally construe the externalist stance in terms of reliabilism. Let me put the point this way. On the pragmatist line we are pursuing here, justiﬁcation attributions play the role of directing inquiry. They are supposed to play an active role of guiding action and belief. In other words, they are concerned with what we ought to believe. But traditional externalism negates this practical aspect of epistemic evaluation. When we render an objective justiﬁcation attribution, we are making a judgment about what by our lights is best-justiﬁed. Thus, we cannot make objective justiﬁcation attributions if facts about whether our beliefs are justiﬁed are in principle inaccessible to us. And as I have already noted, to claim that justiﬁers must be in principle accessible to us is to abandon traditional externalism in favor of some sort of access internalism. So with externalist reliabilism, there may be a fact about whether our beliefs are justiﬁed, but if this fact is not accessible to us, and we cannot make judgments incorporating this fact, then this fact cannot serve the pragmatic goal of guiding inquiry. We cannot judge that this or that theory is the best-justiﬁed because whether a theory is justiﬁed (on externalism) may be a fact that is in principle inaccessible to us. And so we cannot judge what theory is by our lights the best-justiﬁed theory, because justiﬁcation is removed from our ken. To give an example, a fact of the form ‘Double-blind studies are the most

246

Pragmatism and Epistemology

reliable since they remove observer bias in clinical trials’ is a useful fact for guiding inquiry precisely because it is a fact that is known to us. Facts about reliability, however valuable they may be in the abstract, are only pragmatically useful to the extent that they are in principle accessible to us. Thus, the pragmatist can have no truck with traditional externalism. We will have more to say about this as the chapter progresses. What, then, of the connection between justiﬁcation and truth? As I noted above, the strength of traditional externalism was that it preserved a strong connection between justiﬁcation and truth. If we abandon this form of externalism, are we left with anything that deserves the name of epistemology or justiﬁcation? I believe that we are. To see this, we need to understand the proper connection between epistemology and truth. A practice is deﬁned by the goal it pursues: for example, medicine is deﬁned by the goal of preserving and restoring health, or something along those lines. The important thing to note is that this identiﬁcation of a practice based on its deﬁning goals remains valid even if the practice fails to achieve its ends. For example, until recent centuries, physicians could do little to improve the health of their patients, and many of their therapies (such as bloodletting) actually harmed, rather than improving, patients’ health. Does this mean that these physicians were not practicing medicine, and indeed that they were not even physicians in the ﬁrst place? Of course not; the practice of medicine existed in (say) the 16th century, but it was not particularly successful in achieving its ends—namely, the promotion of health. Similarly, our practice of evaluating beliefs, arguments, evidence, and so forth counts as an epistemic practice because it is deﬁned by the goals of understanding, explanation, truth, and instrumental control over nature. Crucially, even if it fails to achieve these goals—even if (say) the skeptical hypothesis is true—this practice is still an epistemic practice, just as the practice of medicine in the 16th century still was really a practice of medicine, even though it failed to achieve its goals.. If the skeptical hypothesis is true, then our epistemic practice fails to achieve its goals—particularly the goal of truth—but it is still an epistemic practice, because, as I argued above with the example of medicine, a practice’s primary orientation is what deﬁnes it. So, if a practice is deﬁned by pursuit of a particular goal, identiﬁcation of that practice is not altered by failure to achieve that deﬁning goal. The analogy here is between treating (medically) and justifying (epistemically). The business of medicine is providing treatments, and a therapy (such as bloodletting) can still count as a treatment even if it is

Pragmatism, Internalism, and Externalism

247

ultimately deleterious to patients’ health. Analogously, the business of epistemology is providing justiﬁcation, and an activity can still count as a justiﬁcation even if it systematically fails to achieve the goal of truth and understanding. Again, the goal of the practice determines the identity of the practice, even if the practice fails to achieve this goal. Thus, externalism must maintain a connection to truth and understanding, and the connection is this: our practice of justiﬁcation must be aimed at achieving true beliefs. This is a different connection than that advocated by traditional externalism, but as I have argued, a practice is deﬁned by the goal at which it aims, regardless of whether it in fact achieves this goal. At this point, the traditional externalist will no doubt argue, ‘But if the skeptical hypothesis is true, then all of our justiﬁcations fail. Thus, there is no guarantee of any connection between justiﬁcation and truth—on your account, how can justiﬁcation count as justiﬁcation? Haven’t you pushed justiﬁcation and truth too far apart?’ The short answer is, ‘No.’ I have not pushed justiﬁcation and truth far apart at all. In an important sense, the connection between justiﬁcation and truth is a necessary connection: as I argued above, our epistemic practice counts as an epistemic practice (as opposed to a moral or medical practice) because of the goals it pursues, among which is the goal of truth. Part of the point of justifying beliefs is to maximize our chance of achieving the truth. Of course, we cannot guarantee that such justiﬁcation will in the end result in truth. But this is just the familiar point that we cannot have certainty. If the traditional externalist wants a guarantee that justiﬁcation necessarily leads to truth and understanding, then the traditional externalist wants certainty—something modern epistemology has demonstrated (I believe) that the epistemologist cannot have. Thus, our epistemology (like much contemporary epistemology, and pragmatist epistemology in particular) is fallibilist in nature. This point—and some of the others I have been making above—is put nicely by Brandom: The expectation or prediction [that a belief will be true, under the circumstances] need not rise to the level of perfect certainty. Although there may well be a use of ‘knows’ that requires such certainty, it was one of the great advances in twentieth-century epistemology prior to reliabilism to realize that such a concept of knowledge not only includes an unrefusable invitation to skepticism, but also is of no use for discussing the achievements of science, and in any case is not obligatory. If our ordinary use of ‘know’ involves such commitments,

248

Pragmatism and Epistemology

that is the best possible reason to replace it by a less committive technical notion that is more useful for our central epistemological purposes. The fact that there are circumstances in which we would have been wrong should not preclude our counting as knowing in the cases where we are in fact right. Our fallibility should not be taken to rule out the possibility of knowledge.19 Notice the reference to ‘our central epistemological purposes,’ and the claim that the relative usefulness of a conception of knowledge bears on whether this conception of knowledge ought to be adopted. Brandom recognizes the point I press above, that a reliabilist conception of epistemology cannot serve the pragmatic goals of our epistemic practice. Though Brandom is talking here about knowledge, his point can be extended just as well to the notion of justiﬁcation: to require justiﬁcation to involve necessity does not serve our epistemological purposes. Thus, our externalist justiﬁcation attributions respond to externalism’s second intuition, the intuition that there must be a close connection between justiﬁcation and truth. I have argued above that there is a close—indeed, necessary—connection between justiﬁcation and truth. The mistake made by traditional externalists is that they have misunderstood the nature of this necessary connection—they have misinterpreted externalism’s second intuition. When the epistemologist demands a necessary connection between justiﬁcation and truth, the traditional externalist has understood this as a requirement that justiﬁcation must necessarily lead to truth (that is, for a process to be justiﬁcation-conferring, it must lead to truth in a certain percentage of cases). But I have argued above that the connection is different, though no less necessary: justiﬁcation must necessarily be aimed at the truth. We can only claim that a theory is objectively justiﬁed if the theory is true as far as we can tell—that is, if our best evidence points to the truth of the claim in question. To understand the connection in the way the traditional externalist has understood it is to return to the bad old days of epistemology, when justiﬁcation was thought to require certainty. Internalism, justiﬁcation, and truth Although I have tried to show that external justiﬁcation attributions preserve a sufﬁcient connection between justiﬁcation and truth (and hence satisfy externalism’s second intuition), we are not out of trouble yet. Remember that externalism’s second intuition was an intuition

Pragmatism, Internalism, and Externalism

249

that internalism did not preserve a sufﬁciently strong tie between justiﬁcation and truth. So one might ask, ‘Are your internalist justiﬁcation attributions really justiﬁcation attributions, or have you severed truth and justiﬁcation at the level of subjective justiﬁcation?’20 But remember how we resolved the problem with regard to external justiﬁcation attributions: the necessary connection between justiﬁcation and truth is that truth is among the goals of justiﬁcation. When we are making an internal justiﬁcation attribution, we are evaluating whether the agent in question adequately employed the facts and epistemic tools that she had at her disposal.21 But of course, a necessary condition on an agent’s being subjectively justiﬁed is that the agent’s practices be oriented toward discovering the truth in the ﬁrst place. If the agent is trying to do something other than discover the truth—say, the agent is seeking beliefs that would make her comfortable and happy, regardless of their truth—then we will obviously not say that this agent is subjectively justiﬁed. So the necessary connection between justiﬁcation and truth that we outlined above is maintained—justiﬁcation must necessarily be aimed at truth. For an agent to be epistemically justiﬁed in any sense—for us to make a positive justiﬁcation attribution, internal or external—it is necessary that the agent be seeking the truth in the ﬁrst place, and that she is using methods that were reasonable for her to use in pursuit of the truth. So both perspectives—internal and external—preserve the tie to truth, explanation, and understanding.

A genuine reconciliation? Clearly, I have only reconciled internalism and externalism if I have reconciled internalism and externalism, and not two other theories masquerading under the names ‘internalism’ and ‘externalism’. Thus, to complete my argument, I must show that I have not pulled a ‘bait and switch’ on the reader. That is, I must show that the pragmatist versions of internalism and externalism are genuine versions of internalism and externalism. Internalism and subjective justiﬁcation One objection that is often made against internalism is that facts about an agent’s epistemic obligations, or whether an agent has done well epistemically, do not supervene on the agent, and so whether a person is justiﬁed or not is not an internal affair at all.22 On the account offered in this book, subjective justiﬁcation attributions evaluate an agent’s performance, relative to a background. But whether or not an

250

Pragmatism and Epistemology

agent’s performance (relative to a background) was good or bad is surely not a fact that supervenes on the agent in question. So, is the account offered here really an internalist account? I think it is. The above-rehearsed arguments for internalism, such as the moral luck analogy, drive us toward a model of subjective justiﬁcation that is recognizably internalist. On this account, a person’s belief will be subjectively justiﬁed in those cases where the person has used the epistemic methods that are available to her and which it is reasonable for her to believe to be reliable. The person’s justiﬁcation cannot be defeated by facts to which the person had no access (here is where internalism’s ﬁrst intuition survives intact). Consider again our hypothetical agent from the beginning of Chapter 7 (let us call her Smith), who lives in a community where astrology is widely practiced and believed, and where evidence of astrology’s unreliability is not readily available. It is reasonable for Smith to think that astrology is a reliable epistemic method; so not only is Smith subjectively justiﬁed in believing that astrology is true, but Smith is also subjectively justiﬁed in believing the predictions made by astrologers (or at least those astrologers who are recognized within the community as being ‘experts’ in their ﬁeld). The important thing to note is that the factors that subjectively justify Smith’s belief—the prevalence of astrology in her community, the testimony of her epistemic peers, the absence of evidence against astrology’s reliability, and so forth—are all things that are within Smith’s perspective. The justiﬁers in question are all within Smith’s perspective. So internalist justiﬁcation attributions really are internalist. Furthermore, the connection to understanding, explanation, and truth (outlined above) makes these attributions genuinely epistemic attributions. It is true that on this account, epistemic norms do not supervene on the agent. Nevertheless, the norms govern justiﬁcation attributions that are licensed by factors that are within the agent’s perspective—facts that are accessible to her, norms with which she has been raised, and so forth. The internalist arguments we rehearsed earlier support considering only factors within the agent’s perspective when making internalist justiﬁcation attributions. They do not support the stronger requirement that the norms themselves must supervene on the agent. This is an unsupportable requirement, one that internalists ought to reject. What is the relation between this version of internalism and strong versions of traditional internalism such as mentalism, a version of internalism defended (among others) by Conee and Feldman?23 Conee and Feldman argue that ‘The justiﬁcatory status of a person’s doxastic attitudes strongly supervenes on the person’s mental states, events, and

Pragmatism, Internalism, and Externalism

251

conditions.’24 Thus, “If any two possible individuals are mentally alike, then they are justiﬁcationally alike, e.g., the same beliefs are justiﬁed for them to the same extent.”25 There are similarities and differences between the two views. First, it may be that on the pragmatist account, all ‘epistemizers’ (to use Alston’s phrase) of a belief must be mental states of the agent. For example, if Smith’s belief in astrology is subjectively justiﬁed because Smith lives in a community where astrology is widely practiced, this may at ﬁrst seem like a justiﬁcation-altering fact that is not a mental state of Smith. However, upon further examination of the case, we might conclude that Smith is subjectively justiﬁed in believing the truth of astronomy just in case the community’s belief in astronomy is reﬂected in Smith’s own mental structure—that is, if the community’s beliefs have impinged on Smith’s mental states so that Smith’s belief in astrology is caused by mental events within Smith, such as Smith’s experience of hearing her epistemic peers discourse frequently on the subject of astrology. Thus, the factors that justify Smith’s belief (or prove other of his beliefs unjustiﬁed) may all, in the ﬁnal analysis, be mental states of Smith. However, there is a distinction to be made between the factors that epistemize Smith’s beliefs and the norms according to which we evaluate Smith’s epistemic performance in terms of how these epistemizing factors bear on Smith’s performance. The former may supervene on Smith; but as noted above, the latter almost certainly do not. First, there are familiar Wittgensteinian reasons for thinking that norms do not supervene on individual agents.26 But also, making the agent the ultimate arbiter of the norms which govern his epistemic performance is ‘wildly permissive,’27 in Alston’s words. As Alston writes, “Practically all beliefs, no matter how shoddy or disreputable, will be justiﬁed on this criterion.”28 Finally, restricting our evaluation of an agent’s epistemic performance to norms that supervene on the agent renders subjective justiﬁcation pragmatically useless. Recall our example (from Chapter 7) of the recreational basketball player: I might say, ‘Jones played well today,’ whereas a professional basketball league talent scout might say, ‘Jones is no good.’ We are both making true judgments, because we are evaluating the player relative to a different set of norms, determined by a different background. I am evaluating Jones’s performance relative to the background of recreational basketball; the talent scouts are evaluating Jones’s performance relative to the background of professional basketball. Do the norms we are using to evaluate Jones supervene on Jones? It seems very implausible that (say) the norms of evaluation for professional-level basketball supervene on any particular

252

Pragmatism and Epistemology

recreational-league basketball player. Similarly in the epistemic case: if I judge that a conspiracy theorist has performed badly, epistemically, I am making a judgment (relative to the background provided by her epistemic community, etc.) that she had failed to live up to certain epistemic norms, that her epistemic performance has not been good. Now, plausibly she does not accept those norms, consciously or unconsciously, so in what sense can these norms be said to supervene on her mental states? They cannot. Again, allowing the norms of evaluation (rather than a belief’s epistemizers) to supervene on the agent being evaluated would only rarely allow our attribution of subjective justiﬁcation to diverge from the agent’s evaluation of his or her own performance. That would render subjective justiﬁcation attributions useless, in large part because of the ‘wild permissiveness’ against which Alston has cautioned us. It may be that the same norms apply to all agents—that is, even if we can judge the recreational basketball player by two different standards, these two standards are the same for all basketball players—in which case Conee and Feldman’s thesis of mentalism would seem to be true of subjective justiﬁcation in the following sense: if we are evaluating two agents relative to the same background and hence using the same norms and standards of evaluation for both agents, then if these two agents are alike mentally, then they are alike in respect of subjective justiﬁcation.

Externalism, reliability, and pragmatism But what about externalist justiﬁcation attributions? Externalist justiﬁcation attributions have little in common with contemporary externalist epistemologies. Thus, one might worry that my proposed reconciliation is a reconciliation in name only, that it reconciles internalism with faux-externalism. I believe, however, that the version of externalism presented in this chapter under the title of ‘externalist justiﬁcation attributions’ is the version of externalism we should endorse. Let me present some reasons for (a) not endorsing traditional externalism, and (b) for regarding my version of externalism as a genuine form of externalism. These two points need to be addressed together. To begin, we should regard the version of externalism presented in this chapter—externalist justiﬁcation attributions—as a genuine form of externalism. After all, at issue in externalist justiﬁcation attributions are propositions and beliefs in the abstract, and not individual agents’ token beliefs or epistemic performances. This might not satisfy the traditional externalist, but it is the most external form of externalism that seems

Pragmatism, Internalism, and Externalism

253

viable and compatible with pragmatism. Consider the matter this way: the pragmatist approach to epistemology (and other sorts of normative discourse) insists that for epistemology to be rationally justiﬁed, it must serve some pragmatic goal. That is, it must somehow be useful to us. But the externalist account of epistemology severs the connection between justiﬁcation and pragmatic utility. If the brain in the vat has no access to facts about whether or not it is justiﬁed, then what use is justiﬁcation? The mere fact that a belief was (or was not) formed by a reliable beliefforming process is pragmatically irrelevant if this fact is inaccessible to us. There is no pragmatically interesting but perspective-less notion of justiﬁcation. For the pragmatist, whether a belief is formed by a reliable beliefforming process must be an important fact. But the only thing that is useful, from the pragmatist objective justiﬁcation perspective, is whether we have good reason to believe that the belief is so-formed. So justiﬁcation cannot consist merely of the fact that a belief was formed via a reliable process, because this fact will often not be of pragmatic use. What is of pragmatic use is information of the form, ‘We have good reason to believe this belief was formed by a reliable belief-forming process,’ or ‘The best evidence currently available suggests that this theory is likely to be true.’ Therefore, justiﬁcation (pragmatically construed) must consist of this sort of information, not of facts about reliability. The latter are important, but they are only part of the story, in a pragmatist epistemology. Thus, traditional externalism is not compatible with pragmatism; it divorces justiﬁcation from the pragmatic goals of inquiry. So if we are to be pragmatists about epistemology (and I argued in Chapter 2 that pragmatism offers a plausible way to justify our epistemic practices), we cannot be traditional externalists. But haven’t we conceded, in discussing the brain in the vat (and other examples), the truth of reliabilism or some traditional form of externalism according to which we make epistemic evaluations based on facts unavailable to an agent? After all, we claimed in Chapter 7 (and will claim again later in this chapter) that the empirical theories of brains in vats or victims of evil demons are not objectively justiﬁed. Thus, for example, statements like ‘Jane is seeing a real tree’ (where ‘Jane’ refers to a particular brain in a vat) are not objectively justiﬁed. But how can we make sense of this claim on anything but a traditional externalist picture? The answer to this question consists of two parts. First, recall that objective justiﬁcation attributions are always made relative to what we

254

Pragmatism and Epistemology

know—they evaluate justiﬁcation by our lights, and as far as we can tell. In the hypothetical brain-in-a-vat scenario, we know (because it is stipulated) that the brain in the vat is not seeing real trees, and that its evidence for the presence of real trees in its visual ﬁeld is ﬁctitious. Thus, we know that the evidence is not justiﬁcation-conferring, and are justiﬁed in judging that the trees that Jane purports to be perceiving are illusory. The second, more crucial part, of our response to this externalist challenge is this: we know that the brain in the vat is using an unreliable belief-forming mechanism precisely because this fact is stipulated as part of the thought experiment. That is why we can apply the externalist reliabilist standard in this case—because it is merely stipulated that we have certain knowledge about the case in question, our epistemic situation is highly artiﬁcial, and we can apply externalist standards that we would be unable to apply to ourselves and our epistemic peers in the real world. For in the real world, our epistemic situation is much less favorable. By the very deﬁnition of externalism, we do not necessarily have access to what justiﬁes our beliefs, and hence (on the traditional externalist account) will often be incapable of being aware of whether or not we are using a reliable belief-forming mechanism. Thus, if we interpret objective justiﬁcation in terms of traditional externalism, then in the real world (with our epistemic limitations and lacking the stipulated knowledge of a hypothetical thought experiment) we simply lack access to facts regarding whether we are objectively justiﬁed or not. Thus, such facts can serve no role in guiding our inquiry, maintaining scientiﬁc consensus, or serving any of the other sorts of pragmatic purposes epistemology is supposed to serve. And the externalist must deny that these objective justiﬁers are accessible; to claim that they are accessible is to repudiate externalism in favor of some form of access internalism. The conclusion to be drawn is this: reliabilism seems like a useful epistemic standard because (misleadingly) it can be applied in highly artiﬁcial circumstances—namely, thought experiments, where our knowledge about whether a belief-forming mechanism is reliable is simply stipulated—but when it comes to real-world applicability, reliabilism (and any form of traditional externalism) is pragmatically useless. ‘Jane is not seeing real trees,’ uttered by us about our ﬁctitious victim of scientiﬁc chicanery, is objectively justiﬁed. But if we were brains in vats, ‘We are not seeing real trees’ would not—could not—be objectively justiﬁed, because the facts that would objectively justify this claim for us are in principle inaccessible to us. This shows the irrelevance of reliabilism to our real-world epistemic situation, as objective justiﬁcation

Pragmatism, Internalism, and Externalism

255

becomes pragmatically irrelevant if we insist on the externalist reliabilist standard. And so the pragmatist must jettison traditional externalism. This brings us to a point I have long postponed. I claimed in Chapter 7 that as ordinary epistemic agents are not capable of bringing their belief systems in conformity with the formal requirements of theories such as coherentism, Bayesianism, and foundationalism, such theories are useless at the level of subjective justiﬁcation and can only be used, if at all, at the level of objective justiﬁcation. But now we can see why these theories are useless at the level of objective justiﬁcation, too. The only way we can know whether a person’s belief system satisﬁes coherentist constraints, or Bayesian constraints, or any other formal constraints, is if this information is simply stipulated. But the real world does not come with such stipulations about matters of fact or logical relations among beliefs or propositions. And so we will not be able to evaluate our theories by these epistemic standards because we lack the cognitive abilities to determine whether any agents (including ourselves) satisfy the formal constraints of these theories. As Mark Lance writes, “Any view which requires the rational to avoid consistency . . . is in an important sense externalist. The reason is that the logical properties of reasonably complex systems . . . are no more transparent to the mind than are, for example, the causal pedigrees considered by reliabilists.”29 And so formal theories of epistemology that place structural constraints on beliefs or propositions are pragmatically useless. They cannot be utilized at the level of subjective or objective justiﬁcation. Despite our rejection of traditional externalism, the pragmatist externalist perspective nevertheless deserves to be called ‘externalism.’ First, it is external relative to the internalist subjective perspective. Second, as noted above, at issue in externalist justiﬁcation attributions are propositions and beliefs in the abstract, and not individual agents’ token beliefs or epistemic performances. External justiﬁcation attributions are external to any individual perspective, and are concerned with the theory or proposition that is supported by the best available evidence. Again, this might not satisfy the traditional externalist, but it is the most external form of externalism that seems viable and compatible with pragmatism. As I have been arguing, pragmatically interesting epistemic moves can only be made within a particular practice. Epistemology, pragmatically construed, is a normative endeavor concerned with how the practice ought to function; and facts that are external to the practice, and which often are in principle inaccessible to the practitioners, can have no bearing on how the practitioners ought to conduct the practice. They are, as we have seen, pragmatically irrelevant.

256

Pragmatism and Epistemology

Let me conclude this section, and provide a ﬁnal argument for regarding our version of externalism as the genuine article, by brieﬂy recapitulating an example of an externalist justiﬁcation attribution from the previous chapter: at time t, we may judge ‘We ought to believe P,’ based on the fact that as far as we can tell P is the best-justiﬁed theory. This is an externalist justiﬁcation attribution; it is about which theory we ought to believe. At time t + n, we may revise our judgment, and judge that Q (which is incompatible with P) is the best-justiﬁed theory. Does that mean we were wrong to claim that P was justiﬁed? It does not necessarily mean this. However, note that in judging our past performance, we are changing perspectives from objective to subjective justiﬁcation: we are saying, in essence, ‘Q is the best-justiﬁed theory, but at time t, we did not go wrong in believing P; it was perfectly rational, given the evidence we had.’ Examples like this one help us see the sense in which externalist justiﬁcation attributions are external. Such attributions are always liable to revision, and we can never say with certainty that the theory we currently hold to be best-justiﬁed will remain bestjustiﬁed in the long run. So externalist justiﬁcation attributions may wither and give rise to new ones, as our practice advances, and as new evidence comes to light. Thus evidence that we do not possess at time t—evidence that is external to our perspective—may ultimately, at t + n, defeat our earlier externalist justiﬁcation attribution, and cause us to regard the justiﬁcation we possessed at time t as merely subjective justiﬁcation. It is in this ﬁnal sense that objective justiﬁcation belongs to the category of pragmatic externalism.

Internalism, externalism, and counterexamples One strength of the current account is that it allows us to deal adequately with various counterexamples which have been raised against internalism and externalism. Let us consider a few representative examples. The clairvoyant The example of the clairvoyant is often used to attempt to demonstrate the inadequacy of the externalist account. In this case, Smith has perfect clairvoyant powers of which he is unaware. Let us suppose that these powers allow him to predict with perfect reliability the winner of the Kentucky Derby. Let us suppose further that Smith’s powers lead him to believe that True Believer will be this year’s Derby winner. Now Smith has no reason (aside from his clairvoyance, of which he is unaware)

Pragmatism, Internalism, and Externalism

257

for believing that True Believer will win this year’s Derby. Nevertheless, his belief was formed by a perfectly reliable belief-forming process. Is Smith’s belief justiﬁed? Internalists claim that the belief is obviously not justiﬁed, but that externalism predicts that it is justiﬁed. Ergo, internalists treat clairvoyance cases as reductios of externalism. What does our pragmatist theory have to say about such cases? The pragmatist would agree that the individual is not subjectively justiﬁed: the clairvoyant judgment has no more subjective validity than a hunch or a guess, as Smith has no reason to rely on his clairvoyant judgments. Doing so is not part of the established practice of his community, and he has no reason to think such judgments are reliable. He would not be performing well, epistemically, if he relied on these judgments. But from the third-person perspective, we (who know that Smith is clairvoyant) may rely on Smith’s judgments. And so we can attribute external justiﬁcation to the proposition that True Believer will win the Derby—and we are correct in making this attribution because we, from a position of external observers, know of Smith’s reliability. (Again, though, note our stipulated knowledge of Smith’s reliability.) And so the pragmatist can concede the ﬁrst premise of the internalist’s reductio: The belief that True Believer will win the Kentucky Derby is justiﬁed from the perspective of externalism. The pragmatist also concedes our internalist’s intuition that Smith is not justiﬁed in holding this belief. But as the pragmatist also argues that internalism and externalism are two different perspectives on justiﬁcation, and not incompatible positions, these two concessions do not contradict each other. Thus, pragmatism explains clairvoyance cases in a way that reconciles internalism and externalism. The fact that pragmatism can explain such cases while preserving both internalism and externalism is another mark in its favor.

The new evil-demon problem A relatively old problem for externalism in general (and reliabilism in particular) is the so-called new evil-demon problem.30 The venerable brain in the vat, or the victim of an evil deceptive demon (à la Descartes), uses belief-forming methods that are systematically unreliable, and hence its beliefs are unjustiﬁed. But (goes the internalist argument) this is extremely counterintuitive; there is surely some sense in which the brain in a vat is justiﬁed in holding the beliefs it does.

258

Pragmatism and Epistemology

Goldman has responded to this objection with what can be called ‘normal-world reliabilism.’ That is, a belief is justiﬁed if it is produced by a process that is reliable in normal worlds (that is, worlds that are like what we take the actual world to be). Thus, on this account, the brain in the vat has justiﬁed beliefs because although its belief-forming processes are in fact unreliable, such processes are reliable in normal worlds. Goldman calls this feature of reliabilism “normal-world chauvinism.”31 Although normal-world chauvinism satisﬁes our intuition that the brain in the vat is (at least in some sense) justiﬁed in its beliefs, it raises other problems. For example, belief-formation processes that are not reliable in normal worlds, but are reliable in possible worlds, are not justiﬁcation-conferring in those possible worlds. Thus, in a world of clairvoyants where clairvoyance is reliable and is recognized as such, it is not a justiﬁcation-conferring process because it is not reliable in normal worlds (i.e., worlds like our own).32 Implausible consequences like this have led Goldman to abandon this version of reliabilism. Goldman’s later attempts to solve the problem are more promising. Goldman distinguishes between strong and weak justiﬁcation, and argues that the brain in the vat is weakly justiﬁed, but not strongly justiﬁed. A belief is strongly justiﬁed if it is “formed (or sustained) by proper, suitable, or adequate methods, procedures, or processes,” whereas weak justiﬁcation means that the belief in question is “blameless” or “nonculpable.”33 Thus, Goldman recognizes (as we have already pointed out) the necessity for two types of justiﬁcation. And of course, this is exactly the solution we propose to this problem. The brain in the vat is subjectively justiﬁed, on our account: she performs epistemically well (given her circumstances) in forming the beliefs that she does. However, when we consider in the abstract the belief, ‘Is Jane [as we shall call our brain in the vat] seeing a real tree right now,’ this belief is not objectively justiﬁed, because we know that the evidence the brain has for the existence of said tree is fabricated. Of course, Goldman interprets strong justiﬁcation in terms of reliabilism, a strategy we have rejected for pragmatic reasons. But the basic move—distinguishing between two types of justiﬁcation—is obviously the solution to the problem that I endorse. Thus, though our pragmatist epistemology has a response to the new evil-demon problem, the problem makes particularly acute the need for more than just the external perspective. The new evil-demon problem shows the necessity of subjective justiﬁcation alongside objective justiﬁcation. It is true that the brain in the vat uses unreliable belief-forming mechanisms, but how can we meaningfully say that the brain in a vat ought to believe differently than it does? The traditional externalist is

Pragmatism, Internalism, and Externalism

259

caught in a dilemma. On the one hand, she can claim that the brain in a vat ought to believe precisely as it does, but that it is still not justiﬁed. Hence, justiﬁcation is removed from the practical realm of guiding inquiry; justiﬁcation is severed from the question of what we ought to believe. On the other hand, the traditional externalist might claim that the brain-in-a-vat’s beliefs are not justiﬁed, and hence it ought to believe differently than it does. This last clause is manifestly absurd—in what sense ought the brain in a vat believe differently than it does? There is the trivial sense, the sense in which the brain’s beliefs are false. But to say that the brain in a vat ought to believe differently than it does just in case its beliefs are false is to collapse justiﬁcation into truth. But we are here concerned with justiﬁcation; and regarding epistemology pragmatically (e.g., as action- and belief-guiding), it is absurd to say that the brain in a vat ought to believe differently than it does.34 Thus, the new evil-demon problem, though easily accommodated by our pragmatist epistemology, makes even more clear the need for the subjective, internalist perspective alongside the objective, externalist one. Of course, the brain in the vat makes its own attributions of objective justiﬁcation, as well as attributions of subjective justiﬁcation. Since we are claiming that the belief ‘Jane is seeing a real tree right now’ is not objectively justiﬁed, should we say that the brain-in-the-vat’s own attributions of objective justiﬁcation to its vat-mates are false? It seems not. Whenever we adopt a perspective that is not our own and evaluate an epistemic performance relative to that perspective and relative to the background determined by that perspective, we are attributing subjective justiﬁcation. And so we may very well say of the brain in the vat that when it attributes objective justiﬁcation to its own (and others’) beliefs about the external world, it does not go epistemically wrong in doing so (even though these beliefs are false). That is, given the brain’s epistemic circumstances, such attributions are wholly understandable. So we are attributing subjective justiﬁcation to the brain’s beliefs. And when the brain itself attributes subjective justiﬁcation to the beliefs of its vat-mates, those are also subjective justiﬁcation attributions—but relative to a different background. Remember, epistemic evaluation can be relativized to any background, and so there is a variety of subjective justiﬁcation attributions we can make, about our epistemic peers, about the brain in the vat, and about the brain-in-the-vat’s epistemic peers. What changes is the assumed background of evaluation, and the standards of evaluation vary with the changing background of evaluation. There are more counterexamples against externalism, but the above should give a model for how we would respond to them. Let us now

260

Pragmatism and Epistemology

turn to examine some alleged counterexamples against internalism in epistemology, and how our pragmatist epistemology can accommodate the intuitions raised by these examples. Brainwashing and the Naïf Ernest Sosa writes that there are certain examples that show that internalism fails to capture important features of justiﬁcation. To demonstrate this, he presents the following two examples: Compare the convictions of someone brainwashed. . . . or the beliefs of a naïf with a crude conception of what justiﬁcation requires in a certain ambit, who acquires his beliefs in ways that he is convinced are methodologically sound, simply because he was raised in a culture where such ways are instilled, so that now, through no fault of his own, our naïf does not properly know what to believe and what not to believe in that ambit.35 Sosa argues, correctly, that there is something amiss in the justiﬁcation of these two people. But, he claims, they are perfectly justiﬁed according to internalism, and so internalism offers an incomplete account of justiﬁcation. Sosa is correct in making both of these claims. But it was precisely a recognition that the internalist and externalist accounts were each incomplete that lead to the current two-perspective pragmatist epistemology. Both the naïf and the brainwashed person are internally justiﬁed—relative to what can be expected of a person in their respective positions, they are performing adequately. But they are not externally justiﬁed. Thus, they are just like Smith, whose belief in astrology was subjectively (but not objectively) justiﬁed, and whom we used as an example to illustrate the existence of two types of justiﬁcation attributions in the ﬁrst place. We recognize that the naïf’s epistemic practices are inadequate in some respects, and so do not adequately support her beliefs. We also realize that the brainwashed person’s beliefs are not supported in a way that would warrant an external justiﬁcation attribution. The case of forgotten justiﬁcation I know that koala bears eat eucalyptus leaves. However, I do not remember where I acquired this knowledge, or how I know it. Yet plausibly, I do know it. So (says the externalist), here is an example of justiﬁed

Pragmatism, Internalism, and Externalism

261

belief (knowledge, even) to whose justiﬁcation I have no access. Thus, internalism is false. However, as we noted in this chapter and in Chapter 7, it is in the spirit of internalism to relativize internalist justiﬁcation attributions to normal human epistemic capacities. In Chapter 7, recall, I argued explicitly that as humans are not cognitively capable of keeping track of the evidence for each of their beliefs, this cannot be required of them epistemically. Thus, to be subjectively justiﬁed in believing that koalas eat eucalyptus leaves, I do not need to remember the evidence on the basis of which I formed this belief. It is only necessary that in forming and maintaining this belief I am performing epistemically well relative to the human norm—that is, relative to what can be expected of normal human cognizers. As I argued in Chapter 7, as humans cannot be expected to remember the justiﬁcations for all of their beliefs, I do not go wrong epistemically in continuing to believe that koalas eat eucalyptus leaves. That is, I am subjectively justiﬁed in holding this belief. So internalism can accommodate forgotten evidence. However, we have cast internalism’s intuitions in terms of facts that are accessible to the agent: our evaluation of an agent should not be altered by facts that are inaccessible to the agent. Doesn’t forgotten evidence count as inaccessible? There are several things to be said in response to this objection. The ﬁrst potential reply is to argue that the agent formed the belief on the basis of evidence to which she had access. Thus, I know koala bears eat eucalyptus leaves because I read it or saw it on TV. Thus, there is a sense in which the evidence is within my perspective, as it is the evidence I used to form my belief in the ﬁrst place. Another more promising reply is to argue that I know that koala bears eat eucalyptus leaves because I know that I formed the belief on the basis of evidence in the ﬁrst place, even if I no longer recall what the evidence was. Thus, by continuing in the belief, I am (as it were) taking the belief on the authority of my past self, who did know the evidential basis of this belief. Thus, memory serves as a form of testimony between my past and present selves. Subjective justiﬁcation for a belief is strengthened if a certain belief is common knowledge and is frequently encountered among one’s epistemic peers. Thus, we can explain why a person can be subjectively justiﬁed in holding a belief for which she has forgotten the evidence. Again, there are more counterexamples against internalism, but (as with the counterexamples to externalism) my goal is not to exhaustively catalogue the counterexamples, but to demonstrate the pragmatist

262

Pragmatism and Epistemology

method for dealing with these alleged counterexamples. Once the model is in hand, other counterexamples can be dealt with in the same way. As a general rule, counterexamples against internalism exploit externalist intuitions and vice versa. The reply, then, is to show how a two-perspective epistemology can accommodate those intuitions at the same time accepting that the target of the counterexample (be it internalism or externalism) has a legitimate place in our scheme of epistemic evaluation, because the intuitions that support the target of the attack can also be accommodated within the two-perspective approach.

Conclusion My diagnosis of the internalist–externalist debate is that a great deal of confusion has resulted from the fact that philosophers have a set of competing but compelling intuitions. The pragmatist solution to this debate is to recognize that both sets of intuitions are legitimate and that each justiﬁes a distinct but legitimate epistemic perspective. The twoperspective approach defended in the past two chapters is very plausible and powerful because it allows us to accomplish things that internalism or externalism by themselves cannot: it explains the legitimacy of both internalist and externalist intuitions; addresses weaknesses in each position by solving counterexamples against both theories; shows us how the Gettier problem is created by the one-perspective approach (and solved by the two-perspective approach); and more. All of this is a powerful argument for our pragmatist epistemology.

Notes 1.

Introduction

1. “Special Issue: The Range of Pragmatism and the Limits of Philosophy,” Metaphilosophy 35:1–2 (January 2004); “Pragmatism,” Canadian Journal of Philosophy, Supplementary Volume XXIV (1999); “Pragmatism and Neopragmatism,” Essays in Philosophy 3:2 (2002). 2. Peirce, “How to Make Our Ideas Clear,” in C. Kloesal (ed.), Writings of Charles S. Peirce. A Chronological Edition, Vol. 3 (1872–78) (Bloomington, IN: Indiana University Press, 1986), p. 273. 3. Bakhurst (1999), p. 232. 4. James (1977), p. 430. 5. Rorty (1989), p. 52, emphasis removed. 6. Ibid., p. 53. 7. Ibid., p. 50. 8. Peirce (1903), quoted in Wiggins (1999), p. 15. 9. Wiggins (1999), p. 15. 10. Wiggins (2004), p. 110. 11. Wiggins, it should be noted, constructs a more charitable construal of Peirce’s account of truth and of the meaning of the above-quoted passage; we will return in detail to Wiggins’s and Peirce’s views of truth in Chapter 4. For another defense of the realist reading of Peirce, see Talisse (2001). For a discussion of much criticism of Peirce’s theory of truth, see, for example, de Waal (1999). 12. James (1977), p. 438, emphasis removed. 13. Peirce (1908), p. 450. For more on James’s theory of truth and Peirce’s ‘seeds of death’ criticism, see Cormier (2001). 14. Hume (1888/1968), p. 469. 15. Mackie (1977), p. 38. 16. Harman (1977). 17. Koons (1998). 18. Brandom (1984) and (1994). 19. Lance and O’Leary-Hawthorne (1997).

2.

Pragmatism, Causal Explanation, and Normative Facts

1. An earlier version of this chapter appeared in Paciﬁc Philosophical Quarterly as ‘Do Normative Facts Need to Explain?’ 2. And indeed this seems to be the view of many particularists. 3. This line of argument has been advanced in Sayre-McCord (1988). I propose to go into greater detail than Sayre-McCord, and discuss what, if not causally explanatory facts, actually justiﬁes our continued participation in moral and epistemic discourse. The main bulk of the chapter is devoted to 263

264

4.

5. 6.

7. 8. 9. 10.

11. 12. 13.

14.

15.

Notes this latter task: a discussion of the pragmatic reasons that justify both moral and epistemic discourse. In the next chapter, we will discuss objections to and difﬁculties with this sort of pragmatic justiﬁcation. Harman and Ayer are emotivists about moral discourse, a position Harman describes as “moderate nihilism.” Mackie is not an eliminativist about morality, but his version of moral discourse is so weakened that it would be proper to describe it as moderate nihilism, too. Harman (1977), p. 4. One of the most convincing works demonstrating this point is, of course, Kuhn (1970). Also good are chapter 3 of Longino (1990) and chapter 2 of Churchland (1979). See, for example, Brink (1989), pp. 182–197. See, for example, Misak (1996), Bakhurst (1999), and Putnam (2004). The terms ‘doctrinal eliminativism,’ ‘explanatory eliminativism,’ and ‘ontological eliminativism’ come from Lance and O’Leary-Hawthorne (1997). Although Jackson and Pettit (1990) argue that higher-level explanations may be necessary even if not causally efﬁcacious. This move is criticized in chapter 8 of Miller (2003). Sellars (1953), p. 222. Lance and O’Leary-Hawthorne (1997), p. 386. In my article ‘An Argument Against Reduction in Morality and Epistemology,’ Koons (2006), I argue at length that normative facts are additional and irreducible elements over and above whatever natural facts might license or co-vary with these facts. Some might claim that all purported examples of normative explanation are merely conceptual analysis. If this is true, it does not hurt my position. I am here examining the types of explanations in which moral facts may ﬁgure, with an eye to conceding that moral facts do not ﬁgure in any interesting way in causal explanations. Since normative explanations are not causal anyhow, a denial that there are genuine normative explanations (as opposed to mere conceptual analysis) is compatible with the concession I am preparing to make to Harman. In any case, Mark Nelson has pointed out to me that this objection assumes that concepts are not normative, and that conceptual analysis is therefore not a form of normative analysis. This assumption is not obviously true. Wright (1992) thinks that an important test for realism is whether a particular discourse enjoys wide cosmological role—that is, whether the states of affairs to which the discourse refers legitimately participate in a wide range of different kinds of explanations. Morality, as we have seen, delivers certain sorts of explanations, but (Wright argues) it does not deliver certain classes of causal explanations. Thus, though moral discourse remains ‘minimally truth-apt’ (i.e., its sentences are candidates for truth and falsity, and have genuine assertoric content), it is not as robustly ‘real’ as, say, scientiﬁc discourse, whose referent states of affairs may legitimately participate in a whole range of explanations, from rational to causal. I think Wright is correct: morality does not live up to the standards of science. I only wish to point out that epistemology suffers from the same narrow cosmological role as morality; and I further wish to show why we might want to continue

Notes

16. 17. 18.

19.

20.

21. 22. 23. 24. 25. 26. 27.

28.

29.

30.

265

participating in forms of discourse that do not display a wide cosmological role. Koons (2006). I borrow this example from Sturgeon (1988), who in turn borrowed it from Bernard De Voto. This assumption is controversial (as in the case of numerals, for example), but again, I am willing to concede claims hostile to my position to show that my position can still prevail in the face of unfriendly assumptions. The term ‘Average American mother’ is not completely analogous to moral terms, because it does not function as a singular term. That is, it can only be substituted into narrowly deﬁned contexts; you cannot, for example, substitute it into the sentence, ‘Where did you see the X?’ However, even if it is not just like a moral term, it does show that ontological and causal explanatory eliminativism do not entail doctrinal eliminativism. The belief that causal explanatory eliminativism entails doctrinal eliminativism can have particularly nasty effects if one believes that only natural kinds are involved in genuine causal explanations. For, if this were the case, then all sentences about artifacts—chairs, shirts, computers, and so forth— would be false. This would call for massive revision of our language, and massive inconvenience—imagine if we couldn’t refer to any artifact whatsoever. We will see later on the relevance of such pragmatic reasons in the debate over doctrinal eliminativism. See chapters 3 and 4 of Kuhn (1970). Kuhn (1959), p. 227. Ibid., p. 231. Ibid., p. 232. Ibid., p. 235. As we will see at the end of Chapter 6, such exclusion is a chief method of sanctioning those who violate epistemic norms. Brandom (1994), for example, argues that normative utterances (logical, epistemic, semantic, moral, and so forth) make explicit norms that are already implicit in the practice. Lance and O’Leary-Hawthorne (1997), alternatively, argue that such utterances serve either to endorse the norms implicit in the practice, or to try to bind the members of the practice to a new set of norms. I am not suggesting that we might be able to infer someone’s epistemological commitments entirely from her non-linguistic behavior. I am merely saying that these commitments are carried out in practice, and that this practice is a manifestation (even if a non-interpretable one) of these commitments. Further, I should add (to avoid rousing the ire of the radical interpreters) that I am not here claiming that we could have epistemic commitments that could not be discovered by a radical interpreter. I am merely denying that non-linguistic behavior alone will always be sufﬁcient evidence for the translator to determine what these commitments are. As an aside, I should note that pragmatic reasons play an important role within science, as well. Every list of criteria for theory choice contains such pragmatic criteria as simplicity and economy. Mark Lance put the point this way.

266

Notes

31. We will see shortly that although the practice itself is pragmatically justiﬁed, individual epistemic attitudes are justiﬁed by epistemic, not pragmatic, reasons. 32. I owe this example to Mark Lance. 33. Some, like Richard Taylor (1988), have argued to the contrary, but it is beyond the scope of this chapter to argue that we are better off playing the moral game. Thus, I will have to assume that we are. 34. I owe this phrase to Duncan MacIntosh. 35. Keep in mind that as a ‘decynicized Hobbesian,’ I am using ‘prudential’ in a non-egoistic sense. 36. This worry was raised by Duncan MacIntosh. 37. Although the case of science suggests an important point: the fact that we pursue a type of discourse for pragmatic reasons does not entail that the discourse must be internally pragmatically structured. After all, the question of whether a particular scientiﬁc inference or belief is rational is a question to be answered from within science, according to scientiﬁc reasons, and not on pragmatic grounds. I discuss this topic in greater detail in the next chapter. 38. I am grateful to Mark Lance for helpful conversations on this topic. Indeed, the example that follows was suggested by him. 39. And, of course, instrumentalists will deny that truth is a valid goal in the ﬁrst place. 40. This worry is similar to John Mackie’s charge that moral properties are epistemically queer (Mackie 1977, pp. 38–42). 41. Harman (1977), p. 4. 42. This worry has been raised most forcefully by David Alm. 43. From comments by David Alm on an earlier draft of Chapter 2 presented at the October 22, 1999, meeting of the Auburn Philosophical Society. 44. I will argue, in the next chapter, for the rational superiority of rule-based over act-based pragmatic systems. 45. Rawls (1955), p. 16. 46. This example was also suggested by David Alm. 47. Peirce (1877), p. 119. 48. Ibid., p. 117. 49. Ibid., p. 118. 50. Ibid. 51. Ibid., pp. 119–120. 52. Ibid., p. 114. 53. Wiggins (2004), p. 93. 54. Ibid., p. 98. 55. Peirce (1877), p. 115. 56. This is the view of Lance and O’Leary-Hawthorne (1997, esp. chapter 3). 57. Gibbard (1990), p. 7. 58. I have noted that the pragmatist might borrow the notion of normative utterances as being primarily legislative in nature from Lance and O’LearyHawthorne (1997). However, this claim about properties is a claim that Mark Lance has explicitly repudiated (conversation with Mark Lance). Thus, although I appeal to their work, it should not be inferred that Lance and O’Leary-Hawthorne would endorse all my uses of their work.

Notes

267

59. This talk of truth requirements should not mislead the reader into thinking that I believe we can give robust truth conditions for moral discourse. It seems likely that sentences of the form ‘P is true if and only if P’ tell us about the role of the truth predicate in our language, and not much about what makes P true. My talk of truth requirements in this paragraph and the next is intended only to bring out the fact that very different kinds of discourse can be true; and it doesn’t follow from the fact that these discourses are different (say, one is primarily legislative and one is primarily descriptive) and that different types of truth are at work in each. 60. One account of moral truth in terms of excellence of reasons is offered in Wiggins (1991). Social practice accounts of morality must always be on the lookout for relativism. A social practice account of normative discourse that avoids relativism (at least the pernicious forms of relativism) is offered in chapter 3 of Lance and O’Leary-Hawthorne (1997). The latter account is unique in that it argues (a) that we cannot give robust truth conditions for moral discourse, and (b) moral discourse is nevertheless truth-apt. We will return to these issues in chapter 4. 61. This analogy was suggested in a different context by Roderick Long. 62. Black (1990), p. 67. 63. Ibid., pp. 68 and 67.

3.

Pragmatism and Rationality

1. Of course, if this constitutive rationality has a moral element, then one might worry this renders the justiﬁcation of morality circular. I address this worry in Chapter 5. 2. Smart (1956), pp. 348–349. 3. Although my case is hardly weakened even if this statement is wrong. The less the individual actions are bearers of fundamental rationality, the stronger is my argument against the rule-worship objection. 4. The following act-utilitarian response was suggested by Mark Nelson. 5. Mark Nelson suggested that the act utilitarian might respond in this way. 6. Hunter (1994), p. 32. 7. Rawls (1971), p. 132. 8. Dean (1997). 9. Ibid., p. 454. 10. Dean (1997), p. 455. 11. Gauthier (1986). 12. Ibid., p. 457. This distinction is similar to Bales’s (1971) distinction between an ethical theory’s account of right-making characteristics of actions and the theory’s decision-making procedure. I will not discuss Bales’s article, as I hold it to be fatally ﬂawed. Bales argues that act utilitarianism could still be a correct account of the right-making characteristics of an action, even if it is not correct as an ethical decision-making procedure. But if act utilitarianism is correct as an account of right-making characteristics, then it seems as though a correct ethical decision-making procedure must (in an individual case) lead us to the utility-maximizing action. But this focus on individual actions is, as I have argued, self-defeating. If utility is what we ought to maximize, we

268

13. 14. 15. 16. 17. 18. 19. 20. 21.

22.

23. 24. 25. 26. 27. 28.

29. 30. 31. 32.

Notes ought not be prejudiced at the outset in favor of act or rule utilitarianism as an account of the right-making characteristics of actions. It may turn out that the utility-maximizing action is not the morally correct one (the focus here being on right-making characteristics)—if the action is forbidden by a utility-maximizing strategy. And I have been arguing for a conclusion similar to this one. Dean (1997), p. 457. Ibid. Ibid., pp. 457–458. See, for example, Gilbert (1989) and (2001). Sugden (2000). Ibid., p. 179. Ibid. Ibid., pp. 180–181. As Sugden points out, it will not help to try to understand this problem from the point of individual rationality by looking for a Nash equilibrium. Not only are there three different Nash equilibria in this problem, but more fundamentally, “the fact that a particular combination of strategies is a Nash equilibrium (even a unique Nash equilibrium) gives neither player a reason to choose to play his part in that equilibrium. All we can say is that, if it were the case that each player expected every other player to play his part in a certain (strict) Nash equilibrium, then each player would have reason to do the same. And that provides us with no escape route from the inﬁnite regress of reasons” (p. 182). Gauthier (1967), pp. 461–462, emphasis removed from original. Gauthier attributes this position to Kurt Baier; it is not clear whether Gauthier himself fully endorses this claim. Ibid., p. 462, emphasis removed from original. Ibid., p. 463. Ibid. Sellars (1992), chapter 7. Velleman (1997). I do not wish to argue that rationality is fundamentally a matter of satisfying one’s preferences. In the next chapter, we will endorse a theory of human ﬂourishing which rejects the idea that preference satisfaction is constitutive of human ﬂourishing, focusing instead on a list of fundamental human interests whose satisfaction constitutes human ﬂourishing. However, to the extent that a preference is derivative upon promotion of an interest, satisfying this preference promotes human ﬂourishing; so for the remainder of this chapter I will speak mostly of group preferences. This is a particularly apt way of speaking, because as a given interest can be satisﬁed in so many ways, we might have a great variety of preferences whose rationality is derivative of the rationality of promoting the interest in question; and so it is natural to speak of acting on this or that preference (as a means of satisfying a particular interest). See Sellars (1956) and chapter 7 of Sellars (1992). Gilbert (2001), p. 113. Ibid., p. 115. Ibid.

Notes 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

43.

44. 45. 46. 47. 48.

49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63.

269

Ibid., p. 116. Ibid., pp. 110–111. Baier (1958), cited in Gauthier (1967). Sugden (2000). Gilbert (2001). Gilbert (1989), p. 37. Gilbert uses the mushroom pickers’ example to illustrate a different point than the one I am illustrating here. Gilbert (1989), p. 376. See Tuomela (1996). The term is, of course, borrowed from Nozick (1974), although it is put to a different use here. The distinction between perfect and imperfect duties is not easy to state precisely. The duty not to lie, though a perfect duty, is negative in the sense that it tells us not to lie. One might say it counsels us always to tell the truth (thereby requiring positive action), but that is not true either, as there is always an alternative to lying or telling the truth, namely, saying nothing. Thoreau (1991), p. 34. Of course, I recognize that there is something ironic about quoting Thoreau in support of the point I am making, as Thoreau thought that our negative duties of non-maleﬁcence were so stringent that they forbade virtually all participation in government and society. However, I take it that this expresses an extreme of moral Puritanism, and is not a view we wish to endorse. Gilbert (2001), p. 117. See, for example, Axelrod (1984). Levy (1997), p. 95. Foot (1985), p. 198. Levy (1997), p. 98. Levy argues that it is in fact act utilitarianism that is inconsistent with the teleological motivation. As we saw earlier, it is act utilitarianism that fails to maximize long-term utility, by allowing free riders, by making cooperation difﬁcult, and so forth. Railton (1984), pp. 156–157. Rawls (1955). Brandom (2000), pp. 88–89. Ibid., p. 88. Carroll (1895), p. 279. Dancy (1993). Hooker (2000), pp. 17–18. See Jackson, Pettit, and Smith (2000). This point was raised by Mark Nelson in private correspondence. Lance and Little (2004), p. 441. Lance and Little (2005), pp. 306–307. Ibid., p. 310. Lance and Little (2004), p. 446. Lance and Little (2005), pp. 310–311. In their article ‘Defeasibility and the Normative Grasp of Context,’ Lance and Little (2004) offer a semantics for defeasible generalizations, centered around the notion of privileged conditions. They deﬁne P as a modal operator, which reads “in privileged conditions,” that functions on generalizations. “Formally, the P operator will simply be a modal operator the semantics of

270

Notes

which is given by designating a range of privileged worlds in a usual possible worlds framework” (p. 443). 64. Lance and Little (2005), p. 313. 65. Lance and Little (2004), p. 307: “Non-standard cases . . . while always perforce deviant, are not always defective.”

4.

Pragmatism, Normativity, and Relativism

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Wiggins (1996), p. 278. Wiggins (1991). John McDowell (1986). Wiggins (1991), p. 67. Ibid. Recall our discussion of this issue from Chapter 2. Wiggins (1991), p. 66. Peirce (1877), pp. 119–120. Ibid., p. 120. Wiggins (2004), p. 105. Ibid., p. 95. C.S. Peirce, ‘Of Reality: Realism, Idealism, and the Ultimate Opinion,’ in Writings of Charles S. Peirce: A Chronological Edition, Volume 3: 1872–1878 (Indiana University Press, 1986), pp. 40–47, at 42. Wiggins (2004), p. 106. Ibid., p. 107. Wiggins elsewhere remarks that he does not intend his account of truth to be a deﬁnition of truth, for he restricts his account to knowable truths. Wiggins writes, “Truth is not itself an epistemological notion, however epistemological we may allow our formulations to be of its marks and attributes. It was for this reason that, in the formulation of the mark of truth given [earlier], I entered an explicit restriction to truths that were knowable” (1996, p. 274). The earlier formulation to which he was referring is this: “If it is true that p then, insofar as it can be known that p, someone can believe that p precisely because p” (1996, p. 273). Although the account offered in this chapter may seem epistemological, in that it offers an account of the truth of normative utterances in terms of excellence of reasons, we will see that this is not so, as I will argue that even reasons that are not known or thought by anyone can be decisive in rendering a particular normative fact true. Actually, it is not entirely clear whether this is true. For example, the question ‘What is the sum of the internal angles of a triangle?’ does not have a unique answer; it depends on which set of geometrical axioms (Euclidean, Riemannian, or Lobachevskian) you use. Of course, one might say that relative to a certain set of axioms, reason does determine a unique answer to this question. But this invites the question of whether there are conclusive reasons to accept one set of axioms as opposed to another. I will not explore this issue here. Finnis (1989). Nussbaum (2000), pp. 78–80.

13. 14. 15.

16.

17. 18.

Notes

271

19. See Chapter 5. 20. Lance and O’Leary-Hawthorne (1997), p. 176. 21. “To say that man is a rational animal, is to say that man is a creature not of habits, but of rules. When God created Adam, he whispered in his ear, ‘In all contexts of action you will recognize rules, if only the rule to grope for rules to recognize. When you cease to recognize rules, you will walk on four feet.’” (Sellars (1950), p. 138). 22. Lance and O’Leary-Hawthorne (1997), pp. 175–176. 23. See Kuhn (1959). 24. One might deny this, but my case is hardly weakened if this denial is correct. After all, relativist worries about pragmatism must be diminished by the conclusion that our interests determine a unique best cooperative strategy. 25. As Mill writes in On Liberty (1978, p. 54), “As it is useful that while mankind are imperfect there should be different opinions, so is it that there should be different experiments of living; that free scope should be given to varieties of character, short of injury to others; and that the worth of different modes of life should be proved practically, when any one thinks ﬁt to try them.” 26. The following historical examples are taken from Kaptchuk (1998). 27. Although as we noted in the previous chapter, even if a norm would be pragmatically superior to the one currently followed by society, it does not automatically follow that one should obey this optimal norm rather than the sub-optimal one which has currency in society. 28. Some readers have disputed my claim that pronunciation is strongly socially dependent in this way. If they are right, then my case is hardly weakened: the less plausible it is to claim that pronunciation is strongly socially dependent, the less plausible it is to claim that morality is strongly socially dependent. 29. I take this example from Lance and O’Leary-Hawthorne (1997). 30. Mark Lance (2000a), p. 128. 31. One must be cautious here. As noted earlier, we should distinguish between whether an action is objectively right, and whether the person was (given her historical context) justiﬁed in thinking the action was right. Thus, even if, say, slavery is immoral, a citizen of ancient Athens might well have been justiﬁed in thinking that it was permissible. 32. Although some communities no doubt regard moral rules as reiﬁed in this way. Certain elements of Catholic practice, for example, resemble reiﬁed morality. Eating meat on Friday used to be permissible, but became immoral. Similarly, a divine command theorist might regard the rules of morality as reiﬁed. Indeed, a Christian who was a biblical literalist might, in viewing the vastly different moralities expressed in the Old and New Testaments, conclude that warlike behavior used to be permissible, whereas now only paciﬁsm is permitted. 33. Quine (1953), p. 43. 34. Lance and O’Leary-Hawthorne (1997), p. 120. 35. Nussbaum (2000), pp. 190–191. 36. Rachels (2003), p. 25. 37. Rorty (1989), p. 52.

272

Notes

5.

Pragmatism, Interests, and Morality

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

See Hampton (1998). Quinn (1993), p. 246, quoted in Scanlon (1998), p. 43. Scanlon (1998). Parﬁt (1984). Arneson (1999). Nozick (1974), pp. 42–43. Arneson (1999), p. 133. Sen (1984), p. 309, quoted in Nussbaum (2000), p. 139. The most developed version of this account is in Nussbaum (2000). Nussbaum (2000), pp. 78–80. Ibid., pp. 71–72. Ibid., p. 76. Nussbaum (1992), p. 223. Nussbaum (2000), p. 83. Ibid., p. 77. Ibid., p. 83. See, for example, Axelrod (1984). The so-called harm principle is usually attributed to Mill. As a legal principle, its most well-known elaboration is probably given by Feinberg (1987). Richardson (1990), p. 15. Ibid., p. 16. Kukla (2000), p. 209, n. 7. I am ignoring, of course, the anthropological theories suggesting that such artistic practices were thought by primitive peoples to aid in the hunt. This is, after all, a myth, not a historical essay. Although it has been argued that there are some circumstances where circular reasoning is acceptable in theoretical justiﬁcation. See, for example, Cling (2002). Well, technically, the ruling body governing the game. But there is a ruling body in soccer to make such decisions only because soccer is a reiﬁed game, a concept we discussed in the previous chapter. This conclusion has been argued for in many places, perhaps best in Quine’s “Two Dogmas of Empiricism.” Two classic attacks on the idea of a theory-neutral observation language are Hanson (1958) and Kuhn (1970). Gregg Osborne pressed this worry. Kant seems to have held this view. See, for example, Singer (1993). Gauthier (1967), pp. 461–462, emphasis removed. Of course, the passage continues, “ . . . yet acting on the system of principles requires that some persons perform disadvantageous acts.” Gauthier concludes that for this reason, morality is not supported by rationality. But of course, this is precisely the position I am trying to refute: I am trying to show that as we have otherregarding interests, and as we can show that cooperative rationality is a viable concept, it is often rational to be moral even if purely self-interested considerations conﬂict with moral considerations. As I have already argued, attempts (like Gauthier’s) to ground morality in rationality, which only

19. 20. 21. 22.

23.

24.

25. 26. 27. 28. 29. 30.

Notes

273

recognize individualistic reasoning (and not cooperative reasoning), are bound to fail.

6.

Pragmatism, Freedom, and Responsibility

1. An earlier version of this chapter (excluding the postscript) appeared in The Philosophical Forum as ‘Is Hard Determinism a Form of Compatibilism?’ 2. Corbin (1996). 3. Smilansky (2002) is perhaps the most sophisticated account of this kind. Smilansky argues that even though libertarian free will is an illusion, it is an essential one. Smilansky is not himself a pragmatist; he refers to his view as ‘illusionism.’ 4. Wittgenstein (1958), pp. 31–32. 5. A good recent anthology on this debate is Moral Particularism, Brad Hooker and Margaret Olivia Little (eds). 6. Recall our discussion of defeasible generalizations (that is, rules that have genuine authority, and which are not mere statistical generalizations, but which nevertheless allow of exceptions) from Chapter 3. 7. Of course, it may be that certain quantum events are uncaused, but as these events do not play a signiﬁcant role in human psychological causation, they may be ignored for the purposes of the free-will debate. 8. See, for example, the section “Draconian Sanctions?” below. 9. See below the section ‘Are We Free or Are We “Free”?’ 10. Harry Frankfurt thinks that freedom and responsibility come apart, but I think that he is driven to this position by some of the counterintuitive results of his overly-formalized deﬁnition of free will. See Frankfurt (1975). 11. Even if the hard determinists value liberty, one might still object that they do not value freedom (in the metaphysical sense), and so there is a point at which the interests of hard determinists and compatibilists diverge. Below (in the section ‘Is Freedom Intrinsically Valuable?’) I argue that both communities (hard determinist and compatibilist) might value freedom/sanctionability only instrumentally; and so their interests will not diverge at this point, either. 12. It is important to remember, as I emphasized in Chapter 3, that pragmatist is not the same as utilitarian. A pragmatically structured practice is, of course, structured so as to serve our interests, but this becomes a utilitarian account only on the assumption that there is some quantiﬁable stuff, ‘goodness,’ which such a practice ought to maximize. I see no reason to think that this assumption is true; and so the account of freedom/sanctionability offered here is a pragmatist account, but not a utilitarian one. (I am grateful to Mark Lance for helpful correspondence on this topic.) 13. One might wonder whether the hard-determinist community would recognize the interest in autonomy and control, which I listed in Chapter 4, given that such a community ostensibly would deny the existence of genuine freedom. I discuss this worry below in the section titled ‘Is Freedom Intrinsically Valuable?’ 14. Rawls (1955), p. 16.

274

Notes

15. One might nevertheless object that if the retributive practice is given a pragmatic justiﬁcation, then the punishment dictated by this practice is not really retributive punishment, but is instead deterrent punishment; and so one might maintain that the hard determinist will not, in fact, punish retributively. I will return to this objection in the section ‘Are We Free or Are We “Free”?’ 16. I will address the question of whether such punishments in the pragmatist community really ought to be considered retributive punishments in the section ‘Are We Free or Are We “Free”?’ 17. Rawls (1955), p. 16. 18. This example was suggested by an anonymous referee. 19. Klein (1995), p. 772. 20. Of course, the hard determinists might not use the phrase ‘moral responsibility,’ having abandoned it (along with ‘free’) in the initial stages of our myth. Obviously, though, what word they use is irrelevant; we are concerned with the meaning of whatever word they end up using to take its place. 21. Alston (1986). 22. Ibid., p. 59. 23. Ibid. 24. Although people are often condemned for having what are taken to be immoral beliefs. I will not discuss here whether this practice is legitimate or not. 25. See, for example, Webb (1993). 26. Webb would endorse the latter formulation. He argues that the justiﬁcation conferred by testimony is basic (i.e., it is not based on any induction regarding the reliability of people’s testimony). Thus, we need not verify that one meets required epistemic standards before admitting her testimony. Instead, each person’s testimony is prima facie justiﬁcation-conferring, unless you can demonstrate some reason (e.g., failure to conform to important epistemic norms) for doubting this person’s testimony. 27. ‘Dear Mr. Mbeki . . . An open letter to the president of South Africa.’

7. Pragmatism and Epistemology 1. Much of this chapter appeared in Episteme as ‘Conservatism, Basic Beliefs, and the Diachronic and Social Nature of Epistemic Justiﬁcation.’ 2. This seems to be the view of Williams (1985) and Baier (1994), and is of course explicit in the view of moral particularists. 3. Audi (1993), p. 18. My distinction between subjective and objective justiﬁcation is also similar to Goldman’s distinction between weak and strong justiﬁcation. See Goldman (1988). 4. Other philosophers have also distinguished between different types of justiﬁcation. Audi (1993) distinguishes between personal and impersonal justiﬁcation; Engel (1992) also argues that there are two different types of justiﬁcation, corresponding to internalism and externalism. These distinctions are somewhat different from the one I am drawing here. 5. We will talk about the importance of a belief system being open to revision later in the chapter.

Notes

275

6. Goldman (1988), p. 52. 7. Ibid. 8. I suspect that the two perspectives coincide in some cases, such as in arguments from authority. I will not discuss this matter here, though. 9. Goldman (1978), p. 510. Goldman (1986, pp. 279–283) argues further that standards of individual rationality should not be tied to what an ‘ideally logical being’ is capable of. 10. Lance (2000b), p. 116. 11. Kaplan (1996), pp. 37–38, quoted in Lance (2000b), p. 116. 12. Compare the criticism of act utilitarianism that humans lack the cognitive and predictive powers to accurately predict the long-term consequences of their actions, and hence lack the ability to abide by act-utilitarian constraints. 13. Lance (2000b), p. 119. 14. Sellars (1997), for example, is often read as endorsing this requirement. 15. Ibid. 16. Brandom (1997), p. 159. 17. Axtell 1997, p. 14. 18. Sklar (1975), p. 375. 19. Ibid., p. 396. 20. Ibid., pp. 396–397. 21. Wittgenstein (1969), §94. 22. See, for example, Wittgenstein (1969), §110. 23. See Kuhn (1959). 24. Although Kuhn has pointed out (1959, 1970) that we should not be too quick to revise our theories in light of evidence that conﬂicts with our theory. 25. Kuhn (1959). 26. Newton’s theory might be adequate to account for much planetary motion, but notoriously fails to accurately predict Mercury’s orbit, as Mercury’s proximity to the sun generates easily observable relativistic effects. 27. Again, see Kuhn (1959). 28. Wittgenstein (1969), §98: “the same proposition may get treated at one time as something to test by experience, at another as a rule of testing.” 29. Sellars (1956), §38. 30. The importance of revisability has been advocated not just by Sellars (1956), but also by Levi (1992), Brandom (1997), Lance and O’Leary-Hawthorne (1997), and others. 31. Wittgenstein (1969), §94. 32. This account is not intended as an exposition of Wittgenstein. I am arguing that our basic beliefs are rational; Wittgenstein, on the other hand, held them to be neither rational nor irrational. 33. The evolutionary analogy will go only so far—after all, evolution does not approach any goal, whereas our empirical inquiry has the goal of knowledge and explanation of the world. Many have taken evolutionary analogies like this more literally, and used this as fodder for instrumentalism in the philosophy of science. Discussion of this aspect of the realism–anti-realism debate would, however, take us too far aﬁeld. 34. Wittgenstein (1958), §217.

276

Notes

35. Again, Wittgenstein thinks these foundational beliefs are arational and not justiﬁed, but I think he is mistaken (for reasons we are now exploring). 36. This is why Mill (1978) writes that we can only know we are right because we allow free inquiry. In On Liberty, he writes, “There is the greatest difference between presuming an opinion to be true because, with every opportunity for contesting it, it has not been refuted, and assuming its truth for the purpose of not permitting its refutation. Complete liberty of contradicting and disproving our opinion is the very condition which justiﬁed us in assuming its truth for purposes of action; and on no other terms can a being with human faculties have any rational assurance of being right” (p. 18). If we do not allow challenges to orthodoxy, then orthodoxy ceases to be rational. It is only rational as long as it is before the tribunal of experience; when a belief is given permanent reprieve from challenges, then it ceases to be justiﬁed. 37. If a system became immune to challenge and revision, individuals subscribing to that system might be perfectly well-justiﬁed in continuing to have the beliefs dictated by the system in question, and thinking that the system is rational. But the system itself ceases to be rational. We will discuss this further later in the chapter. 38. Lance and O’Leary-Hawthorne (1997), p. 120. 39. This objection was raised by an anonymous editor. 40. Dawkins (1996), p. 139. 41. Popper (1959), p. 41, quoted in Campbell (1974), p. 415. 42. Campbell (1974), p. 421. 43. Kitcher (1993), pp. 95–96. 44. Ibid., p. 105. 45. Ibid., p. 92, italics in original. 46. Some philosophers have argued, though, that conservatism is supported by pragmatic reasons. Kuhn (1959) argues that conservatism speeds the progress of science. If these philosophers are right, then conservatism is supported by both pragmatic and epistemic principles. I will not deny this, but I will insist that conservatism is at least supported by the latter type of reasons. 47. See, for example, Goldstick (1971) [“ . . . independently of any such empirical grounds the bare fact that some proposition has been believed by us up to the present should be a consideration in its favor” (p. 186)]; Goldstick (1976); Foley (1982) [“ . . . a proposition acquires a favorable epistemic status for a person simply by being believed by him” (p. 165)]; Christensen (1994) [“ . . . an agent is in some measure justiﬁed in maintaining a belief simply in virtue of the fact that the agent has that belief” (p. 69)]; Adler (1996) [“ . . . believing that p is a reason for belief or continued belief that p” (p. 80)]. The above quotes merely represent the authors’ characterization of epistemic conservatism; most of the authors cited in fact reject conservatism as an epistemic principle. Harman (1986) is one philosopher who supports the principle of conservatism, however. 48. Wittgenstein (1969), §111. 49. An anonymous editor helped in the formulation of this point. 50. I have already argued in this chapter (and will further argue in Chapter 8) that given the fact that an individual cannot evaluate the rationality of these

Notes

277

55.

basic propositions, we should judge individual rationality not on the basis of the structure of an individual’s set of beliefs, but instead based on how the agent revises his or her beliefs. Wittgenstein (1969), §286. Some have argued that revision is crucial to all beliefs, not just basic ones. Mark Lance and John O’Leary-Hawthorne (1997) write, in a passage I quoted earlier, “A practice could in effect adopt the positivist proposal of treating a whole bunch of claims as de jure unchallengeable . . . But we would not be tempted to adopt such a practice ourselves. Such a practice seems to encourage—even be constitutive of—dogmatism, preclude dialogue, induce cognitive sterility, and all at no obvious gain” (p. 120). I will not pursue this issue here, but will instead conﬁne my argument to the importance of revisability for basic beliefs. I owe this objection to an anonymous referee. I am not entirely sure this is the case. If the reﬂective members of the community are aware that they recently came into existence, and were created with these beliefs, they might question the source of their beliefs about the world and come, legitimately, to doubt them. Axtell (1997), p. 14.

8.

Pragmatism, Internalism, and Externalism

51. 52.

53. 54.

1. See, for example, Dancy (1992), Engel (1992), Sennett (1992), and Audi (1993). 2. I am grateful to Mark Lance for pointing out this parallel, and for the example that follows. 3. Bonjour (1986), pp. 104–105. 4. David Armstrong (1973) introduces externalism precisely to solve the inﬁnite-regress problem. Bonjour (1986) discusses externalism, and Armstrong’s externalism in particular, speciﬁcally as a solution to the regress problem. 5. Bonjour (1986), p. 102. 6. Although in later writings Bonjour rejects coherentism and comes to accept some version of foundationalism, he remains consistent in his commitment to internalism. See Bonjour (1999). 7. Bonjour (1999), p. 122. 8. Many contemporary philosophers have thought they could press externalism into the service of a naturalistic outlook. Thus, we see externalism is a thread that runs through various versions of naturalized epistemology. However, the intuitions I am discussing are intuitions directly connected to the question of whether to understand justiﬁcation internally or externally; this naturalistic motivation, on the other hand, is not directly about this question, but is instead about the sort of properties that naturalistically minded philosophers ought to countenance. Thus, I hesitate to call it one of externalism’s motivating intuitions. 9. Bonjour (1986), p. 105. 10. See Kuhn (1959).

278

Notes

11. Although Kuhn has pointed out (1959, 1970) that we should not be too quick to revise our theories in light of evidence that conﬂicts with our theory. 12. Keith Lehrer (1971) and Ernest Sosa (1999) formulate versions of this principle, and demonstrate how it leads to skepticism. This principle is implicit in Cartesian skeptical arguments. 13. Engel (1992), p. 139. Engel also understands the externalist perspective in reliabilist terms, which I reject later in this chapter. 14. Brandom (1997), pp. 158–159. 15. Many have argued that our perceptual beliefs have grounds, not evidence. The above point remains intact if we substitute ‘grounds’ for ‘evidence.’ 16. These points count against those who think that the sole purpose of epistemic discourse is to maximize truth and minimize falsity. It has been pointed out that if all we cared about were truth, we could sit around all day and calculate addition problems, thereby guaranteeing ourselves a limitless supply of true (indeed, necessarily true) beliefs. But of course, we care about more than just truth—we want explanation and understanding. The ﬁnal goal I mention above—instrumental control over and mastery of nature—is particularly telling. If instrumental control over nature is one of the goals of inquiry, then truth is at best an intermediate goal on the road to this ﬁnal end. I think this demonstrates fairly conclusively that truth is not the only goal of epistemic discourse. 17. Cf. Brandom (1998). I owe my formulation of this point—that our assessments should not be seen as external to the game of giving and asking for reasons—to Brandom. 18. See, for example, Engel (1992) and Sennett (1992). 19. Brandom (2000), p. 210, note 1. 20. A similar objection against the notion of personal or subjective justiﬁcation is raised by Reiter (1998). 21. The ﬁnal clause—‘that she had at her disposal’—is important if we are to retain the internalist ﬂavor of internal justiﬁcation attributions. 22. See, for example, Alston (1986); Sosa (1999), pp. 150–151; and Comesaña (2005). 23. Conee and Feldman (2001). 24. Ibid., p. 234. 25. Ibid. 26. Indeed, Wittgenstein seems to deny that norms supervene even on entire communities, even if you take into account all past and future behavior. But this issue need not detain us here. 27. Alston (1988/1996), p. 61. 28. Ibid., p. 62. 29. Lance (2000b), p. 119. 30. See, for example, Cohen (1984). 31. See Goldman (1986). 32. See, for example, Leitgeb (2004), p. 335. 33. Goldman (1988), p. 52–53. 34. When I say, for example, ‘You ought to believe P,’ I am not reverting to the subjective perspective. To illustrate with a non-epistemic example, I might say, ‘You ought to try this wine.’ I am not saying that you have performed badly by not trying it, or that your failure to believe that you ought to try

Notes

279

it results from some failure of rationality. In other words, I am not making a subjective claim. Rather, I am making a claim from the objective perspective: I am saying that your tasting this wine is an action that objectively ought to occur; I am not in any way evaluating your performance as adequate or inadequate. 35. Sosa (1999), p. 151.

Bibliography Adler, Jonathan E. 1996: ‘An Overlooked Argument for Epistemic Conservatism,’ Analysis 56:2 (April), pp. 80–84. Alston, William P. 1986: ‘Internalism and Externalism in Epistemology,’ Philosophical Topics 14:1, pp. 185–226. Reprinted in Linda Martín Alcoff (ed.), Epistemology: The Big Questions (Malden, MA: Blackwell Publishers, 1998). Armstrong, David 1973: Belief, Truth and Knowledge (Cambridge: Cambridge University Press). Arneson, Richard J. 1999: ‘Human Flourishing Versus Desire Satisfaction,’ Social Philosophy and Policy 16:1 (Winter), pp. 113–142. Audi, Robert 1993: The Structure of Justiﬁcation (Cambridge, UK and New York: Cambridge University Press). Axelrod, Robert M. 1984: The Evolution of Cooperation (New York: Basic Books). Axtell, Guy 1997: ‘Recent Work on Virtue Epistemology,’ American Philosophical Quarterly 34:1 (January), pp. 1–26. Baghramian, Maria 2004: Relativism (London and New York: Routledge). Baier, Annette C. 1994: Moral Prejudices (Cambridge, MA: Harvard University Press). Baier, Kurt 1958: The Moral Point of View: A Rational Basis of Ethics (Ithaca, New York: Cornell University Press). Bakhurst, David 1999: ‘Pragmatism and Moral Knowledge,’ in Misak (1999), pp. 227–252. Bales, R. Eugene 1971: ‘Act-Utilitarianism: Account of Right-Making Characteristics or Decision-Making Procedure?’ American Philosophical Quarterly 8:3 (July), pp. 257–265. Berger, Joseph 1987: ‘The Papal Visit: Catholics, in Poll, Admire Pope but Disagree,’ The New York Times (10 September). Black, Robert 1990: ‘Moral Scepticism and Inductive Scepticism,’ Proceedings of the Aristotelian Society 30, pp. 65–82. Bonjour, Laurence 1986: ‘Can Empirical Knowledge Have a Foundation?’ in Paul K. Moser (ed.), Empirical Knowledge: Readings in Contemporary Epistemology (Savage, Maryland: Rowman and Littleﬁeld Publishers, Inc.), pp. 95–115. ——— 1999: ‘The Dialectic of Foundationalism and Coherentism,’ in John Greco and Ernest Sosa (eds), The Blackwell Guide to Epistemology (Malden, MA and Oxford, UK: Blackwell Publishers), pp. 117–142. Brandom, Robert B. 1979: ‘Freedom and Constraint by Norms,’ American Philosophical Quarterly 16:3 (July), pp. 187–196. ——— 1984: ‘Reference Explained Away,’ The Journal of Philosophy 81:9 (September), pp. 469–492. ——— 1994: Making It Explicit: Reasoning, Representing, and Discursive Commitment (Cambridge, MA: Harvard University Press). ——— 1997: ‘Study Guide,’ in Wilfrid Sellars (ed.), Empiricism and the Philosophy of Mind (Cambridge, MA: Harvard University Press), pp. 119–181. 280

Bibliography

281

——— 1998: ‘Insights and Blindspots of Reliabilism,’ Monist 81:3, pp. 371–392. Reprinted as Chapter 3 of Brandom (2000), pp. 97–122. ——— 2000: Articulating Reasons: An Introduction to Inferentialism (Cambridge, MA and London, England: Harvard University Press). Brink, David O. 1989: Moral Realism and the Foundations of Ethics (Cambridge, MA Cambridge University Press). Campbell, Donald 1974. ‘Evolutionary Epistemology,’ in Paul Arthur Schilpp (ed.), The Philosophy of Karl Popper, The Library of Living Philosophers Volume XIV Book 1 (La Salle, Illinois: Open Court), pp. 413–463. Carroll, Lewis 1895: ‘What the Tortoise Said to Achilles,’ Mind New Series 4:14 (April), pp. 278–280. Christensen, David 1994: ‘Conservatism in Epistemology,’ Noûs 28:1, pp. 69–89. Churchland, Paul 1979: Scientiﬁc Realism and the Plasticity of Mind (Cambridge: Cambridge University Press). Cling, Andrew D. 2002: ‘Justiﬁcation-Affording Circular Arguments,’ Philosophical Studies 111:3 (December), pp. 251–275. Cohen, Stewart 1984: ‘Justiﬁcation and Truth,’ Philosophical Studies 46, pp. 279–296. Comesaña, Juan 2005: ‘We Are (Almost) All Externalists Now,’ Philosophical Perspectives 19, pp. 59–76. Conee, Earl and Richard Feldman 2001: ‘Internalism Defended,’ in Hilary Kornblith (ed.), Epistemology: Internalism and Externalism (Cambridge, MA: MIT Press), pp. 231–260. Copp, David (ed.), 2006: Oxford Handbook of Ethical Theory (New York: Oxford University Press). Corbin, John Fowler 1996: ‘A Pragmatic Defense of Free Will,’ Journal of Value Inquiry 30:1–2 (April), pp. 247–260. Cormier, Harvey 2001: The Truth is What Works: William James, Pragmatism, and the Seed of Death (Lanham, Maryland: Rowman and Littleﬁeld Publishers, Inc.). Dancy, Jonathan 1992: ‘Externalism for Internalists,’ in Enrique Villanueva (ed.), Philosophical Issues Volume 2: Rationality in Epistemology (Atascadero, CA: Ridgeview Publishing Company), pp. 93–114. ——— 1993: Moral Reasons (Oxford: Blackwell). Dawkins, Richard 1996: Climbing Mount Improbable (New York: W.W. Norton and Company). Dean, Richard 1997: ‘A Defence of Constrained Maximization,’ Dialogue 36, pp. 453–467. ‘Dear Mr. Mbeki . . . An open letter to the president of South Africa,’ Editorial, Nature 404:6781, p. 907. Devine, Philip E. 1989: Relativism, Nihilism, and God (Notre Dame, IN: University of Notre Dame Press). De Waal, Cornelis 1999: ‘Eleven Challenges to the Pragmatic Theory of Truth,’ Transactions of the Charles S. Peirce Society 35:4 (Fall), pp. 748–766. Douglas, R. Bruce, Gerald M. Mara, and Henry S. Richardson (eds), 1990: Liberalism and the Good (New York: Routledge). Engel, Mylan, Jr, 1992: ‘Personal and Doxastic Justiﬁcation in Epistemology,’ Philosophical Studies 67:2 (August), pp. 133–150. Feinberg, Joel 1987: Harm to Others: The Moral Limits of the Criminal Law (New York: Oxford University Press).

282

Bibliography

Finnis, John 1989: Natural Law and Natural Rights (Oxford: Clarendon Press). Foley, Richard 1982: ‘Epistemic Conservatism,’ Philosophical Studies 43:2, pp. 165–182. Foot, Philippa 1985: ‘Utilitarianism and the Virtues,’ Mind 94:374 (April 1985), pp. 196–209. Frankfurt, Harry 1975: ‘Three Concepts of Free Action: Part II,’ Proceedings of the Aristotelian Society, Supplement 49, pp. 113–125. French, Peter A., Theodore E. Uehling, Jr, and Howard K. Wettstein (eds), 1988: Midwest Studies in Philosophy XIII: Ethical Theory: Character and Virtue (Notre Dame, IN: University of Notre Dame Press). Galbraith, John Kenneth 1958: The Afﬂuent Society (Boston: Houghton Mifﬂin Company). ——— 1967: The New Industrial State (New York: New American Library). Gauthier, David P. 1967: ‘Morality and Advantage,’ The Philosophical Review 76:4 (October), pp. 460–475. ———1986: Morals by Agreement (New York: Oxford University Press). Gettier, Edmund L. 1963: ‘Is Justiﬁed True Belief Knowledge?’ Analysis 23, pp. 121–123. Gibbard, Allan 1990: Wise Choices, Apt Feelings (Cambridge, MA: Harvard University Press). Gilbert, Margaret 1989: On Social Facts (Princeton, NJ: Princeton University Press). ——— 2001: ‘Collective Preferences, Obligations, and Rational Choice,’ Economics and Philosophy 17, pp. 109–119. Goldman, Alvin I. 1978: ‘Epistemics: The Regulative Theory of Cognition,’ The Journal of Philosophy 75:10, pp. 509–523. ——— 1986: Epistemology and Cognition (Cambridge, MA: Harvard University Press). ——— 1988: ‘Strong and Weak Justiﬁcation,’ in James E. Tomberlin (ed.), Philosophical Perspectives 2: Epistemology (Atascadero, California: Ridgeview Publishing Company), pp. 51–69. Goldstick, D. 1971: ‘Methodological Conservatism,’ American Philosophical Quarterly 8:2 (April), pp. 186–191. ——— 1976: ‘More on Methodological Conservatism,’ Philosophical Studies 30, pp. 193–195. Grover, D.L., J. Camp, and N.D. Belnap, Jr 1975: ‘A Prosentential Theory of Truth,’ Philosophical Studies 27, pp. 73–125. Hampton, Jean E. 1998: The Authority of Reason (Cambridge: Cambridge University Press). Hanson, Norwood Russell 1958: Patterns of Discovery: An Inquiry into the Conceptual Foundations of Science (Cambridge: Cambridge University Press). Hare, R.M. 1952: The Language of Morals (New York: Oxford University Press). Harman, Gilbert 1977: The Nature of Morality (New York: Oxford University Press). Chapter 1 reprinted as Chapter 6 in G. Sayre-McCord (ed.), 1988: Essays on Moral Realism (Ithaca: Cornell University Press), pp. 119–124. ——— 1986. Change in View (Cambridge, MA: MIT Press). Hooker, Brad 2000: ‘Moral Particularism: Wrong and Bad,’ in Hooker and Little (2000), pp. 1–22. ——— and Margaret Olivia Little (eds), 2000: Moral Particularism (Oxford: Clarendon Press).

Bibliography

283

Hume, David 1888/1968: A Treatise of Human Nature, L.A. Shelby-Bigge and P.H. Nidditch (eds) (Oxford: Oxford University Press). Hunter, Daniel 1994: ‘Act Utilitarianism and Dynamic Deliberation,’ Erkenntnis 41, pp. 1–35. Jackson, Frank and Philip Pettit 1990: ‘Program Explanation: A General Perspective,’ Analysis 50, pp. 107–117. Jackson, Frank, Philip Pettit, and Michael Smith 2000: ‘Ethical Particularism and Patterns,’ in Hooker and Little (2000), pp. 79–99. James, William 1977: The Writings of William James: A Comprehensive Edition, John J. McDermott (ed.) (Chicago: The University of Chicago Press). Kaplan, Mark 1996: Decision Theory as Probability (Cambridge: Cambridge University Press). Kaptchuk, Ted J. 1998: ‘Intentional Ignorance: A History of Blind Assessment and Placebo Controls in Medicine,’ Bulletin of the History of Medicine 72:3, pp. 389–433. Kitcher, Philip 1993. The Advancement of Science (New York: Oxford University Press). Klein, Martha 1995: ‘Responsibility,’ in Ted Honderich (ed.), The Oxford Companion to Philosophy (Oxford: Oxford University Press), pp. 771–772. Koons, Jeremy Randel 1998: ‘A Conditional Defense of Moral Realism’ (PhD Dissertation, Georgetown University). ——— 2000: ‘Do Normative Facts Need to Explain?’ Paciﬁc Philosophical Quarterly 81:3 (September), pp. 246–272. ——— 2002: ‘Is Hard Determinism a Form of Compatibilism?’ The Philosophical Forum 33:1 (March), pp. 81–99. ——— 2005: ‘Conservatism, Basic Beliefs, and the Diachronic and Social Nature of Epistemic Justiﬁcation,’ Episteme 2:3 (October), 203–218. ——— 2006: ‘An Argument against Reduction in Morality and Epistemology,’ Philosophical Investigations 29:3, pp. 250–274. Kripke, Saul 1977: ‘Identity and Necessity,’ in S. Schwartz (ed.), Naming, Necessity, and Natural Kinds (Ithaca, New York: Cornell University Press). ——— 1982: Wittgenstein on Rules and Private Language (Oxford: Basil Blackwell). Kuhn, Thomas S. 1959: ‘The Essential Tension,’ in C.W. Taylor (ed.), The Third (1959) University of Utah Research Conference on the Identiﬁcation of Scientiﬁc Talent (Salt Lake City: University of Utah Press), pp. 162–174. Reprinted in Kuhn (1977), pp. 225–239. ——— 1970: The Structure of Scientiﬁc Revolutions, 2nd edition (Chicago: University of Chicago Press). ——— 1977: The Essential Tension: Selected Studies in Scientiﬁc Tradition and Change (Chicago: The University of Chicago Press). Kukla, Rebecca 2000: ‘Myth, Memory and Misrecognition in Sellars’ Empiricism and the Philosophy of Mind,’ Philosophical Studies 101, pp. 161–211. Lance, Mark 1990: ‘Rules, Practices and Norms,’ in Souren Teghrarian, Anthony Seraﬁni, and Edward M. Cook (eds), Ludwig Wittgenstein: A Symposium on the Centennial of His Birth (Wakeﬁeld, NH: Longwood Academic), pp. 77–86. ——— 2000a: ‘The Word Made Flesh: Toward a neo-Sellarsian View of Concepts, Analysis, and Understanding,’ Acta Analytica 15:25, pp. 117–135. ———— 2000b: ‘The Best is the Enemy of the Good: Bayesian Epistemology as a Case Study in Unhelpful Idealization,’ in Niall Shanks and Robert B. Gardner

284

Bibliography

(eds), Poznan Studies in the Philosophy of Science and the Humanities 71:Logic, Probability and Science (Amsterdam: Rodopi), pp. 112–135. Lance, Mark and M. Little 2004: ‘Defeasibility and the Normative Grasp of Context,’ Erkenntnis 61:2–3 (November), pp. 435–455. ——— 2005: ‘Defending Moral Particularism,’ in James Dreier (ed.), Contemporary Debates in Moral Theory (Blackwell), pp. 305–321. ——— 2005: ‘Particularism and Anti-Theory,’ in Copp (2006) pp. 567–594. Lance, Mark Norris and J. O’Leary-Hawthorne 1997: The Grammar of Meaning: An Exploration of the Normative Character of Semantic Discourse (New York: Cambridge University Press). Lehrer, Keith 1971: ‘Why Not Skepticism?’ Philosophical Forum 2:3, pp. 283– 298. Reprinted in George S. Pappas and Marshall Swain (eds), 1978: Essays on Knowledge and Justiﬁcation (Ithaca, NY: Cornell University Press), pp. 346–463. Leitgeb, Hannes 2004: Inference on the Low Level: An Investigation into Deduction, Nonmonotonic Reasoning, and the Philosophy of Cognition (Dordrecht, The Netherlands: Kluwer). Levi, Isaac 1992: The Fixation of Belief and Its Undoing: Changing Beliefs Through Inquiry (New York: Cambridge University Press). Levy, Sanford S. 1997: ‘Utilitarianism Alternatives to Act Utilitarianism,’ Paciﬁc Philosophical Quarterly 78, pp. 93–112. Longino, Helen E. 1990: Science as Social Knowledge (Princeton, NJ: Princeton University Press). Lovibond, Sabina and S.G. Williams (eds), 1996: Essays For David Wiggins: Identity, Truth and Value Aristotelian Society Series Vol. 16 (Oxford: Blackwell Publishers). Mackie, J.L. 1977: Ethics: Inventing Right and Wrong (New York: Penguin Books). McDowell, John 1983: ‘Aesthetic Value, Objectivity, and the Fabric of the World,’ in E. Schaper (ed.), Pleasure, Preference and Value (Cambridge: Cambridge University Press), pp. 1–16. ——— 1986: ‘Critical Notice of Ethics and the Limits of Philosophy by Bernard Williams,’ Mind 95:379 (July), pp. 377–386. Mill, John Stuart 1978: On Liberty, Elizabeth Rapaport (ed.) (Indianapolis, IN: Hackett Publishing Company, Inc.). Miller, Alex 2003: An Introduction to Contemporary Metaethics (Cambridge: Polity Press). Misak, Cheryl 1996: ‘Pragmatism, Empiricism and Morality,’ in Lovibond and Williams (1996), pp. 201–218. ——— (ed.), 1999: Pragmatism: Canadian Journal of Philosophy Supplementary Volume 24 (Calgary: University of Calgary Press). ——— (ed.), 2004: The Cambridge Companion to Peirce (Cambridge: Cambridge University Press). Miscevic, Nenad 1989: Rationality and Cognition: Against Relativism-Pragmatism (University of Toronto Press). Nozick, Robert 1974: Anarchy, State, and Utopia (New York: Basic Books). Nussbaum, Martha 1992: ‘Human Functioning and Social Justice: In Defense of Aristotelian Essentialism,’ Political Theory 20:2 (April), pp. 202–246. ——— 2000: Women and Human Development: The Capabilities Approach (Cambridge: Cambridge University Press). Parﬁt, Derek 1984: Reasons and Persons (Oxford: Oxford University Press).

Bibliography

285

Peirce, Charles Sanders 1877: ‘The Fixation of Belief,’ in Peirce (1992), pp. 109–123. ——— 1908: ‘A Neglected Argument for the Reality of God,’ in Peirce (1998), pp. 434–450. ———1992: The Essential Peirce: Selected Philosophical Writings, Volume I (1867– 1893), Nathan Houser and Christian Kloesel (eds) (Bloomington, IN: Indiana University Press). ——— 1998: The Essential Peirce: Selected Philosophical Writings, Volume II (1893– 1913), Nathan Houser and Christian Kloesel (eds) (Bloomington, IN: Indiana University Press). Popper, Karl 1959: The Logic of Scientiﬁc Discovery (London: Routledge). Putnam, Hilary 1977: ‘Meaning and Reference,’ in S. Schwartz (ed.), Naming, Necessity, and Natural Kinds (Ithaca, New York: Cornell University Press). ———2004: The Collapse of the Fact/Value Dichotomy and Other Essays (Cambridge, MA: Harvard University Press). Quine, W.V.O. 1953: ‘Two Dogmas of Empiricism,’ From a Logical Point of View (New York: Harper Torchbooks). Quinn, Warren S. 1986: ‘Truth and Explanation in Ethics,’ Ethics 96:3, pp. 524–544. ——— 1993: Morality and Action (Cambridge: Cambridge University Press). Rachels, James 2003: The Elements of Moral Philosophy, 4th edition (New York: McGraw-Hill). Railton, Peter 1984: ‘Alienation, Consequentialism, and the Demands of Morality,’ Philosophy and Public Affairs 13 (Spring), pp. 134–171. Rawls, John 1955: ‘Two Concepts of Rules,’ The Philosophical Review 64:1, pp. 3–32. ——— 1971: A Theory of Justice (Cambridge, MA: Harvard University Press). Reiter, David 1998: ‘Engel on Internalism and Externalism in Epistemology,’ Erkenntnis 49:2, pp. 175–184. Richardson, Henry S. 1990: ‘The Problem of Liberalism and the Good,’ in Douglas, Mara, and Richardson (1990), pp. 1–28. Rorty, Richard 1989: Contingency, Irony, and Solidarity (New York: Cambridge University Press) Ryan, Sharon 2003: ‘Doxastic Compatibilism and the Ethics of Belief,’ Philosophical Studies 114, pp. 47–79. Sayre-McCord, Geoffrey 1988: ‘Moral Theory and Explanatory Impotence,’ in G. Sayre-McCord (ed.), Essays on Moral Realism (Ithaca, NY: Cornell University Press). Scanlon, T.M. 1998: What We Owe to Each Other (Cambridge, MA: Harvard University Press). Sellars, Wilfrid 1950: ‘Language, Rules, and Behavior,’ in Sydney Hook (ed.), John Dewey: Philosopher of Science and Freedom (Dial Press), pp. 289–315. Reprinted in Sellars (1980), pp. 129–155. ——— 1953: ‘A Semantical Solution of the Mind-Body Problem,’ Methodos 5, pp. 45–82. Reprinted in Sellars (1980), pp. 219–256. ——— 1956: ‘Imperatives, Intentions, and the Logic of ‘Ought,” Methodos 6, pp. 226–268. ——— 1980: Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars, Jeffrey F. Sicha (ed.) (Atascadero, CA: Ridgeview Publishing Company).

286

Bibliography

——— 1992: Science and Metaphysics: Variations on Kantian Themes, The John Locke Lectures for 1965–1966 (Atascadero, CA: Ridgeview Publishing Company). ——— 1997: Empiricism and the Philosophy of Mind (Cambridge, MA: Harvard University Press). Sen, Amartya K. 1984: Resources, Values, and Development (Cambridge, MA: Harvard University Press). Sennett, James F. 1992: ‘Toward a Compatibility Theory for Internalist and Externalist Epistemologies,’ Philosophy and Phenomenological Research 52:3 (September), pp. 641–655. Singer, Peter 1993: Practical Ethics, 2nd edition (Cambridge: Cambridge University Press). Sklar, Lawrence 1975: ‘Methodological Conservatism,’ The Philosophical Review 84:3 (July), pp. 374–400. Smart, J.J.C. 1956: ‘Extreme and Restricted Utilitarianism,’ Philosophical Quarterly 6:25 (October), pp. 344–354. Smilansky, Saul 2002: Free Will and Illusion (New York: Oxford University Press). Sosa, Ernest 1999: ‘Skepticism and the Internal/External Divide,’ in John Greco and Ernest Sosa (eds), The Blackwell Guide to Epistemology (Malden, MA: Blackwell Publishers, Inc.), pp. 145–157. Stace, Walter T. 1952: Religion and the Modern Mind (London: MacMillan). Steinfels, Peter 1994: ‘Catholics Question Article of Faith,’ The New York Times (5 June). Sturgeon, Nicholas L. 1988: ‘Moral Explanations,’ in G. Sayre-McCord (ed.), Essays on Moral Realism (Ithaca, NY: Cornell University Press), pp. 229–255. Sugden, Robert 2000: ‘Team Preference,’ Economics and Philosophy 16, pp. 175–204. Talisse, Robert B. 2001: ‘On the Supposed Tension in Peirce’s “Fixation of Belief”,’ Journal of Philosophical Research XXVI, pp. 561–569. Taylor, Richard 1988: ‘Ancient Wisdom and Modern Folly,’ in French, Uehling, and Wettstein (eds), Midwest Studies in Philosophy XIII(Notre Dame, Indiana: University of Notre Dame Press), pp. 54–63. Thoreau, Henry David 1991: ‘Civil Disobedience,’ in Hugo Adam Bedau (ed.), Civil Disobedience in Focus (London and New York: Routledge), pp. 28–48. Tuomela, Raimo 1995: The Importance of Us (Stanford, CA: Stanford University Press). ——— 1996: ‘Rational Cooperation and Collective Goals,’ ProtoSociology 8–9, pp. 280–314. Velleman, J. David 1997: ‘How to Share an Intention,’ Philosophy and Phenomenological Research 57:1 (March), pp. 29–50. Webb, Mark Owen 1993: ‘Why I Know About As Much As You: A Reply to Hardwig,’ Journal of Philosophy 90:5 (May), pp. 260–270. Wiggins, David 1991: ‘Moral Cognitivism, Moral Relativism and Motivating Moral Beliefs,’ Proceedings of the Aristotelian Society 91, pp. 61–85. ——— 1996: ‘Reply to Cheryl Misak,’ in Lovibond and Williams (1996), pp. 276–279. ——— 1999: ‘C.S. Peirce: Belief, Truth, and Going from the Known to the Unknown,’ in Misak (1999), pp. 9–29.

Bibliography

287

Wiggins, David 2004: ‘Reﬂections on Inquiry and Truth Arising from Peirce’s Method for the Fixation of Belief,’ in Misak (2004), pp. 87–126. Williams, Bernard 1985: Ethics and the Limits of Philosophy (Cambridge, MA: Harvard University Press). Wittgenstein, Ludwig 1958: Philosophical Investigations, trans. G.E.M. Anscombe (New York: Macmillan Publishing Co., Inc.). ——— 1969: On Certainty, G.E.M. Anscombe and G.H. von Wright (eds), Denis Paul and G.E.M. Anscombe, transl. (New York: Harper and Row). Wright, Crispin 1992: Truth and Objectivity (Cambridge, MA: Harvard University Press).

Index action, see rationality Adler, Jonathan E., 276n47 Alm, David, 266n42, 266n43, 266n46 Alston, William, 180–1, 251–2, 278n22 amoralist, 68–70 Aristotle, 152 Armstrong, David, 277n4 Arneson, Richard, 127 astrology, rationality of belief in, 8, 191–2, 196, 239, 250 Audi, Robert, 191, 274n4, 277n1 Axelrod, Robert M., 269n45, 272n17 Ayer, A.J., 14 Baier, Annette, 274n2 Baier, Kurt, 69, 268n22 Bakhurst, David, 1, 264n8 Bales, R. Eugene, 267n12 basic beliefs, see belief, basic Bayesianism, 73, 197–8, 255 belief basic, 212–23 impermissibility of practical justiﬁcation of, 33–5 instability of practical justiﬁcation of, 37–9 Blackburn, Simon, 15 Black, Robert, 44–5 Bonjour, Laurence, 235–9, 277n4 brain in vat, 253–4, 257–60 see also evil demon brainwashing (counterexample to internalism), 260 Brandom, Robert, 4, 85–6, 100, 200, 243, 247–8, 265n27, 275n30, 278n17 Brandt, Richard B., 61, 64 Brink, David, 264n7 Campbell, Donald, 216 Carroll, Lewis, 86

causal explanation, see explanation, causal Chamberlin, Wilt, 106, 107, 110, 111 Christensen, David, 276n47 Churchland, Paul, 264n6 circular justiﬁcation, permissibility of, 139–41, 144–6 clairvoyant (counterexample to externalism), 256–7 Cling, Andrew D., 272n23 Clinton, Bill, 229–30 Cohen, Stewart, 278n30 coherentism (epistemology), 101, 189, 197, 199–200, 237–8, 255 Comesaña, Juan, 278n22 compatibilism, 153–5, 160ff complexity of belief systems, 100–4 Conee, Earl, 250–2 conservatism (epistemic), 100, 203–8, 212–23, 240 cooperation, see rationality Copernicanism, 21 Corbin, John Fowler, 152 Cormier, Harvey, 263n12 Dancy, Jonathan, 87, 277n1 Dawkins, Richard, 56–7, 216 Dean, Richard, 56–7 desires and reasons for action, 123–6 determinism, 156, 160 DeVoto, Bernard, 265n17 De Waal, Cornelis, 263n11 diachronic justiﬁcation, see justiﬁcation, diachronic nature of dogmatism, 207–8, 212–23, 229 double-blind clinical trials, 102 doxastic voluntarism, 7, 179–82, 236, 242 Dummett, Michael, 41 duties, positive and negative, 74, 148 288

Index 289 Eddington, Arthur Stanley, 230–1 eliminativism doctrinal, 19, 20–1, 25, 26, 29, 41 explanatory, 16, 40 ontological, 20, 40 Engel, Mylan, Jr, 241–2, 274n4, 277n1 epistemology, 189ff evidence forgotten evidence as counterexample to internalism, 260–1 evil demon, 224, 228, 257–60 see also brain in vat; New Evil Demon Problem excellence of reasons, see truth and excellence of reasons exclusion from epistemic community, 182–6, 202, 210 explanation causal, 14, 16, 17, 18, 21 non-causal, 17–18 normative, 18 rational, 17 vindicatory, 92, 94, 96, 117 explanatory requirement, 14, 15–20, 26–7, 29 and epistemology, 19–20 see also Harman, Gilbert expressivism, see normative discourse externalism, epistemic, 8, 234ff pragmatist rejection of traditional externalism, 245–6, 252 externalist justiﬁcation attributions, see justiﬁcation, objective

Frankfurt, Harry, 273n10 free will, 7, 152ff

fallibilism, 104, 214, 244, 247–8 Feinberg, Joel, 272n18 Feldman, Richard, 250–2 Finnis, John, 98 ﬂourishing desire-satisfaction theories, 126–8 hedonistic theories, 126–7 objective-list theories, 127–9 see also interests, basic Foley, Richard, 276n47 Footballers’ Problem, 58–62, 67–8 Foot, Philippa, 77 foundationalism (epistemology), 189, 199, 255

incompatibilism, 156 interests, 7, 30, 62–3, 98ff afﬁliative or other-regarding, 62–4, 75, 91, 134, 147 basic, 128, 130–4 derivative, 130–4 non-basic, 138–9 internalism, epistemic, 8, 234ff internalist justiﬁcation attributions, see justiﬁcation, subjective

Galilei, Galileo, 37–8, 108 Gauthier, David, 56–7, 63, 65, 150 Gettier problem, 231–2 Gibbard, Allan, 15, 40 Gilbert, Margaret, 58, 66–9, 72, 74–5, 268n16 goblins (thought experiment), 224, 226–8 Goldman, Alvin I., 19, 192, 197, 258n3 Goldstick, Daniel, 276n47 Hampton, Jean, 123 Hanson, Norwood Russell, 272n26 hard determinism, 153–5, 160ff Harman, Gilbert, 3, 5, 14–19, 29n47 Hawking, Stephen, 25 Hawthorne, John, see John O’Leary -Hawthorne Hill, Austin Bradford, 102 Hobbes, Thomas, 44, 141, 144 Hooker, Brad, 87–8, 273n5 Hubble, Edwin, 212 Humean theory of normative reasons, see desires Hume, David, 3, 19 Hunter, Daniel, 55

Jackson, Frank, 264n10, 269n56 James, William, 2, 3 joint commitment, 65–7

290

Index

justiﬁcation context-sensitivity of subjective justiﬁcation, 194–9, 202–3 deontic conception, 201–2 diachronic nature of, see revision of moral and epistemic systems objective, 191–4, 208ff, 242–8, 252–62 subjective, 191–208, 223ff, 238–42, 248–52, 256–62 and truth, 237, 244–9 type required for knowledge, 229–32 see also circular justiﬁcation, permissibility of; myths, justiﬁcation-conferring Kant, Immanuel, 74, 148n28 Kaplan, Mark, 198 Kaptchuk, Ted J., 271n26 Kitcher, Philip, 216–17 knowledge, 229–32 Kuhn, Thomas S., 21–3, 210–11, 264n6, 272n26, 275n24, 276n46, 278n11 Kukla, Rebecca, 141 Lance, Mark, 4, 88–90, 108, 113, 197–8, 199, 215, 255n27, 265n30, 266n32, n38, n58, 267n60, 273n12, 275n30, 277n52 Lehrer, Keith, 278n12 Leitgeb, Hannes, 278n32 Levi, Isaac, 275n30 Levy, Sanford S., 76 Lewis, Carl, 202 libertarianism (free will), 154 Lind, James, 102 Little, Margaret, 88–90, 273n5 Longino, Helen, 264n6 Long, Roderick, 267n61 McDowell, John, 6, 31, 32, 41, 93–7, 105, 117 Macintosh, Duncan, 266n34 Mackie, John, 3, 14n40 McNaughton, David, 31 maximization, straightforward vs. constrained, 56–7 Mbeki, Thabo, 184–5

Mhlongo, Sam, 185 Miller, Alex, 264n10 Mill, John Stuart, 171, 272n18, 276n36 Misak, Cheryl, 93, 264n8 Modest Probabilism, 198 Moore, G.E., 31 moral luck, 235 moral responsibility, see responsibility, moral mushroom pickers (Gilbert example), 72 myths, justiﬁcation-conferring, 141–4 naïf (counterexample to internalism), 260 natural law theory of morality, 135 Nature magazine, 185 Nelson, Mark, 264n14, 267n4, n5, 269n57 New Evil Demon Problem, 257–60 non-monotonic inference, see rules, defeasible normative discourse expressivism, 40, 42 as legislative discourse, 40, 42 whether descriptive, 40–4 Nozick, Robert, 127, 269n41 Nussbaum, Martha, 98, 114, 127–30 O’Leary-Hawthorne, John, 4, 113, 215n27, 266n58, 267n60, 275n30, 277n52 Osborne, Gregg, 272n27 Parﬁt, Derek, 126 particularism, moral, 87–8, 96, 156 Peirce, Charles Sanders, 1, 3, 38–9, 94–5 perception, moral and epistemic, 31–3 Pettit, Philip, 264n10, 269n56 Popper, Karl, 216 practical and theoretical justiﬁcation, see belief practice (set of rules), see rules see also social practice pragmatic reasons and epistemology, 6, 21–5 and morality, 25–31

Index 291 preferences adaptive preferences, problem of, 127–8 collective preference, 65–7 and reasons for action, 125–6 Prisoner’s Dilemma, 60–2, 76 punishment to enforce cooperative strategy, 150 objections to pragmatist account of, 167–70, 171–5, 178 see also sanctions Putnam, Hilary, 264n8 Quine, W.V.O., 112–13, 272n25 Quinn, Warren, 124 Rachels, James, 117 Railton, Peter, 77 rationality act-based (non-strategic), 50–8 coerced participation in cooperative strategies, 69–71 conﬂicts between individual and cooperative, 73–5 cooperative, 6, 58–65, 131–5 fundamental vs. derivative, 49–52 individual, 58–65 instrumental or means-end, 46–7, 51, 61 objections against strategic cooperative rationality, 75–82 opting out of cooperative strategies, 68–71 prudential, 27–8, 35–6, 67–8, 136, 141–6, 190 strategic, 6, 51–8 Rawls, John, 34–5, 55, 82, 140, 144, 167–8, 170 reasons timelessness of, 107–11 see also truth and excellence of reasons reason, see rationality Reiter, David, 278n20

relativism of circumstances, 116–17 cross-cultural, 113–16 epistemology and, 210 historical, 105–12 traditionally associated with pragmatism, 2–3 reliabilism, 189, 203, 245, 254–5, 258 see also externalism, epistemic responsibility epistemic, 7, 179–82, 236 moral, 7, 177–8 revision of moral and epistemic systems, 100, 103, 112–13, 206–7, 212–23, 227–8, 241–2, 243–4 Richardson, Henry, 140 Rorty, Richard, 2, 117 Ross, W.D., 90 rules applicability and complexity of, 55–6, 73, 78–9 defeasible, 84–90, 146–7 and practices, 33–5, 47 reiﬁcation of, 110–11 see also rationality rule-worship objection, 6, 47–50, 77, 82–3 Ruritanians (thought experiment), 223–6, 228 sanctions, 155–66 see also punishment Sayre-McCord, Geoffrey, 263n3 Scanlon, T.M., 123–6 Schiller, F.C.S., 2, 3 Sellars, Wilfrid, 17, 65–6, 200, 212–13, 243n14, 275n30 Sen, Amartya, 98, 128 Sennett, James F., 277n1 Singer, Peter, 6, 149 Sklar, Lawrence, 204 Smart, J.J.C., 6, 45, 47–8, 50, 54 Smilansky, Saul, 152 Smith, Michael, 269n56 social practice, 98–105, 109, 112 Sosa, Ernest, 260, 278n12, n22 Stahl, Georg, 195 strategies, see rationality

292

Index

Sturgeon, Nicholas, 265n17 Sugden, Robert, 58–60, 69n21 Talisse, Robert B., 263n11 Taylor, Richard, 266n33 Thoreau, Henry David, 74 truth-aptness of normative discourse, 42–4 truth and excellence of reasons, 6, 92–7, 117 Tuomela, Raimo, 73 underdetermination of morality and epistemology by interests, 98–104, 111 utilitarianism act, 34–5, 52–6, 83, 167–8, 275n12

as different from pragmatism, 121–3, 147–50 rule, 34–5, 47, 75–9, 167–8 van Fraassen, Bas, 18 Velleman, David, 65 vindicatory explanation, see explanation, vindicatory Webb, Mark Owen, 274n5, 274n6 Wiggins, David, 3, 6, 39, 93–7, 105, 117n60 Williams, Bernard, 274n2 Wittgenstein, Ludwig, 155, 205, 213–15, 220–2 Wright, Crispin, 5, 264n15