The Ethics of Belief and Beyond: Understanding Mental Normativity (Routledge Studies in Epistemology) [1 ed.] 0367245507, 9780367245504

Citation preview

The Ethics of Belief and Beyond “Specialists in the ethics of action and the nature of practical reason now routinely interact with specialists in epistemology and the normativity of belief to address broader issues in the normativity of all kinds of mental states. As both felds learn from another, new questions and new avenues for inquiry come into view. This is a welcome collection of new essays by frstrate philosophers in the emerging feld of mental normativity.” —Peter J. Graham, University of California, Riverside, USA This volume provides a framework for approaching and understanding mental normativity. It presents cutting-edge research on the ethics of belief as well as innovative research beyond the normativity of belief – and towards an ethics of mind. By moving beyond traditional issues of epistemology the contributors discuss the most current ideas revolving around rationality, responsibility, and normativity. The book’s chapters are divided into two main parts. Part I discusses contemporary issues surrounding the normativity of belief. The chapters here cover topics such as control over belief and its implication for the ethics of belief, the role of the epistemic community for the possibility of epistemic normativity, responsibility for believing, doxastic partiality in friendship, the structure and content of epistemic norms, and the norms for suspension of judgment. In Part II the focus shifts from the practical dimensions of belief to the normativity and rationality of other mental states – especially blame, passing thoughts, fantasies, decisions, and emotions. These chapters illustrate how we might approach an ethics of mind by focusing not only on belief, but also on debates about responsibility, rationality, and the normativity of involuntary mental states. The Ethics of Belief and Beyond paves the way towards an ethics of mind by building on and contributing to recent philosophical discussions in the ethics of belief and the normativity of other mental phenomena. It will be of interest to upper-level students and researchers working in epistemology, ethics, philosophy of action, philosophy of mind, and moral psychology. Sebastian Schmidt is a postdoctoral researcher in philosophy at the University of Zurich. His research in epistemology and metaethics focusses on responsibility and reasons for attitudes. Gerhard Ernst holds a chair for philosophy at the Friedrich-AlexanderUniversität Erlangen-Nürnberg (Germany). Since 2018 he is President of the Deutsche Gesellschaft für Philosophie (German Society for Philosophy). His main research areas are epistemology and metaethics. His works include Das Problem des Wissens (The Problem of Knowledge) (2002) and Die Objektivität der Moral (The Objectivity of Morality) (2008).

Routledge Studies in Epistemology Edited by Kevin McCain

University of Alabama at Birmingham, USA

and Scott Stapleford

St. Thomas University, Canada

Pragmatic Encroachment in Epistemology Edited by Brian Kim and Matthew McGrath New Issues in Epistemological Disjunctivism Edited by Casey Doyle, Joseph Milburn, and Duncan Pritchard Knowing and Checking An Epistemological Investigation Guido Melchior Well-Founded Belief New Essays on the Epistemic Basing Relation Edited by J. Adam Carter and Patrick Bondy Higher-Order Evidence and Moral Epistemology Edited by Michael Klenk Social Epistemology and Relativsm Edited by Natalie Alana Ashton, Martin Kusch, Robin McKenna and Katharina Anna Sodoma Epistemic Duties New Arguments, New Angles Edited by Kevin McCain and Scott Stapleford The Ethics of Belief and Beyond Understanding Mental Normativity Edited by Sebastian Schmidt and Gerhard Ernst

For more information about this series, please visit: www.routledge.com/ Routledge-Studies-in-Epistemology/book-series/RSIE

The Ethics of Belief and Beyond Understanding Mental Normativity Edited by Sebastian Schmidt and Gerhard Ernst

First published 2020 by Routledge 52 Vanderbilt Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2020 selection and editorial matter, Sebastian Schmidt and Gerhard Ernst; individual chapters, the contributors The right of Sebastian Schmidt and Gerhard Ernst to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging in Publication Data A catalog record for this book has been requested ISBN: 978-0-367-24550-4 (hbk) ISBN: 978-0-429-28497-7 (ebk) Typeset in Sabon by Apex CoVantage, LLC

Contents

Acknowledgements 1

Introduction: Towards an Ethics of Mind

vii 1

S E B A S TI A N S CH MIDT

PART I

New Perspectives on Belief Normativity

21

A Doxastic Agency and Responsibility

22

2

Implications of the Debate on Doxastic Voluntarism for W. K. Clifford’s Ethics of Belief

23

M A RTI N A L I N DN E R

3

Believing as We Ought and the Democratic Route to Knowledge

47

M ATTH E W C H RISMA N

4

Responsibility for Doxastic Strength Grounds Responsibility for Belief

71

B E N O I T G A U LTIE R

B Reasons for Belief

86

5

87

Relationships and Reasons for Belief L I N D S AY C R AWFO RD

6

Instrumental Reasons for Belief: Elliptical Talk and Elusive Properties A S B J Ø R N S TEGL ICH - P E TE RSE N A N D MATTIAS SK IPPER

109

vi

Contents

7 Suspension of Judgment, Rationality’s Competition, and the Reach of the Epistemic

126

ERROL LORD

PART II

Facets of an Ethics of Mind

147

A Responsibility, Reasons, and Rationality

148

8 Responsibility for Attitudes, Object-Given Reasons, and Blame

149

S E B A S TI A N S CH MIDT

9 Two Kinds of Rationality

176

G E R H A R D E R N ST

B The Ethics of Blame, Fear, Decision, Passing Thought, and Phantasy

191

10 The Ethics of Blame: A Primer

192

D . J U S TI N C OATE S

11 How Safe Should We Feel? On the Ethics of Fear in the Public Sphere

215

S A B I N E A . D ÖRIN G

12 Determining the Future

234

M ATTH E W S O TE RIO U

13 Silence and Salience: On Being Judgmental

256

N E A L A . TO GN A ZZIN I

14 The Value of a Free and Wandering Mind

270

M I R I A M S C HL E IFE R McCO RMICK

List of Contributors Index

289 292

Acknowledgements

This volume owes its existence, frst and foremost, to its contributors. We discussed our ideas with each other for the frst time at the conference Ethics of Mind: Responsibility, Normativity, and Rationality that took place at the Friedrich-Alexander-Universität Erlangen-Nürnberg (Germany) in June/July 2017. This was now almost three years ago. We thank our contributors warmly for their reliability and patience throughout the publication process. We are grateful to Lisa Marani for co-organizing this workshop with us. We would also like to thank Benjamin Kiesewetter, Conor McHugh, Anne Meylan, and David Owens for contributing to the workshop by commenting on frst drafts of our papers and by presenting their most recent research in this area. We are grateful also to all the other participants at the workshop, especially our students in our (Lisa Marani’s and Sebastian Schmidt’s) graduate class on the ethics of mind in summer 2017, for their helpful comments and their interest. Special thanks to Andreas Lugauer for going over the fnal manuscript with us. Thanks also to Christian Seidel for his very helpful advice. For their generous fnancial support, we would like to thank the Thyssen-Stiftung. We would also like to thank Kevin McCain and Scott Stapleford for considering and accepting our volume for the Routledge Studies in Epistemology series. We also thank Allie Simmons, Ramachandran Vijayaragavan, Kerry Boettcher, Brian Frastaci, as well as the entire staff from Routledge who were involved in the publication of this volume for so patiently leading us throughout the editing process.

1

Introduction Towards an Ethics of Mind Sebastian Schmidt

1.

The Problem of Mental Normativity

Our attitudes – our beliefs, desires, emotions, and intentions – are not actions. They are conceived of as mental states we are in, rather than as things we actively perform. Intuitively, we are not responsible merely for being in a state – at most, we are responsible for causing a state or for failing to avoid it. And yet we respond to and evaluate our attitudes in ways which are similar to the ways in which we respond to and evaluate our actions. We think that we ought to believe in human-induced climate change, and we even consider it to be appropriate to criticize others if they fail to believe in it. A malicious desire, like the desire for another’s suffering, can rightly provoke not only disapproval, but also resentment or indignation. An emotion like anger might turn out to be unjustifed, and we might owe an apology to the person who was the target of our hostile emotion. And merely intending to become a better person is often already worthy of praise or credit.1 Indeed, it is a widely shared assumption also in contemporary epistemology and metaethics that not only our actions but also our attitudes are governed by norms, and that we are, in some sense, responsible for whether we comply with these norms. In epistemology, there is a long tradition that aims at providing a justifcation for our beliefs about an external world. Skeptical challenges – as posed by the thought experiments of the Cartesian Demon (Descartes 1641) or the Brain in a Vat (Putnam 1981) – call into doubt that we have reasons for our everyday beliefs (like your belief that you are reading this text right now). Replying to such skeptical challenges – justifying our everyday beliefs – is sometimes viewed as the central motivation for engaging in epistemology.2 It is thus not farfetched at all to conceive of epistemology as a normative discipline analogous, somehow, to ethics. How far we can take this analogy is, of course, a question open to reasonable dispute. Especially Clifford (1877) and James (1896) took the analogy so far as to speak of an ethics of belief: they discussed how moral and prudential considerations bear on the norms of belief, thereby motivating contemporary discussions about what kinds of reasons determine what we ought to believe.3

2

Sebastian Schmidt

In metaethics (broadly construed), an extensive and specialized discussion about the norms of rationality arose within the last decades. John Broome (2013) argued that being rational consists in having coherent attitudes, rather than in responding correctly to (apparent) reasons.4 According to Broome, we are, for example, rationally required to (roughly) intend those means that we believe are necessary to achieve our ends.5 This is a requirement to adjust our intentions and beliefs to each other. It is not presented as a requirement to perform certain actions in order to intentionally ensure by indirect means that we are coherent (for example, by meditation). Rather, rational requirements are meant to govern our attitudes directly. This also holds for accounts that conceive of rationality as the capacity to respond correctly to one’s reasons. For insofar as these accounts are concerned with responding to reasons for attitudes, they do not take themselves to be concerned with prudential or moral reasons for managing our attitudes.6 It seems that we do not (or cannot) weigh rational requirements governing our attitudes against, say, prudential considerations that seem to favour an attitude. Take, for example, the consideration that an eccentric billionaire promises you a lot of money as a reward for causing yourself to believe something that is unsupported by your evidence. A common view holds that in such cases, there are two normative requirements in place: a rational requirement not to have the unsupported belief, and a practical requirement to cause yourself to have it.7 Such cases are currently debated, and some of our contributions will question the picture according to which epistemic reasons, or other object-given reasons for attitudes, constitute their own normative domain that is independent from the practical realms of prudence and morality.8 Yet the general framework in which philosophers currently conceive of rational requirements is to treat them as separate from, and quite unrelated to, practical requirements: prudence and morality govern our voluntary conduct; rationality governs our attitudes. Morality and prudence concern action; rationality concerns the mind.9 However, philosophers have also recently discussed a line of thought that calls into doubt that our attitudes are the proper object of any normative evaluation whatsoever.10 Roughly, the line of thought goes as follows. Our beliefs, desires, emotions, and intentions are not subject to the will in the way our actions are. We cannot decide to believe in the way we can decide to act.11 When someone offers me a huge reward for believing, in the absence of adequate evidence, that there is an even number of stars in the universe, I will not be able to believe it. This is true even if I want to believe it because I want to have the reward. Similarly, it does not seem possible to desire, feel, or intend something merely because someone offers me a reward for doing so. It seems that we cannot believe that p for just any reason that shows it worth believing that p. It seems that we cannot adopt a mental state or an attitude for just any reason that shows it worth being adopted.12

Introduction

3

By contrast, if someone offered me a reward for performing an action for which I have no reward-independent reason, like the action of raising my arm right now, I could just do it and collect the reward. We usually can perform an action for any reason that shows the action worth doing: our actions are, as I would like to put it, exercises of voluntary control. We exercise voluntary control when we act: being able to act for reasons is a fundamental kind of freedom. Arguably, the fact that actions are free in this way is what explains why it can be true that we sometimes ought to do one thing rather than another: if raising my arm was not an exercise of voluntary control, it would not make sense to say that I ought to raise it. The rising of my arm would be a mere natural event outside the space of reasons that could not be evaluated in normative terms. In the absence of voluntary control, there would be no obligations to act, and no reasons to act in a certain way rather than in another. Maybe there would not even be such a thing as an action.13 Furthermore, it seems that we could not appropriately be blamed and praised for what we do if what we do was not an exercise of voluntary control. At least in this sense, Ought implies Can. This raises a philosophical puzzle. If being subject to norms and reasons requires voluntary control, then it becomes mysterious how anything can be subject to norms that is not under our voluntary control. How can epistemology then be concerned with the justifcation of our beliefs? And how can metaethics be concerned with normative requirements governing our beliefs and intentions? While contemporary philosophy assumes that there is mental normativity, we at the same time seem to lack an explanation of how there can be such a thing. So, how can it be that we ought to believe, desire, feel, or intend something, that we have reasons for mental states, and that we can be praised and blamed for them? Call this the problem of mental normativity.14,15

2.

Locating Our Approach: How the Ethics of Belief Motivates an Ethics of Mind

I take the following two questions as providing a helpful framework for joining the debates that I just sketched: 1. How can we be responsible for our attitudes? 2. What attitudes should we have? These are the fundamental questions of an ethics of mind.16 The frst question is concerned with explaining how there can be such a thing as an ethics of mind – how our practice of holding each other accountable for our beliefs, emotions, desires, and intentions can be justifed given that these attitudes are not under our voluntary control. Proponents in this debate either argue that attitudes are, contrary to frst impression, controlled in

4

Sebastian Schmidt

a similar way as actions are controlled (“voluntarily”);17 or they argue that we can exercise a form of direct but non-voluntary control over our attitudes, i.e. a form of direct control that is different from the control that we exercise when we act;18 or they deny that they are directly controlled in any way, and thus either deny some sense of “Ought implies Can,”19 or say that we are never directly responsible for our attitudes, but only indirectly insofar as we control our attitudes through our actions.20 By providing an explanation of our mental responsibility in this way, such accounts aim at providing a foundation for understanding the norms that underlie our practice of holding one another accountable for our attitudes.21 The second question is central to the debates about reasons for belief and about attitudinal rationality. It is concerned with the precise nature of the norms that govern our attitudes. If we can make sense of mental normativity and responsibility, then what are the standards or requirements that govern our attitudes, and what are the kinds of reasons for which we can hold attitudes, and how do we distinguish them? Are all reasons for attitudes object-given (or “of the right kind”), or are there state-given reasons (that is, “of the wrong kind”) as well?22 How do these reasons determine what we ought to believe, desire, feel, intend, etc.?23 And what is the normative force of these “mental Oughts,” which include the standards of epistemic rationality?24 This overall framework for discussing mental normativity, provided by our two main questions, is mainly inspired by contemporary debates in the ethics of belief. The philosophical debate about whether we can be responsible for what we believe and what we ought to believe is fourishing like never before. Questions which preoccupied Clifford and James are discussed by contemporary scholars under the headings “Ethics of Belief” (Feldman 2000; Matheson/Vitz 2014), “Reasons for Belief” (Reisner/Steglich-Petersen 2011), “The Aim of Belief” (Chan 2013), and “Epistemic Norms” (Littlejohn/Turri 2014). Central questions include: • • • • •

•

Is only evidence for or against p relevant for whether one ought to believe that p, or are there state-given reasons which favour having a belief, e.g., facts about how valuable it would be to believe that p?25 If only evidence is relevant to this question – can this fact be explained by conceptual features of belief, like belief’s “aim of truth”?26 Can we weigh epistemic and practical reasons against each other so as to determine what we ought to believe, all things considered?27 Is it even possible to believe for practical reasons?28 Do we have any meaningful control over what we believe? Given that we do not control our beliefs as we control our actions (“voluntarily”), how can we explain our responsibility for believing, especially given that we are often subject to the criticism of epistemic irrationality?29 How do we have to interpret epistemic norms, and especially their (apparent) normative force?30

Introduction

5

Discussions surrounding the ethics of belief are partly an upshot of what is often mentioned as the practical turn in contemporary epistemology. Early indicators for this movement were virtue epistemology and debates about pragmatic encroachment with respect to the analysis of knowledge.31 Currently, the practical turn is shaped not only by the ethics of belief, but also, for example, by discussions about the value of knowledge and true belief,32 as well as by debates in social epistemology,33 including feminist epistemology.34 The present volume is in line with this practical turn in that its frst part addresses the ethics of belief. There we discuss topics that currently shape and even broaden the scope of the practical turn, like control over belief and its implication for the ethics of belief (Lindner), the role of the epistemic community for the possibility of epistemic normativity (Chrisman), responsibility for believing and doxastic strength (Gaultier), the doxastic norms within relationships, and whether they require us to be doxastically partial towards, say, our friends (Crawford), the structure, content, and force of epistemic norms (Skipper/Steglich-Petersen), as well as practical reasons for suspending judgment (Lord). We summarize the contributions of Lindner, Chrisman, and Gaultier under the title “Doxastic Agency and Responsibility,” and the contributions of Crawford, Steglich-Petersen/Skipper, and Lord under “Reasons for Belief.” While the frst debate is more concerned with the possibility of an ethics of belief (compare question 1 of the ethics of mind), the latter debate is concerned with the nature and structure of an ethics of belief (compare question 2). The aim of this frst part is to bring both debates closer to each other, especially because the two broad topics they discuss are connected in important ways: how we explain the possibility of epistemic normativity will shape our understanding of epistemic norms. For example, if we argue that belief is subject to a robust form of direct control, then a pragmatist account about what we ought to believe might seem more plausible.35 By contrast, if we are doxastic involuntarists, i.e., if we think that belief is not directly controlled at all but nevertheless subject to a robust form of normativity, then an evidentialist picture might seem more attractive.36 Our volume, however, also reaches beyond epistemology, and even beyond the ethics of belief. Our second part is introduced by two contributions that frame the ethics of mind more generally. The frst subpart discusses the frst main question of an ethics of mind: how can we be responsible for our attitudes? (Schmidt) The second one is concerned with an aspect of the second main question: how should we understand theoretical and practical rationality? (Ernst) We summarize these two contributions under the label “Responsibility, Reasons, and Rationality.” Especially in the second subpart, our focus will shift from the practical dimensions of belief to issues revolving around reasons, responsibility, and agency with respect to other mental states or attitudes – especially

6

Sebastian Schmidt

the ethics of blame (Coates), the ethics of fear in public contexts (Döring), the agency we exercise in deciding (Soteriou), our responsibility for our passing thoughts (Tognazzini), and the limits of our mental responsibility when it comes to our fantasies and mind-wanderings (McCormick). We summarize these contributions under the title “The Ethics of Blame, Fear, Decision, Passing Thought, and Phantasy.” Both our subparts illustrate how we might approach an ethics of mind in contemporary philosophy: either “top-down” by deriving insights relevant for mental normativity from already existing debates (frst subpart), or “bottom-up” by considering the normative dimensions of specifc mental phenomena and thus providing the building blocks for an ethics of mind (second subpart). In this second main part, some of our contributions are not (or not primarily) contributions to epistemology, even if epistemology is understood broadly as the study of the normativity of belief. Rather, they are more generally contributions to an ethics of mind, or to the discussion about the rationality of attitudes, or about what mental states or attitude we ought to have. Here our outlook is motivated both by general debates about the nature and normativity of rationality, where the focus is commonly not only on belief but also on intention,37 by debates about the rationality of emotion,38 as well as by recent research on, more generally, responsibility for attitudes.39 Epistemologists can proft from research in these areas because there are parallels between the normativity of belief and the normativity of other attitudes – most strikingly the distinction between two kinds of (purported) reasons for attitudes: object-given and state-given, or “right kind” and “wrong kind” of reasons (on references, see note 22). Turning their attention beyond belief and towards attitudes in general will provide epistemologists with fresh arguments and ideas they would otherwise remain oblivious to. We discuss the most current ideas revolving around the normativity of mental states and especially the normativity of belief in the frst part. By focusing primarily on belief normativity, we hope to draw some insightful lessons from a fourishing debate which can then be used for working on the normativity of mental states more generally. Going beyond the practical turn in epistemology, our volume reaches towards an ethics of mind.

3.

The Contributions

Our opening contribution in Chapter 2 by Martina Lindner connects the locus classicus of the ethics of belief – W. K. Clifford’s paper – with the contemporary discussion about doxastic control. Clifford argues for the evidentialist principle that one should never believe upon insuffcient evidence. His position differs from most contemporary evidentialist approaches in that he takes this principle to be a moral norm backed by a consequentialist justifcation. It is a widely held view that Clifford fails to establish his ethics of belief – not least because, as critics argue,

Introduction

7

a consequentialist moral framework commits one to an ethics of belief that is (more or less radically) different from Clifford’s position. Lindner shows that examining the question whether Clifford or these critics are right eventually amounts to an empirical investigation; however, she also argues that this philosophical dead end can be avoided by drawing on results from the philosophical debate on doxastic voluntarism. Even though the question can thus not ultimately be answered, Lindner claims that Clifford’s ethics of belief fares better than his critics’ positions with regard to the results from this debate. Furthermore, the widely accepted moderate critical position does not fare much better than the radical one. Chapter 3 by Matthew Chrisman then proposes a specifc account of how doxastic control can explain our responsibility for belief and how this motivates a social approach towards epistemological issues. In the attempt to understand the norms governing believers, epistemologists have tended to focus on individual belief as the primary object of epistemic evaluation. However, norm governance is often assumed to concern, at base, things we can do as a free exercise or manifestation of our agency. Yet believing is not plausibly conceived as something we freely do but rather as a state we are in, usually as the mostly automatic or involuntary result of cognitive processes shaped by nature, bias, and ideology. In this chapter, Chrisman sketches a response to this tension. This response is based on rejecting the traditional theoretical focus on an individual’s particular beliefs as the primary object of normative epistemic evaluation. If we shift our focus from the particular beliefs of individuals to the community and its information managing practices, we may lessen the tension between norm governance and automaticity and involuntariness in a way that construes autonomous cognitive agency as a resultant of rather than a precondition for our norm-governed epistemic sociality. In Chapter 4, Benoit Gaultier then concludes our frst subpart by approaching the issues of control and normativity. He is asking how it is possible for deontic evaluations of beliefs to be appropriate if we do not have voluntary control over our beliefs. Gaultier argues that we should reject the claim that we can have indirect control over beliefs in virtue of the basic voluntary control we have over our actions. We have another kind of indirect control over beliefs: we can demonstrate doxastic strength or, on the contrary, doxastic weakness when forming our beliefs. That is, we can resist or, on the contrary, fail to resist the infuence of some of our conative attitudes. And in the same way that we take our actions to be open to blame when they result from having demonstrated weakness of will even though this does not consist in doing something at will, we take our beliefs to be open to blame and hence subject to deontic evaluation, when they result from having demonstrated doxastic weakness even though this does not consist in doing something at will. Chapter 5 by Lindsay Crawford opens our part on reasons for belief. The central dispute between evidentialists and pragmatists about reasons

8

Sebastian Schmidt

for belief concerns whether or not non-evidential considerations can be reasons for belief. In recent work, some pragmatists about reasons for belief have made their case for pragmatism by appealing, in part, to a broad range of cases in which facts about one’s relationships with signifcant others (friends, romantic partners, and the like) appear to give one non-evidential reasons to have beliefs skewed in their favor. Crawford explores whether and how such relationships might provide non-evidential reasons for belief. After frst making the case against the evidentialist’s claim that there cannot be non-evidential reasons for belief, she turns to examine different accounts of how our personal relationships might ground reasons for doxastic partiality. These are accounts on which reasons for doxastic partiality are grounded in facts about what attitudes are partly constitutive of being a good signifcant other to someone, and accounts on which reasons for doxastic partiality are grounded in reasons, more broadly, to beneft one’s signifcant others. Crawford argues that our personal relationships do not provide non-evidential reasons for belief, even if the pragmatist is right that there can be non-evidential reasons for belief. In the following Chapter 6, Asbjørn Steglich-Petersen and Mattias Skipper defend epistemic instrumentalism against a challenge from alleged asymmetries between epistemic and practical reasons. Epistemic instrumentalists think that epistemic normativity is just a special kind of instrumental normativity. According to them, you have epistemic reason to believe a proposition insofar as doing so is conducive to certain epistemic goals or aims – say, to believe what is true and avoid believing what is false. Perhaps the most prominent challenge for instrumentalists in recent years has been to explain, or explain away, why one’s epistemic reasons often do not seem to depend on one’s aims. This challenge can arguably be met. But a different challenge looms: instrumental reasons in the practical domain have various properties that epistemic reasons do not seem to share. In their contribution, Skipper and Steglich-Petersen offer a way for epistemic instrumentalists to overcome this challenge. Their main thesis takes the form of a conditional: if we accept an independently plausible transmission principle of instrumental normativity, we can maintain that epistemic reasons in fact do share the relevant properties of practical instrumental reasons. In addition, we can explain why epistemic reasons seem to lack these properties in the frst place: some properties of epistemic reasons are elusive, or easy to overlook, because we tend to think and talk about epistemic reasons in an “elliptical” manner. Our fnal contribution on doxastic normativity in Chapter 7 explores the boundaries of epistemic normativity. There Errol Lord argues that we can understand these better by thinking about which mental states are competitors in rationality’s competition. He argues that belief, disbelief, and two kinds of suspension of judgment are competitors. Lord shows

Introduction

9

that there are non-evidential reasons for suspension of judgment. One upshot is an independent motivation for a certain sort of pragmatist view of epistemic rationality. Our second main part on the ethics of mind more broadly opens with my own proposal for explaining responsibility for attitudes in Chapter 8. There I argue that the problem of responsibility for attitudes is best understood as a puzzle about how we are responsible for responding to our object-given reasons for attitudes – i.e., how we are responsible for being (ir)rational. The problem can be solved, I propose, by understanding the normative force of reasons for attitudes in terms of blameworthiness. I present a puzzle about the existence of epistemic and mental blame which poses a challenge for the very idea of reasons for attitudes. We are left with three options: denying that there are any reasons for attitudes, opting for pragmatism about reasons for attitudes, or arguing that the challenge rests on a misunderstanding of the normative force of reasons for attitudes. I fnally suggest a version of the last strategy. We can understand the normative force of reasons for attitudes, and thereby solve the problem of mental responsibility, by acknowledging that the way we blame each other for irrationality is different from the way we blame each other when we failed to respond correctly to reasons for action. Gerhard Ernst’s Chapter 9 then tries to clarify the nature of rationality. He does this by distinguishing two fundamentally different kinds of rationality: rationality in the “adjustment-sense” and rationality in the “evaluation-sense.” A person is rational in the adjustment-sense if her mental states are well adjusted to each other, i.e. if her beliefs, emotions, and intentions ft together (in a sense Ernst explains); a person is rational in the evaluation-sense if she has evaluative beliefs which are adequate on the basis of her non-evaluative beliefs. Our “rational response system,” as Ernst calls it, is concerned with keeping our mind unifed by adjusting our mental states to each other and with evaluating what we believe. On the basis of the distinction between rationality in the adjustment-sense and rationality in the evaluation-sense one can understand what is right about the view that rationality consists in responding adequately to reasons and what is right about the view that rationality consists in being consistent (broadly understood). Our fnal subpart on the normativity of specifc attitudes is opened by Justin Coates’ contribution on the ethics of blame in Chapter 10. It is widely held that if an agent is not morally responsible for her action – i.e., if she is not deserving of blame – then we have a (decisive) reason to refrain from blaming her. But though this is true, the fact that someone is deserving of blame isn’t clearly suffcient for there to be most allthings-considered reason for blaming that person. Other considerations bear on this question as well. Coates offers an account of some of these considerations – particularly those that can serve as deontic constraints on blame. He also offers a reply to those skeptical of the “ethics of

10

Sebastian Schmidt

blame” on the grounds that such theorizing invariably appeals to the “wrong kind of reasons.” We then turn to the ethics of emotions in Chapter 11 – more specifcally, to Sabine Döring’s account of the ethics of fear in public contexts. The question is why it is that objective safety level and subjective feeling of safety may come apart. Answering this question requires an analysis of the nature of fear in the public sphere since feeling safe means to feel that one avoids the frightening, i.e. the threats or dangers that one perceives in the world. Döring argues that the fact-resistance fear might display in the public sphere is due to the characteristic function that fear fulflls in this sphere. In the public sphere, fear is typically embodied by “fear narratives,” i.e. by real-life structuring stories of local groups which thereby respond to feelings of powerless fear and insecurity. By specifying what the threat is and how to deal with it in the right way, fear narratives reduce felt insecurity. Because of this and by thus enabling a feeling of superiority which may even turn the fear narrative into a narrative of enthusiastic heroism, fear in the public sphere does itself satisfy subjective preferences. This is why objective safety level and subjective feeling of safety may come apart. The question is how we could and should handle this. In Chapter 12, Matthew Soteriou considers the kind of agency, and thus responsibility, that is involved in deciding to act, rather than in acting itself. One of his aims is to trace out connections between the notion that we occupy a tensed temporal perspective from which we regard the future as open, and the notion that we occupy a deliberative standpoint from which we act under the idea of freedom. A further aim is to suggest that identifying connections between the psychology of self-determination and our temporal psychology should be central to an account of the sort of mental agency we exercise in deciding to act, and hence central to an account of what makes us responsible for our decisions. After having considered ways in which aspects of the psychology of our temporal perspective may contribute to explaining our capacity to exercise agency in deciding to act, Soteriou considers how our capacity to make decisions can contribute to explaining some of the distinctive features of the temporal perspective on the future that we occupy. In particular, he considers and responds to the suggestion that when it comes to our perspective on our future, there is a potential tension between the standpoint of theoretical reason and the standpoint of practical reason. Neal Tognazzini then turns to how we evaluate and judge one another for our passing thoughts in Chapter 13. He explores the concept of judgmentalism: what it is and why it’s morally problematic. After criticizing an account offered by Gary Watson, Tognazzini argues for a broader understanding of what it is to be judgmental, encompassing not just the overall beliefs that we form about someone else, but also the very pattern of our thoughts about those with whom we are involved in interpersonal

Introduction

11

relationships. The thesis is that to care about someone is to be oriented towards them, or to see them through a particular mental lens, in a way that produces a particular pattern of salience and silence. That is: caring about someone (at least ideally) has the effect of making some features of that person particularly salient, and silencing or screening off other features from one’s consciousness. One is aptly described as judgmental when one’s thoughts do not display this sort of pattern, indicating a failure to fully adopt the orientation that constitutes properly caring about the person. Our fnal Chapter 14 by Miriam Schleifer McCormick delineates the limits, or at least one limit, of the ethics of mind. Many theorists, including McCormick herself, have argued that some states of mind are appropriate targets of certain reactive attitudes even if they cannot be directly controlled. McCormick now worries that the scope of agency can be widened too far so that no area of mind is beyond the reach of appropriate assessment and judgement. She begins with the intuition that there is, or ought to be, a domain of the mind that is completely free of normative assessment, where you are safe to let your thoughts and images go wherever they take you without concern that you are doing anything wrong, where praise and shame do not apply. McCormick begins by offering an example of the kind of state she thinks should be beyond normative judgment; she argues that certain kinds of wakeful fantasies are on par with sleeping dreams. If one shares McCormick’s view that there is a “free” domain of the wakeful mind, then what she is doing can be seen as clarifying why such states exempt them from judgment. If one does not share this intuition, then what McCormick is doing can be seen as specifying what criteria would be needed for a kind of state (or domain) to be free in this sense. And then some may argue that no wakeful fantasies satisfy these criteria. McCormick addresses those arguments and argues that if the fantasies as characterized are appropriate targets of normative assessment, then it will be very diffcult to exempt dreams of sleep, as well as other exercises of imagination. Of course, some people (like Augustine and surprisingly many others) will not mind this result. McCormick doesn’t think then that is the end of the discussion, stalemate and parting of intuitions, for she argues that a case can be made for the value of having a realm of imagination that is beyond the reach of any kind of judgment.

4.

Outlook

These last papers which were concerned with the normativity of specifc attitudes (Chapters 10–14) can be read as exploring various facets of an ethics of mind. They might thus provide building blocks for a more encompassing theory of mental normativity that has yet to be spelled out. Our contributions on the possibility of an ethics of mind (Chapters

12

Sebastian Schmidt

3, 4, and 8) as well as our chapters on what attitudes we ought to have, that is, on issues surrounding reasons and rationality (Chapters 2, 5, 6, 7, 8, and 9), are meant to inspire such an ethics of mind also on a more abstract level – i.e., without focusing on any specifc attitude. We hope that our volume motivates such approaches and brings epistemologists, (meta)ethicists, and philosophers of mind to a fruitful exchange.40

Notes 1. The reactions might come close to or might be sometimes identical with Strawson’s (1962) reactive attitudes (resentment, indignation, guilt, gratitude). 2. See, for instance, Williams (1999: 35). For a good reader about contemporary classics on skepticism about the external world, see DeRose/Warfeld (1999). 3. While evidentialists deny that anything besides evidential considerations – i.e., facts indicating the truth of a proposition – can be reasons for belief, pragmatists argue that also practical considerations – i.e., prudential or moral considerations that count in favour of holding a specifc belief – can count as reasons for belief and thus determine what we ought to believe. For some infuential cases for evidentialism, see Adler (2002), Kelly (2002), Shah (2006), Velleman (2000a), and Wedgwood (2002). Recent pragmatists include McCormick (2015), Reisner (2009), and Rinard (2015, 2017, 2019a, 2019b), as well as our contributors Crawford, Steglich-Petersen/Skipper, and Lord (see Chapters 5–7 in this volume; esp. Lord makes an extensive case for a version of pragmatism when it comes to rationally suspending judgment). It is noteworthy that some epistemologists think that not only reasons are relevant for determining what we ought to believe: pragmatic encroachers (see note 31 following for references) argue that what counts as suffcient evidence is determined by the practical stakes at play in believing; not all of them would accept that the practical stakes can enter as reasons to believe (cf. also the distinction between thresholds partialism and reasons partialism in Crawford’s contribution, Chapter 5 of this volume, p. 92). 4. The account of rationality as responding correctly to one’s apparent reasons – i.e., the reasons one would have if one’s beliefs were true – is associated with Parft (2001, 2011) and was further spelled out by Schroeder (2009) (for instrumental rationality) and Way (2009, 2010). For a proposal of how to reconcile rationality as coherence and rationality as responding correctly to reasons, cf. the contribution by Ernst (Chapter 9). 5. “Rationality requires of us that, if we intend an end and believe some means is necessary to that end, then we intend the means. This has to be a requirement of rationality. A person is necessarily irrational if she does not intend whatever she believes is a necessary means to an end she intends” (Broome 2005: 2). 6. See Kiesewetter (2017) and Lord (2018), who argue that no formulation of the coherentist requirements of rationality can make it intelligible why we ought to be rational or why we have reason to be rational; see also Kolodny (2005), who motivated their accounts. Kiesewetter and Lord instead propose that we should understand rationality as responding correctly to our (available) reasons, i.e., the reasons that are epistemically accessible or possessed by us. Note that these accounts are quite different from the accounts of rationality as responding correctly to apparent reasons mentioned in note 4 earlier. I present an argument in favour of rationality as reasons-responsiveness in Schmidt (forthcoming) as response to Broome (forthcoming).

Introduction

13

7. Wedgwood (2017: 41–52) argues that in such cases, there are two independent requirements at play: although we ought to believe what is prudentially best according to the “wrong kind of reasons,” we ought to believe what is epistemically rational to believe according to the “right kind of reasons.” There is no question, according to Wedgwood, what to believe “all things considered.” See Meylan (ms) and Rinard (2017, 2019a, 2019b) as well as Crawford (this volume, Chapter 5, Section 2) for some disagreement. I argue in Schmidt (2017) that there is a sense in which we “ought to believe” what is best to believe in the same way as we ought to do what is best to do. See also the contribution by Coates (Chapter 10, Section 3) for some arguments that the “all-things-considered” question is meaningful, especially when it comes to when we ought to blame. 8. Martina Lindner, Matthew Chrisman, as well as Benoit Gaultier all discuss the ethics of belief as being concerned with how we infuence, manage, or control our beliefs through prior activity (see Chapters 2–4), though Lindner ultimately views the control that we have over our beliefs as being very limited. I argue in my contribution (Chapter 8) that we need to keep rational and moral criticism distinct in order to understand responsibility for attitudes. 9. See McCormick (2015: 1–3) for how contemporary philosophers see a disparity between norms of belief and norms of action, so that belief requires its own kind of ethics (see esp. Adler 2002). 10. In the doxastic realm, this line of thought is often traced back to Williams (1970) and Alston (1988). Adams (1985) is often cited as one of the frst ones to defend our responsibility for attitudes more generally. 11. For some arguments see Bennett (1990), Hieronymi (2006), Schmidt (2016), Scott-Kakures (1994), and Williams (1970). A famous defense of the possibility of deciding to believe is Ginet (2001). McCormick (2015, ch. 6) argues that our beliefs are sometimes subject to the same kind of control as our actions are (guidance control). 12. Cf. Hieronymi’s defnition of voluntariness (2006: 45–49), which I adopt here. She is following Bennett’s (1990) defnition. Prima facie, a lack of voluntary control might be controversial in the case of intentions. See Kavka (1983), however, for why intentions are as involuntary as any other mental state or attitude. See Soteriou (this volume, Chapter 12) on how we can be said to control our decisions (which might be construed as formings of intentions). 13. However, John Hyman (2015) argues that even inanimate things, like the sun, can act (it is shining). I reserve “action” or “to act” here for full-blooded agency that is intentional under a description. 14. I use “mental normativity” here as referring to the phenomenon of normative standards governing our attitudes. One might also wish to point out that there are mental actions that are subject to normative standards. Mental actions, however, do not give rise to the problem of mental normativity because they are, per defnition, under our voluntary control. 15. On more detailed versions of this problem, see our contributions by Chrisman (Chapter 3, Section 2) and Schmidt (Chapter 8, Section 3). Most contemporary philosophers phrase the puzzle in terms of responsibility: How can we be responsible for our mental states in the absence of voluntary control over them? McHugh (2017) points out that the puzzle arises for norms and reasons if we assume some kind of “Ought implies Can” principle. See also Alston (1988) and Owens (2000) for the puzzle in terms of norms and reasons. Somewhat oversimplifying, I will here view evaluations expressed in blame, praise, or various forms of reactive and moral attitudes (resentment,

14

16. 17.

18. 19.

20.

21.

22.

23.

24.

Sebastian Schmidt indignation, gratitude, guilt) as one form of normative evaluation among others which seem to presuppose voluntary control. Adams already spoke of an “ethics of motives, and more generally of states of mind” (1985: 12). See esp. Ginet (2001), but also McCormick (2015), who refrains from calling the kind of control she has in mind for both action and belief “voluntary control.” She endorses an account for responsibility for hope that is along the lines of her account of doxastic control in McCormick (2017). See, among others, Boyle (2011), Hieronymi (2006, 2008, 2009, 2014, ms), Montmarquet (1992), Raz (2011, esp. his ch. 5 and part III), and Smith (2005). See especially Adams (1985) and Owens (2000, 2017). Owens only denies “Ought implies Can” for beliefs and emotions, but not for intentions. According to Owens, we are responsible for our intentions in virtue of having refective control over them – a control we also exercise over actions, but which is different from voluntary control nevertheless. Direct responsibility for belief has most explicitly been denied by Meylan (2013, 2017) and Peels (2017). Also Chrisman’s account (2008, 2018, this volume) emphasizes forms of indirect control, but he distances himself from accounts that view doxastic agency as too external to belief. In this volume, Gaultier (Chapter 4) argues that doxastic responsibility is to be explained in terms of indirect control. However, his position is original in that it denies that the relevant form of indirect control is voluntary control. This is the idea of Meylan (2013), titled Foundations of an Ethics of Belief. There are of course positions that do not neatly ft into one of the categories I present here. Next to Gaultier’s position in this volume (cf. last note), the position of Graham (2014) is to be mentioned, who thinks that we are never directly responsible for our actions, but only for our attitudes. He does not explicitly ground this direct responsibility in control and thus most likely belongs to the same category as Adams and Owens (see note 19). For a discussion of the distinction between object-given and state-given (or right-kind and wrong-kind) reasons, see Gertken/Kiesewetter (2017). The labels “right kind” and “wrong kind” of reasons have their origin in a debate about ftting attitude accounts of value (see Rabinowicz/Rønnow-Rasmussen 2004). For some doubts about whether this distinction is applicable to intentions, see Heuer (2018). See Schmidt (Chapter 8 of this volume, Section 2) for some initial characterization of the two kinds of reasons. For proposals of how to weigh practical reasons for belief against epistemic reasons, see Reisner (2008, ms) and Steglich-Petersen/Skipper (forthcoming). McCormick (2017) proposes how to weigh different kinds of reasons for hope with one another. For a critical voice on weighing both kinds of reasons, see, e.g., Papineau (2013: 70). There has been previous work on the kinds of norms that govern blaming-attitudes (see Coates/Tognazzini 2013; Coates 2016). In this volume, Coates pursues the ethics of blame further (Chapter 10). Recently, there have been arguments against the “normativity” of epistemic reasons, see Papineau (2013), Rinard (2015), Glüer/Wikforss (2018), Maguire/ Woods (forthcoming, ms), and Mantel (2019); for a critical discussion, see Paakkunainen (2018). Such arguments can be understood as questioning the normative force of (what we deem to be) reasons for belief. Kiesewetter (ms) defends the normativity of epistemic reasons. The worry that epistemic reasons are not normative partly stems from a general worry with considering object-given reasons as normative reasons (see Maguire 2018 for some doubts about the existence of reasons for affective attitudes; see my contribution in Chapter 8 of this volume for a discussion of an argument against

Introduction

25.

26. 27. 28.

29. 30.

31. 32. 33.

34.

35. 36. 37. 38.

15

the existence of object-given reasons more generally). For a recent positive proposal about the signifcance of evidence, and thus of epistemic normativity, see Mitova (2017, ch. 10), whose account is motivated by Velleman’s (2000b, 2009). We might want to distinguish between two questions here (for this distinction, see also Crawford, Chapter 5, p. 92): First, do pragmatic factors infuence the threshold of when to count a belief as epistemically justifed or rational? (See Fantl/McGrath 2002; Kim/McGrath 2019). Secondly, are practical considerations indicating a belief’s value reasons for belief? (See note 3 earlier). See esp. Velleman (2000a), Owens (2003), and Steglich-Petersen (2006, 2009, 2013). On weighing different kinds of reasons, see note 23 earlier. Ginet (2001), McCormick (2015, esp. chs. 3 and 6), Reisner (2008, 2009), and Rinard (2015, 2017, 2019a, 2019b) think that it is possible to believe for practical reasons; Bennett (1990), Hieronymi (2006), Schmidt (2016), ScottKakures (1994), and Williams (1970), among others, think that it is not. The fourth and ffth question are often seen as interrelated, for many authors argue that we are responsible for beliefs insofar as we are able to control them. See notes 17, 18, and 20 earlier. As pointed out by, for example, Raz (2011: 41–45), epistemic reasons are not related to value in the way practical reasons seem to be. If we do not necessarily realize any value by following our epistemic reasons, how can these reasons possess normative force? See note 24 on the present discussion about the normativity of epistemic reasons. Steglich-Petersen advances an understanding of epistemic reasons as instrumental reasons, which might safeguard their normativity and at the same time preserve their relation to value (see esp. Steglich-Petersen 2011, Skipper/Steglich-Petersen this volume). On virtue epistemology, see esp. Zagzebski (1996) and Sosa (2007/2009). On pragmatic encroachment, see esp. Fantl/McGrath (2002), Hawthorne (2003), and Kim/McGrath (2019). See, e.g., the volume of Pritchard et al. (2010). One topic currently under discussion in social epistemology is, next to discussions about power relations (see next note), peer disagreement (see Frances 2014 for an introduction). On social epistemology more broadly, see, e.g., Craig (1990), Gilbert (1989), Goldman (1999), Haddock et al. (2010), and Matheson/Vitz (2014). On feminist epistemology as a branch of social epistemology, see Anderson (1995). See also the work of Fricker (2007) on power relations in the epistemic realm. For a more recent collection, see Grasswick (2011). See also Grasswick (2014) for a recent discussion of epistemic norms in the context of feminist epistemology. For this sort of connection between the debates, see esp. how the two parts of McCormick (2015) are set up. See, e.g., the account of Adler (2002). See notes 4–6 for references. See esp. de Sousa (1987). Issues under discussion include whether emotions can make it rational to act against one’s best judgment (Arpaly 2000), or whether recalcitrant emotions, i.e., emotions which persist in the face of one’s judging that they are not appropriate, can be rational (Roberts 2003, 91–93; Brady 2007, 2009; Tappolet 2012; 2016, 31–45), and, more generally, discussions about cases in which our emotions confict with our judgment (Helm 2001, ch. 5; Döring 2009, 2010). On the evaluative dimensions of emotions, see also Price (2015, chs. 6 and 7).

16

Sebastian Schmidt

39. Here especially the works of Pamela Hieronymi (2006, 2008, 2014) and Angela Smith (2005, 2015a, 2015b) on responsibility for mental states are to be mentioned. Adams (1985) became a classical point of reference for this discussion. Furthermore, various authors argued that we can neither choose our emotions (Oakley 1992, 130–132), nor our intentions (Kavka 1983; Owens 2000, 81–82; Hieronymi 2006) at will. 40. I am grateful to Gerhard Ernst and Marie van Loon for very helpful comments.

References Adams, Robert M. (1985): “Involuntary Sins”, The Philosophical Review 94, 3–31. Adler, Jonathan (2002): Belief’s Own Ethics, Cambridge, MA: MIT Press. Alston, William P. (1988): “The Deontological Conception of Epistemic Justifcation”, Philosophical Perspectives 2, 257–299. Anderson, Elizabeth (1995): “Feminist Epistemology: An Interpretation and a Defense”, Philosophical Topics 23, 27–58. Arpaly, Nomy (2000): “On Acting Rationally Against One’s Best Judgment”, Ethics 110, 488–513. Bennett, Jonathan (1990): “Why Is Belief Involuntary?”, Analysis 50, 87–107. Boyle, Matthew (2011): “‘Making Up Your Mind’ and the Activity of Reason”, Philosopher’s Imprint 11, 1–24. Brady, Michael S. (2007): “Recalcitrant Emotions and Visual Illusions”, American Philosophical Quarterly 44, 273–284. Brady, Michael S. (2009): “The Irrationality of Recalcitrant Emotions”, Philosophical Studies, 145, 413–430. Broome, John (2005): “Have We Reason to Do as Rationality Requires? A Comment on Raz”, Journal of Ethics and Social Philosophy 1, 1–8. Broome, John (2013): Rationality Through Reasoning, Chichester: Blackwell. Broome, John (forthcoming): “Rationality versus Normativity”, Australasian Philosophical Review 5. Chan, Timothy (ed.) (2013): The Aim of Belief, New York: Oxford University Press. Chrisman, Matthew (2008): “Ought to Believe”, The Journal of Philosophy 105, 346–370. Chrisman, Matthew (2018): “Epistemic Normativity and Cognitive Agency”, Noûs 52, 508–529. Clifford, William K. (1877): “The Ethics of Belief”, in: The Ethics of Belief: And Other Essays, New York: Prometheus Books 1999, 70–96. Coates, D. Justin (2016): “The Epistemic Norm of Blame”, Ethical Theory and Moral Practice 19, 457–473. Coates, D. Justin/Tognazzini, Neal (eds.) (2013): Blame: Its Nature and Norms, New York: Oxford University Press. Craig, Edward (1990): Knowledge and the State of Nature, Oxford: Clarendon Press. DeRose, Keith/Warfeld, Ted (1999): “Introduction: Responding to Skepticism”, in: Skepticism: A Contemporary Reader, ed. by DeRose/Warfeld, London: Routledge, 1–26. Descartes, René (1641): Meditations on First Philosophy, Indianapolis: Hackett Publishing 1993.

Introduction

17

De Sousa, Ronald (1987): The Rationality of Emotion, Cambridge, MA: MIT Press. Döring, Sabine (2009): “The Logic of Emotional Experience: Non-Inferentiality and the Problem of Confict without Contradiction”, Emotion Review 1, 240–247. Döring, Sabine (2010): “Why Be Emotional?”, in: The Oxford Handbook of Philosophy of Emotion, ed. by Goldie, New York: Oxford University Press, 283–301. Fantl, Jeremy/McGrath, Matthew (2002): “Evidence, Pragmatics, and Justifcation”, Philosophical Review 111, 67–94. Feldman, Richard (2000): “The Ethics of Belief”, Philosophy and Phenomenological Research 60, 667–695. Frances, Bryan (2014): Disagreement, Cambridge: Polity Press. Fricker, Miranda (2007): Epistemic Injustice: Power and the Ethics of Knowing, New York: Oxford University Press. Gertken, Jan/Kiesewetter, Benjamin (2017): “The Right and the Wrong Kind of Reasons”, Philosophy Compass 12. Gilbert, Margaret (1989): On Social Facts, London: Routledge. Ginet, Carl (2001): “Deciding to Believe”, in: Knowledge, Truth, and Duty: Essays on Epistemic Justifcation, Responsibility, and Virtue, ed. by Steup, New York: Oxford University Press, 63–75. Glüer, Kathrin/Wikforss, Åsa (2018): “Reasons for Belief and Normativity”, in: The Oxford Handbook of Reasons and Normativity, ed. by Star, New York: Oxford University Press, 575–599. Goldman, Alvin I. (1999): Knowledge in a Social World, New York: Oxford University Press. Graham, Peter A. (2014): “A Sketch of a Theory of Moral Blameworthiness”, Philosophy and Phenomenological Research 88, 388–409. Grasswick, Heidi (ed.) (2011): Feminist Epistemology and Philosophy of Science: Power in Knowledge, Dordrecht: Springer, 133–156. Grasswick, Heidi (2014): “Understanding Epistemic Normativity in Feminist Epistemology”, in: The Ethics of Belief: Individual and Social, ed. by Matheson/Vitz, New York: Oxford University Press, 216–243. Haddock, et al. (eds.) (2010): Social Epistemology, New York: Oxford University Press. Hawthorne, John (2003): Knowledge and Lotteries, New York: Oxford University Press. Helm, Bennett W. (2001): Emotional Reason: Deliberation, Motivation, and the Nature of Value, Cambridge: Cambridge University Press. Heuer, Ulrike (2018): “The Relevance of the Wrong Kind of Reasons”, in: Normativity: Epistemic and Practical, ed. by McHugh/Way, New York: Oxford University Press, 47–66. Hieronymi, Pamela (2006): “Controlling Attitudes”, Pacifc Philosophical Quarterly 87, 45–74. Hieronymi, Pamela (2008): “Responsibility for Believing”, Synthese 161, 357–373. Hieronymi, Pamela (2009): “Two Kinds of Agency”, in: Mental Actions, ed. by O’Brian/Soteriou, New York: Oxford University Press, 138–168. Hieronymi, Pamela (2014): “Refection and Responsibility”, Philosophy and Public Affairs 42, 3–41.

18

Sebastian Schmidt

Hieronymi, Pamela (ms): Minds That Matter, unpublished manuscript. Hyman, John (2015): Action, Knowledge, and Will, New York: Oxford University Press. James, William (1896): “The Will to Believe”, in: The Will to Believe: And Other Essays in Popular Philosophy, and Human Immortality, Dover: Dover Publishing 1956, 1–31. Kavka, Gregory S. (1983): “The Toxin Puzzle”, Analysis 43, 33–36. Kelly, Thomas (2002): “The Rationality of Belief and Some Other Propositional Attitudes”, Philosophical Studies 110, 163–196. Kiesewetter, Benjamin (2017): The Normativity of Rationality, New York: Oxford University Press. Kiesewetter, Benjamin (ms): “Are Epistemic Reasons Normative?”, unpublished manuscript. Kim, Brian/McGrath, Matthew (2019): Pragmatic Encroachment in Epistemology, London: Routledge. Kolodny, Niko (2005): “Why Be Rational?”, Mind 114, 509–563. Littlejohn, Clayton/Turri, John (eds.) (2014): Epistemic Norms: New Essays on Action, Belief, and Assertion, New York: Oxford University Press. Lord, Errol (2018): The Importance of Being Rational, New York: Oxford University Press. Maguire, Barry (2018): “There Are No Reasons for Affective Attitudes”, Mind 127, 779–805. Maguire, Barry/Woods, Jack (forthcoming): “The Game of Belief”, The Philosophical Review. Maguire, Barry/Woods, Jack (ms): “All Genuine Reasons for Belief Are the Wrong Kind of Reasons for Belief”, unpublished manuscript, April 2018 version. Mantel, Susanne (2019): “Do Epistemic Reasons Bear on the Ought Simpliciter?”, Philosophical Issues, 214–227. Matheson, Jonathan/Vitz, Rico (2014): The Ethics of Belief: Individual and Social, New York: Oxford University Press. McCormick, Miriam S. (2015): Believing against the Evidence: Agency and the Ethics of Belief, London: Routledge. McCormick, Miriam S. (2017): “Rational Hope”, Philosophical Explorations 20, 127–141. McHugh, Conor (2017): “Attitudinal Control”, Synthese 194, 2745–2762. Meylan, Anne (2013): Foundations of an Ethics of Belief, Frankfurt a.M.: Ontos. Meylan, Anne (2017): “The Consequential Conception of Doxastic Responsibility”, Theoria 83, 4–28. Meylan, Anne (ms): “Doxastic Divergence and the Problem of Comparability: Pragmatism Defended Further”, unpublished manuscript. Mitova, Veli (2017): Believable Evidence, Cambridge: Cambridge University Press. Montmarquet, James (1992): Epistemic Virtue and Doxastic Responsibility, Lanham: Rowman & Littlefeld. Oakley, Justin (1992): Morality and the Emotions, London: Routledge. Owens, David J. (2000): Reason without Freedom: The Problem of Epistemic Normativity, London: Routledge. Owens, David J. (2003): “Does Belief Have an Aim?”, reprinted in his Normativity and Control, New York: Oxford University Press, 51–68.

Introduction

19

Owens, David J. (2017): Normativity and Control, New York: Oxford University Press. Paakkunainen, Hille (2018): “Doubts about ‘Genuinely Normative’ Epistemic Reasons”, in: Metaepistemology, ed. by McHugh, et al., New York: Oxford University Press, 122–140. Papineau, David (2013): “There Are No Norms of Belief”, in: Chan 2013, 64–79. Parft, Derek (2001): “Rationality and Reasons”, in: Exploring Practical Philosophy: From Action to Values, ed. by Egonsson, et al., Aldershot: Ashgate, 17–39. Parft, Derek (2011): On What Matters, Vol. 1, New York: Oxford University Press. Peels, Rik (2017): Responsible Belief, New York: Oxford University Press. Price, Carolyn (2015): Emotion, Cambridge: Polity Press. Pritchard, Duncan, et al. (2010): The Nature and Value of Knowledge: Three Investigations, New York: Oxford University Press. Putnam, Hilary (1981): “Brains in a Vat”, in his Reason, Truth, and History, Cambridge: Cambridge University Press, 1–21; reprinted in DeRose/Warfeld (eds.) (1999): Skepticism: A Contemporary Reader, 27–42. Rabinowicz, Wlodek/Rønnow-Rasmussen, Toni (2004): “The Strike of the Demon: On Fitting Pro-Attitudes and Value”, Ethics 114, 391–423. Raz, Joseph (2011): From Normativity to Responsibility, New York: Oxford University Press. Reisner, Andrew (2008): “Weighing Pragmatic and Evidential Reasons for Belief”, Philosophical Studies 138, 17–27. Reisner, Andrew (2009): “The Possibility of Pragmatic Reasons for Belief and the Wrong Kind of Reasons Problem”, Philosophical Studies 145, 257–272. Reisner, Andrew (ms): “Comparing Pragmatic and Alethic Reasons”, Chapter II of Reisner: The Pragmatic Foundations of Theoretical Reason, unpublished manuscript. Reisner, Andrew/Steglich-Petersen, Asbjørn (eds.) (2011): Reasons for Belief, Cambridge: Cambridge University Press. Rinard, Susanna (2015): “Against the New Evidentialists”, Philosophical Issues 25, 208–223. Rinard, Susanna (2017): “No Exception for Belief”, Philosophy and Phenomenological Research 94, 121–143. Rinard, Susanna (2019a): “Equal Treatment for Belief”, Philosophical Studies 176, 1923–1950. Rinard, Susanna (2019b): “Believing for Practical Reasons”, Noûs 53, 763–784. Roberts, Robert C. (2003): Emotions: An Essay in Aid of Moral Psychology, Cambridge: Cambridge University Press. Schmidt, Sebastian (2016): “Können wir uns entscheiden, etwas zu glauben? Zur Möglichkeit und Unmöglichkeit eines doxastischen Willens”, Grazer Philosophische Studien 93, 571–582. Schmidt, Sebastian (2017): “Why We Should Promote Irrationality”, Grazer Philosophische Studien 94, 605–615. Schmidt, Sebastian (forthcoming): “Rationality and Responsibility”, Australasian Philosophical Review 5. Schroeder, Mark (2009): “Means-End Coherence, Stringency, and Subjective Reasons”, Philosophical Studies 143, 223–248. Scott-Kakures, Dion (1994): “On Belief and the Captivity of the Will”, Philosophy and Phenomenological Research 54, 77–103.

20

Sebastian Schmidt

Shah, Nishi (2006): “A New Argument for Evidentialism”, The Philosophical Quarterly 56, 481–498. Smith, Angela M. (2005): “Responsibility for Attitudes: Activity and Passivity in Mental Life”, Ethics 115, 236–271. Smith, Angela M. (2015a): “Attitudes, Tracing, and Control”, Journal of Applied Philosophy 32, 115–132. Smith, Angela M. (2015b): “Responsibility as Answerability”, Inquiry 58, 99–126. Sosa, Ernest (2007/2009): A Virtue Epistemology: Apt Belief and Refective Knowledge, Vols. 1/2, New York: Oxford University Press. Steglich-Petersen, Asbjørn (2006): “No Norm Needed: On the Aim of Belief”, The Philosophical Quarterly 56, 499–516. Steglich-Petersen, Asbjørn (2009): “Weighing the Aim of Belief”, Philosophical Studies 145, 395–405. Steglich-Petersen, Asbjørn (2011): “How to Be a Teleologist about Epistemic Reasons”, in: Reisner/Steglich-Petersen 2011, 13–33. Steglich-Petersen, Asbjørn (2013): “Transparency, Doxastic Norms, and the Aim of Belief”, Teorema 32, 59–74. Steglich-Petersen, Asbjørn/Skipper, Mattias (forthcoming): “An Instrumentalist Account of How to Weigh Epistemic and Practical Reasons for Belief”, Mind. Strawson, Peter F. (1962): “Freedom and Resentment”, Proceedings of the British Academy 48, 1–25. Tappolet, Christine (2012): “Emotion, Perception and Emotional Illusions”, in: Perceptual Illusions: Philosophical and Psychological Essays, ed. by Calabi, London: Palgrave Macmillan, 205–222. Tappolet, Christine (2016): Emotions, Values, and Agency, New York: Oxford University Press. Velleman, J. David (2000a): “On the Aim of Belief”, reprinted in his The Possibility of Practical Reason, Oxford: Clarendon Press 2009, 244–281. Velleman, J. David (2000b): The Possibility of Practical Reason, Oxford: Clarendon Press. Velleman, J. David (2009): How We Get Along, Cambridge: Cambridge University Press. Way, Jonathan (2009): “Two Accounts of the Normativity of Rationality”, Journal of Ethics and Social Philosophy (Discussion Note), 1–8. Way, Jonathan (2010): “Defending the Wide-Scope Approach to Instrumental Reason”, Philosophical Studies 147, 213–233. Wedgwood, Ralph (2002): “The Aim of Belief”, Philosophical Perspectives 16, 267–296. Wedgwood, Ralph (2017): The Value of Rationality, New York: Oxford University Press. Williams, Bernard (1970): “Deciding to Believe”, reprinted in his Problems of the Self, Cambridge: Cambridge University Press 1973, 136–151. Williams, Michael (1999): “Skepticism”, in: The Blackwell Guide to Epistemology, ed. by Greco/Sosa, Oxford: Blackwell, 35–69. Zagzebski, Linda (1996): Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge, Cambridge: Cambridge University Press.

Part I

New Perspectives on Belief Normativity

A

Doxastic Agency and Responsibility

2

Implications of the Debate on Doxastic Voluntarism for W. K. Clifford’s Ethics of Belief Martina Lindner

1.

Introduction

One of the most infuential, but also one of the most controversial, positions in the debate on an ethics of belief is the one which is usually considered as its origin – the position of William Kingdon Clifford, who famously argues for his Principle that “it is wrong always, everywhere, and for anyone, to believe anything upon insuffcient evidence” (Clifford 1877: 186). Clifford’s radical position has been criticized for various reasons; in this text I will deal especially with those kinds of criticism that are concerned with the fact that Clifford argues within a consequentialist framework. Such criticism is based on the general idea that a consequentialist framework commits one to an ethics of belief that is quite different from Clifford’s position. But though such critics are united in the rejection of Clifford’s Principle that it is wrong to believe upon insuffcient evidence, they may be categorized according to how radically they differ from Clifford’s ethics of belief. While proponents of what I will call the Permissibility Account hold that within such a consequentialist framework it is sometimes permitted to believe upon insuffcient evidence, proponents of what I will call the Requirement Account argue that in some cases it is even required to do so because it leads to better consequences. In the frst part of this text I will have a closer look at Clifford and at his critics. I will then show that considering whether Clifford or his critics are right leads the philosopher to an unsatisfying result: it amounts to empirical questions and, what is more, to rather diffcult ones. Therefore, in the second part of my text I will try to sidestep these empirical questions and I will investigate whether a line of argumentation within the realm of philosophical inquiry could – though not ultimately decide – clarify the discussion and help to decide which conception is most plausible. More precisely, I will draw on the debate on doxastic voluntarism, i.e. the debate on whether and in how far people can voluntarily control their beliefs. My drawing on this debate is due to the fact that this debate is relevant to all conceptions of an ethics of belief insofar as they have to deal with the question of whether their normative framework requires

24

Martina Lindner

voluntary control of beliefs. We will gain three – more or less surprising – results: frstly, none of the three conceptions is immediately refuted by the claims I derive from the debate on doxastic voluntarism; secondly, Clifford’s conception of an ethics of belief nevertheless fares better with regard to these claims than both the Requirement Account and the Permissibility Account. And thirdly, perhaps most surprisingly, the Permissibility Account – the account that is by far the most popular among Clifford’s critics – does not fare much better than does the Requirement Account with regard to the debate on doxastic voluntarism.

2.

W. K. Clifford’s Ethics of Belief – An Overview

William Kingdon Clifford’s ethics of belief is – as Andrew Chignell noted – “chiefy remembered for two things: a story and a principle” (Chignell 2018, n. pag.). Before we have a closer look at his aforementioned Principle and its justifcation, let us have a look at the story. A shipowner is in doubt as to whether his rather old and run-down ship which is to set sail the next day is still seaworthy. But he succeeds in overcoming these doubts – more precisely, he “stif[es]” (Clifford 1877: 178) them, directing his thoughts away from his doubts to the many voyages his ship has safely come through and to an all-protecting providence. So he fnally “acquire[s] a sincere and comfortable conviction that his vessel [is] [. . .] thoroughly safe and seaworthy” (ibid.: 177). But it turns out that his belief is false: the ship sinks and all passengers lose their lives. Is the shipowner guilty of the passengers’ death? Clifford claims that he is – even though “he did sincerely believe in the soundness of his ship” (ibid.: 178): “[T]he sincerity of his conviction can in no wise help him, because he had no right to believe on such evidence as was before him” (ibid., italics in original). To put Clifford’s thought in contemporary terms we could say that he is morally to blame for acquiring an epistemically unjustifed belief (cf. Wood 2002: 3). But Clifford does not restrict his moral condemnation of epistemically unjustifed believing to cases where the welfare of many people is at risk. Believing upon insuffcient evidence is never harmless, Clifford argues, and so he formulates his aforementioned Principle: “it is wrong always, everywhere, and for anyone, to believe anything upon insuffcient evidence.” It might strike us at once that Clifford’s universal Principle is unduly extensive: Certainly, in some cases believing upon insuffcient evidence has bad consequences – but in many cases it seems to be quite harmless. One reason why the shipowner’s belief seems to be morally signifcant is that he immediately acts on it. But what about beliefs we do not act on? And what about beliefs which cannot result in harmful actions even if they prove wrong? When we take a closer look at Clifford’s text, we fnd two different lines of argumentation in support of his universal Principle.

Doxastic Voluntarism and Clifford

25

Firstly, Clifford maintains that although a belief may not be immediately acted on, it is “stored up for the guidance of the future” (Clifford 1877: 181), future actions may be based on it. Furthermore, Clifford explains that in our “aggregate of beliefs” no belief is “isolated from the rest” (ibid.): “No real belief, however trifing and fragmentary it may seem, is ever truly insignifcant; it prepares us to receive more of its like, confrms those which resembled it before, and weakens others” (ibid.). So every belief I acquire – even if I never directly act on it – indirectly infuences my actions, because it infuences the acquisition of other beliefs that may themselves infuence my actions. Moreover, other people may rely on my belief and act on it. So even if I can make sure that I will never act on it, I must not acquire an epistemically unjustifed belief because it may lead to harmful actions indirectly. Let us call this line of argumentation the argument from the aggregate of beliefs. Clifford’s second line of argumentation covers even those beliefs that cannot result in harmful actions if they prove wrong. He argues that even in those cases acquiring a belief upon insuffcient evidence is harmful because it fosters a harmful disposition: credulity. Every time I acquire a belief without suffcient evidence, I nurture a disposition to be credulous (cf. ibid.: 185–186). Let us call this line of argumentation the argument from credulity. Before we turn to Clifford’s critics now, let us have a closer look at the type of norms Clifford advocates here and at how he argues for them.

3.

Clifford’s Evidentialist Ethics of Belief – Moral Norms Within a Consequentialist Framework

Clifford’s reference to the harmful consequences of believing upon insuffcient evidence shows that he argues within a framework of moral normativity. More precisely, Clifford’s aforementioned Principle is a moral norm with a consequentialist justifcation (cf. Wood 2002: 23). So although Clifford argues for an evidentialist ethics of belief, the structural framework of his conception (the kind of norms he is advocating and their justifcation) is quite different from current evidentialists’ approaches to the ethics of belief. Following a distinction by Miriam McCormick, one can discern that evidentialists in the current debate usually ground their position either on conceptual considerations, claiming that evidentialism follows directly from the concept of belief, or on the existence of epistemic norms as an autonomous normative domain.1 So even though the content of Clifford’s ethics of belief is evidentialist, the structural framework of his position comes close to a pragmatist account – and thus to an account of opponents of evidentialism (cf. Mitova 2008b: 479). For although Clifford and the pragmatists of the current debate differ fundamentally with regard to the content of their respective doxastic norms – pragmatists famously reject the view that belief should only be

26

Martina Lindner

based on evidential considerations – they both think that evidential warrant is not in itself signifcant for justifying a belief. What should guide us in our belief adoption is the same kind of practical considerations that govern actions (e.g. McCormick 2015) or states (Rinard 2015: 132) – whereby Clifford restricts the relevant considerations to moral ones. No epistemic norms are introduced as norms sui generis. When McCormick in her defense of pragmatism claims that “the only way to make sense of epistemic value as a good to be promoted is to link it, or ground it, in the practical or moral” (McCormick 2015: 38) and when Susanna Rinard argues that “a consideration C constitutes a genuine reason to believe P if and only if it is a pragmatic consideration in favor of believing P” (Rinard 2015: 220) whereas she “include[s] moral considerations under the heading ‘pragmatic’” (ibid.: 217), Clifford shares their basis for the assessment of beliefs: moral considerations. Thus, Veli Mitova characterizes Clifford as a “closet pragmatist” (Mitova 2008b: 481). On the one hand, with his moral consequentialist justifcation of evidentialism, Clifford can sidestep a notorious problem for evidentialists in the current debate: to explain how evidence gains its normative force. But on the other hand, his moral consequentialist framework opens the way for other kinds of problems for a proponent of evidentialism – especially those that have to do with the morally relevant consequences of believing upon insuffcient evidence: if a critic could show that the consequences are different from what Clifford expected, Clifford’s universal evidentialist Principle would be undermined. In the next paragraph, I will consider how Clifford might be criticized on consequentialist grounds. We will see that such criticisms of Clifford’s position on consequentialist grounds prove quite strong. However, if Clifford’s arguments could stand these consequentialist criticisms, then his position might be of more interest to current debates on the ethics of belief and deserve more attention from proponents of evidentialism.

4.

Criticism of Clifford’s Position on Consequentialist Grounds – The Permissibility Account and the Requirement Account

Let us frst consider an alternative conception of an ethics of belief which claims (as many positions of Clifford’s critics do) that believing upon insuffcient evidence is wrong in many cases but that there are nevertheless cases in which it is permissible to believe upon insuffcient evidence, i.e. that there are exceptions to Clifford’s Principle.2 Let us call such an alternative conception of an ethics of belief a Permissibility Account.3 A Permissibility Account seems to be quite plausible within a consequentialist framework. It might strike even the most sympathetic reader of Clifford’s text that Clifford might be right in many aspects but probably goes too far in condemning every instance of epistemically unjustifed

Doxastic Voluntarism and Clifford

27

belief. Clifford’s Principle that believing upon insuffcient evidence is impermissible in any case depends on his contention that every instance of epistemically unjustifed believing has – either directly or indirectly – severely bad consequences. But it seems to be quite obvious that there are cases in which the consequences of an epistemically unjustifed belief are not as bad as Clifford argues (or where there are no bad consequences at all). Take my belief that my aunt gave me a teddy for my ffth birthday. Suppose that my vague memory cannot justify that belief. Surely, I will never act on that belief. Why should one blame me for that belief? And even if it proves wrong, it seems to be harmless for myself as well as for anybody else. To be sure, Clifford anticipates cases like these and attempts to show that here too harmful consequences are lurking. But it is at least doubtful whether Clifford’s arguments are persuasive in such cases. Take the argument from credulity: Is Clifford right in claiming that every single epistemically impermissible belief makes me credulous? Of course, Clifford may be right – but he need not. At least, he cannot treat it as an indisputable fact in his argumentation. It is quite clear that one could fnd many cases like this one for which Clifford’s Principle does not seem to be plausible. And it is his consequentialist framework that causes this problem. Worse still for Clifford, his consequentialist framework paves the way for an even more radical criticism: Up to this point, we have just asked whether there are exceptions to Clifford’s Principle, i.e. cases in which the consequences of epistemically unjustifed belief are not as bad as Clifford argues. But couldn’t believing upon insuffcient evidence even have positive consequences in some cases? And if believing upon insuffcient evidence has positive consequences in some cases – isn’t one even required to believe upon insuffcient evidence in these cases? Take a patient’s belief that he will recover from a severe illness. Let us suppose that his frm belief in his recovery would not make him neglect seeking medical treatments but would instead be an incentive to be informed about new developments in this feld. Suppose that his frm belief can increase the probability that he will actually recover (although it is still less than 50%) and suppose that his belief gives himself and his family strength and confdence – shouldn’t we say that the belief that he will recover has positive consequences although it would be acquired upon insuffcient evidence? As we have seen, Clifford’s arguments to the effect that every epistemically unjustifed belief has negative consequences do not seem to be persuasive here. But even if we assume for the sake of argument that Clifford is right in claiming that if the patient believes that he will recover, he will corrupt his aggregate of beliefs and make himself credulous – don’t the positive consequences override the negative ones in cases like that? If there are really cases in which the positive consequences of believing an epistemically unjustifed proposition override the negative ones (and in which the consequences of believing that proposition are

28

Martina Lindner

better than the consequences of not believing it), then it would be quite plausible that a consequentialist conception of an ethics of belief would not only permit but even require belief in such cases. So it is obviously Clifford’s consequentialist justifcation of his Principle that lets his ethics of belief fall prey to this criticism.4 Let us call such kind of radical criticism of Clifford’s ethics of belief, a position that includes the claim that one is sometimes required to believe upon insuffcient evidence, a Requirement Account.5 Now let us consider more precisely what we have to investigate for our central question whether Clifford or his moderate or radical critics have developed the most plausible view. If one assumes a consequentialist framework, one has to investigate what such a consequentialist framework really commits one to. What epistemic principles follow from such a framework? Does it commit one to never believing upon insuffcient evidence as Clifford thinks? Or does it commit one to “usually” not believing upon insuffcient evidence but does allow for exceptions? Or does it require one to believe upon insuffcient evidence in some cases? Note frst that within Clifford’s consequentialist framework it is just due to contingent facts that he can argue for his universal evidentialist Principle (recall that his arguments rest on the fact that every instance of believing upon insuffcient evidence spoils our aggregate of beliefs and makes us credulous). So in order to establish his Principle, Clifford has to show that believing upon insuffcient evidence does, as a matter of fact, lead to harmful consequences in any case. But for his critics the situation is just the same: They have to show that there are cases in which believing upon insuffcient evidence does not have any negative consequences, i.e. that there are cases in which Clifford’s arguments do not persuade, or – if one acknowledges that there are negative consequences like those that are mentioned by Clifford – that the positive consequences outweigh the negative ones. As we have seen, Clifford’s critics can bring forward quite convincing cases where Clifford’s rigid ethics of belief seems out of place. But, as we have also seen, Clifford anticipates such cases and endeavours to show that here too his ethics of belief is applicable. In order to decide which conception of an ethics of belief is right, one has to answer empirical questions like the following: Does acquiring a belief upon insuffcient evidence spoil our aggregate of beliefs? And to which degree does it spoil it? Does acquiring a belief upon insuffcient evidence make us credulous? And to which degree? What amount of harm is to be expected? In order to see whether possible good consequences could override the negative ones, we have to deal with these empirical – and rather diffcult – questions. If deciding whether Clifford or his critics are right is therefore tantamount to answering these empirical questions, then we would have to leave the realm of philosophical inquiry. And that would be a rather

Doxastic Voluntarism and Clifford

29

unsatisfying result for the philosopher – all the more so as an issue is concerned that is relevant for several philosophical disciplines such as epistemology, ethics, and the philosophy of religion. To avoid this unsatisfying result and to try to discuss this issue within the realm of philosophical inquiry, I will investigate whether a line of argument which is based on results of the debate on doxastic voluntarism could allow us to decide which conception is more plausible.6

5.

The Debate on Doxastic Voluntarism – Relevant Results for Our Central Question

The debate on doxastic voluntarism – the investigation of whether and in how far we can control our beliefs – is crucial to the ethics of belief since all conceptions of an ethics of belief have to deal with the question of whether their normative framework requires voluntary control of beliefs. Sometimes it is argued that Clifford’s and his critics’ positions – or generally all conceptions of an ethics of belief – are “on a par” with regard to the debate on doxastic voluntarism.7 This claim is based on the thought that an ethics of belief makes sense only if my will can affect the way I form my beliefs – as otherwise norms for the acquisition of beliefs would be pointless. This is certainly right, but it does not follow that all conceptions of an ethics of belief are really “on a par” because each conception requires a specifc kind of voluntary control of the formation of beliefs and so it is possible that results of the debate could speak in favour of one conception and against another. In the remainder of this text I will try to show that Clifford’s ethics of belief fares better than his critics’ positions with regard to the results of the debate on doxastic voluntarism. I will frst consider Clifford’s ethics of belief and the Requirement Account. Later I will turn to the Permissibility Account, and I will show that either it faces just the same problems as the Requirement Account or it can be refuted along Clifford’s lines. The debate on doxastic voluntarism is conducted intensively and controversially.8 Needless to say, there are almost no defnite results that are accepted by all participants of this debate. But nevertheless, I think that many participants might accept my following outline of some results of this debate that are relevant here.9 As both Clifford’s conception and the Requirement Account imply that one is able to infuence one’s belief formation, I will start with possible means for infuencing our beliefs. Then I will consider in how far making use of these means is effective; i.e. I will consider whether consciously making use of these means can lead to acquiring the desired belief. Then I will investigate whether we can control our beliefs. When we consider how our belief formation works, we can distinguish two kinds of possible means for infuencing our beliefs. Firstly, our belief formation can be infuenced by an unbiased investigation on the given

30

Martina Lindner

topic. We can read diverse books on that topic, we can talk to people harbouring different views, or we can question our own preconceptions concerning that topic. Such an investigation may be conducted more or less consciously, deliberately and thoroughly, ranging from a long-term examination to a short refection whether something is the case. It is quite clear that such an investigation can infuence what we believe; hence, we may consider all parts of an investigation as means for infuencing our beliefs. What is important here is that these means do not point to a particular belief beforehand. They can lead to a gain in evidence or to a different view on the evidence one already has, and the evidence itself might then suggest a belief. Therefore, I will refer to them as evidenceoriented means. Secondly, our beliefs can be infuenced by factors apart from such an evidence-oriented investigation. Obviously, we can infuence our beliefs also by manipulative means: we can focus only on evidence that speaks in favour of some proposition, we can talk only to those people who share our opinions on a given subject, we can read all and only those books that support a given view, or we can even consult a hypnotist who might predispose us to subconsciously acquire the desired belief.10 Unlike evidence-oriented means, manipulative means are made use of to acquire a particular belief, and the means are chosen with regard to the aim of acquiring that desired belief.11 Since “beliefs aim at truth” (Williams 1973: 136) and evidence might be regarded as a “guide to truth” (Kelly 2016, n. pag.), when we consciously form beliefs – and that is just what we do when we make use of either kind of means – we will “usually” believe according to the evidence. So we can assume that making use of evidence-oriented means (to acquire the belief that fts the evidence best) will normally be effective more often than making use of manipulative means (to acquire a particular belief). But let us investigate more closely in how far these two kinds of means can be consciously made use of if one wants to infuence one’s belief management (as such infuence on one’s belief management is required by each conception of an ethics of belief). Evidence-oriented means are unproblematic in this respect: obviously, we can not only consciously make use of them in our belief formation; the thus acquired belief will usually be even more stable if we are aware of the fact that we have acquired it in an unbiased investigation into the topic. By contrast, making use of manipulative means faces a diffculty in this respect. Manipulative means are obviously connected to non-evidential considerations in belief formation and to pragmatic arguments for belief, i.e. arguments not to the effect that a certain proposition is true but that having a certain belief is benefcial (cf. Jordan 1996: 409; Rinard 2015: 221, footnote 2): manipulative means may serve in rendering these pragmatic arguments effective in giving rise to the desired belief. But many authors point to the ineffectiveness of pragmatic arguments on our belief formation when

Doxastic Voluntarism and Clifford

31

contrasting them to evidential considerations. While strong evidence in favour of p would normally result in my belief that p, being offered a large amount of money if I believe that q – and thus being faced with a pragmatic argument for believing that q – is usually ineffective (cf. e.g. Alston 1989: 122). But in how far are pragmatic arguments ineffective for acquiring the belief they point to? To begin with, it would be wrong to think that pragmatic considerations can never be effective, that making use of manipulative means can never lead to the belief one seeks to acquire. Alfred Mele asserts that even intentionally making use of these means may lead to the belief one aims at (cf. Mele 1987: 133). Yet although Mele is primarily concerned with defending the possibility of acquiring a belief for nonevidential reasons and illustrating scenarios in which this happens, he also implicitly assumes that one cannot consciously hold a belief while thinking that one believes it for pragmatic reasons. In a Pascalian wager scenario he describes, a person that has adopted a belief consciously for non-evidential reasons now regards the belief as supported by the evidence: As more time passes, he [i.e. the Pascalian bettor] may become very pleased with himself for having hit upon this plan of “deceiving himself” (which, of course, he would no longer describe in this way); for, he might think, if he had not done this, he would never have seen the light. Mele 1987: 133, emphasis mine Mele obviously thinks that this scenario would otherwise lose its plausibility.12 There are many authors who argue for such a refective instability in believing for pragmatic reasons, i.e. for the claim that one cannot consciously hold a belief and at the same time think that one believes it for pragmatic reasons. Firstly, such a claim can be found in two loci classici of the debate on doxastic voluntarism: Bernard Williams holds that if I had “acquire[d] a ‘belief’ irrespective of its truth” then “I could not [. . .] in full consciousness, regard this as a belief of mine, i.e. something I take to be true” (Williams 1973: 148). Barbara Winters argues that “it is impossible to believe that one believes p and that one’s belief of p originated and is sustained in a way that has no connection with p’s truth” (Winters 1979: 243). Furthermore, Ward E. Jones argues for a version of the claim that allows for degrees of belief and puts forth a “First-Person Constraint on Doxastic Explanation” (Jones 2002: 217): If I am at all convinced of a non-epistemic explanation of my particular, presently-held belief that p, then the stronger this conviction, the weaker will be either (a) my belief that p, or (b) my belief that I believe that p. Ibid.: 226

32

Martina Lindner

He thus concludes that pragmatic arguments “must somehow cover their own tracks. They must lead to belief without the believer being aware that they are doing, or have done, so” (ibid.: 229).13 Anne Meylan argues that “I cannot consider both that the mental state that I acquired is a belief and that I acquired this mental state for a non-epistemic reason” (Meylan 2013: 47).14 Summing up, we gain the following: claim A:15 If I am conscious of my belief being based on non-evidential considerations, then I cannot consider it as a belief anymore. More precisely, either I will stop believing it outright, or I will continue to believe it without being conscious that I have retained the belief (i.e. I will not consciously consider it as a belief). I think that for our investigation it is worthwhile examining the second disjunct of this claim more closely. What role can a belief that I cannot consciously consider as a belief play? I think that role is different from the role of full-blown beliefs both in our decisions how to act and what (else) to believe: If I become aware of the fact that my belief X is based on non-evidential considerations, this would change X’s role in my aggregate of beliefs. Even though I might retain belief X, I would not consciously consider it as a belief. Perhaps I would consider it as a wish. Or I would consider it as a prejudice that I cannot get rid of. In any case, it would change the way I would consciously deal with it – I would not deal with it like I would deal with a full-blown belief. Consciously, I would rather not base my decisions to act on X. I would rather not infer other beliefs from it. If it contradicts some other belief Y which is based on evidential considerations, then I would consciously not feel inclined to give up that belief Y. So although the role of a belief of which I know that it is based on pragmatic considerations (and thus do not consciously consider as a belief) might not differ from the role of full-blown beliefs with regard to decisions we come to without refection, its role differs considerably with regard to conscious decisions. As the role of beliefs for our decisions is crucial for both Clifford’s and the two rival conceptions of an ethics of belief, I will return to these considerations later. Let us now turn to the question of whether we can control our beliefs. It is quite obvious that making use of manipulative means in belief formation may be effective – but it need not. I can infuence my belief formation by taking the aforementioned steps (e.g. reading all and only those books that support the desired belief) – but whether I end up with that desired belief is not up to me. And even if I have been successful in acquiring a belief by making use of manipulative means, refective instability threatens the thus-acquired belief as we typically make use of manipulative means when we try to acquire a belief on pragmatic grounds. So although I may sometimes be able to infuence my beliefs by manipulative means, I cannot thus control my beliefs.16

Doxastic Voluntarism and Clifford

33

But the lack of control that has become apparent here is not restricted to cases in which we make use of manipulative means – I cannot ultimately control my beliefs by making use of evidence-oriented means either. To see that, we have to work out what an effective control by making use of evidence-oriented means would exactly mean. Recall that we do not make use of evidence-oriented means to acquire a particular belief but to acquire the very belief that is best supported by the evidence. Suppose that the shipowner in Clifford’s story did inquire thoroughly into the seaworthiness of his ship; suppose that he did everything he could to believe according to the evidence. Whether he fnally ends up with the belief that is best supported by the evidence17 is – again – not up to him.18 However, it is sometimes argued that we can control our beliefs, and this position is defended by pointing out that there are states that cannot be controlled directly but that are nevertheless widely recognized to be under our indirect control. Take the state of being able to speak Italian. Although there is nothing I can do to be able to speak Italian right now, there are steps I can take to make sure that I can speak Italian in, say, a few months. It may be argued that we can infuence our beliefs to the same extent as we can infuence states like being able to speak Italian. And so, as the latter is generally thought to be not just an infuence but a kind of control, we can also indirectly control our beliefs.19 This consideration is especially relevant for the use of evidenceoriented means. Making use of evidence-oriented means to acquire the belief that is best supported by the evidence is comparable to attending a language course to acquire command of Italian – in both cases the means are exactly those that commonly lead to the state one aims at. Of course, in both cases there could be hindrances to arrive at the aim: I may have considered all the relevant evidence and nevertheless not end up with the belief that is best supported by it. Or I may try very hard to learn Italian but nevertheless not end up being able to speak it (as I may have absolutely no talent for learning languages). And yet, we may say that we can control these states since we can control all the relevant steps that commonly lead to our aim. Whether our circumstances “join in” and fnally let us arrive at our aim is not relevant to the ascription of control. But I think this comparison is fawed. Whereas the steps I should take to learn Italian normally lead to acquiring command of it unless the circumstances do not “join in” or unless I have an “unusual” mental constitution, the same does not hold in the case of belief. Many authors (see e.g. Mele 1987: 115) point to the fact that our belief formation is often and commonly infuenced by non-epistemic factors – to such an extent that we cannot rely on acquiring the belief that is best supported by the evidence even if we have considered the relevant evidence as best as we could. I think the infuence of evidence-oriented means on our beliefs is rather comparable with our infuence on states like being in fear of fying. There are steps that are thought to lead to getting rid of that fear. We are able to undertake these steps but whether we lose our fear is not up to

34

Martina Lindner

us. And so we cannot, neither directly nor indirectly, control whether we are in fear of fying – just as we cannot control whether we end up with the belief we aim at. Summing up, we cannot control our beliefs, neither by making use of evidence-oriented means nor by making use of manipulative means. So we gain claim B: We can voluntarily infuence our beliefs but we cannot control them. Claim B can further clarify the disjunction in claim A. Our considerations on claim B imply that I cannot control the giving up of a certain belief by making myself aware of the fact that I lack evidence for it. Being aware of a lack of evidence for a belief does not ultimately determine my giving up that belief. Thus, it is quite plausible that claim A allows for the case that one does not lose a belief although one knows that it is based solely on non-evidential considerations (but one would not consciously consider it as a belief in that case). Analogously, claim B implies that making myself aware of the cogency of evidential considerations does not ultimately determine what belief I end up with.

6.

Implications of the Debate on Doxastic Voluntarism for Clifford’s Ethics of Belief and for the Requirement Account

With these claims from the debate on doxastic voluntarism we can now turn to our central question: which conception of an ethics of belief is more plausible with regard to these claims? As I have mentioned, I will frst consider Clifford’s ethics of belief and its radically antagonistic concept, the Requirement Account, and then later turn to the moderate alternative to Clifford’s conception, the Permissibility Account. Both Clifford’s ethics of belief and the Requirement Account are not just normative in the sense that they tell us which beliefs are justifed or unjustifed; they also reclaim to exert a guiding function, as they tell us how we should manage our beliefs: they tell us which beliefs we should adopt and which beliefs we should give up. So both accounts require our ability to observe their claims; they require our ability to manage our beliefs according to them. But both accounts also require our ability not to believe according to their claims – their guiding function would be lost if our belief-forming mechanisms were just so constituted that one would always believe according to their demands. So if for one account one could show that it is impossible either to observe its demands or to violate them, this account would be ruled out from the start. We have seen that with the aforementioned claims from the debate on doxastic voluntarism neither Clifford’s ethics of belief nor the

Doxastic Voluntarism and Clifford

35

Requirement Account are thus ruled out. It is, of course, possible that by means of an unbiased investigation into a topic I will end up with the belief that is best supported by the evidence – it is clearly not ruled out from the start that one can follow Clifford’s Principle. And it is also possible that by making use of manipulative means one ends up with a benefcial belief although that belief is not backed by the evidence – and so it is not ruled out from the start that one could follow the demands of the Requirement Account. And the fact that it is not ruled out from the start that one can follow the demands of the Requirement Account shows that it is possible to violate Clifford’s Principle – and vice versa. But although neither account is ruled out from the start, the claims from the debate on doxastic voluntarism might yet speak in favour of one account. We will now have a closer look at in how far the aforementioned claims can buttress either of the accounts. We start with claim A: If I am conscious of my belief being based on non-evidential considerations, then I cannot consider it as a belief anymore. More precisely, either I will stop believing it outright, or I will continue to believe it without being conscious that I have retained the belief (i.e. I will not consciously consider it as a belief). Let us frst consider the Requirement Account because it seems to be more directly concerned with this claim, as according to this position believing upon insuffcient evidence is sometimes required and as believing upon insuffcient evidence can often be regarded as believing that is based on non-evidential considerations. Although claim A does not directly rule out the possibility that one can adopt a belief for non-evidential reasons, I think this claim shows that the Requirement Account is quite different from what we usually expect from an ethics. I think we usually expect from moral norms that the following process applies: I realize what moral norms require me to do; I can take the appropriate steps to fulfll their requirements and – what is crucial here – after performing the action I can refect on what I have endeavoured to fulfll that requirement and I can evaluate whether I have fulflled it. But, if claim A is right, this is not feasible for the Requirement Account, since every belief that I have acquired according to the demands of the Requirement Account is threatened by refective instability. For if I refect on how I have fulflled that requirement then I am conscious of a non-epistemic explanation of my belief. So according to claim A either I will lose my belief that p outright, or I will retain the belief but I will not consciously consider it as a belief. So if I refect on how I have fulflled that requirement, then either I will lose my belief and so the hoped-for positive consequences of that belief cannot ensue, or I will not realize that I have adopted the desired belief and hence I will not realize that I have fulflled the requirement. Therefore either I cannot refect on how I have fulflled the requirement, or I

36

Martina Lindner

will not fulfll it at all, or I will not be able to realize that I have fulflled it. These considerations cannot suffce to refute the Requirement Account, but I think it is weakened insofar as these considerations show that it is quite different from what we usually expect from moral norms. And now for Clifford’s position. Firstly, Clifford’s position is – unlike the Requirement Account – not weakened by claim A. But here, I think, fairness requires that this should not count in its favour because the claim applies only to such cases of belief formation that are interdicted by Clifford’s position anyway. Nevertheless his position is strengthened by this claim; namely, this claim shows how a potential vulnerability of Clifford’s position can be – at least partly – removed: Clifford’s Principle does not only require one to adopt only beliefs that are supported by the evidence but also to give up beliefs one already has if they are not backed by suffcient evidence. But how is this possible, if beliefs are, according to claim B, not under our control? For, if that is the case, my decision to give up an unjustifed belief might possibly not be effective. And therefore, Clifford’s Principle might possibly require what we cannot fulfll. But claim A shows how that vulnerability could partly be removed. If I – by the thorough and impartial investigation that Clifford demands – become convinced that my belief that p is based on non-evidential considerations, then I will not consider it as a belief anymore. Either I will stop believing it outright, or I will continue to believe it but cannot consciously consider it as a belief anymore. In the frst case, I would stick to Clifford’s rule; in the second case, at least the impact of this propositional attitude on my actions would not be as strong as if it were a full-blown belief and thus the consequences would probably not be as harmful. As explained, consciously, I would rather not base my decisions to act on it. And it would probably spoil my aggregate of beliefs much less than a full-blown belief would, for I would not consciously infer other beliefs from it. So the negative consequences of this attitude are less problematic in comparison to the negative consequences of a full-blown belief.20 Yet – because I ultimately lack control of my beliefs – I would not be able to completely fulfll the requirement of Clifford’s Principle, namely, to believe only what is backed by suffcient evidence. But the lack of control over beliefs proves problematic for the Requirement Account as well. We will now turn to claim B to examine this issue more closely. Claim B states that we can voluntarily infuence our beliefs but cannot control them. So we now have to examine what kind of voluntary impact on belief formation is required by each conception. At frst, it might seem that both conceptions equally require at least indirect voluntary control. In both conceptions the positive consequences could only ensue if the required belief is adopted. The mere commitment to the respective norm, the mere attempt to stick to this norm in the formation of a belief will not suffce if we do not end up with the required belief. Take the Requirement Account frst and take for clarifcation once again the patient’s situation.

Doxastic Voluntarism and Clifford

37

According to the Requirement Account he ought to believe that he will recover. But suppose that although he does everything he can to adopt the required belief, he does not end up with it. But only belief in his recovery has the hoped-for positive consequences; the mere attempt to adopt the belief that he will recover has no better outcome than making no attempt to change his belief at all. Now take Clifford’s conception and the shipowner case. This case is not about adopting a required belief but about not adopting a forbidden belief, but the diffculty with claim B is just the same as for the Requirement Account in the frst place. According to Clifford’s ethics of belief the shipowner should thoroughly investigate and he should not believe upon insuffcient evidence. Suppose now that the shipowner sticks to Clifford’s Principle. If his investigation shows that the belief that his ship is seaworthy is not backed by the evidence, he should not acquire this belief. But here, too, even if he investigates patiently and thoroughly and does everything he can to ensure that he believes according to the evidence, he cannot control with what belief he will fnally end up. Perhaps the shipowner’s desire to avoid additional costs leads him to misinterpret the evidence, to underestimate what counts against the belief that the ship is seaworthy, and to overestimate what counts for that belief.21 And this would result in just the same negative consequences for the passengers as if he had not paid any attention to Clifford’s Principle at all. So it seems as if claim B is equally problematic for Clifford’s position and for the Requirement Account. But I think Clifford is nevertheless better off with regard to this claim because of two reasons. The frst arises out of Clifford’s specifc argumentation. Whereas the aforementioned considerations concerning claim B apply to Clifford’s argument from the aggregate of beliefs, they do not apply to his argument from credulity. Let us consider once again the shipowner case and let us suppose – as set out earlier – that he thoroughly investigates the matter but nevertheless ends up with the belief that his ship is seaworthy. Then Clifford’s argument from the aggregate of beliefs would not work because although the shipowner would have acted according to Clifford’s ethics of belief, he would have adopted a wrong and therefore dangerous proposition in his aggregate of beliefs. But nevertheless the way he proceeded in the formation of his belief – investigating the matter thoroughly – would have positive consequences because it would weaken his tendency to be credulous. So the argument from credulity would still hold even in such cases, and so claim B is much less problematic for Clifford’s position as it is for the Requirement Account. The fact that in such cases Clifford can make use of an argument that refers to the way one proceeds in belief formation and not just to the content of the belief is, of course, due to the fact that Clifford’s conception – unlike the Requirement Account – does not require one to adopt some particular belief.

38

Martina Lindner

The other reason why claim B is not as problematic for Clifford’s position as it is for the Requirement Account applies less to specifc features of Clifford’s argumentation and more to its universal evidentialism and would therefore be relevant even if one rejected the argument from credulity (and, as I have explained, the argument from credulity is not beyond any reasonable doubt). Even if both conceptions in a way suffer from the fact that the acquisition of a belief cannot be controlled but only infuenced, one could ask for which conception an attempt to infuence one’s beliefs would be more effective. So in which conception is it more likely that one succeeds in adopting the required belief? We have seen that making use of evidence-oriented means fts our “usual” way of consciously forming beliefs best and so, one may infer, the “success rate” is higher for making use of evidence-oriented means than of manipulative means. And it is quite clear that in order to stick to Clifford’s ethics of belief making use of evidence-oriented means is required, whereas the Requirement Account calls for the use of manipulative means. So the chances to adopt the required belief are higher for Clifford’s conception than they are for the Requirement Account. This consideration can be bolstered by looking back at claim A. Claim A says that a belief can only be consciously held if one does not think that one believes it for non-evidential reasons. For a belief that has been acquired upon non-evidential considerations this is possible only if either one forgets about the role of non-evidential reasons in the acquisition of the belief (cf. Jones 2002: 229; Foley 1994: 41; Meylan 2013: 65) or if one is aware of the fact that one has acquired the belief for non-evidential reasons but thinks that one has obtained evidence for it after its acquisition (cf. Winters 1979: 246–248; Meylan 2013: 65, footnote 93). I will consider each of the cases in turn. Let us frst consider the case that one forgets about the role of nonevidential considerations. How could one make sure that one succeeds in forgetting about the role of these non-evidential factors? Which of the manipulative means can ensure that? Of course, there might be means that are reliable in this respect, but these means would be quite radical ones like undergoing hypnosis or taking certain drugs – and obviously such radical means can only be used in a small number of cases. As regards less radical manipulative means, there may be cases in which one makes use of them and afterwards forgets having done so. But such cases occur by chance and surely they are far from frequent – remember that one has consciously made use of these means. Now take the case in which one is aware of the fact that one has acquired the belief for non-evidential reasons but thinks that one has obtained evidence for it after its acquisition. To start with, it is quite plausible that there are cases in which one obtains evidence after acquiring the belief – some authors argue that there are even cases in which evidence for a proposition can only be obtained if one already believes that proposition.22 But, surely, one cannot depend on obtaining evidence afterwards – such

Doxastic Voluntarism and Clifford

39

cases arise by chance. And, in any case, if one thinks one has obtained evidence afterwards, regardless of whether one really has, refective instability lurks at another level. If one thinks that one has obtained evidence for a belief after its acquisition, while at the same time one is aware of the role of non-evidential factors in its acquisition, one will probably always be rather suspicious of that belief. For in such cases it is always possible that – because of one’s irrational tendency to search for confrmation of one’s beliefs rather than for disproof – one misinterprets something as evidence for that belief although, actually, it does not support that belief at all (cf. Mele 1987: 125–126). And so it is doubtful whether one can consciously consider the acquired belief as a belief and thus whether it can really amount to a full-blown belief. We see that in both these cases whether I end up with the desired belief is accidental; I cannot rely on it.23 Thus, one is not only unable to make sure that one fulflls the demands of the Requirement Account – even attempting to do so will often be futile as there are no means to fulfll the demands that are both reliable and applicable in more than just a few cases. However, in Clifford’s conception one could assume that the chance to adopt the required beliefs is quite good. As I have explained, Clifford’s Principle fts the “usual” way in which one would consciously form beliefs best. And not only are the means one could make use of – like inquiring thoroughly and impartially – not restricted by claim A; claim A even shows that in situations in which one is aware of a lack of evidential warrant the belief concerned is not a full-blown belief – and thus that situations in which one cannot believe according to the evidence are not paradigmatic instances of belief. So although one might not be able to live up to Clifford’s Principle in such cases, the consequences would not be as severe as with full-blown beliefs. And as the prevention of severe consequences grounds Clifford’s Principle, the fact that one sometimes cannot fulfll Clifford’s Principle is, in my eyes, not overly problematic for Clifford’s conception since in these cases the consequences are not as severe as for paradigmatic instances of belief. To sum up, even if the claim that we cannot control but only infuence our beliefs has turned out problematic for both Clifford’s ethics of belief and the Requirement Account, it is less problematic for Clifford’s conception. Taken together, these considerations show that Clifford’s ethics of belief fares better with regard to the results of the debate on doxastic voluntarism than does the Requirement Account.

7.

Implications of the Debate on Doxastic Voluntarism for the Permissibility Account

Let us now turn to the Permissibility Account. The Permissibility Account seems to be much more plausible than both alternative accounts in the frst place. But we should have a closer look at how it might precisely work. If the Permissibility Account is right, there are cases in which the

40

Martina Lindner

positive consequences of believing upon insuffcient evidence override the negative ones and in these cases one is permitted to adopt the belief. Imagine that Paula is quite a good football player who thinks about whether she may adopt the belief that she will do excellently at the next match.24 She knows that this belief will make herself as well as her team not only much happier but also much more confdent and might thus even increase their chances to win the match. She knows that she and her team have worked hard to improve their performance, so they deserve to be successful – which prompts her to think that there are morally relevant consequences speaking in favour of adopting the belief that she will do excellently. Now she tries to decide whether her evidence supports this belief – and if it does not, whether the negative consequences of believing upon insuffcient evidence could be overridden by the positive ones. She realizes that the evidence is in no way suffcient for her belief – she has not been in really good shape for some time and she knows that being under pressure (because of the importance of the match) will diminish her prospects of performing well. So she makes up her mind as to whether she may nevertheless adopt the belief because of its positive consequences. She knows that she has to take into account all aspects of her situation. So she ponders on the evidence to see to which degree exactly it speaks against her desired belief. She tries to quantify the beneft she and the others might gain from her belief. She considers the question in how far this belief could affect her aggregate of beliefs. After long and careful analysis she comes to the conclusion that believing upon insuffcient evidence is permitted in that case. Now she knows that she may surrender to the pragmatic arguments in favour of believing that proposition. But probably that would not work anymore now: the more conscious she is of the fact that evidential considerations do not speak in favour of that proposition while pragmatic arguments do, the less probable – in light of our considerations on claim A – it is that she is really able to acquire that belief. And her long and careful investigation has made her exceedingly conscious of the fact that this belief is not backed by the evidence. So the Permissibility Account faces just the same diffculties as the Requirement Account does, although it may have appeared much more plausible at frst glance. But a proponent of the Permissibility Account may allege that it has an advantage over the Requirement Account: whereas the Requirement Account requires one to adopt certain beliefs, the Permissibility Account just permits their adoption. So the proponent of the Permissibility Account can argue that although one may not be able to consciously adopt an epistemically unjustifed belief, the Permissibility Account allows for cases in which one already has some tendency to believe a certain evidentially unjustifed proposition. Our considerations have shown that it is possible to have such a tendency to believe a proposition that is not supported by the evidence – for example in cases in which one has not really

Doxastic Voluntarism and Clifford

41

made up one’s mind yet about the relevant evidence. The Permissibility Account permits one to give in to that tendency then. So it seems as if the Permissibility Account might be feasible if it is considered as some kind of laissez-faire policy in belief-formation: if I have a tendency to believe an unsupported proposition and if the positive consequences of believing override the negative ones, then I may give in to that tendency. But this argument cannot support the Permissibility Account – at least, it can be shown to be applicable only in rare cases. The Permissibility Account cannot per se support a laissez-faire policy since for such a policy Clifford’s arguments are obviously fatal. For although Clifford’s assertion that evidentially unjustifed believing has bad consequences in any case is questionable, there are defnitely cases (e.g. Clifford’s shipowner case) where believing upon insuffcient evidence leads to dramatically harmful consequences. And if I do not thoroughly investigate whether adopting a particular evidentially unjustifed belief has such bad consequences, I will never detect these harmful cases. So a proponent of the Permissibility Account can in no way dispense with a thorough investigation of the matter. And so the permission to believe is only relevant in cases in which my tendency to believe “survives” both a thorough investigation with the result that evidence does not support my belief, and my awareness that it is only practical considerations that lead me to believe. But as we have seen, cases in which my tendency to believe “survives” are not only rare; they also occur purely by chance (which also implies that we cannot even rely on their occurrence when it comes to particularly important beliefs). So although these considerations cannot ultimately reject the Permissibility Account, the randomness of its applicability counts strongly against the Permissibility Account.

8.

Conclusion

Let us now gather the results of our investigation. In trying to answer our central question which conception of an ethics of belief – Clifford’s or his critics’ – is the most plausible, we have reached a philosophical dead end: a number of diffcult empirical questions concerning the consequences of believing upon insuffcient evidence on which the answer to our central philosophical question depends. We have tried to evade such a complicated empirical inquiry and to stay in the realm of philosophy by examining whether the philosophical debate on doxastic voluntarism could help us answer our central question. We end up with a somewhat mixed result. On the one hand, we cannot ultimately answer our central question: neither is Clifford’s conception not affected by the debate on doxastic voluntarism, nor are Clifford’s and his critics’ conceptions “on a par” with regard to this debate. On the other hand, we have seen that Clifford’s conception defnitely fares better with regard to some important and mostly well-accepted results of this debate.

42

Martina Lindner

But what have we really gained? Doesn’t the answer to our central question still depend on the aforementioned empirical questions concerning the consequences of believing upon insuffcient evidence? Could we ever achieve any defnite result without answering them? I think, we did achieve a result that is independent from the answer to these empirical questions. Suppose that the empirical questions concerning the consequences of believing upon insuffcient evidence would all have to be answered in favour of the critics’ positions, i.e. believing upon insuffcient evidence would not, or only to a small degree, corrupt our aggregate of beliefs and would not, or only to a small degree, make us credulous. Is it thus clear that the Permissibility Account or the Requirement Account is right? By no means, I think. Take the Permissibility Account frst: if the empirical questions would have to be answered in its favour, then the Permissibility Account would permit believing upon insuffcient evidence in a lot of cases (as there are a lot of cases in which the positive consequences would override the negative ones) – but we have seen that in many cases one would not be able to make use of that permission, since acquiring a belief upon insuffcient evidence is possible only in rare cases. Similar considerations tell against the Requirement Account, for it would require us to believe against the evidence quite frequently. But as we have seen, it is scarcely possible to infuence one’s belief formation so as to adopt a belief that is not supported by the evidence. And since no one is obligated beyond their ability, an account of an ethics of belief that requires us to believe against the evidence quite frequently is not feasible.25

Notes 1. McCormick calls these two lines of argument for evidentialism “Conceptual Defenses of Evidentialism” and “Normative Defenses of Evidentialism” (McCormick 2015: 15–51). 2. For proponents of such a view see e.g. James (1896), Haack (2001), and McCormick (2015). 3. More precisely, I will call any conception of an ethics of belief that includes the claim that there are cases in which it is permissible to believe upon insuffcient evidence a Permissibility Account. The conception does not have to be based on a consequentialist framework. But I think that a Permissibility Account is persuasive especially within such a framework, and so I will restrict my investigation to the consequentialist variant (to which I will refer as “the Permissibility Account” in what follows). Analogous terminological remarks apply to my use of “Requirement Account.” 4. Jack Meiland argues against Clifford’s position exactly with reference to their shared argumentative framework. He holds that every evidentialist ethics of belief that rests on a practical justifcation is itself “open to a critique based also on ‘practical’ reasons” (Meiland 1980: 530). Mitova (2008a) attempts to show that epistemic norms cannot be justifed pragmatically but concedes a special status to Clifford’s position (cf. ibid.: 151). 5. A “full-fedged” Requirement Account is rarely found – Meiland (1980) is one of the few authors who argue for one. Recently, Rima Basu (2019) has developed a similar account which includes the claim that one is sometimes required not to believe upon what evidentialists might regard as suffcient evidence.

Doxastic Voluntarism and Clifford

43

6. Brian Zamulinski also makes use of the debate on doxastic voluntarism to argue for Clifford’s conception and against some kind of Permissibility Account, but his argument depends on the plausibility of Clifford’s claim that believing upon insuffcient evidence makes us credulous (Zamulinski 2002: 445). Thus, it relies on one of those empirical claims in Clifford’s text that are – as I have stated before – highly controversial and diffcult to verify. That is why I do not make use of Zamulinski’s argument here. 7. Cf. Meiland (1980: 528–529), where he argues that the question of whether we can voluntarily control our beliefs is a problem for all conceptions of an ethics of belief alike. See also Huss (2009: 250). 8. In my outline, I partly draw on classical texts of the debate on doxastic voluntarism. The current debate is not in all its facets applicable here. For my central question I am concerned with whether we can acquire or give up a belief by willing to do so. So I leave aside such facets of the current debate like whether on a compatibilist account – in contrast to a libertarian – belief is free (cf. e.g. Ryan 2003) or whether we can be held responsible for our beliefs even though we cannot believe at will (cf. e.g. McHugh 2013). 9. In case you do not agree with my outline of the debate on doxastic voluntarism, you can treat my arguments as some kind of hypothetical consideration: If I were among those who accept these claims, then I could infer such-and-such results for the debate on Clifford’s ethics of belief. Of course, it might be more desirable to base my arguments only on generally accepted claims – but that would exclude reference to any current philosophical debate. 10. That I call this kind of means “manipulative” stems from Alfred Mele’s usage of the term in his treatment of self-deception. Mele states that our desire that some proposition p is true can lead us “to manipulate (i.e., to treat inappropriately) a datum or data relevant, or at least seemingly relevant, to the truth value of p” (Mele 1987: 127). Such manipulation may, for example, consist in “selective evidence-gathering” (ibid.: 126) – and so in making use of just that kind of means that I have described here. 11. One and the same action, e.g. talking to proponents of a given view or reading a certain book, may serve both as an evidence-oriented means and as a manipulative means; to which class it should be counted depends on the particular intention, i.e. whether it is part of a thorough investigation or whether it is made use of to acquire a particular belief. Anne Meylan (2013: 68) puts forth a classifcation of kinds of control with regard to belief (“theoretical control” and “Pascalian control”) that is analogous to my classifcation of means for the acquisition of belief. 12. For a similar story to illustrate the possible effcacy of pragmatic considerations and a similar constraint with regard to the possibility of consciously holding a belief while thinking that one believes it for pragmatic reasons, see Cook (1987). 13. Even Richard Foley argues quite similarly that manipulative means can only be effective if one is not conscious of having based one’s belief on them (cf. Foley 1994: 41). 14. Meylan conceives of this assertion as a reformulation of Williams’s. 15. I cannot argue for this claim at length here. For further arguments in support of it (as well as for other authors who hold it) see the references mentioned earlier. For instance, Meylan defends her claim by showing that it accounts for the phenomenon of transparency with regard to beliefs that has been prominently analyzed by Nishi Shah (cf. Meylan 2013: 56; see also Shah 2006). 16. It is for example William Alston who points out that the volitional effect we have on our beliefs is no “control” but just an “infuence” (1989: 134–137).

44

17.

18.

19. 20.

21.

22.

Martina Lindner Meylan attempts to show, contra Alston, that although making use of the kind of means that I have called manipulative may not always be effective or may even be ineffective in most cases, one can be said to exercise control over one’s belief (Meylan 2013: 79–82). But, as far as I see, Meylan’s arguments are not relevant for my considerations here. Meylan is primarily concerned with showing that we can be held responsible for our belief acquisition; so she attempts to establish a conception of control that accounts for such an ascription of responsibility (cf. ibid.: 10). But I am primarily concerned here with the question of whether making use of manipulative means is effective and so the chances to succeed in acquiring the desired belief are crucial for my question of whether we can control our beliefs. Both Alston’s considerations on this point and Meylan’s critique of these considerations do not seem to be applicable to my answer to the question of whether we can control our beliefs by making use of evidence-oriented means. Alston states that control can only be ascribed to us if we could successfully make use of evidence-oriented means to acquire a particular belief – and he shows that we cannot do so (cf. Alston 1989: 130–132). Meylan holds that for the ascription of control it is suffcient to be able to control the acquisition of a “determinable” belief concerning a given topic – but not the acquisition of a “determinate” belief that such-and such is the case concerning that topic (cf. Meylan 2013: 70–78). But I am concerned here with whether we can make sure that we always believe according to the evidence at all, i.e. whether we can make sure that no non-evidential factors infuence our beliefs. Alfred Mele’s explanation of “akratic belief” can further clarify why we ultimately lack control over our beliefs. The term “akratic belief” refers to a phenomenon in which a lack of control over my beliefs is apparent, namely, to the phenomenon that I may be convinced that all things considered I should not believe a certain proposition but nevertheless believe it. According to Mele akratic belief is possible because “the motivational force of a reason” for belief “may not be in alignment with the agent’s evaluation of that reason” (Mele 1987: 115). Such an argument is presented e.g. by Alston (1989: 134), who does not himself embrace it. With regard to beliefs one already has there is an inconsistency in Clifford’s position. Clifford himself points to the diffculty of giving up a belief one already has: “No man holding a strong belief on one side of a question, or even wishing to hold a belief on one side, can investigate it with such fairness and completeness as if he were really in doubt and unbiased” (Clifford 1877: 181). Yet, he does not abstain from formulating his Principle in a way that obviously also includes beliefs one already has. So if the aforementioned vulnerability of Clifford’s conception can be partly removed by the previous considerations on claim A, these considerations may be part of a rational reconstruction of Clifford’s position that seeks to overcome this inconsistency. This clarifcation of the case draws on Mele’s explanation of self-deception. Mele claims that one’s desire that p may lead one to misinterpret data which do not support p as supporting p and vice versa (cf. 1987: 125–126). Such misinterpretation may be intentional or unintentional (cf. ibid.: 128–129). That the shipowner cannot believe according to the evidence here although he wants to may be due to unintentional misinterpretation of the data. A well-known example is William James’s contention that in some cases of religious belief evidence can only be obtained with at least some prior

Doxastic Voluntarism and Clifford

45

inclination to believe: “We feel [. . .] as if evidence might be forever withheld from us unless we met the hypothesis [i.e. the proposition that a personal God exists] half way” (James 1896: 31). 23. Cf. William Alston’s examination of “whether, or to what extent, it is in our power to carry out an intention to take up a certain propositional attitude” (Alston 1989: 136, emphasis mine) in which he states that we have no control over most of our beliefs “and that for the others we have, at most, some spotty and unreliable control of the long-range sort” (ibid., emphasis mine). 24. Rinard alleges a similar example for her claim that there are “cases in which purely pragmatic, non-evidential considerations do seem to constitute good reasons for belief” (Rinard 2015: 210). 25. I would like to thank Sebastian Schmidt and Winfried Schröder for helpful comments on an earlier draft of this paper.

References Alston, William P. (1989): “The Deontological Conception of Epistemic Justifcation”, in his Epistemic Justifcation: Essays in the Theory of Knowledge, Ithaca: Cornell University Press, 115–152. Basu, Rima (2019): “The Wrongs of Racist Beliefs”, Philosophical Studies 176, 2497–2515. Chignell, Andrew (2018): “The Ethics of Belief”, in: The Stanford Encyclopedia of Philosophy (Spring 2018 Edition), ed. by Zalta, URL = . Clifford, William Kingdon (1877): “The Ethics of Belief”, in: Clifford: Lectures and Essays, Vol. 2, ed. by Stephen/Pollock, London: Macmillan 1879, 177–211. Cook, J. Thomas (1987): “Deciding to Believe without Self-Deception”, The Journal of Philosophy 84, 441–446. Foley, Richard (1994): “Pragmatic Reasons for Belief”, in: Gambling on God: Essays on Pascal’s Wager, ed. by Jordan, Lanham: Rowman & Littlefeld, 31–46. Haack, Susan (2001): “‘The Ethics of Belief’ Reconsidered”, in: Knowledge, Truth, and Duty: Essays on Epistemic Justifcation, Responsibility, and Virtue, ed. by Steup, New York: Oxford University Press, 21–33. Huss, Brian (2009): “Three Challenges (and Three Replies) to the Ethics of Belief”, Synthese 168, 249–271. James, William (1896): “The Will to Believe”, in: James: The Will to Believe and Other Essays in Popular Philosophy, Cambridge, MA: Harvard University Press 1979, 13–33. Jones, Ward E. (2002): “Explaining Our Own Beliefs: Non-Epistemic Believing and Doxastic Instability”, Philosophical Studies 111, 217–249. Jordan, Jeff (1996): “Pragmatic Arguments and Belief”, American Philosophical Quarterly 33, 409–420. Kelly, Thomas (2016): “Evidence”, The Stanford Encyclopedia of Philosophy (Winter 2016 Edition), ed. by Zalta, URL = . McCormick, Miriam Schleifer (2015): Believing against the Evidence: Agency and the Ethics of Belief, London: Routledge.

46

Martina Lindner

McHugh, Conor (2013): “Epistemic Responsibility and Doxastic Agency”, Philosophical Issues 23, 132–157. Meiland, Jack (1980): “What Ought We to Believe?”, reprinted in: The Theory of Knowledge: Classical and Contemporary Readings, ed. by Pojman, Belmont: Wadsworth 2003, 526–536. Mele, Alfred R. (1987): Irrationality: An Essay on Akrasia, Self-Deception, and Self-Control, New York: Oxford University Press. Meylan, Anne (2013): Foundations of an Ethics of Belief, Frankfurt: Ontos. Mitova, Veli (2008a): “Why Pragmatic Justifcations of Epistemic Norms Don’t Work”, South African Journal of Philosophy 27, 141–152. Mitova, Veli (2008b): “Why W. K. Clifford Was a Closet Pragmatist”, Philosophical Papers 37, 471–489. Rinard, Susanna (2015): “Against the New Evidentialists”, Philosophical Issues 25, 208–223. Ryan, Sharon (2003): “Doxastic Compatibilism and the Ethics of Belief”, Philosophical Studies 114, 47–79. Shah, Nishi (2006): “A New Argument for Evidentialism”, The Philosophical Quarterly 56.225, 481–498. Williams, Bernard (1973): “Deciding to Believe”, in his Problems of the Self, Cambridge: Cambridge University Press, 136–151. Winters, Barbara (1979): “Believing at Will”, The Journal of Philosophy 76, 243–256. Wood, Allen W. (2002): “W. K. Clifford and the Ethics of Belief”, in his Unsettling Obligations: Essays on Reason, Reality and the Ethics of Belief, Stanford: CSLI Publications. Zamulinski, Brian (2002): “A Re-Evaluation of Clifford and His Critics”, The Southern Journal of Philosophy 40, 437–457.

3

Believing as We Ought and the Democratic Route to Knowledge Matthew Chrisman

1.

Introduction

Here is one way to think of the theory of knowledge and some of its recent history: Knowledge is a relation an individual can stand in to true propositions, and epistemology’s job is to fgure out what this relation is like. To do so, we start out with the platitude that knowing p requires believing p truly, and then we go off searching for the extra ingredient that makes the difference between true belief and knowledge. In the heyday of naturalized epistemology, it was tempting to think this extra ingredient could be some nonnormative (usually modal) feature of the belief, such as its reliability, sensitivity, or safety. But these modal properties come in degrees, so we might need a normative relation after all (something like being as reliable, sensitive, or safe as the relevant kind of belief ought to be). Moreover, another strand in the theory of knowledge has long stressed the connection between knowledge and normative standards for belief such as being reasonable, rationally supported by the evidence, warranted, and/or the manifestation of intellectual virtue. This encouraged a normative approach to the theory of knowledge, which is my starting point here. Assuming knowledge requires true belief, epistemology’s job would then seem to be one of fguring out what normative standard an individual believer must meet in order for her particular true belief to count as knowledge. And coming to this via the “extra ingredient” idea has encouraged a lot of contemporary epistemology to follow a methodology of contrasting cases. That is, we consider a case of some individual believer who we intuitively think knows that p and contrast this with cases of some other individual believer in the same true proposition who we intuitively think doesn’t know that p; then we try to determine what the normatively relevant difference is between the cases. On this backdrop, my aim in this paper is twofold. Ultimately, in Sections 6–7, I want to contribute to the motivation to and development of an alternative approach to the theory of knowledge, which has emerged in recent years, centering on the social role of knowledge in communities of interdependent believers with diverse perspectives on the world,

48

Matthew Chrisman

rather than a methodology of contrasting cases about individual believers. It’s not that I think intuitions about those cases are worthless, but I suspect they are symptoms of a deeper albeit more implicit understanding of knowledge developed through our lived experience as members of epistemic communities, where various normative standards are at play in regulating our economy of shared knowledge. If we can make this social epistemological understanding more explicit, I think we will have a better shot at an illuminating theory of knowledge. Initially in Sections 2–5, in order to explain why I think the more individualistic approach to the theory of knowledge characteristic of traditional epistemology is problematic, I want to consider an important objection to the normative approach to the theory of knowledge, pursued through the methodology of contrasting cases. Although this will need to be refned later, the objection is basically that, if we’re going to reasonably say (in the way characteristic of genuine normative evaluation) that you ought to do something, then it must be up to you whether you do it; but what you believe isn’t (typically or perhaps ever) up to you, the individual believer. So there is a false presupposition of can in the idea that there are ways individuals ought to believe that would make the difference between an individual knowing and not knowing. If you’re an epistemologist engaged in developing a theory of knowledge, especially if you are pursuing the normative approach, then you are probably already thinking of ways to respond to this objection. Maybe there’s a kind of autonomous activity closely related to belief that underwrites our holding individual believers to a standard of what they epistemically ought to believe despite much belief being automatic, involuntary, or otherwise outside the scope of autonomous action. Or maybe there’s a distinctive “purely evaluative” kind of normativity that doesn’t presuppose any kind of autonomy, and epistemic norms should be assimilated to that kind of normativity. In my experience, however, making out such responses is not as easy as it looks. So, in the central parts of this paper, I want to consider some individualistic ways of responding to this ‘ought’ implies ‘can’ objection that I fnd attractive in various respects but then explain why I don’t fnd any of them fully satisfactory. The take-home message will be that they all leave out a kind of autonomous agency we enjoy only through our sociality. In political philosophy, this idea of a kind of autonomous agency enjoyed only through sociality is often attributed to Rousseau. I won’t get into any of the details of Rousseau’s political morality here, but my organizing intuition is that recognizing a kind of intellectual autonomy based in our sociality (alongside Rousseau’s notion of a kind of political autonomy based in our sociality) can help us to make better sense of normative standards a belief must meet in order to count as knowledge, even while granting that most individual belief is automatic, involuntary, or otherwise outside the scope of autonomous action.

Believing as We Ought

49

Even with recently increased interest in social epistemological issues, the standard (conservative) order of approach is to start with individual knowers and try to fgure out what makes them knowers; then on this basis we to seek to answer social epistemological questions about things like the nature of knowledge through testimony, highly complex group knowledge, knowledge extended through technology, and social/political structures needed to recognize, protect, and promote knowledge. I am interested in all of these questions, but I suspect the order of approach needs to be reversed: frst social epistemology, then individual epistemology. And I intend my suggestion that there is a kind of intellectual autonomy based in our sociality to provide important groundwork for reversing it, thereby suggesting an alternative methodology for the theory of knowledge.

2.

The ‘Ought’ Implies ‘Can’ Objection to Norms of Belief

If we are pursuing the theory of knowledge individualistically and normatively, then we assume that there are true propositions one ought to believe and true propositions one ought not to believe; and this ‘ought’ is understood to be distinctively epistemological – i.e. not moral, prudential, or anything else. Believing as one ought, in this sense, is a normative standard one must meet in order to count as knowing.1 This is why one central way to criticize someone epistemologically is to suggest that they believe something that they (epistemologically) ought not to believe. The main challenge to this idea, raised prominently by Williams (1970) and Alston (1988), relies on the assumption that genuine normative demands on someone to φ presuppose that φ-ing is the sort of thing they can decide to do, or voluntarily do, or freely exercise their will in doing, or something like that. I’ll consider challenges to this assumption later, and we should note that the problem looks slightly different depending on how one refnes it; but the initial point is that believing doesn’t seem to be like autonomously doing something. For one thing, believing is not typically conceived as something one does but rather a way that one is. That’s because belief is typically and for good reason assumed to be a mental state rather than a mental action or activity.2 But there’s a deeper worry here based in a picture of our mental lives that is radically at odds with the (caricature) picture sometimes thought to be implicit in a lot of traditional epistemology. The supposedly traditional picture is one whereby the paradigmatic cognitive activity is deliberation, which is conceived as a process of considering one’s evidence, or what one takes to be reasons, for and against a proposition p in the course of deciding whether to believe p. On this picture, although belief is stative, it is understood to be the product of something active: deliberation, i.e. consciously weighing up and responding to reasons. But

50

Matthew Chrisman

as Williams and Alston suggest and as much recent work in cognitive psychology and behavioral economics seems to confrm, a more realistic picture of our mental lives is one according to which most beliefs are clearly not the product of anything like conscious weighing up and responding to reasons.3 It may be tempting to respond by arguing that the mental states reached in ways other than deliberation are proto-beliefs or posited degrees of confdence; or maybe they are instances of “animal belief” posited as representational states useful for explaining some behavior but distinguishable from the fully human beliefs that are the subject of traditional epistemology and the possible objects of normative evaluation. For example, de Sousa argues that we should separate the kind of subjective confdence that fgures in prediction and explanation of behavior, both animal and human, and the mental states classically called ‘belief.’ He writes, [t]he Bayesian theory provides us with a way of describing the mechanism of nonverbal deliberation in humans and other animals. At a more sophisticated level, our own reliance on language adds incalculably to the scope and complexity of our deliberations, by providing us with a stock of sentences accepted as true. This is what the classical notion of belief is designed to capture. de Sousa 1971: 57–58 And Evans infuentially distinguished between ‘information states’ and ‘beliefs,’ writing, It is as well to reserve ‘belief’ for the notion of a far more sophisticated cognitive state: one that is connected with (and, in my opinion, defned in terms of) the notion of judgement, and so, also, connected with the notion of reasons. Evans 1982: 124 However, these “levels of mentality”4 seem to me to cut across possible objects of knowledge or at least objects of what we should be trying to understand when developing a theory of knowledge. We could of course reserve the word ‘belief’ for mental states only sophisticated and linguistically capable reasoners acquire through deliberation on reasons, but then we will just need another word for the mental state, which, when true, can count as knowledge provided that its possessor meets some normative standard. So, I’m inclined instead to recognize lots of different levels or kinds of belief. Moreover, even among the kinds of belief only accessible to sophisticated and linguistically capable humans, it remains highly implausible that most of them are the product of deliberation about reasons.

Believing as We Ought

51

To see why, consider the following. Most beliefs about our perceptible environment, even sophisticated ones supposedly available only to linguistically capable beings, appear to be automatic and unchosen (e.g., ask yourself whether you deliberated over whether to believe there are words on this page). Indeed, as Hume (1739/1740, I.iv) recognized long ago, it is actually quite important for our survival for most ordinary beliefs, especially those about useful things in our immediate environment, to be shielded from the slow and highly fallible processes of human deliberation.5 A further part of this picture is that beliefs, especially those about humdrum though unperceived facts, are stereotypically formed via intuition or testimony without anything like explicit deliberation (e.g., ask yourself whether you at some point chose to believe that the currency of Mexico is the peso and did so for reasons you weighed up or even considered). Forming beliefs automatically and intuitively rather than based on deliberation is a cognitive necessity; without it we’d commonly suffer paralyzing processing overloads. And, to pick out a different kind of example, contemporary political epistemology and cognitive psychology is increasingly coming to understand how many of our beliefs are strongly infuenced by typically unconscious heuristics and biases. Moreover, what we regard as evidence for and against controversial propositions seems to be tacitly though strongly framed by background ideologies on which we rarely refect explicitly. Nothing in this picture of human cognitive psychology implies that we never form beliefs by deliberating on evidence or reasons, nor does it entail that it is impossible to decide to believe, or to believe voluntarily, or to exercise one’s will in believing. Of course, we philosophers are very familiar with being unsure about some argument or objection, thinking about it for a while, discussing it with colleagues and students, and fnally deciding what should be believed. And there may be special cases of belief that are susceptible to explicit decision. For example, Velleman (1985) argues that deciding to perform some action, such as to take a walk, is identical to forming the belief that one will perform the action; and surely in normal circumstances one can voluntarily decide to take a walk, and indeed exercise one’s will in so deciding. So, if Velleman is right (which I don’t mean to assert that he is), then at least that kind of belief is subject to autonomous decision. In a different vein, McCormick (2011, 2015) has argued that, in cases where the evidence isn’t conclusive, one can act on practical reasons in freely choosing to believe, e.g., that there is a God or that one’s spouse is innocent. So, if McCormick is right (which I don’t mean to assert that she is), it would appear that some beliefs can be voluntarily chosen. To a proponent of the previous picture of human cognitive psychology, however, these kinds of cases will at best look like exceptions that prove the rule, not paradigm cases of belief formation. And, if that’s right, then epistemologists pursuing the normative approach to the theory of

52

Matthew Chrisman

knowledge still face a challenge stemming from the apparent involuntariness of belief. We want the theory of knowledge to apply to all knowledge, or at least all paradigm cases. Yet, in a lot of paradigm cases, it doesn’t look as if we have the kind of voluntary control, freedom, autonomy, or whatever that is presupposed when we hold individual believers to normative standards about what they ought (epistemologically) to believe. As I see things, this ‘ought’ implies ‘can’ objection is the fundamental challenge to a normative approach to the theory of knowledge, which is one of the dominant approaches in contemporary epistemology.6 Next, I want to consider three responses to the challenge, all of which have some attraction but none of which is fully satisfactory, in my view. The frst response can be classifed as seeking to identify a kind of genuine epistemological norm or ‘ought’ which doesn’t however imply individual believers have autonomy over what they believe. The second and third responses can be classifed as seeking to identify a kind of autonomous activity individuals engage in that is intimately related to belief but different from paradigm autonomous action.

3.

Norms of Belief as Evaluative Rather Than Prescriptive

The frst response to the objection that I want to consider rejects the implication from ‘ought’ to ‘can.’ Of course, we know that the word ‘ought’ is used to make probabilistic claims that have nothing to do with things like decision, voluntary control, or autonomy. The weather reporter’s suggestion that the storm ought to hit land by nightfall doesn’t impute any agency to the storm, and it is not a normative evaluation of the storm in the way that claims about what people epistemologically ought to believe are normative evaluations of believers. So a proponent of this response needs a genuinely normative sense of ‘ought’ that nonetheless doesn’t imply ‘can’ in the sense of being something up to the person evaluated. Is there such a sense of ‘ought’? We might fnd something like this sense of ‘ought’ in the suggestion that, while paradigmatic norms of action are prescriptive, some other norms are “evaluative” rather than prescriptive (cf. McHugh 2012a, 2012b). What this means is that the function of these norms in our thinking is to set a standard for evaluating something rather than prescribing action. However, it can’t be any old standard that is an evaluative norm. Being as tall as George Clooney is a standard we can use to evaluate heights, but this isn’t (all by itself) a normative standard; a normative standard is something like being as tall as one ought to be in order to play the leading role in some movie. What does it take, in general, for a standard to be “normative” in this evaluative rather than prescriptive sense? I don’t really know. In fact, I dislike the idea of so-called evaluative norms and would prefer to insist

Believing as We Ought

53

on a stricter separation of normative notions from evaluative notions. But I think there is something right about the idea that some ‘ought’ claims advert to a standard of something’s being good or ideal in some respect, without prescribing that thing’s doing anything. Applied to epistemological ‘ought’s then, the idea is to argue that when we make claims about what someone ought to believe, we are adverting to a “normative” standard for belief regarding what would make a belief good or ideal in some particularly epistemological respect, without prescribing that belief. And this fts well with a normative approach to the theory of knowledge which seeks to identify the “normative” standard, whose satisfaction makes the difference between cases of true belief and cases of knowledge. Moreover, this idea offers a different route for responding to the ‘ought’ implies ‘can’ objection. For we often deploy these so-called “normative” standards in evaluating things without implying that those things are within the purview of voluntary control, free choice, autonomy. Human intestines ought to maintain a range between 4.0–7.0 pH. It’d be silly to object to this claim by arguing that human intestines don’t freely choose their pH values or can’t manifest autonomous agency in maintaining them; and, while humans may have some indirect control over the pH value of their digestive system through their dietary choices, the oughtclaim does not usually seem like an application of a normative standard meant to indirectly guide dietary choice. Similarly, maybe there are kinds of beliefs human minds ought to maintain (given some body of evidence), and we should understand this claim as adverting to an evaluative rather than prescriptive epistemological norm. If that’s right, then maybe we could respond to the objection to norm governance of belief by maintaining that the relevant ‘ought’s don’t imply ‘can,’ in the relevant sense, and so there is no problem with accepting the idea that belief is not voluntary, freely chosen, or autonomous but nonetheless subject to epistemological norms. Despite not liking the notion of a “merely evaluative” norm, I like this response to the objection so much that I’ve published a paper defending a variant of it (Chrisman 2008). What I particularly like is the fact that, because beliefs are states, claims about what someone ought to believe seem apt to treat as claims about how someone ought to be rather than what someone ought to do. Moreover, if we conceive of knowledge as true belief that also meets some standard of being what someone ought (epistemologically) to believe, I think it seems right to think of this oughtclaim not as prescribing belief formation but as evaluating a belief already formed for how well it meets some standard. I’m sensitive, however, to the worry that belief is importantly different from the pH values of our intestines. So-called evaluative norms applied to the latter are purely physiological and can be identifed by considering the biological function of intestines and determining what pH values

54 Matthew Chrisman promote that function. Can we do something similar to identify evaluative norms for belief? Maybe, but it is much less clear what function beliefs perform and so more diffcult to derive claims about what one ought to believe from a view about the function of belief. Sure, we can say beliefs are for storing information that may be useful in navigating the world in pursuit of the satisfaction of our desires. But I doubt that all beliefs are for performing that function (moral beliefs, religious beliefs, philosophical beliefs, and pure mathematical beliefs all seem like salient counterexamples). And, anyway, epistemological evaluations of belief are not sensitive to whether or not maintaining them promotes desire satisfaction; they’re sensitive to considerations about what epistemological reasons there are to believe something, which are the kinds of reasons that relate to knowledge, not practical success in satisfying desires. Partly because of this contrast, in my earlier paper, I argued that epistemological ‘ought-to-believe’s were mainly interesting for how they related to prescriptive claims about what a believer’s epistemic community, which might include him/herself but which also includes other people, ought to do about the believer having the beliefs he/she ought to have.7 In this way of thinking of things, evaluative ‘ought’s about belief get replaced with prescriptive ‘ought’s about inquiry and instruction. I still think there is a kernel of truth in this idea, and it fgures in the more social sort of approach to the theory of knowledge that I want to motivate later. As far as it has been developed so far, however, I now worry that the idea that claims about what someone ought to believe advert to norms about how inquirers and instructors ought to behave risks collapse into to some version of the indirect control response I want to discuss next. For it sounds as if ‘ought-to-believe’s inherit their normativity from their indirect connection to various prescriptive norms involved in the governance of actions of inquiry and instruction.

4.

Indirect Control

The next response I’ll consider is an expansion and refnement of the supposedly traditional picture of our mental lives. The idea is to argue that, as long as paradigmatic belief is subject to indirect voluntary control, even if this doesn’t often come in the form of deliberation, that’s enough to secure the kind of autonomous agency over belief needed to answer the objection. In more detail, someone pursuing this line of thought might argue that many states are like belief in that one often just fnds oneself in them: e.g. states like being ft, being rich. But we should recognize an “indirect” kind of control we exercise over such states. We understand more or less how to get into one of these states and how to get out of these states if that’s what we take ourselves to have reason to do. (Of course, doing so is not always easy, and we are maybe subject to weakness of will in our attempts to do so.) If it’s right that there are

Believing as We Ought

55

states we more or less know how to get into, it shows that being in a state like one of these is the sort of thing one can decide to do, or do voluntarily, or exercise one’s will in doing. In sum, the response is that being ft and being rich are not literally things one does in some active way, but they are possible products of things one does in response to (perceived) reasons for doing so and so objects of indirect voluntary control.8 Does this provide the resources needed to rebut the ‘ought’ implies ‘can’ objection to holding individual believers to normative standards? I think we should recognize that there are cases where one does not literally decide to believe but one does decide to actively do something in order to make up one’s mind about something. I already mentioned the seemingly special case of belief formed through explicit deliberation about reasons for and against some proposition. But maybe we should also count beliefs as indirectly voluntary when they are formed via other activities one engages in to fnd things out: e.g., investigating the source of some sound, looking things up on Wikipedia, listening to the news, reading a book, etc. As such, it might begin to look like we have something more than just an exception to prove the rule. If these actions are all indirect ways to exercise control over one’s beliefs, then perhaps we do have enough “autonomous doxastic activity” on the scene to make sense of subjecting believers to normative standards in evaluating whether or not they have knowledge. Like I said, I fnd something attractive in each of the responses to the ‘ought’ implies ‘can’ challenge I’ll discuss here, but I’m not completely satisfed with any of them. Regarding this response, I have two objections. My frst objection is that there is an important difference between belief and the other states given as examples of indirect voluntary control. The normative claims “I ought to be ft” or “I ought to be rich” can guide activity that even if not always easy to execute has the predictable result of one’s being in the state. Belief, it seems, doesn’t work this way. At least, it is not usual to think “I ought to believe p” and for this normative thought to guide activity resulting in forming the belief p.9 That’s sort of how Pascal suggested one might come to believe in God, but this was for explicitly practical reasons, making the ‘ought’ one of prudence rather than epistemology. Normally, things go the other way around when belief is subjected to epistemic norms. One believes p and then can refect on whether that’s a proposition one ought (epistemically) to believe, perhaps adducing considerations that explain why this is a belief one ought to maintain. Similarly, except in pathological cases like compulsive obsession, if one thinks that one ought (epistemically) not to believe p, that is not anything like the beginning of a plan to get out of the state of believing p but rather the event whereby one ceases to believe. So even though there are states over which we exercise a kind of indirect voluntary control, suggesting that it makes sense to say that one ought to be in them, belief doesn’t seem to interact in the same way with voluntary control in response to norm guidance.

56

Matthew Chrisman

One might object that I’ve misconstrued the way norms guide activity that results in belief. Of course one doesn’t usually think, for any specifc proposition p, “I ought to believe p” and then set about to acquire that belief. But don’t we often think things of the form “I wonder what I ought to believe about x” and then set about investigating x, which we know will result in various beliefs and which we hope to be the beliefs we ought to have about x? Yes, that’s certainly something we sometimes do. But I’d suggest it’s not – and this is my second objection to the indirect-control response to the ought-implies-can worry – how most beliefs are formed. I don’t really know how to quantify “most” here, but investigation into a topic x with the aim of forming the beliefs about x that one ought to have strikes me as only slightly more common than explicit deliberation about reasons for and against some proposition. When you see someone walking by or you’re told that a package will arrive tomorrow or you understand that racism is enjoying political mainstreaming, these are ways of coming to believe something without deliberation or investigation.10 I think epistemologists can and should seek to understand what’s going on in cases of explicit deliberation and investigation; and it seems to me that there is an interesting form of indirect voluntary control over our beliefs in such cases. But my present suggestion is that other cases which do not involve active deliberation and investigation are also important and central cases of belief formation, and we should demand that a theory of knowledge and doxastic autonomy treat them as such.11

5.

Active Beliefs

The fnal response to the ‘ought’ implies ‘can’ objection that I want to consider is to argue that there is a sense in which believing itself is active after all. And once we recognize the distinctive activity in believing, we will be in a position to understand a kind of autonomous activity that is importantly different from autonomous action but nevertheless suffcient to underwrite holding believers to normative standards. Earlier I pointed out that belief is typically conceived as a mental state, a way one is rather than something one does. So how could belief itself nevertheless be active? The basic idea in this response is to argue that to understand what it is to believe p, we have to appreciate the activities of mind wrapped up with maintaining a belief.12 One doesn’t typically (ever?) pick up and hold a single belief isolated from other beliefs, impressions, and actions. When one believes p, one is defeasibly disposed to immediately form beliefs about the obvious consequences of p if the question of their truth is relevant or salient. One is defeasibly disposed, if one has any cause to refect on it, to feel convinced of p and take there to be evidence for this conviction; one is also defeasibly disposed to distrust those who assert not-p and to be surprised if one has an experience as

Believing as We Ought

57

if not-p. One is also defeasibly disposed to perform various cognitive actions such as to treat p as evidence in deliberations and investigations into other things, rely on the truth of p in various forms of practical reasoning, and refect on the grounds for believing p if one discovers this belief is inconsistent with something else one believes.13 As indicated, all of those dispositions are defeasible, and there are surely others than the ones I just mentioned. But the key point is that our grip on paradigmatic cases of believing p seems to depend crucially on appreciation of some complex disposition to engage in various mental and nonmental activities involved in maintaining the belief p. So, even if we want to insist (in full metaphysical precision) that belief is a state and so not a paradigm activity, we should recognize that it is a state we understand only via a tacit understanding of a complex disposition to engage in intimately related activities. And, according to this second response to the objection, that’s enough to make belief active by proxy in a way that most states are not. As long as maintaining a belief is something one does actively, then it can make perfectly good sense to hold people to normative standards about what they ought to believe. For believing as one ought to is the same thing as maintaining the beliefs one ought to maintain. As before, I think this response to the ‘ought’ implies ‘can’ objection is instructive, but I am not fully satisfed. To begin to explain why, let me point out that even if we count belief itself as active “by proxy” of the activity of maintaining belief, that doesn’t show that we decide to believe, voluntarily believe, or exercise our wills in believing, or anything similar, which is what the ‘ought’ implies ‘can’ objection appears to demand. To see this, notice that there are lots of states we understand only (or at least primarily) by reference to correlated dispositions to activities.14 What is it to be frightened, for instance, except to be disposed to fee, perspire, and have a faster heart rate, form impressions of danger, etc.? Or what is it to be awake except to be disposed to consciously experience certain kinds of things, engage in certain kinds of brain activity, etc.? I don’t know the right account of being frightened or being awake. But my point in mentioning these ideas is that, even if it’s true that we can make sense of these states only via an appreciation of complex dispositions to engage in the activities intimately related to them, that doesn’t show that we typically decide when to be in one of these states or that being in one of those states is voluntary in some sense or that being in one of these states involves some kind of exercise of our will. We can, of course, voluntarily choose to do things that we know to eventuate in being in these states (e.g. watch the horror flm to induce fright, drink the strong cup of coffee to maintain wakefulness). Modeling doxastic autonomy in this fashion, however, is back to the indirect control response to the objection explored in the previous subsection. More to the point, however, even if believing p is to be understood by essential reference to the activities wrapped up in maintaining the belief p,

58

Matthew Chrisman

a proponent of the picture of our mental lives dominant in contemporary cognitive psychology and behavioral economics is still going to argue that a whole lot of the beliefs that we maintain are automatic, intuitive, and/or the products of implicit biases and ideological framing. Perhaps, they might grant, when you see that there are words on this page, you’re engaged in a lot of cognitive and other activity in maintaining the belief that there are words on this page. But still, taken as a whole complex of conditional dispositions, it’s not up to you whether or not to engage in this activity, so why can we hold you to a normative standard regarding this activity? Why think this activity manifests your autonomous agency? Similarly, maybe my maintaining the belief that the currency of Mexico is the peso involves activating a complex disposition to various activities, some of which are clear cases of voluntary action, such as exchanging money for pesos when I cross the border into Mexico. But still, one might reasonably insist that it’s not up to me whether to maintain this belief, so why do I get held to a normative standard regarding it? More abstractly, the worry about the active belief response to the objection is that, although some of the activities dispositionally connected to believing p may be voluntary or things we choose to do, that doesn’t really address the objection, unless one can choose to maintain or can voluntarily maintain or can exercise one’s will in the very activity of maintaining a belief. And, so far at least, it doesn’t seem that one can do this in very many of the paradigm cases of belief.15 In response, one might object that I’ve mischaracterized the activity of maintaining belief or at least the bit of it where voluntary control gets into the picture. For there seems to be a potential mental activity one can voluntarily engage in with respect to any belief. One can notice what one believes about some topic, and then subject that belief to further scrutiny. Even Descartes, arguably, wasn’t suggesting in the Meditations that we can start with a blank slate and then go about adding beliefs to it at will. Rather he is more plausibly read as suggesting that we enjoy the freedom to take a refective stance towards each of our beliefs, gathering further evidence for and against it, and deliberating about whether it is a belief we ought to have. If we count as autonomous only those beliefs for which one has actually engaged in such refective activity, then the sphere of doxastic autonomy will be too small. If, however, maintaining belief somehow involves a stance towards one’s beliefs whereby this kind of refective activity is always possible (even if rarely actual), we may have hit on a conception of the activity of maintaining a belief that extends the sphere of doxastic autonomy wide enough to rebut the ‘ought’ implies ‘can’ objection.16 I think this is the best version of the active-belief response to the objection, and I am very sympathetic. However, I suspect Descartes was wrong if he thought it generally humanly possible to sit alone by the proverbial freplace and notice from a third-personal perspective what

Believing as We Ought

59

beliefs one has about any topic, in order to then subject them to further scrutiny, identifying the bases of these beliefs and considering the support they provide. This is a more general challenge for the idea that we are autonomous in maintaining our beliefs because we always can subject anything we believe to rational refection and investigation. Even identifying what one believes about many topics requires intercourse with other people and is sometimes very diffcult, possibly impossible, to do accurately.17 After all, if most belief is automatic, intuitive, biased, and/or ideologically framed, there’s no guarantee one can just, so to speak, ask oneself what one believes and get a straightforward answer. Moreover, many of the concepts we deploy in believing are externalist, in the sense that their application conditions depend on connections to things in the world including the thoughts and actions of other people. And even more importantly, when it comes to scrutinizing a belief, I think one usually needs to leave the armchair. Of course, one needs to leave the armchair to gather further evidence about empirical facts, but the world doesn’t present its facts to the lonely explorer unvarnished by the natural and social infuences on our belief forming processes. So one needs to leave the armchair and link minds with other believers, in order to go out and investigate the world. This is part of why I believe we need a more socialized approach to the theory of knowledge and understanding the kind of doxastic autonomy enjoyed in certain kinds of epistemic communities. I turn to that in the next section. I want to take a step back to the theory of knowledge approached via a normative standard, but I will suggest this approach is best pursued not in terms of a standard individual beliefs must meet but rather in terms of the role doxastic standards play in generating a kind of autonomy available only through our sociality.

6.

Doxastic Autonomy as Social à la Rousseau

So far I have been exploring an objection to an approach to the theory of knowledge which assumes that the key question in differentiating true belief from knowledge is one about the normative standard believers must meet in order for one of their beliefs to count as knowledge. The objection, in short, is that holding believers to normative standards assumes that believers exercise autonomous agency in holding their beliefs, but this assumption is wrong. Or at least it is mostly wrong: when we look at any individual belief, it will usually be the product of automatic, intuitive, biased, and/or ideologically infuenced cognitive processes; and that’s enough to undermine any assumptions of doxastic autonomy. Although I think the responses considered earlier have some merit, I want to suggest that a full response requires reconceptualizing doxastic autonomy in a more social key. I propose we consider moving away from a focus on individuals and what they can or cannot freely believe to a

60

Matthew Chrisman

focus on communities and the way norm-governance within the context of our sociality facilitates different kinds of autonomy. My idea is to base a fuller understanding of doxastic autonomy in the sort of epistemic sociality involved in our collective enforcement of particular sorts of rules. This is based on an analogous understanding of practical autonomy that emerged in Rousseau’s (1762) response to Hobbes’ (1651) account of the genealogy of political morality. So next I’ll give a thumbnail sketch (with no pretensions of exegetical preciseness) of that important moment in the history of philosophy. Famously, Hobbes imagined pre-moral humanoids to be free in the sense that they are not subject to any moral rules that could be enforced with legitimacy. He then tried to explain how such pre-moral humanoids could have moved from the state of nature into a society where at least someone was in a position to legitimately enforce something recognizable as moral norms. The basic strategy in his explanation was to identify the “natural” freedoms a characteristic individual pre-moral human would be rational to give up in exchange for the security that comes with submission to the social enforcement of a set of rules. And on this basis, he suggested that we should think of ourselves (contemporary and fully social humans) as having autonomously albeit implicitly accepted the legitimacy of social enforcement of those rules, which do limit our freedom. In response, Rousseau granted that pre-moral humanoids would possess a kind of “natural” freedom that we contemporary and fully social humans lack, but he argued that such freedom is different from genuine autonomy. This is because the activities we fully socialized humans characteristically engage in and want to engage in autonomously can only be conceptualized in ways that already presuppose commitment to collective projects that are partly constituted by (tacit) acceptance of norm governance. For example, in Rousseau’s view, genuinely autonomous humans are ones who can do things like get married, play a football game, purchase a vacation home, sing in a choir, say “#metoo,” convict a criminal, fght climate change – activities all of which require cooperation with other people recognized as moral agents to whom one gives the authority to enforce various norms. In this way, Rousseau conceived of the kind of freedom characteristic of autonomy as a possible result of rather than precondition on our sociality as humans subject to the authority of moral rules. There is signifcant discussion in the history of political philosophy about Hobbes’ and Rousseau’s contrasting approaches to political autonomy, and my intention isn’t to contribute to this discussion here. Rather I mention this contrast here in the context of thinking about the kind of autonomy presupposed in holding believers to normative standards because I think it may provide a new perspective on what that kind of autonomy looks like. And this perspective might shed new light on the role of normative governance of belief in the theory of knowledge.

Believing as We Ought

61

More specifcally, in thinking about doxastic autonomy, I think it might prove useful to refect on the kinds of social arrangements that make for more and less autonomous communities of believers. This is in parallel to (though obviously also deeply intertwined with) ideas we might have about the kinds of social arrangements that make for more and less autonomous communities of moral agents. For instance, within a society of people structured wholly as masters and slaves, the slaves almost completely lack practical autonomy. Similarly, we might think that wherever some people in a community are wholly dominated in thought through indoctrination, these people almost completely lack doxastic autonomy.18 On the other side, however, in a society where there are no enforceable norms for how people treat each other, complete anarchy reigns, and we shouldn’t regard this kind of society as fostering practical autonomy. Similarly, if there are no enforceable norms for how people treat each other intellectually, something akin to cognitive anarchy might reign and we wouldn’t then regard this community as mastering enough common understanding of reality to foster doxastic autonomy. That quick sketch is highly unrealistic given strong evolutionary and social forces that seem to prevent extreme domination or complete anarchy, in both the practical and doxastic realms. I hope, however, that the sketch can help to frame an investigation into the elements of a social arrangement that affect how much practical and doxastic autonomy we think there is in a group of people. In order to understand this kind of autonomy, we need to fnd the middle ground between domination and anarchy, in both the practical and doxastic realms, both of which I’m now conceiving of as essentially social. And this middle ground lies, I want to suggest, in the social arrangements that allow for collective pursuit of common practical and intellectual aims, and these social arrangements are ones constituted by (implicit) acceptance of norm governance. What are the common practical and intellectual aims? In its most abstract formulation we might conceive of these aims as living well together and understanding the world (again, obviously, deeply intertwined). Maybe that is too abstract to be anything but a placeholder for a theory, but we can begin to make progress on that theory in the case of collective intellectual aims by noticing that humans (essentially?) have diverse perspectives on the world and different capacities to expertise; and we improve our understanding of the world by sharing our diverse perspectives and integrating our respective expertise. To put it simplistically, you see things from there, I see things from here, we collectively understand what we see by fnding a way to share each other’s perspectives. You have the capacity to learn about x, I have the capacity to learn about y, we both better understand the world when we each go out and learn about one of these things and then fnd a way to integrate what we have learned in a common understanding of some aspect of the world. Of course, it’s not only or even mainly about you and me. Our epistemic lives

62

Matthew Chrisman

involve many overlapping communities of people who share perspectives and integrate expertise as part of improving collective understanding of the world.19 I also want to suggest that this sharing of diverse perspectives and integrating respective expertise depends on mutual and collective albeit tacit submission to epistemic norms. In trusting others in one’s epistemic community, we accept the legitimacy of various epistemic norms. To share perspectives requires more than mere combination; we must apply standards for addressing tensions or incompatibilities in the diverse perspectives. To integrate differing areas of expertise, we must apply standards for organizing various bodies of knowledge along various explanatory and justifcatory dimensions. These standards are epistemological norms. I’m not claiming that there’s one universal set of norms with absolute epistemic authority. Epistemic communities might be usefully evaluated as more and less autonomous depending on how well the norms they mutually (though tacitly) accept promote collective understanding of the world. But there could be different parts or aspects of the world whose understanding is better or worse promoted by different sets of epistemic norms, and there could be different paths to equally good understanding of the world. However, as we move towards the extremes of intellectual domination and intellectual anarchy in imagining different possible epistemic communities, it is less and less plausible that there are mutually accepted epistemological norms in these communities. And the lack of such epistemological norm governance threatens even the most basic or piecemeal collective understanding of the world. On this kind of picture, knowledge is spread out in complicated ways across communities of believers whereby individuals’ knowledge-conferring justifcation for belief often depends on things understood only by other people. And while it might make sense to think of how much of reality particular individuals understand, the degree to which reality is understood and the process of improving that understanding is a collective matter.20 What does this tell us about the kind of doxastic autonomy presupposed by holding believers to normative epistemic standards? My suggestion is that we shouldn’t think of such norms as presupposing some conceptually antecedent sort of capacity to decide to believe or voluntarily choose our beliefs or freely exercise our will in an individual belief. Rather, following Rousseau’s treatment of practical autonomy as emerging from self-submission to normative governance that constitutes collective action, as contrasted with natural freedom, I want to suggest that believers are autonomous (to the degree that they are) in virtue of the way their beliefs contribute to a collective epistemic project of understanding of the world. Participation in this project is constituted by implicit acceptance of the authority of some mutually enforceable epistemological norms. The normative epistemological standards applying to

Believing as We Ought

63

belief on this picture are not prescriptions that could or should guide a believer’s belief-formation. This is a way of agreeing with an aspect of the “evaluative”-norms response to the ‘ought’ implies ‘can’ objection. However, I think it needs to be taken further into a social conception of the nature and role of holding beliefs to (evaluative) normative standards. We may not (usually) get to choose what we believe, but we do get to choose whether to testify to others, contribute what we believe to some group’s overall understanding of some issue, check our beliefs against technological-cum-social repositories of collective knowledge, and appeal to others in appreciating the ideologies and biases infuencing belief formation and information processing in ourselves and the institutions in which we have power. Those are the sorts of autonomous activities that don’t even make sense from the individualistic perspective but which seem to me to be quite central to what it is to be someone who knows something – at least in most of the cases where it might make sense to ask whether someone believes as they ought to believe. This is a way of agreeing with aspects of the indirect-control and active-belief responses to the ‘ought’ implies ‘can’ objection, but I’m suggesting that they too need to be taken further in the social conception of the relevant actions and activities intimately connected to belief. Earlier I expressed a worry about conceiving of claims about what someone ought to believe as indirectly about what they ought to do to get themselves to believe something. The worry was, roughly, that the relevant ought-to-believe claims can guide the believer’s activity only in cases where one takes something like an alienated or therapeutic stance towards one’s own belief. In the normal case, one’s belief that p is antecedent to one’s belief that p is what one ought to believe. But the situation is obviously different with claims about what someone else ought to believe and claims about what “we” ought to believe. A professor can think that her students ought to believe that Kant wrote the First Critique, and there’s nothing mysterious or pathological about that thought guiding her actions of instruction. A political commentator can think that we (members of her epistemic community) ought to believe that racism is enjoying political mainstreaming, and there’s nothing mysterious or pathological about that thought guiding her actions of investigative reporting. Rousseau argued that it was through tacit agreement with a social contract that humans free themselves from our animalistic passions and urges and thereby achieve the kind of autonomy needed to pursue the sorts of robust projects characteristic of a good human life. These projects depend on our ability to pursue the common good, which in turn depends on our ability to understand moral norms as rules we give unto ourselves. By extending this conception of autonomy and norm-governance from morality to epistemology, I think we can make sense of a different – and I submit more interesting – sphere of doxastic autonomy from the ones

64

Matthew Chrisman

often discussed in debates about doxastic voluntarism and normative epistemology. It is tacit agreement with a social epistemic contract that humans free themselves from the limitation of knowledge to things in their immediate environment and thereby achieve the kind of intellectual autonomy needed to pursue the intellectual projects characteristic of a good human life. Just like our practical projects (and as an integral part of them) these epistemic projects depend on our ability to pursue common knowledge and transmit it culturally, which in turn depends on our ability to understand epistemic norms as rules given unto ourselves.

7.

Conclusion: Towards Social Epistemological Approaches to the Theory of Knowledge

At the beginning of this paper, I mentioned the methodology of contrasting cases that has dominated the theory of knowledge, in both its nonnormative and normative guises. By extending the Rousseauean approach to political autonomy to intellectual autonomy, I think we’re in a position to start thinking about other methodologies. In some fairly generic and so not terribly informative way, we might begin with the idea that knowledge is true belief one epistemologically ought to have. Then the more social understanding of knowledge encouraged by the conception of doxastic autonomy just outlined would encourage us to think about this ‘ought’ in terms of what one can contribute to the collective project of understanding the world.21 That is to say, broadly speaking, what one ought epistemologically to believe should be defned in terms of what norms of belief adherence to which will improve our collective understanding of the world. This is the proposed normative difference between mere true belief and knowledge. How do we evaluate which ways of believing will improve our collective understanding of the world? Contrasting cases may help, but I’d suggest that a more illuminating methodology will be to think about the sorts of belief we want to encourage in fellow members of our epistemic community as part of sharing diverse perspectives and integrating different areas of expertise. This might include various ways of believing, and different ways might be appropriate for understanding different aspects of reality or within different epistemic communities depending on their histories and structures. But some of the areas I think we should think about more are testimony, memory, extended information systems, just treatment of other believers. More specifcally, if we’re investigating the epistemological standards believers should be held to in thinking about whether they count as knowers, we should think about the role such knowledge plays in an economy of information within a community of believers. Framed this way, we might come up with claims like the following: One ought to believe in ways one should stand behind in testimony (cf. Goldberg 2010,

Believing as We Ought

65

2018). One ought to believe in ways that cohere not only with refectively accessible knowledge but also technologically and socially accessible knowledge (cf. Goldman 1999). One ought to believe in ways that avoids epistemic injustice, i.e. in a way that meets the social obligation not to discount the testimony of others because of prejudice against the type of person they are (cf. Fricker 2007; Anderson 2012). One ought to believe in ways that acknowledge fallibility in the face of implicit biases. One ought to believe in ways that are self-aware about the unavoidable ideological framing of various issues and the challenges this poses for sharing perspectives. Do these ‘ought’s imply the ‘can’ of autonomous activity? The more social approach to the theory of knowledge I am exploring here suggests that the degree to which we hold each other accountable to such social rules (and the kind of sanction/reward structure attaching to that accountability) is a key part of what makes some communities of believers more autonomous and others less. On this picture, particular belief lies outside the scope of individual choice, and the activity of maintaining a belief may be mostly automatic, intuitive, and even driven by implicit biases and ideologies. But we should think of that activity as embedded in and structured by social rules constituting epistemic communities. And when these social rules promote mutual understanding of the world, they are the sorts of norms, conformity to which fosters the very kind of doxastic autonomy presupposed in saying that someone (epistemically) ought to believe something.22

Notes 1. This is consistent with the idea that, for each true proposition, there is no all-things-considered requirement to know it. I think there are all-thingsconsidered requirements to know some things; however, the kind of normativity I’m focussing on here is domain-specifc and sometimes appealed to in making sense of the difference between knowledge and other true belief. The idea is to consider things people believe and ask whether they ought to believe those things or something different, but only from the perspective of determining whether those beliefs are instances of knowledge. 2. I make this case in more detail in Chrisman (2012, 2017). See also Steward (1997) and Hunter (2001) for important defenses of the idea that beliefs are states a person is in rather than states inside of a person. 3. I have in mind the vast literature stemming from Tversky/Kahneman (1973). But see also Mercier/Sperber (2017) for a picture of reasoning that doesn’t depend on the system-1-vs. system 2 distinction and still portrays human belief formation as mostly unreasoned. 4. Stevenson (2002) argues that there are six different levels of belief and that it is mainly an uninteresting verbal question whether we reserve the term ‘belief’ for the highest levels. I’m inclined to agree and think that belief at each level is a possible object of knowledge. So the theory of knowledge approached as a normative project will need to explain how normative standards can be applied to all of these different kinds of mental states, whether or not we call them all ‘belief.’

66

Matthew Chrisman

5. Moreover, the psychological law-governed and predictable effects of having a belief seem to be more correlated with beliefs one does not consciously deliberate to or about. Cf. Mandelbaum (2016) for a discussion of implicit biases and an interpretation of psychological data supporting this suggestion. 6. Williamson’s (2000) “knowledge-frst” program might be seen as an alternative approach in contemporary epistemology. However, it takes knowledge as primitive and so eschews the project of explaining the difference between true belief and knowledge. Moreover, it carries similar normative commitments in that it sets out knowledge as a normative standard for belief, which raises many of the same questions I will discuss here but in a different key. I set this rich program to one side in what follows. 7. Cf. Kauppinen (2018) for a similar idea linking the genuine normativity of standards for belief to the sorts of accountability appropriate in the believer’s epistemic community for those who violate the standard. 8. See Meylan (2015, 2017) for a powerful defense of this idea. 9. Cf. Nolf (2014) for a more realistic account of how normative evaluation of belief might sometimes affect one’s stance towards this belief and the formation of other beliefs. I regard this as a special case and by no means the usual way that belief is formed. 10. Cf. Nolf (2018: 67–70) for a related argument against the idea that most doxastic agency is exercised in manipulation of one’s environment, e.g. through exposing oneself to various pieces of evidence or putting into place environmental safeguards against believing wrongly. 11. Objection: Perhaps beliefs, which are indirectly controlled through deliberation or investigation, are not statistically common but still paradigmatic or normal. After all, for any proposition p one believes, can’t one go about investigating whether p is true or deliberate about whether p is true? If yes, we might think that the possibility of indirect control via deliberation or investigation into whether p is true is partly constitutive of a mental state’s being the belief that p, even if most beliefs are not ones actually formed through deliberation or investigation. Response: I’m not so sure one can, even in principle, subject everything one believes to rational refection characteristic of deliberation and investigation. Maybe some beliefs are ineffable. And even if one can, in principle, deliberate or investigate whether p for any p one believes, I suspect one will often need social input to determine what one believes. (More on this later, when I consider a similar idea that one can in principle rationally scrutinize any belief one holds.) Moreover, this objection trades on diffcult issues about how to understand what constitutes a mental state’s being a belief. Since we can’t open up the mind, identify the beliefs, and investigate their nature, I suspect this question is ultimately a task for conceptual engineering. That is to say that we face a question about what conception of belief is best suited for our theoretical purposes. I’m content to recognize a diversity of theoretical purposes for the concept of belief, and maybe the conception of belief advanced by this objection is well suited for some interesting theoretical purpose. However, I’m also inclined to think that this conception is not well suited for the purpose of understanding knowledge in its full generality. Meylan (2017: 14–15, 19–21) argues that we exercise indirect agency far more often than it seems and that we can foresee many doxastic consequences of our actions, and I don’t mean to deny this. It’s just that many of what I see as paradigmatic cases of knowledge (e.g. via perception or testimony or gradual settling into a view that just seems to make sense of a lot of things) don’t seem to me to be like that. Cf. Owens (2000: 87) for a different sort of critique of the indirect agency response to the objection.

Believing as We Ought

67

12. See Boyle (2011) for defenses of the idea that belief is an active state. I address Boyle’s arguments more directly in Chrisman (2018). See also Hunter (2018) for criticism of Boyle and criticism of the view I outline in that paper. Here I’m considering what I take to be an attractive descendent of Boyle’s idea. It’s different from the similar ideas foated in Chrisman (2018) that doxastic agency is located most centrally in the activity of maintaining a whole system of beliefs and Nolf (2018) that doxastic agency is located most centrally in the cognitive processes manifesting our belief-regulating dispositions. In both cases, these views focus on whole belief systems rather than individual beliefs, whereas Boyle focusses on individual beliefs. The picture I develop later is an extension into the social sphere of the idea of locating doxastic agency at the level of systems of belief, which I now view as inherently social in beings like us. Hieronymi (2009) posits a sui generis kind of activeness involved with continually settling a question for oneself that she thinks belief manifests. One way to make sense of this idea is as the view that continually settling a question whether p for oneself is a way of engaging in the activity of maintaining a belief. If that’s what Hieronymi has in mind, it is a version of the idea I discuss in the previous text. In note 15 later, I consider a different way of making sense of that idea based in Hieronymi’s idea of “evaluative control”; see also Setiya (2008) for different critical discussion of Hieronymi’s attempt to make sense of the activeness of belief. 13. Schwitzgebel (2002) develops an account of belief as a complex dispositional state. He argues that there is a stereotype of dispositions associated with various beliefs, and one can manifest the stereotypical dispositions involved in believing p to a greater and lesser degree. 14. See Hunter (2018), who makes a similar point. 15. Some might want to argue that we are nonetheless active in maintaining a belief in a way that distinguishes belief from other mental states such as headaches. For example, Hieronymi (2006) argues that we enjoy “evaluative control” over our beliefs insofar as our maintaining them is responsive to “constitutive reasons” which bear on the truth of their content, which she distinguishes from the “managerial control” involved in responding to “extrinsic reasons” for having various beliefs, i.e. acting so as to come to have various beliefs because it will be good to have them. Since such evaluative control is not a kind of control we have over headaches, does this mark out a distinctive form of mental activity? Sure, but I’m less convinced that it marks out a kind of agency, as most reasons-responsive belief still seems to me to be intuitive, automatic, and unchosen. Moreover, I suspect that there are mental states such as being frightened or being bored that are often responsive to constitutive reasons without manifesting the sort of agency presupposed by holding someone to genuine normative standards. When we say that someone ought not be frightened by the thunderclap, we don’t blame them for being frightened like we blame someone for not believing the obvious consequences of what they already believe. Cf. McCormick (2018), who allows that evidence-responsiveness might not be enough to secure doxastic agency but goes on to argue that we can, in some intuitive cases, believe in response to non-evidential reasons which is characteristic of paradigmatic exercises of agency and which would distinguish believing from being frightened or being bored. 16. Keeling (forthcoming) suggests that one can, at least in principle, determine why one believes what one believes for any belief one holds. She contends that, even if we do not exercise choice or control in forming a belief, we can always ask for anything we believe: “Why (do I) believe p?” And she thinks taking up the deliberative stance and answering “Because q” is a matter of

68

17. 18.

19.

20. 21. 22.

Matthew Chrisman choosing the basis for our continuing to believe p. This is different from the diagnostic stance, whereby one might try to explain what caused one to have the belief in the frst place. And the answer from the deliberative stance is something one can autonomously do (choose the basis) which is also causally effcacious (insofar as one does indeed continue to believe p). Cf. Bem (1970), Gopnik/Meltzoff (1994), Lycan (2008), Dretske (2004), and Carruthers (2009) for various arguments against the idea that we can generally introspect what we believe. This is related to the republican conception of freedom as nondomination (see especially Pettit 1997, ch. 2 and Pettit 2012, ch. 1); however, I go on to suggest that one also lacks doxastic autonomy in an intellectual anarchy. Cf. also Medina’s (2013) appeal to the value of epistemic friction for resisting the threats of epistemic domination. This may wrongly suggest I think everyone in a community has to have the aim of understanding the same part of the world and that the strategies contributing to this aim are always consensual and cooperative. The overlapping nature of various diverse epistemic communities already suggests that there may be a common aim only in some very abstract sense. Moreover, there are surely different competing strategies for pursuing understanding of any part of the world. Cf. Weisberg/Muldoon (2009) on research “followers” and research “mavericks.” Millgram (2015, ch. 2) presents an especially pessimistic account of this. See Nguyen (2018) for a more hopeful response. Cf. Craig (1990) and Rosenberg (2002) for similar attempts to reconceive the methodology for developing a theory of knowledge. For helpful feedback on this paper, I’d like to thank Graham Hubbs, Sophie Keeling, Sebastian Köhler, Sebastian Schmidt, Nick Treanor, and Sam Wilkinson, as well as the audience at the Ethics of Belief workshop in Erlangen in 2017.

References Alston, William (1988): “The Deontological Conception of Epistemic Justifcation”, Philosophical Perspectives 2, 257–299. Anderson, Elizabeth (2012): “Epistemic Justice as a Virtue of Social Institutions”, Social Epistemology 26, 631–673. Bem, Daryl (1970): Beliefs, Attitudes, and Human Affairs, Belmont, CA: Brooks/ Cole. Boyle, Matthew (2011): “‘Making Up Your Mind’ and the Activity of Reason”, Philosophers’ Imprint 11, 1–24. Carruthers, Peter (2009): “How We Know Our Own Minds: The Relationship between Mindreading and Metacognition”, Behavioral and Brain Sciences 32, 121–182. Chrisman, Matthew (2008): “Ought to Believe”, Journal of Philosophy 105, 346–370. Chrisman, Matthew (2012): “The Normative Evaluation of Belief and the Aspectual Classifcation of Belief and Knowledge Attributions”, Journal of Philosophy 109, 588–612. Chrisman, Matthew (2017): “Performance Normativity and Here and Now Doxastic Agency”, Synthese, 1–9. Chrisman, Matthew (2018): “Epistemic Normativity and Cognitive Agency”, Noûs 52, 508–529.

Believing as We Ought

69

Craig, Edward (1990): Knowledge and the State of Nature: An Essay in Conceptual Synthesis, New York: Oxford University Press. de Sousa, Ronald (1971): “How to Give a Piece of Your Mind: Or, the Logic of Belief and Assent”, Review of Metaphysics 25, 52–79. Dretske, Fred (2004): “Knowing What You Think vs. Knowing That You Think It”, in: The Externalist Challenge, ed. by Schantz, Berlin: Walter de Gruyter, 389–399. Evans, Gareth (1982): The Varieties of Reference, New York: Oxford University Press. Fricker, Miranda (2007): Epistemic Injustice: Power and the Ethics of Knowing, New York: Oxford University Press. Goldberg, Sanford (2010): Relying on Others: An Essay in Epistemology, New York: Oxford University Press. Goldberg, Sanford (2018): To the Best of Our Knowledge: Social Expectations and Epistemic Normativity, New York: Oxford University Press. Goldman, Alvin (1999): Knowledge in a Social World, New York: Oxford University Press. Gopnik, Alison/Meltzoff, Andrew (1994): “Minds, Bodies and Persons: Young Children’s Understanding of the Self and Others as Refected in Imitation and ‘Theory of Mind’ Research”, in: Self-Awareness in Animals and Humans, ed. by Parker, et al., New York: Cambridge University Press, 166–186. Hieronymi, Pamela (2006): “Controlling Attitudes”, Pacifc Philosophical Quarterly 87, 45–74. Hieronymi, Pamela (2009): “Believing at Will”, Canadian Journal of Philosophy 35, 135–187. Hobbes, Thomas (1651): Leviathan, ed. by Missner, London: Routledge 2016. Hume, David (1739/1740): Treatise of Human Nature, ed. by Norton/Norton, New York: Oxford University Press 2002. Hunter, David (2001): “Mind-Brain Identity and the Nature of States”, Australasian Journal of Philosophy 79, 366–376. Hunter, David (2018): “The Metaphysics of Responsible Believing”, Manuscrito 41, 1–31. Kauppinen, Antti (2018): “Epistemic Norms and Epistemic Accountability”, Philosophers’ Imprint 18, 1–16. Keeling, Sophie (forthcoming): “The Transparency Method and Knowing Our Reasons”, Analysis. Lycan, William (2008): “Phenomenal Intentionalities”, American Philosophical Quarterly 45, 233–252. Mandelbaum, Eric (2016): “Attitude, Inference, Association: On the Propositional Structure of Implicit Bias”, Noûs 50, 629–658. McCormick, Miriam (2011): “Taking Control of Belief”, Philosophical Explorations 14, 169–183. McCormick, Miriam (2015): Believing against the Evidence: Agency and the Ethics of Belief, London: Routledge. McCormick, Miriam (2018): “Responding to Skepticism about Doxastic Agency”, Erkenntnis 83, 627–645. McHugh, Connor (2012a): “Epistemic Deontology and Voluntariness”, Erkenntnis 77, 65–94. McHugh, Connor (2012b): “The Truth Norm of Belief”, Pacifc Philosophical Quarterly 93, 8–30.

70

Matthew Chrisman

Medina, José (2013): The Epistemology of Resistance, New York: Oxford University Press. Mercier, Hugo/Sperber, Dan (2017): The Enigma of Reason, Cambridge, MA: Harvard University Press. Meylan, Anne (2015): “The Legitimacy of Intellectual Praise and Blame”, Journal of Philosophical Research 40, 189–203. Meylan, Anne (2017): “The Consequential Conception of Doxastic Responsibility”, Theoria 83, 4–28. Millgram, Elijah (2015): The Great Endarkenment: Philosophy for an Age of Hyperspecialization, New York: Oxford University Press. Nguyen, C. Thi (2018): “Expertise and the Fragmentation of Intellectual Autonomy”, Philosophical Inquiries 6, 107–124. Nolf, Kate (2014): “Why Is Epistemic Evaluation Prescriptive?”, Inquiry: An Interdisciplinary Journal of Philosophy 57, 97–121. Nolf, Kate (2018): “Moral Agency in Believing”, Philosophical Topics 46, 53–74. Owens, David (2000): Reason without Freedom: The Problem of Epistemic Normativity, London: Routledge. Pettit, Philip (1997): Republicanism: A Theory of Freedom and Government, New York: Oxford University Press. Pettit, Philip (2012): On the People’s Terms: A Republican Theory and Model of Democracy, Cambridge: Cambridge University Press. Rosenberg, Jay (2002): Thinking about Knowing, New York: Oxford University Press. Rousseau (1762): The Social Contract, ed. by Dunn, New Haven: Yale University Press 2002. Schwitzgebel, Eric (2002): “A Phenomenal, Dispositional Account of Belief”, Noûs 36, 249–275. Setiya, Kieran (2008): “Believing at Will”, Midwest Studies in Philosophy 32, 36–52. Stevenson, Leslie (2002): “Six Levels of Mentality”, Philosophical Explorations 5, 105–124. Steward, Helen (1997): The Ontology of Mind, New York: Oxford University Press. Tversky, Amos/Kahneman, Daniel (1973): “Availability: A Heuristic for Judging Frequency and Probability”, Cognitive Psychology 5, 207–232. Velleman, J. David (1985): “Practical Refection”, Philosophical Review 94, 33–61. Weisberg, Michael/Muldoon, Ryan (2009): “Epistemic Landscapes and the Division of Cognitive Labor”, Philosophy of Science 76, 225–252. Williams, Bernard (1970): “Deciding to Believe”, in his Problems of the Self, Cambridge: Cambridge University Press 1973. Williamson, Timothy (2000): Knowledge and Its Limits, New York: Oxford University Press.

4

Responsibility for Doxastic Strength Grounds Responsibility for Belief Benoit Gaultier

1.

The Problem

The question I’m interested in is: How is it possible for deontic evaluations of beliefs such as the claim that a subject S should not have believed that p at a given time t to be appropriate? This question arises because of the “argument from doxastic involuntarism,” which concludes that such evaluations cannot be appropriate: 1. Our beliefs can be properly subject to negative deontic evaluation iff it’s within our power not to form them. 2. It’s within our power not to form our beliefs iff we have voluntary control over them. 3. The fact that we cannot believe at will in the way we can act at will – in short, doxastic involuntarism – implies that we do not have voluntary control over our beliefs. 4. Therefore, beliefs cannot be properly subject to deontic evaluation; such evaluations, when it comes to beliefs, are inappropriate. The argument from doxastic involuntarism is valid, and I will take for granted that its highly counterintuitive conclusion cannot but be false. This conclusion is, as Sharon Ryan writes, “shocking,” because it implies that “it is never accurate to say that one should have believed differently,” which is an expression that we use “in a literal sense, very often” (Ryan 2003: 49) – i.e. that we very often use deontically rather than for expressing the evaluative judgement that it would have been good or better or ideal that one believes differently. To motivate this assumption, I’ll simply mention the argument that if it is appropriate for one to say that S should not have φ-ed in certain circumstances when her action of φ-ing is enkratic – i.e. results from her

72

Benoit Gaultier

belief that it is perfectly adequate for her to φ in these circumstances – then either of the following judgements has to be appropriate: a. S should not have come to that belief. b. S should not have acted in accordance with it. But (b) comes down to judging that S should have exhibited the kind of irrationality in which akrasia consists (i.e. φ-ing while believing one should not φ), which sounds particularly odd.1 Judgement (a) is much more natural: S should not have come to that belief. Now, (a) implies that beliefs can be properly subject to deontic evaluations. So, if it is admitted that enkratic actions can be properly subject to such evaluations, the same goes with the beliefs from which these enkratic actions result.2 I’ll assume that premise (1) is true because I’ll assume that the “ought implies can” principle is true and that (1) directly derives from it. More exactly, I’ll assume that this principle is true when formulated so as to escape the objection that one can have the obligation to φ at t even when one cannot φ at t but could have made it the case that one can φ at t. For example, I can have the obligation to repay a debt on Monday even if on Monday I cannot honour it but could have honoured it (for example, if I had not spent all my money in bars this weekend). A formulation of the “ought implies can” principle that escapes this objection is: If one ought to φ at t, then one can φ at t, or can put oneself, before t, in a position to φ at t. Correlatively: If one should have φ-ed at t, then one could φ at t or could have put oneself, before t, in a position to φ at t. Premise (1) should then be restated in the following way: 1.* The beliefs we form at any time t can be properly subject to deontic evaluation iff (a) it is within our power, at t, not to form them, or (b) it was within our power, before t, to do things that would have led us not to form them at t. About premise (3) (“The fact that we cannot believe at will in the way we can act at will implies that we do not have voluntary control over our beliefs”) it could be argued that it is false because there are cases in which I clearly have voluntary control over my believing certain things even if I do not believe these things at will in the way I can act at will. Suppose I do not believe in extraterrestrial intelligence, and suppose I know that if I take a certain pill or read a certain book, I will believe after that in extraterrestrial intelligence. Or suppose I know that I have the

Responsibility for Doxastic Strength

73

following special ability: if I say “I believe in extraterrestrial intelligence” three times in a row and then add “Abracadabra!”, I will believe after that in extraterrestrial intelligence. In such cases I clearly have voluntary control over my believing in extraterrestrial intelligence even if I do not believe that at will in the way I can act at will. Admittedly, I can act at will in a certain way so that, as a result, I believe a certain thing. But being able to do certain things T at will does not consist in being able to do some other things T* at will with the result that T obtains. For one to be able to do certain things at will, one has to be able to do these things without having to do some other things. This means that the things one can do at will are those actions that are often characterised in the literature as basic actions. So, in the case just indicated, I do not believe in extraterrestrial intelligence at will in the way I can act at will, but it clearly seems correct to say that I have voluntary control over my believing in extraterrestrial intelligence (in particular in the variant in which I have the Abracadabra ability). And this directly speaks against premise (3). Since what can be done at will are basic actions, it can then be argued that being able to do a certain thing at will is having basic voluntary control over the doing of that thing. Premise (3) should then be restated as follows: 3.* The fact that we cannot believe at will in the way we can act at will implies that we do not have basic voluntary control over the formation of our beliefs. After having restated premises (1) and (3) as I just did, premise (2) should then be modifed as follows: 1.*

The beliefs we form at any time t can be properly subject to deontic evaluation iff (a) it is within our power, at t, not to form them, or (b) it was within our power, before t, to do things that would have led us not to form them at t. 2*a. It is within our power, at t, not to form the beliefs we form at t iff we have basic voluntary control over our beliefs. 2*b. It was within our power, before t, to do things that would have led us not to form these beliefs at t iff we have basic voluntary control over doing these things. 3.* The fact that we cannot believe at will in the way we can act at will implies that we do not have basic voluntary control over the formation of our beliefs. But from these premises, it is obviously invalid to conclude: 4. Therefore, beliefs cannot be properly subject to deontic evaluation; such evaluations, when it comes to beliefs, are inappropriate.

74 Benoit Gaultier Indeed, from premises (1*) and (2*b), it follows that if we have basic voluntary control over doing, before t, things that would have led us not to form the beliefs we form at t, then these beliefs can be properly subject to deontic evaluation. And it follows from (3) that we can have basic voluntary control over our actions. So, it seems that even if it is not within our power, at t, not to form the beliefs we form at t because we do not have basic voluntary control over our beliefs, they can be properly subject to deontic evaluation because, before t, we can have basic voluntary control over actions that, if done, would have led us not to form these beliefs at t. In short, it seems that beliefs can be properly subject to deontic evaluation because we can have indirect control over them in virtue of the basic voluntary control we have over our actions.3 However, I’ll argue that this claim – let’s call it Indirect Voluntary Control – should be rejected and that it is another kind of indirect control that makes our beliefs properly subject to deontic evaluation.4

2.

Why We Should Reject Indirect Voluntary Control

Let’s admit that the judgement that S should not have believed that p at t is grounded on the judgement that S should have done things before t that would have led her, at t, not to form the belief that p – for instance, attentively reading easily accessible documents related to the issue of whether p rather than just taking a quick look at one and then leaving the library and going to the beach all day long. Let’s also admit that S has basic voluntary control over these actions that would have led her not to form certain beliefs at t (her way of inquiring and of examining the evidence, in the main). Suppose now that these actions are enkratic: they result from S’s belief that there’s no need for her to inquire further and to examine the evidence with more attention. Then for these actions to be properly subject to deontic evaluation it is necessary that this belief from which they result can itself be properly subject to deontic evaluation. Now, arguing that this belief can itself be properly subject to deontic evaluation on the condition that it results from actions from S over which she has basic voluntary control would obviously lead to an infnite regress. Therefore, it seems that, at some point, our deontic evaluations of beliefs must not be grounded on deontic evaluations of actions based on beliefs from which they enkratically follow. This argument appears in the literature in a more particular form in the context of discussions addressing the question of when one can be blamed for (all-things-considered) wrong actions the wrongness of which one is ignorant. In these discussions devoted to articulating the “epistemic condition for moral responsibility,” one argument from Gideon Rosen (2004) and Michael Zimmerman (1997) has received much attention. It can be reconstructed as follows:

Responsibility for Doxastic Strength

75

Saying that 5. When S is ignorant that her action (or omission) A is wrong or falsely believes that A is not wrong, S is blameworthy for A iff S is blameworthy for this ignorance I or this belief B, and 6. S is blameworthy for I or B iff they result from a blameworthy action (or omission) A′ and 7. S is blameworthy for A′ iff S is blameworthy for her false belief B′ that A′ is not wrong, or for her ignorance I′ that A′ is wrong leads to an infnite regress because, following this line of thought, one will have to say that 8. S is blameworthy for B′ or I′ iff they result from a blameworthy action (or omission) A′′ and 9. S is blameworthy for A′′ iff S is blameworthy for her false belief B′′ that A′′ is not wrong, or for her ignorance I′′ that A′′ is wrong and so on, ad infnitum. Therefore, according to Rosen, we should rather say that 5.* When S is ignorant that her action (or omission) A is wrong or falsely believes that A is not wrong, S is blameworthy for A iff S is blameworthy for her false belief B that A is not wrong, or for her ignorance I that A is wrong, and 6.* S is blameworthy for B or I iff they result, at some point, from a blameworthy action (or omission) A′ which is so because S does A′ while believing that A′ is wrong – i.e. that she should not do A′. In other words, for Rosen, “the only possible locus of original responsibility is an akratic act” (Rosen 2004: 307) for which the agent is directly culpable; “every blameworthy action must be either itself an akratic action or the causal upshot of one” (Rudy-Hiller 2017). In short, “blameworthiness requires akrasia” (Wieland 2017: 13). And because for Rosen, due to “opacity of mind,” “we are never warranted in confdent attributions

76

Benoit Gaultier

of clear-eyed akrasia [. . .] we are not warranted in confdent attributions of responsibility or blame for bad actions” (FitzPatrick 2008: 593). As a consequence, “few, if any, of our judgements of culpability are justifed” (Rudy-Hiller 201). The “question about akrasia,” as Matthew Talbert calls it, then is “whether a person might be blameworthy for her wrongdoing even if, in the etiology of her action, she never acted wrongly by her own lights” (Talbert 2017: 47). Rosen’s conclusion has often been resisted by rejecting (5). Because we are not interested here in deciding when a subject is blameworthy for a wrong action (or omission) of hers, but in deciding when a subject is responsible for her beliefs, many of the solutions that have been advanced in the literature to refute Rosen’s conclusion do not apply to our regress that concerns our responsibility for our beliefs in general. This regress can be formulated in the following way. Saying that 10. S is responsible for her (true or false) belief B iff it enkratically results from an action (or omission) A′ for which S is responsible and 11. S is responsible for A′ iff S is responsible for the (true or false) belief B′ from which A′ enkratically results leads to an infnite regress. Let’s try to apply to this latter regress Rosen’s answer to the former: 12. S is responsible for her (true or false) belief B iff it results, at some point, from an action (or omission) A′ for which S is responsible because S does A′ while believing that A′ is wrong, i.e. that she should not do A. Rosen’s claim – and its consequences – that every blameworthy action must be either itself an akratic action or the causal upshot of one have often been judged to be too revisionary to be plausible. But things are even worse when it comes to (12): saying that a subject is responsible for her beliefs when they are the causal upshot of an akratic action – and hence that a belief that does not result from an akratic action cannot be deontically evaluated – does not have any plausibility.5 How then to escape the regress? A natural thought is that it is possible to escape it by introducing moral and intellectual vices (or virtues) or morally and intellectually vicious (or virtuous) tendencies into the picture (Montmarquet 1993; FitzPatrick 2008, 2017). The idea, to put it in a nutshell, is to say that for any proposition p, a subject is responsible for her belief that p because she is responsible for the way in which she inquired into p from which this belief results – how

Responsibility for Doxastic Strength

77

much evidence she collected, how much time she spent (re)examining it, and how much attention she devoted to it, etc. – and this depends on her moral and intellectual character. There are, however, at least two problems with this proposal. First, it is not obvious that we are responsible for our moral and intellectual character – even if it results from previous voluntary actions from us; indeed, they themselves result from the intellectual and moral standpoint we occupy at a certain moment, and it is not obvious that we are responsible for our occupying this standpoint at that moment. Second, supposing we can be held responsible for our moral and intellectual character, what about the beliefs that result from those actions that do not fow from or are not manifestations of it? (For instance, when an open-minded person demonstrates dogmatism on a certain issue on a certain occasion, leading her to examine very superfcially the new evidence at her disposal on this issue at that time; or when a generally enkratic person occasionally stops inquiring into whether p while judging that she should do the contrary.) Should we say that, ipso facto, we are not responsible for them, so that an open-minded person cannot be held responsible, and so cannot be blamed, for the beliefs she sometimes forms and holds out of dogmatism? That would be strongly counterintuitive. It follows from the foregoing that there does not seem to be much hope in grounding our responsibility for our beliefs only on actions over which we have voluntary control. Rik Peels follows another path (although this might not be immediately obvious): “whether or not one believes responsibly depends on whether one has acted in accordance or contrary to one’s beliefs about one’s intellectual obligations” (Peels 2017: 198). This means that one can be blamed for one’s beliefs not only if they result from actions or omissions one occurrently believes to be wrong given one’s intellectual obligations, but also from actions one non-occurrently or tacitly believes to be wrong given these obligations (which makes, in the latter case, these actions non-akratic stricto sensu, one might want to argue). Similarly, Philip Robichaud (2014) argues that one’s beliefs are also blameworthy when they result from a way of inquiring that one takes to have suffcient, though not decisive, reasons to be wrong (which also makes these actions non-akratic, strictly speaking).6 This may be correct, but it does not tell us why or in virtue of what we can be held responsible for these actions and the beliefs resulting from them. A fortiori, it does not tell us why or in virtue of what we can be held responsible for beliefs not resulting from such akratic or quasi-akratic (as we might call them) actions. The same goes for William FitzPatrick’s claim that our beliefs are blameworthy when they result from actions that result themselves from “indulging vicious tendencies” (FitzPatrick 2017: 42). That said, these views suggest nonetheless an important point: our responsibility for our beliefs has to do with things – being (quasi-)akratic or not, indulging vicious tendencies or not – over which we do not have

78

Benoit Gaultier

basic voluntary control, that we do not do at will. It’s time at present to explore this path.

3.

Another Kind of Indirect Control

Let’s start with an idea that Philippe Chuard/Nicholas Southwood (2009) have emphasised: for a subject S to have, at t, the ability (and opportunity) to φ, S does not have to have the ability to φ voluntarily; so, “ought implies can” does not mean “S ought to φ only if S has voluntary control over φ-ing” (Chuard/Southwood 2009: 620). Could then S have at a certain time t the ability and opportunity not to form the belief she formed at that time t while not having basic voluntary control over the formation of this belief? A natural way of understanding the negative deontic evaluation that S should not have believed that p at t is that S’s doxastic performance in forming the belief that p at t was inferior to what it could have been given S’s cognitive or intellectual abilities at t. S had the ability (and opportunity), at t, to do better than she did by forming the belief that p, and this makes it possible for one to blame S for having formed it. Let us for instance consider two subjects who both form the same belief on the basis of the same fallacious reasoning, but whose intellectual abilities differ in the way Jessica Brown imagines: One of the subjects has the intellectual ability to see for herself that the reasoning is fallacious and moreover has been trained to avoid it. By contrast, the other subject does not have the intellectual ability to see for herself that the relevant fallacy is a fallacy, and has not been taught that the reasoning is fallacious. Brown 2018: 6 As she remarks, “It would be inappropriate and unfair to rebuke and be angry with the second low-ability subject who cannot see for herself that it is fallacious reasoning, and has not been taught any different” (ibid.). Our different reaction to the low and high-ability subjects indicates that for us “differences in a subject’s ability to X makes a difference to the degree of their responsibility and blameworthiness” (ibid.), which makes it “preferable to adopt an account of blameworthy belief on which the degree of blame can be affected by one’s abilities” (ibid.: 17). Let us make more explicit the sense in which the doxastic performance of the high-ability subject could have been better: her cognitive or intellectual abilities enabled her to better process the evidence she had at her disposal – which would have led her not to respond doxastically to it by forming the belief in question. In other words, by processing her evidence in the way she did, which led her to form this belief, the subject was below the level of doxastic performance that her cognitive or

Responsibility for Doxastic Strength

79

intellectual capacities enabled her to reach then. She, in this sense, could have believed better than she did. However, for it to be appropriate to say, in cases of doxastic underperformance, that the subject should have believed better than she did, and for it to be appropriate to blame her for such underperformances, it is necessary that she can be held responsible for the intervening factor that explains them. What then could be this factor (a) that can make a subject’s doxastic performance at t inferior to the level of doxastic performance her cognitive or intellectual abilities enable her to reach at t, and (b) for which she can be held responsible?7 As a matter of fact, we take certain cases of belief formation to be instances of culpable doxastic weakness, as we might call them: when what we want, or do not want, to be the case interferes in our processing of the evidence in such a way that what we then believe to be the case culpably differs from what we would otherwise have believed on the basis of this evidence. In such cases, the subject processes the evidence at her disposal in such a way that this results in her forming a belief (a) that, given her intellectual abilities, she would not have formed if she had not demonstrated such a weakness, and (b) that is less likely to be true than the belief she would otherwise have formed. To put it more precisely: when a subject’s conative attitudes – her wishes, desires, hopes, preferences, etc. – interfere in her processing of the evidence and we take the resulting beliefs to be beliefs that, given her cognitive abilities, she would not have formed had these conative attitudes not interfered, our reactive attitude in such situations turns out to be the following: we judge that it was within her power to demonstrate doxastic strength, as we might call it, rather than, as we say, letting her hopes and desires corrupt her judgement and determine her beliefs. We do not judge it to be always easy to demonstrate doxastic strength, but it seems that we never judge this to be something that the subject involved could not have demonstrated. And when a subject demonstrates doxastic weakness, the resulting beliefs are less likely to be true than those she would have formed if this had not happened, because conative attitudes do not aim at tracking the way the world is but have, on the contrary, a world-to-mind direction of ft. Not all cases of doxastic weakness make the subject involved blameable and the beliefs concerned beliefs she should not have formed though. In other words, not all cases of doxastic weakness are in themselves cases of culpable doxastic weakness. But they make the subject concerned open to blame in the sense that if someone (herself included) (a) takes it to be a moral or epistemic wrong to demonstrate doxastic weakness (e.g. on the ground that this is to self-undermine one’s own judgement), or

80

Benoit Gaultier

(b) deplores the consequences of this or that instance of doxastic weakness, the fact that we take it to be always within one’s power – whatever one’s intellectual character traits and however diffcult that may be in certain circumstances – to demonstrate doxastic strength always leaves room for holding the subject concerned responsible for this wrong or these consequences. In order to evidence this fact and illustrate its consequence, let’s consider a case put forward by Béla Szabados (with the aim of showing that not all instances of self-deception are (morally) wrong), and let’s (rather) ask this question about this case: does it show that there are instances of doxastic weakness that are not open to blame? Suppose a woman’s son has apparently been killed in an accident. It is not absolutely certain that he has been killed, but there is good evidence that he has. Now it would be quite natural for such a woman to deceive herself into thinking that her son is alive. In fact, we would be surprised if she did not cling to such a belief until the evidence is conclusively against it. [. . .] It is her love and hope for her son that makes her cling to her belief so desperately. [. . .] It is by no means clear that the moral qualities that she exhibits in deceiving herself are not admirable. [. . .] The woman might be said to have failed in her regard and respect for the truth; or she might be said to have failed because she, by deceiving herself, betrayed herself as a rational agent. [. . .] [But] no morally sensitive person can possibly accept such a perspective without ceasing to be a morally sensitive person. [. . .] Respect for rationality and truth are important moral considerations – but they are not all-important! Morality cannot be reduced merely to respect for rationality and truth. Szabados 1974: 29 This case does not show that there are instances of doxastic weakness that are not open to blame. Suppose this woman’s belief – formed out of doxastic weakness – that her son is still alive leads her other son to regain hope while he was trying to get used to the idea that his brother has been killed in an accident. Suppose this regain of hope leads him to be so shocked when he learns, two days after, that his brother is well and truly dead that he commits suicide. And suppose her mother knows that without this regain of hope due to her belief formed out of doxastic weakness he would not have committed suicide. To those telling her that, contrary to what she says, it is not her fault that he did this, we can expect her to object that she could have not formed this belief and hence that she is, at least partly, responsible for his suicide. Obviously, in such a context it would be extremely harsh from

Responsibility for Doxastic Strength

81

anyone (herself included) to blame her for having formed this belief. But its being formed out of doxastic weakness does leave room for holding her(self) responsible for her son’s suicide, and so for her rejecting as untrue the comforting and well-intentioned words that she cannot be held responsible for it. She could then argue that the fact that her belief that her son is alive resulted from a lack of doxastic strength does make her open to blame, and that while those who would blame her for that (even only inwardly) would be morally guilty of harshness, insensitiveness, and lack of compassion,8 there is no moral wrong in blaming herself for her lack of doxastic strength that resulted in her second son’s suicide.9 Moreover, it could be argued that this woman would have been even more admirable if, in addition to her love for her son that led her to believe that he is still alive (because this is something that, due to her love, she strongly desires), she had demonstrated doxastic strength – i.e. if her judgment had been insensitive to such a strong desire. Szabados is inclined to hold the opposite because he seems to think that forming in these circumstances the belief that her son is probably dead either means: (a) that she does not strongly desires that he is alive, or (b) that she has a desire for truth that surpasses her desire that he is alive – which, arguably, would be morally wrong, and so would make the belief that her son is probably dead morally wrong. But there is a third option: she strongly desires that her son is alive, and more than she desires truth; but because of her doxastic strength, her judgment is insensitive to this desire, so that she forms the belief that he is probably dead. This belief would not be open to blame, contrary to her belief that he is alive, formed out of doxastic weakness. This apparent reactive attitude of ours – taking it to be always within our power to demonstrate doxastic strength rather than letting ourselves go to doxastic weakness (even if, in certain circumstances, this would be extremely diffcult) – is not more incongruous and demanding than our reactive attitude vis-à-vis weakness of will: just as with doxastic strength, we do not judge it to be always easy to demonstrate strength of will, but it also seems that we never judge this to be something it was impossible for the subject involved to demonstrate.10 And in the same way that we take our actions to be open to blame when they result from having demonstrated weakness of will even though, on any account of weakness of will,11 demonstrating weakness or strength of will does not consist in (or reduce to) basic voluntary actions from us, we take our beliefs to be open to blame when they result from having demonstrated doxastic weakness even though demonstrating doxastic weakness or strength does not consist in (or reduce to) basic voluntary actions from us. That such beliefs are open to blame is not more mysterious than for such actions to be so. In neither case does the kind of control we have consist in doing something at will in the sense in which I can generally raise my hand at will, but this does not prevent this control to be real.

82

Benoit Gaultier

It consists, when we display doxastic strength or strength of will, in succeeding in resisting the infuence of some of our conative attitudes – which is not something we do at will, but in which our agency is involved. According to Michael Brent (2014), it seems that for Richard Holton (2003) it is only when we display strength of will and so succeed in resisting the force of certain desires that a distinctive causal role of us as agents in the performance of our actions can be observed – while, when we do not do that, our desires, beliefs and intentions and other motivational factors suffce to explain this performance. But it could be argued that this rather applies to the formation of our beliefs: it is only when we display doxastic strength – or, on the contrary, fail to do so – that a distinctive causal role of us as agents in the formation of our beliefs can be observed, making us responsible for them. While our control over our actions can also consist in having basic voluntary control over them, things are different when it comes to our beliefs: we do not have over them another form of control. Even though this control is indirect because it consists in controlling our beliefs by doing something else than forming them – i.e. by resisting the infuence of some of our desires (which, once again, involves our agency but does not consist in belief-based actions from us, with the consequence that the regress problem mentioned before no longer arises) – it does not precede their formation: demonstrating doxastic weakness or strength is not something that takes place before the formation of the concerned beliefs, just as demonstrating weakness or strength of will is not something that takes place before the doing of the concerned actions. Therefore, beliefs formed out of doxastic weakness are beliefs it was within our power not to form at the moment we formed them, just as actions done out of weakness of will are actions it was within our power not to do at the moment we did them, even though this does consist in having basic voluntary control over these beliefs or these actions. This is why beliefs are open to blame and hence can be properly subject to deontic evaluation. And it is also why blaming our beliefs is not always a misleading way of blaming preceding actions of us from which they result.12

Notes 1. When “should” is taken in its ordinary deontic sense. In the evaluative – arguably much less ordinary – sense of “should,” there clearly seem to be situations where people should have been akratic. When “should” is taken in its evaluative sense, it’s not odd to say that tyrants, genocidal maniacs, serial killer, terrorists, and the like should not have acted enkratically – i.e. in accordance with their belief that it was perfectly adequate for them to kill, harm, terrorize, etc. – and should have been akratic. 2. On this line of reasoning, see also (Littlejohn 2013, Rosen 2003). 3. On such views, see Meylan (2017). 4. Indirect Voluntary Control seems to be accepted by almost all the authors discussing the question of whether we can be held responsible for our beliefs.

Responsibility for Doxastic Strength

5. 6. 7.

8. 9.

83

But some of them hold that we have, in addition, a direct but non-voluntary control over our beliefs, or want to account for the responsibility we have over them in terms of reasons-responsiveness. I won’t discuss these views here. Whatever their merits, it seems to me that there is something deeply unsatisfactory with them: according to these views, we can be held responsible for our beliefs in virtue of the way in which they are formed, either given the very nature of belief, or given some properties of the mechanisms at play in the formation of our beliefs. In both cases, whether or not we can be held responsible for this or that belief of ours is not grounded on something that depends on us, on our agency. And I would argue that if our agency is not involved, it is problematic to hold us responsible for our beliefs. To the objection that our agency is well and truly involved because epistemic agency should not be conceived in the way in which we conceive practical agency, that these two forms of agency radically differ, I would simply respond as follows: our intuition that our agency has to play a role is the intuition that something like our ordinary practical agency has to play a role, so that it is not following this intuition to say that a totally different form of agency is involved in the formation of our beliefs. Contrary, for instance, to the claim a subject is blameworthy for her beliefs iff they are the causal upshot of an akratic action – which, while highly implausible, does not seem not to make any sense, contrary to (12). See also, on this issue, Levy (2009). George Sher holds that an agent, let us call him Paul, is responsible for “an act of whose wrongness or foolishness he is unaware [when] his failure to recognize the act as wrong or foolish falls short of satisfying some applicable standard” that is “set precisely by that agent’s current cognitive capacities” (Sher 2009: 21, 110). Since this standard is so set, Paul’s failure to satisfy it has to be explained, for him to be responsible for the failure, by a factor that is external to his current cognitive capacities. This factor is, according to Sher, “the interaction of some combination of his constitutive attitudes, dispositions, and traits” that “make him the person that he is” (ibid.: 88). For Sher, “[t]hese attitudes, traits, and dispositions include [. . .] the underlying mechanisms, both physical and nonphysical, that shape his emotional reactions and the speed and accuracy of his inferences and decision-making” (ibid.: 86). But it could be argued that such things contribute to determining Paul’s current cognitive capacities and hence the standard below which he is responsible for his failure to recognize the act concerned as wrong or foolish. As a consequence, these things cannot make him responsible for this failure. Which means that it is perfectly possible that a subject is blameworthy for something and, at the same time, that blaming her for the thing in question is morally wrong. Let’s note in passing that it may happen that a subject’s evaluation of the evidence is very careful and that she enkratically doxastically responds to this evaluation by forming and holding a belief whose content and degree are in line with this evaluation, but that she nonetheless is open to blame for her belief because it is formed and held out of doxastic weakness. It may happen for instance that I strongly desire that my good old colleague Martin’s friendship is sincere, that whether this is true or not is so important to me that I spend hours and hours considering the question, anxiously reviewing all the evidence I have and devoting all my attention to it, and that my conclusion that his friendship is sincere is formed out of doxastic weakness: without this strong desire, I would have immediately realised that the evidence clearly indicated the opposite. This means that when a belief is formed and held in an epistemically responsible way (as we might call it), this does not guarantee

84

Benoit Gaultier

that it is not formed and held out of doxastic weakness, and so does not guarantee that it is not open to blame. 10. In line with this claim, Alfred Mele writes that he “would not be surprised if many, and even most, lay respondents to a story in which it is made clear that an agent is moved by an irresistible desire would, if asked, say that he displays some weakness of will [. . .] and [. . .] think that if the agent’s will were not weak, he would be able to resist the desire and that the desire’s being irresistible by him entails some weakness of will” (Mele 2010: 403). So, when James Beebe (2013) reports that experimental subjects are happy to attribute “weakness of will to agents whose actions stem from compulsion” even when it is described as “uncontrollable” and “overwhelming,” this does not necessarily imply that they take these actions to be absolutely compelled or irresistible. 11. Whether in terms of beliefs and desires only, in terms of beliefs, desires and intentions, or in terms of such factors plus will-power. See e.g. Holton (2003) for an overview of these different accounts and a defence of the latter. 12. This paper has decisively improved after crucial suggestions and criticisms from Anne Meylan. As a result, the amount of contestable inferences it contains should have fallen signifcantly. I want to thank her warmly for that. Thanks also to Sebastian Schmidt for his very useful comments.

References Beebe, James (2013): “Weakness of Will, Reasonability, and Compulsion”, Synthese 190, 4077–4093. Brent, Michael (2014): “Understanding Strength of Will”, in: New Advances in Causation, Agency, and Moral Responsibility, ed. by Massimo/Caputo, Cambridge: Cambridge Scholars Press, 165–178. Brown, Jessica (2018): “What Is Epistemic Blame?”, Noûs, DOI: 10.1111/ nous.12270. Chuard, Philippe/Southwood, Nicholas (2009): “Epistemic Norms without Voluntary Control”, Noûs 43, 599–632. Fitzpatrick, William J. (2008): “Moral Responsibility and Normative Ignorance: Answering a New Skeptical Challenge”, Ethics 118, 589–613. Fitzpatrick, William J. (2017): “Unwitting Wrongdoing, Reasonable Expectations, and Blameworthiness”, in: Responsibility: The Epistemic Condition, ed. by Robichaud/Wieland, New York: Oxford University Press, 29–46. Holton, Richard (2003): “How Is Strength of Will Possible?”, in: Weakness of Will and Practical Irrationality, ed. by Tappolet/Stroud, New York: Oxford University Press, 39–67. Levy, Neil (2009): “Culpable Ignorance and Moral Responsibility: A Reply to FitzPatrick”, Ethics 119, 729–741. Littlejohn, Clayton (2013): “The Unity of Reason”, in: Epistemic Norms, ed. by Littlejohn/Turri, New York: Oxford University Press, 135–154. Mele, Alfred (2010): “Weakness of Will and Akrasia”, Philosophical Studies 150, 391–404. Meylan, Anne (2017): “The Consequential Conception of Doxastic Responsibility”, Theoria 83, 4–28. Montmarquet, James (1993): Epistemic Virtue and Doxastic Responsibility, Lanham: Rowman & Littlefeld.

Responsibility for Doxastic Strength

85

Peels, Rik (2017): Responsible Belief, New York: Oxford University Press. Robichaud, Philip (2014): “On Culpable Ignorance and Akrasia”, Ethics 119, 137–151. Rosen, Gideon (2003): “Culpability and Ignorance”, Proceedings of the Aristotelian Society 103, 61–84. Rosen, Gideon (2004): “Skepticism about Moral Responsibility”, Philosophical Perspectives 18, 295–313. Rudy-Hiller, Fernando (2017): “A Capacitarian Account of Culpable Ignorance”, Pacifc Philosophical Quarterly 98, 398–426. Ryan, Sharon (2003): “Doxastic Compatibilism and the Ethics of Belief”, Philosophical Studies 114, 47–79. Sher, George (2009): Who Knew? Responsibility without Awareness, New York: Oxford University Press. Szabados, Béla (1974): “The Morality of Self-Deception”, Dialogue 13, 25–34. Talbert, Matthew (2017): “Akrasia, Awareness, and Blameworthiness”, in: Responsibility: The Epistemic Condition, ed. by Robichaud/Wieland, New York: Oxford University Press, 1–28. Wieland, Jan Willem (2017): “Introduction”, in: Responsibility: The Epistemic Condition, ed. by Robichaud/Wieland, New York: Oxford University Press, 47–63. Zimmerman, Michael (1997): “Moral Responsibility and Ignorance”, Ethics 107, 410–426.

B

Reasons for Belief

5

Relationships and Reasons for Belief Lindsay Crawford

1.

Introduction

What we believe about our friends, family members, and romantic partners (in short, our signifcant others)1 constitutes an important component of how we relate to them more broadly. Being disposed to believe and resist believing certain things about our signifcant others, in ways that generally tend to favor those signifcant others, is no less a part of what makes our relationships with our signifcant others distinctive than our particular modes of interaction with them and on-going patterns of concern for them. Familiar examples of this sort of special treatment in belief abound. A friend might be particularly resistant to revising her beliefs about a close friend in response to unfattering reports about his recent behavior. A parent may be more inclined to believe her son’s passionate denial of a serious accusation than she would be if the accused were not her son. A spouse may view the likelihood of his partner’s success at a risky business venture much more favorably than he otherwise would. These are all just some of the ways in which we exhibit doxastic partiality when it comes to what we believe about our signifcant others. Doxastic partiality consists, more broadly, in having generally favorable beliefs about a signifcant other that you would not have about an otherwise similarly situated insignifcant other, about whom you have the same evidence; and resisting having generally unfavorable beliefs about a signifcant other, in ways that you would not resist having beliefs about an otherwise similarly situated insignifcant other about whom you have the same evidence.2 To say that doxastic partiality is a component of how we distinctively relate to our signifcant others is not yet to make any particular normative claim about how doxastic partiality ought to fgure into those relationships. Does anything normative follow? There are, of course, a number of importantly different normative claims that may or may not follow that are worth distinguishing at the outset. One might wonder, for example, whether exhibiting doxastic partiality when it comes to our

88

Lindsay Crawford

signifcant others is a praiseworthy or admirable tendency of ours, or at least not something seriously criticizable. Or one might wonder whether the standards by which one reasons about another person may permissibly vary, depending on whether or not that person stands in a particular relationship with that other person. Here, my focus will be on the question of whether our personal relationships can provide reasons for belief – specifcally, reasons for beliefs that would exhibit doxastic partiality on our part toward our signifcant others.3 Can our relationships give us reasons to believe well of our signifcant others? To believe what they tell us? To think that they’ll succeed? I begin, in section 2, by making the case that these questions are not ones that can be quickly answered in the negative by an appeal to the prevailing view that only evidential considerations can be reasons for belief.4 Call this view about reasons for belief evidentialism. If evidentialism is true, then our relationships with our signifcant others cannot themselves be the sort of thing that can provide reasons to form beliefs about those signifcant others. I argue that the evidentialist cannot categorically block non-evidential considerations from counting as reasons for belief on the ground that such considerations cannot enter into deliberation about what to believe in the right sort of way. If that is so, then it remains an open question whether our personal relationships can provide reasons to be doxastically partial when it comes to our signifcant others. If we did have reasons for doxastic partiality, what would those reasons consist in? What would explain their normative force? In Sections 3 to 5, I examine some explanatory options recently developed by those who defend the claim that our personal relationships can require doxastic partiality (whom I’ll refer to here as partialists). The frst of these explanatory options aims to ground reasons for doxastic partiality in facts about what attitudes are partly constitutive of being a good signifcant other to someone. The basic thought here is that if doxastic partiality is partly constitutive of being a good signifcant other to someone, then that provides reason for an agent who stands in that relation to another person to exhibit doxastic partiality. The second of these explanatory options aims to ground reasons for doxastic partiality in reasons, more broadly, to beneft our signifcant others. If being doxastically partial toward our signifcant others can beneft them, then that provides reasons for doxastic partiality. I argue that these explanatory options fail. Our personal relationships do not give us non-evidentially reducible reasons for belief, even if the pragmatist is right that there can be non-evidential reasons for belief. In Section 6, I turn to the broader question of whether denying that our personal relationships can give us non-evidentially reducible reasons for doxastic partiality puts pressure on us to deny that our personal relationships can give us reasons for partial treatment of any kind. I argue that there may be ways to defend reasons for partial treatment without having to defend reasons for doxastic partiality.

Relationships and Reasons for Belief

2.

89

An Initial Challenge: Evidentialism About Reasons for Belief

Consider the following case. Your close friend Stan, who is also a member of your cohort in the competitive MFA program you both attend, harbors signifcant doubts about whether he has the artistic talent to succeed. In the course of confessing his doubts, it becomes clear to you that Stan is looking for your reassurance, some sign of validation. But not just your reassurance; Stan would not be comforted by what he took to be empty words on your part. He wants you to think that he’s got the talent to succeed, even if he doesn’t express that desire directly. Let’s also suppose that Stan is especially vulnerable to you and your view of his talents, in the sense that his ability to stick it out through the more diffcult periods of the MFA program importantly depends on his sense that you think he has the talent to succeed. Our question about this case is: can your friendship with Stan make a difference to what you ought to believe about whether Stan has what it takes? More specifcally, can your friendship with Stan provide some reason for you to believe that he has what it takes – a reason you would not otherwise have, if Stan were not your friend?5 Here’s a natural response: “Giving Stan whatever pep-talk he needs to help him see through his self-doubt is one thing,” an evidentialist might say. “But whether you should actually believe that Stan has what it takes is quite another. Your friendship may give you all kinds of practical reasons to treat Stan in various ways that help reassure him and lift his spirits, but your friendship couldn’t give you reason to believe that Stan has what it takes.”6 Why not? For many, this response is just a straightforward upshot of a broader evidentialist view about reasons for belief. According to the evidentialist, reasons to believe that p are exclusively evidential considerations that bear on the likely truth or falsity of whether p. On this view, considerations that bear on whether believing p would beneft me, or whether it would help me achieve certain of my goals, and the like, are not reasons to believe p. If our relationships themselves could provide reasons for or against having certain beliefs about our signifcant others, those reasons would be non-evidential and so ruled out by the evidentialist’s restriction on reasons for belief. Why accept evidentialism about reasons for belief? A prominent argument for evidentialism runs as follows.7 First, there is a plausible deliberative constraint on reasons, on which only considerations that an agent can follow, and in the right sort of way, in deliberation about whether to φ can be reasons to φ.8 This is a very general constraint on what reasons to φ must be like. What counts as following those reasons, and in the “right sort of way,” in coming to φ or refrain from φ-ing, depends on what sort of φ-ing we’re talking about, as well as the nature of rational

90

Lindsay Crawford

deliberation about whether or not to φ. Second, non-evidential considerations that tell in favor of a belief are not considerations that an agent can follow, in the right sort of way, in deliberation about what to believe. The “unfollowability” of non-evidential considerations in deliberation about what to believe, combined with the deliberative constraint on reasons, jointly entail the claim that non-evidential considerations cannot constitute reasons for belief. Reasons for belief are exclusively evidential. A number of philosophers have recently challenged this argument by targeting the evidentialist’s unfollowability claim.9 One way to bring out that challenge is to offer up an episode of deliberation about what to believe in which an agent is sensitive to non-evidential considerations bearing on what to believe, and then try to determine what exactly makes such non-evidential considerations unfollowable in the right way. An episode of Pascalian deliberation is an obvious choice to zero in on. In deliberating about whether or not to believe that God exists, you consider the Pascalian reason for believing that God exists that having that belief has a higher expected value than failing to believe that God exists (so long as you assign a non-zero probability to God’s existence). Moved by this sort of consideration, you conclude that you ought to believe that God exists. What makes this sort of consideration unfollowable in the right sort of way? Not the fact that such a consideration could not function as a premise in deliberation about what to believe. It’s certainly functioning as a premise in an episode of some kind of deliberation about whether to believe that God exists. One may, of course, wish to distinguish here between such practical deliberation about what to believe, on the one hand, and theoretical deliberation on the other. Theoretical deliberation involves considerations functioning as premises in a particular way: by making the conclusion of the bit of deliberation more likely true. And to be sure, the fact that believing in God has the highest expected value does not make God’s existence more likely true; Pascalian deliberation is not theoretical deliberation. But the evidentialist cannot respond by drawing this distinction and then insisting that only by being able to function as a premise in theoretical reasoning can a consideration count as a reason for belief. That would obviously be question-begging. We would need some independent reason to think that theoretical reasoning about what to believe is to be privileged over practical reasoning about what to believe when we’re testing whether a particular consideration can meet the deliberative constraint on reasons. There are different moves available to the evidentialist at this point. The most obvious one is to point out that theoretical deliberation is deliberation that properly concludes in a belief, while practical deliberation about whether to believe concludes in something like an intention to get yourself to believe.10 The evidentialist can then appeal to something like the following general principle: Deliberation about whether or not to φ just is the kind of deliberation that concludes in φ-ing (or not φ-ing).

Relationships and Reasons for Belief

91

We would then have independent reason to privilege theoretical deliberation over practical deliberation when we evaluate whether a particular consideration satisfes the deliberative constraint on reasons. Deliberation about whether to believe p is just theoretical deliberation. But this general principle is false. Deliberation about whether to have children, whether to make the cross-country move, and the like are not episodes of deliberation that conclude in, e.g., having (or not having) children, or making (or not making) the cross-country move. But they are, nevertheless, episodes of deliberation about whether to have children, whether to make the cross-country move, and so on. And so we can’t rely on that general principle to provide an independent reason why theoretical deliberation ought to be privileged over practical deliberation in this context. There may be other alternatives available to the evidentialist to make the case that theoretical deliberation ought to be privileged over practical deliberation. But note that even if the evidentialist can fnd an independent reason to privilege theoretical deliberation over practical deliberation when using the deliberative constraint as a test on whether a certain consideration can count as a reason, so long as the evidentialist countenances practical deliberation about whether to intend to believe, or to get oneself to believe, then the evidentialist is committed to offering what looks either like weirdly fragmented advice to an agent deliberating about what to believe.11 Suppose that after having been riveted by your recent lecture on Pascal’s wager in your Introduction to Philosophy class, your student seeks your professional advice about what she ought to believe. She makes a compelling case that there is good evidence that her roommate dislikes her, but she also makes a compelling case to you that she would be quite a bit better off if she refrained from believing that her roommate dislikes her. So, she asks: “Should I believe my roommate dislikes me, because that’s what the evidence suggests? Or should I not believe that she dislikes me, because that would make me feel better?” Having just made the distinction in class between theoretical deliberation about what to believe, and practical deliberation about whether to get yourself to have a belief, you might advise her in the following way: “Well, if you’re asking whether you should believe that your roommate dislikes you, then yes. That said, you absolutely should do what it takes to get yourself not to believe that your roommate dislikes you.”12 But this advice is impossible to follow. The student can’t follow this advice to get herself to not believe p and believe p, just as one can’t follow advice to do what it takes to get oneself not to eat the cake and eat the cake. Insofar as it counts against a normative view that it licenses advice that is impossible to follow, this point will count against evidentialism. If the student at this point reasonably asks which one of these is more important – what she ought to believe, or what she ought to get herself not to believe – it might be tempting to insist that there

92 Lindsay Crawford simply is no question here about which one is more important than the other.13 The problem is that going this way signifcantly weakens the evidentialist’s position. Here’s what one might have thought was a fair way to gloss the evidentialist’s view: that agents ought to heed their evidence when it comes to forming and revising their beliefs. But even this, it turns out, is not something that the evidentialist can maintain, so long as the evidentialist allows for the possibility that an agent may very well have some practical reason to get herself to believe against the evidence. To be clear, this discussion does not exhaust the space of possible responses available to the evidentialist. Nor does it conclusively establish pragmatism about reasons for belief. The point is to clear space for the sort of investigation we will be taking up, in the following sections, into whether or not relationships can provide reasons for belief. The focus in what follows will be on different defenses of partialism, according to which, broadly speaking, facts about our personal relationships can affect what we ought to believe. Note, also, that many recent defenses of pragmatism about reasons for belief appeal to the same cases that partialists use to motivate the intuition that relationships can make a difference to what we ought to believe.14 In addition to the independent value of examining different defenses of partialism, it will also be worth looking more closely at whether the partialist’s key cases really are cases that the pragmatist can use to motivate the broader case for pragmatism about reasons for belief. Before we continue, it will help to explicitly distinguish between two kinds of partialist views. According to what we might call thresholds partialism, our personal relationships can affect the evidential-support threshold that evidential considerations must meet in order to be suffcient to require certain beliefs about a signifcant other.15 One might think, for example, that the costs or practical stakes associated with having a false belief about a signifcant other are higher than the costs or practical stakes associated with having a false belief about an insignifcant other. Someone who advocates for thresholds partialism need not be committed to thinking that our personal relationships can provide reasons for belief. One might hold, for example, that only evidential considerations can be reasons for belief, but that relationships can affect just how much evidence agents need to have in order for it to be the case that those agents are normatively required to form or revise certain beliefs about their signifcant others. According to what we might call reasons partialism, by contrast, our personal relationships can provide non-evidential normative reasons to form certain beliefs about a signifcant other – reasons that can add up to normatively require those beliefs. It’s the prospects for reasons partialism that I will explore in what follows.

Relationships and Reasons for Belief

3.

93

The Constitutive Claim About Relationships and Doxastic Partiality

If our personal relationships could provide reasons for belief about our signifcant others, what would those reasons be? In her defense of doxastic partiality in friendship, Stroud (2006) notes that the good friend’s reason seems to be simply that the person in question is her friend. But that someone is your friend is not itself a relevant epistemic reason [. . .] to form different beliefs about him than you would about anyone else. Stroud 2006: 513. But we should be careful not to try and identify the kind of reasons that personal relationships can provide by focusing on the content that those reasons might have. After all, that Stan is your friend might constitute evidence that Stan’s got what it takes, if, say, you had a track record of cultivating friendships with all and only the graduate students who happened to have had what it took to succeed. If that were the case, then the fact that Stan is your friend might provide some evidence that Stan has what it takes. In order to identify the kind of reason that some consideration might provide for φ-ing, we need to identify the facts in virtue of which that consideration counts as a reason to φ. We can now sharpen our main question. In virtue of what is the fact that someone is your signifcant other a reason to have a particular belief about that person? The frst option I want to consider answers this question in the following way. (To simplify matters, I’ll put the basic strategy here specifcally in terms of friendship.) The fact that someone is your friend can be a reason to be doxastically partial to that friend, in virtue of the following fact: that it is constitutive of being a good friend to someone that one is doxastically partial to that person. Defenses of partialism are grounded in various appeals to constitutive claims about the attitudes of a good friend. It is said that generally believing well of a friend, or giving them the beneft of the doubt, or viewing the prospects of a friend’s success favorably, and so on, are all just part of what it is to be a good friend to that person. As Stroud (2006) puts it, Just as a certain way of behaving concurs with our intuitive picture of the good friend, so also, I wish to claim, does a certain way of believing where our friends are concerned. A good friend does not defend her friend outwardly [. . .] while inwardly believing the worst of her friend. [. . .] [T]he good friend is prepared to take her friend’s part both publicly and, as it were, internally. What I wish to suggest [. . .] is that the differential doxastic practices of the good friend are a constitutive feature of friendship. Stroud 2006: 50516

94

Lindsay Crawford

Why should we think that doxastic partiality is a constitutive feature of being a good signifcant other (e.g., being a good friend)? The constitutive claim about friendship is often grounded in the claim that being practically partial toward a signifcant other, without being correspondingly doxastically partial, would constitute a form of insincerity. And being insincere in one’s friendships is something that makes a person worse, qua good friend, than that person would be if that person were sincere in one’s friendships. So, if one wants to allow that certain ways of treating a friend are constitutive of being a good friend to that person, one will have to allow that there are certain ways of believing that are constitutive of being a good friend to them, on pain of having to allow that being a good friend could be compatible with insincerity.17 To get clearer on this point, consider our central case involving Stan. Suppose you want to maintain that being Stan’s friend might give you a reason to reassure him, e.g., by telling him that his doubts more likely stem from imposter syndrome than from a clear-eyed assessment of his talent. The claim is that, insofar as you want to concede that being Stan’s friend might give you a reason to tell Stan that his doubts more likely stem from imposter syndrome, you must also concede that being Stan’s friend gives you a reason to actually believe that his doubts more likely stem from imposter syndrome. Otherwise, you’re committed to maintaining that being a good friend to Stan could be compatible with telling Stan things that you don’t actually believe, thereby leading him to think that you have a certain view about him that you do not actually have. A better friend would be someone who both tells Stan that his doubts stem from imposter syndrome and who actually believes that they do.18 There are various ways one might try to put pressure on the constitutive claim about the attitudes of a good friend.19 But for now, I will grant the claim that it is partly constitutive of what it is to be a good signifcant other to someone that one exhibit certain forms of doxastic partiality to that person. I will argue in this section that even if we grant the constitutive claim, that fact cannot be that in virtue of which your friendship with someone can be a reason for belief. Consider a different kind of relationship you might stand in with respect to another person: the relationship of being that person’s frenemy. There are plausibly certain attitudes constitutive of what it is to be a good frenemy to someone. Perhaps it is just part of what it is to be a good frenemy to someone that you perfectly conceal your distain for that person in order to lead that person to believe that you believe better of them than you in fact do. (A merely mediocre frenemy might be someone whose negative views of their rival are, despite their best efforts, not always perfectly hidden.) The fact that it is partly constitutive of being a good or excellent frenemy to someone that you perfectly conceal your distain for them so as to lead them on, does not give you, Fred’s frenemy, a reason to perfectly conceal your distain for Fred so as to lead him on.

Relationships and Reasons for Belief

95

Of course, frenemyship is not exactly a valuable kind of relationship. One might suggest, then, that facts about what is constitutive of being a good participant in a relationship that is not valuable cannot be a source of reasons, but facts about what is constitutive of being a good participant in a valuable kind of relationship can be. But I’m not sure this response is satisfactory. Consider an agent who is a participant in a perfectly valuable kind of relationship but who feels increasingly estranged from her signifcant other. Let’s say that Joyce is in a romantic relationship with John, but her investment in that relationship is waning. There is nothing problematic about their relationship; Joyce is just becoming a different person, with different desires and priorities. She is reevaluating her relationship with John. Even if it is partly constitutive of what it is to be a good partner to someone that one prioritize that romantic relationship in certain ways, that fact alone does not appear to give Joyce a reason to prioritize her relationship in those ways with John. She can perfectly well recognize that she is John’s romantic partner, and even acknowledge that good partners prioritize their relationship in various ways, and still coherently ask herself whether that fact really is a reason for her to do the things she’s always done for John. Would it help to maintain that an agent must not only occupy a morally permissible relationship but also identify with it? I’m not sure that amendment will help, either. After all, one can mistakenly care about or identify with the role they occupy. For example, John, who is as-yet unaware of Joyce’s waning investment in their relationship, may be mistaken for continuing to care about or identify with that relationship. In that case, it’s not clear that the fact that John is Joyce’s romantic partner gives him a reason to continue treating Joyce in the ways constitutive of being a good romantic partner to someone. One can imagine a friend of John’s who is aware of Joyce’s growing estrangement advising John that he has reason to reconsider how he prioritizes Joyce in his life or that he has reason not to put in the extra effort into their relationship. One might wonder whether Joyce and John really are in a relationship anymore. After all, we do sometimes speak of an estranged partner who slowly but eventually leaves their partner as someone who “left the relationship a long time ago.” The suggestion might be that Joyce and John fail to have reasons in the previously-mentioned cases because they no longer actually exist in the relevant sort of relationship that could generate reasons. The fact that John is my [Joyce’s] partner cannot be a reason for Joyce to have certain attitudes about John because John is not, in fact, her partner any longer, in the normatively relevant sense. But can attitude-dependent relationships be so tenuous? In her discussion and defense of the constitutive claim about friendship, Stroud herself is careful to note that such claims “need not be interpreted as tolerating no lapses or exceptions on the friend’s part, on pain of immediately exiting the friendship-relation altogether” (2006: 502). And I think Stroud

96

Lindsay Crawford

and other partialists who defend the constitutive claim are right to reject such a narrow restriction on normatively relevant relationships. After all, it’s part of the partialist’s goal to defend a picture on which even imperfect partners and friends and family members can come to see those relationships as reason-giving. A view according to which only people who exist in these attitude-dependent relationships can have reasons to be doxastically partial would mean that few, if any, of us can have such relationship-dependent reasons. It’s not clear, then, that the constitutive claim can explain how the fact that one stands in a relationship with someone else can serve as a reason for belief. We need, it seems, more than just the constitutive claim in order to explain how the fact that one stands in a personal relationship with another person can be a reason for certain beliefs about that person.

4.

Believing to Beneft

Let’s widen our focus. In virtue of what might your relationship with someone be a reason to do certain things for that person? How might a relationship be a reason for practical partiality? Let’s begin with the following simple observation. It’s plausible to think we have some reason to do things that will beneft our signifcant others, and some reason to avoid doing what will harm them. As Keller (2018) notes, just as how we treat a signifcant other can stand to beneft or harm that person, what we believe about our signifcant other can also stand to beneft or harm that person. For example, believing that your spouse will succeed in quitting smoking might be just what he needs to steel his resolve in his effort to quit. Doubting a friend’s ability to maintain a monogamous relationship may encourage her to take an overly cynical view toward her capacity for self-restraint that negatively affects her romantic pursuits. These are just some of the many familiar ways in which a signifcant other can be affected, positively or negatively, by what we believe about them. The suggestion, then, might be that the fact that someone is your signifcant other can be a reason for you to form (or resist forming) certain beliefs about them – reasons you would not have to form (or resist forming) beliefs about an otherwise identically situated insignifcant other – in virtue of the fact that, frst, that what you believe about your signifcant others can beneft them or harm them, and second, that you have a more fundamental reason to do what contributes to a signifcant other’s beneft and to avoid doing what can lead to their harm. Keller (2004, 2018) defends a view along these lines. On Keller’s view, there are different kinds of reasons for belief – familiar epistemic reasons, on the one hand, and reasons that have to do with the beneft to another person that having such a belief would provide, on the other.

Relationships and Reasons for Belief

97

When we assess your beliefs, we can ask whether you have the beliefs that beneft me. We can also ask whether your beliefs are grounded in the evidence. [. . .] At least in principle, there can be conficts between two kinds of reasons for belief: reasons arising from considerations of others’ interests, on the one hand, and reasons arising from considerations of evidence and truth, on the other. Keller 2018: 23 On his view, facts about our personal relationships are supposed to explain why reasons of the former kind arise specifcally in the context of our personal relationships: “Sometimes, you do have a good reason to manage your beliefs in response to considerations of another person’s interests, because sometimes, that reason is generated by a special relationship that the two of you share” (ibid.: 25). Of course, we all stand to be affected in various ways by what anyone believes about us, not just our signifcant others. Why, then, would the fact that I’m in a personal relationship with someone make a difference to the reasons I have to form beliefs (or resist forming beliefs) in ways that will beneft them (or avoid leading to their harm)? Why wouldn’t we just have reasons to form (or resist forming) beliefs in ways that would beneft (or avoid leading to the harm of) just anyone? Keller claims that “[i]f you have power over me, to exercise through your beliefs, then you have even more power over the people close to you” (ibid.: 23). A natural way of thinking about what it means for you to have “more power” over a signifcant other than someone else is just to see the signifcant other as being more dependent, in some sense, on what you do or believe about them, in ways that insignifcant others are not. If your signifcant others are especially dependent on you for their wellbeing, in ways that insignifcant others are not, then it might seem that the fact that you have a personal relationship with someone can make a difference to what you have reason to believe. But personal relationships do not neatly track relations of power or dependence. If we think that it’s relations of power or dependence that fundamentally matter, there is no reason to think that these will line up with personal relationships. Strangers may very well be especially dependent on our view of them in ways that our signifcant others are not. And there will be cases where we fnd relations of dependence between two individuals in which there is intuitively no reason for one to exhibit doxastic partiality toward the other. For example, a candidate for admission to a university may be especially dependent on the admissions committee members’ view of her abilities. Indeed, the candidate, unknown to the committee personally, is plausibly much more dependent on their view of her than on the view of her actual signifcant others. But the fact that the candidate is especially dependent on the committee’s beliefs about

98

Lindsay Crawford

her does not, presumably, give those members a reason to be doxastically partial toward her. Here’s a different but related suggestion worth developing further, also in Keller (2018). Our relationships with our signifcant others allow us to enjoy and beneft from certain “special goods.” And these special goods, in Keller’s view, can help explain how we can have reasons to be doxastically partial toward our signifcant others. As Keller puts it, [Y]ou can have good reasons to form beliefs for the sake of another person, even where that can mean forming beliefs that are not based on the evidence. In particular, the reasons concerned can arise from certain distinctive goods of friendship. Sometimes, in forming beliefs for someone else’s sake, you act as a good friend. Ibid.: 26 Keller’s (2013) discussion of reasons for various forms of partiality defnes a good as a distinctive or special good “if a full description of its nature implies that there is only one person, or only a few people, by whom it could be provided” (106). These special goods are “all goods that come only within certain relationships and can be provided only by the people with whom those relationships are shared” (107). For example, the good you provide your friend when you visit him in the hospital is not the good, to him, of just any old person’s presence at his bedside. Rather, the good you provide your friend is the special good of providing him your companionship – something you are in a privileged position to provide, in virtue of your relationship to him. For Keller, there are also certain special goods that you are in a privileged position to provide to your signifcant others by being doxastically partial toward them. One of those special or distinctive goods includes support: it is a good thing to have someone in your life who will encourage you and who will be there to see you through diffcult times. Another is openness: it is a good thing when a signifcant other can feel comfortable making herself vulnerable to you. For Keller, being able to provide these and other special goods to our signifcant others is often dependent on what we believe about them. Your ability to support your friend may depend, in some cases, on believing that she will succeed in a diffcult venture, or it may depend, in others, on having a slightly infated view of her abilities. Likewise, your ability to offer a friend the opportunity to be genuinely open with you may depend, for example, on resisting the urge to view his new creative project critically, or it may depend on believing positively of his overall moral character when he privately shares with you certain recent missteps in his love life. So perhaps what gives one a reason to have certain beliefs about one’s signifcant others is the fact that having such beliefs would enable one to establish or maintain these distinctive goods with those signifcant others – special goods they could only get from your beliefs about them.

Relationships and Reasons for Belief

99

But it’s clear that the fact that there is some special good that only you could provide to another person, by having certain beliefs about that person, need not give you a reason to provide it. An obsessed fan might have an interest in establishing various special goods that would come with having an intimate romantic relationship with the singer he idolizes – goods that the idolized singer is in a privileged position to provide, given the fan’s singular focus on her – but the singer has no reason to show the fan partiality.20 Maybe what matters is that our signifcant others can justifably expect these special goods, in ways that other people who might have an interest in these special goods cannot. But I’m not sure that justifed expectations can help ground reasons for partiality. Return to the case of John and Joyce. John may be justifed in expecting his estranged partner Joyce’s support and openness, if her growing estrangement is as-yet unknown to him. But if Joyce is reevaluating her relationship with him, the mere fact that John is justifed in expecting her support and openness does not give Joyce a reason to have the attitudes on which John’s ability to access those goods partly depends. To be clear, I think that expectations we have about our signifcant others can play a signifcant role in explaining some of the key intuitions to which partialists often appeal. The point here is just that expectations about the attitudes and treatment of a signifcant other, even when those expectations are seemingly justifed, need not be reason-generating. The conditions I’ve identifed here (relations of dependence, a person’s having an interest in establishing and maintaining the special goods distinctive of a personal relationship) cannot be what explains how a personal relationship can provide (non-evidentially reducible) reasons for belief. This is because those conditions, if they’re doing the normative work, would effectively make the appeal to personal relationships unnecessary. This is not going to be of any help to a partialist who wanted to locate some special feature of our personal relationships that would explain how those relationships distinctively give rise to reasons for doxastic partiality.

5.

Reasons and Reason-Modifers

I want to consider the following objection to the use of cases in the previous section to show that reasons for doxastic partiality are not grounded in reasons to beneft our signifcant others. The objection is that these cases do not really show that there are no reasons to be doxastically partial. Even if it’s not the case that the singer, or the admissions committee member, or Joyce ought to show their respective signifcant others doxastic partiality, perhaps the singer, the admissions committee member, and Joyce each have some reason, perhaps of very little weight, to show these people doxastic partiality. The basic idea here would be that in

100

Lindsay Crawford

all of these cases, the fact that having beliefs that would beneft someone provides at least some reason to exhibit doxastic partiality toward them, even if that reason is massively outweighed. On this sort of picture, non-evidential reasons for belief come cheap. One then needs to look to further features of context to determine whether these reasons can be suffcient to make it the case that one ought to have that belief.21 I want to grant that it is open to a pragmatist about reasons for belief to develop a view on which facts about whether or not having some particular belief would beneft another person are suffcient to generate reasons for belief. But this would not help to explain what the partialist wants to explain: that there is something distinctive about our personal relationships that provides reason to be doxastically partial. Indeed, reasons for doxastic partiality fall out of this picture entirely. You don’t have reasons to be doxastically partial toward those to whom one is specially related if every person is someone you have some reason to have generally favorable beliefs about, insofar as those beliefs stand to beneft that person. Perhaps the idea might be that we all have some reasons to have beliefs that beneft others, and our personal relationships with others add to their weight. There are different ways of thinking about how personal relationships might add to the weight of these reasons. Most straightforwardly, relationships might add to the weight of these reasons by adding another reason to the mix – a reason provided by the relationship itself. This doesn’t yet tell us anything about what exactly that reason is, though. In effect, this view brings us back to the beginning of our inquiry – what is this non-evidential consideration, provided by the relationship itself, that constitutes a reason for belief? Alternatively, one might hold that relationships add to the weight of these reasons by modifying their weight.22 In an adaptation of Lord’s (2016) account of reasons for partiality more broadly, one might hold that non-evidential reasons for belief about our signifcant others are provided by facts about whether or not those beliefs would beneft our signifcant others, and that the relationship intensifes the strength of those reasons. In other words, reasons to have beliefs that beneft others can be found anywhere, but those reasons are stronger in the context of a relationship. They’re the sort of things that can add up to be suffcient in the context of that relationship. What explains how a personal relationship intensifes the reasons that are already there? No doubt there are different explanatory possibilities here. Again, borrowing from Lord’s (2016) account, one might try to explain why it is that personal relationships intensify reasons by appealing to something like the following idea: that doxastic partiality (along with partiality more broadly) to a signifcant other is part of what intimacy and vulnerability to that person requires, and that these, in turn, are necessary components of living a good life. If that is so, then a reason

Relationships and Reasons for Belief

101

to have a belief that would beneft a signifcant other of yours would be weightier than that same reason would be if that person were not a signifcant other. There are, however, some worries to have about this account. First, let’s grant that being intimate with and vulnerable to a signifcant other are elements of the good life, and that intimacy and vulnerability require doxastic partiality on our part. Even so, it would not follow that these facts would intensify whatever particular reasons there are to be doxastically partial in particular instances. Vulnerability to and intimacy with a signifcant other are not things that are plausibly threatened by a departure from partiality in any particular instance. If that is so, then it’s not clear why our ability to be intimate with and vulnerable to our signifcant others would strengthen or intensify whatever non-evidential reason there is for belief in any particular case. The worry, in other words, is that it is not clear how the general importance of having relationships of intimacy with and vulnerability to others in our lives is going to help explain the weight of a reason to be doxastically partial in any particular instance, even if it’s the case that intimacy and vulnerability in general require some degree of doxastic partiality. Second, I think there may be good reasons to deny that genuine intimacy with and vulnerability to another person really does require something like doxastic partiality. Consider vulnerability, for example. Being vulnerable to another person involves, in part, exposing oneself to the possibility of a certain amount of risk, e.g., the possibility of being let down or hurt by that person. I take it that the riskiness of vulnerability is part of what makes vulnerability to another person so valuable in a personal relationship. But to form a belief about a signifcant other in order to beneft her, rather than form a belief about her in a way that responds to what one takes to be her features, seems precisely to avoid taking on that kind of risk – the risk, e.g., of being disappointed by that person.23 There is, of course, a sense in which believing well of another person – say, putting aside your skepticism about a friend’s new creative project – can allow that person to make herself more vulnerable to you. But the vulnerability here is not reciprocal. One wonders, then, whether vulnerability in the truly reciprocal sense is something that is compatible with doxastic partiality, let alone something that requires it. There may be other ways of trying to develop an account of how our relationships modify the non-evidential reasons for belief that arise in contexts beyond those of our relationships. One will want to make sure, however, that an account of what explains the modifying force of a relationship can be used to explain the weight of reasons in particular instances, and that the account of that relationship’s value is actually compatible with being doxastically partial toward one’s signifcant other.

102

6.

Lindsay Crawford

Reasons for Partiality, Doxastic and Practical

In this section, I want to address in a bit more detail an important aspect of the partialist’s defense of the claim that our personal relationships can give us (non-evidentially reducible) reasons to be doxastically partial – one that I mentioned briefy in Section 3, and which raises broader questions about partiality in personal relationships. The partialist takes it as common ground between her and her interlocutor that personal relationships can give us other kinds of reasons for partial treatment – reasons to act on behalf of a signifcant other; reasons to be motivated to act in particular ways; reasons to emotionally respond and attend to certain aspects of our signifcant others; and so on. The partialist holds that if you want to maintain that personal relationships can provide these other kinds of reasons for partiality, then you ought to accept that our personal relationships can give us reasons for doxastic partiality, too. And if you deny that our personal relationships can give us reasons for doxastic partiality, you ought to deny that our relationships can give us other kinds of reasons for partiality. Reasons of partiality come as a kind of package deal. Let’s examine the package-deal claim a bit more closely. Can one deny that our personal relationships can give us reasons to be doxastically partial, without having to deny that our relationships give us reasons for other forms of partiality? It would take us far beyond the scope of the paper to canvass the various ways philosophers have tried to defend the idea that we can have reasons to show our signifcant others partial treatment, and determine whether these particular views could consistently maintain that we do not have reasons to show our signifcant others doxastic partiality. Instead, I want to focus on a couple of ways that partialists have argued for the package-deal claim, and try to make the case that we may have resources to reject it. (To be clear, my goal here is only to assess the package-deal claim; I do not aim to defend any particular view of reasons of partiality more broadly.) The main argument in defense of the package-deal claim is one I briefy sketched in Section 3. The basic idea here goes as follows. Suppose you want to maintain that a friendship can give one reasons to extend partial treatment to a friend, and deny that that friendship can give one reasons to be correspondingly doxastically partial to that friend. It is possible, in other words, for your friendship to give you reason to act as if p, even though it does not give you reason to believe p. If so, then a perfectly good friend, responding to all of the reasons that his friendship provides, may act as if p, while failing to believe p. But a perfectly good friend, then, would be treating his friend insincerely. And we “do not conceptualize the good friend as manifesting a split of this kind between behavior and belief when it comes to her friend” (Stroud 2006: 505). So if you want to maintain that personal relationships can provide reasons to act

Relationships and Reasons for Belief

103

on behalf of a friend, while holding that relationships do not provide reasons for beliefs that correspond to those actions, then you will have to maintain either that a friend is not at all worse as a friend for being insincere, or that our personal relationship can provide reasons for belief. It is diffcult to get a handle on what it means to act in a way that does not “correspond” to one’s beliefs, and how this can give rise to an objectionable form of insincerity. It’s not clear what beliefs “correspond” to my picking up a friend from the airport, for example, such that responding to whatever reasons I have to pick my friend up from the airport, without responding to reasons for the beliefs that “line up” with that action, would constitute a form of insincerity. The best cases for the partialist to focus on here are cases in which the relevant action is an assertion: e.g., one asserts that p while believing not-p. To motivate the partialist’s intuition, consider the following two variants on our case involving Stan. In the frst case, you tell Stan that his doubts about his talent more likely stem from imposter syndrome than a clear-eyed assessment of his talents. Moreover, you actually believe this. In the second case, you tell Stan that his doubts about his talent more likely stem from imposter syndrome than a clear-eyed assessment of his talents. But this time, you don’t believe it. Let’s suppose, in fact, that you think he’s probably seeing things quite clearly; he is glimpsing the fact that he really doesn’t have what it takes. Are you worse as a friend in the second case, compared with the frst case? The intuition that there is something worse about the second case compared with the frst case is a powerful one. But it’s an intuition we may be able to accommodate without having to grant that your friendship with Stan gives you a reason to be doxastically partial toward him. Let’s grant that there is something worse about the second case. It’s plausible to think that what makes it worse is the fact that you intentionally lead Stan to believe that you have a certain view of things that you don’t actually have. The thought is that friendship is in tension, not with some kind of belief-action misalignment per se, but with intentionally leading a friend to believe things that you don’t actually believe. But if that’s what makes the second case worse than the frst case, then there’s no more reason to think that you have reason to be doxastically partial to Stan than that you do not have reason to tell Stan what you tell him. A second line of defense for the package-deal claim is grounded in the observation that what we believe about a signifcant other is deeply bound up with how we treat them and how we are motivated to act on their behalf. Along these lines, some partialists hold that, absent a compelling reason to think there is some deep difference between these different domains, we should allow that if we can have reasons in one domain, we have reasons in all of them. Hazlett puts the point particularly clearly: [P]artial love requires partiality in action and in motivation; but there is no principled way to concede that while denying that partial

104

Lindsay Crawford love requires partiality in belief. For what makes the domain of belief special, that would insulate it from the kinds of partiality that seem required in the domains of action and motivation? Why would we expect it to be special, in this way? There is no reason to expect that. So partial love requires partiality in belief. Hazlett 2013: 93

Of course, if evidentialism about reasons for belief were correct, then we would have a principled way to distinguish between doxastic partiality and other forms of partial treatment. But even if we don’t grant evidentialism, there may be other ways to mark an important distinction here. The following brief remarks are offered only as an incomplete sketch of how one might begin to develop a response to the partialist’s packagedeal claim on grounds that do not presuppose evidentialism. One way to draw a distinction is to note that beliefs are shareable in a way our actions and motivations are not. To say that a belief is “shareable” is to say that another person can come to believe what you believe. One way to share a belief with another person is to share your reasons for belief with that person, which she can then take as her own reasons for belief. If the partialist is right that our personal relationships can provide reasons for belief, these are not reasons you could in principle share with another person, in the specifc sense that those reasons could not become another person’s reasons for believing what you believe. For example, a friendship-generated reason you have to believe that Stan has what it takes is not a reason that anyone else who isn’t similarly related to Stan could come to take as a reason to believe that Stan has what it takes. Maintaining that our personal relationships can give us reasons for beliefs about our signifcant others may make it diffcult, then, to see how those beliefs could be eligible to play the same role in testimonial transmission as the rest of our beliefs. There are also broader, related questions here about the connection between belief and assertion. If our relationships give us reasons for belief about our signifcant others, are these beliefs ones that we can justifably assert to people who are not participants in those relationships? It’s not clear that assertion could be justifable, if our reasons aren’t shareable in the relevant sense. There is more to say here. The broader suggestion I want to offer for now is that one might fnd principled grounds for rejecting the package-deal claim by looking to the ways that our beliefs about our signifcant others stand to play important roles well beyond the contexts of the relationships in which those beliefs arise and, according to the partialist, are rationalized.

7.

Conclusion

If we reject evidentialism about reasons for belief and so grant that at least some of our reasons for belief consist in considerations that are not evidentially reducible, we need to determine what exactly those reasons

Relationships and Reasons for Belief

105

consist in, and in virtue of what they can be reasons for belief. My focus here has been specifcally on reasons for belief that our personal relationships might be thought to provide, partly because pragmatists about reasons for belief often appeal to defenses of doxastic partiality in the literature to motivate their case for pragmatism. I think we face certain important, distinctive challenges in trying to develop a view on which our personal relationships could provide reasons for belief – challenges that we may not face when trying to explain how we could have other kinds of non-evidentially reducible reasons for belief (perhaps reasons of prudence or self-interest). In particular, if one wants to provide an account of how our relationships can give us reasons for belief, one would need to fnd a way to restrict the scope of reasons for belief, in ways that would help illuminate how our relationships could appropriately ground or explain those reasons. In my view, the partialist has not met this challenge.24

Notes 1. Following Paul/Morton (2018), I use the term “signifcant other” to refer to those to whom we are related in personal relationships, including friendship, romantic partnerships, familial relationships, and the like. 2. This defnition of doxastic partiality is in keeping with other contemporary discussions of doxastic partiality in, e.g., Stroud (2006), Keller (2004, 2018), and Hazlett (2013), among others. 3. In what follows, I will sometimes use “reasons to be doxastically partial” or “reasons for doxastic partiality” as shorthand for reasons for beliefs that would exhibit doxastic partiality toward our signifcant others. 4. Some philosophers claim that whatever sort of normative pressure our relationships might exert on our beliefs couldn’t come in the form of reasons for belief, precisely because they think that only evidential considerations could serve as reasons for belief (see, e.g., Paul/Morton (2018) and Arpaly/Brinkerhoff (2018). Goldberg’s (2018) discussion of the different practical and epistemic reasons that our friendships provide also presupposes evidentialism about reasons for belief. As I will argue later, we should not grant this evidentialist assumption that so frequently shapes the discussion of the normative signifcance of doxastic partiality in personal relationships. 5. The question here is not whether you should have various beliefs about whether he’ll be fortunate enough to encounter and avoid the various elements of good and bad luck outside of his control that will play a role in determining his success (like whether he’ll be in the “right place at the right time,” and so on). Whether or not Stan succeeds is not something that is entirely up to him. What Stan wants to know is whether you think he has what it takes to succeed. 6. To be clear, this is not the only response available to the evidentialist. Later in this section I examine other kinds of responses an evidentialist might have to this case, including one on which we must distinguish between having a reason to believe that Stan has what it takes and having a reason to get yourself to believe that Stan has what it takes. 7. See esp. Shah (2006) and Kelly (2002) for a development of this defense of evidentialism. See also Parft (2011), Raz (2011), and Kolodny (2005), among others.

106

Lindsay Crawford

8. Maguire/Woods (forthcoming) also frame the use of the deliberative constraint in this argument for evidentialism in terms of the “followability” of reasons. 9. See, e.g., Rinard (2015, 2019a, 2019b), McCormick (2015, ch. 2), and Leary (2017). 10. This is the move Shah (2006) makes in his discussion of Pascalian deliberation. 11. See Rinard (2019a) for more on this point. 12. In fact, an evidentialist who doesn’t want to commit himself to a view about what is practically rational for the student to do might respond in a different way: “Well, if you’re asking whether you should believe that God exists, then no. As to what you should get yourself to believe, I’m not in a position to say. On that you might consult someone who has a more informed view about practical rationality.” But this response runs into the second of the two problems I raise against the frst bit of advice. 13. This is a position adopted by Feldman (2000: 691–694). 14. See, e.g., Rinard (2015, 2019a, 2019b), Maguire/Woods (forthcoming), McCormick (2015), and Reisner (2018). 15. See Kawall (2013) and Paul/Morton (2018) for more on what I’m calling thresholds partialism. Thresholds partialism can be developed in ways broadly similar to various forms of “pragmatic encroachment” accounts of epistemic justifcation in the literature. 16. Hazlett (2013) makes a similar point: “someone who did not exhibit partiality in action or partiality in motivation would not be a genuine, true, or real friend. [. . .] Just as failures of partiality in action and motivation strike us, intuitively, as betrayals of friendship, so do failures of partiality in belief. We expect our friends to think well of us, we think that they ought to think well of us, and to the extent that they fail to do so, they strike us as less than genuine friends” (90). 17. In his response to the suggestion that friendship gives us reasons to act but not reasons for belief, Hazlett responds that “friendship, on this proposal, seems to require an unappealing kind of insincerity. [. . .] If we reject the idea that friendship ever requires this kind of insincerity, then we should conclude that (some) friendships really do require (a disposition towards) partiality bias” (2013: 92). We fnd a similar appeal to the insincerity point in a number of other defenses of partialism. Stroud notes that “[a] good friend does not defend her friend outwardly [. . .] while inwardly believing the worst of her friend. We do not conceptualize the good friend as manifesting a split of this kind between behavior and belief when it comes to her friend. Rather, the good friend is prepared to take her friend’s part both publicly and, as it were, internally” (2006: 505). In a similar vein, Keller responds to the objection that friendship can require us to act in ways that do not line up with our beliefs in the following way: “I doubt [. . .] that this kind of pretend approval is what we really want in a good friend. [. . .] You want a friend who’s on your side, not one who’s good at faking it” (2004: 335). 18. I will return to this defense of the constitutive claim in Section 6. 19. See, e.g., Arpaly/Brinkerhoff (2018). 20. One might think that the singer has at least some pro tanto reason to show the fan partiality, albeit one that is massively outweighed by all the other reasons there are not to show the fan partiality. I consider this sort of view in the following section. 21. Thanks to Mark Schroeder and others for pressing me on this. 22. For more on the distinction between reasons and reason-modifers, see Dancy (2004).

Relationships and Reasons for Belief

107

23. On a much broader objection to partialism that is related to this one, see Crawford (2019). 24. For helpful questions and feedback, I would like to thank audience members at the symposium on practical reasons for belief at the 2019 American Philosophical Association Central Division Meeting, and participants at the 2019 Philosophy Desert Workshop. Special thanks to Nico Cornell, Kathryn Lindeman, Eliot Michaelson, David Plunkett, Sebastian Schmidt, and Amy Sepinwall.

References Arpaly, Nomy/Brinkerhoff, Anna (2018): “Why Epistemic Partiality Is Overrated”, Philosophical Topics 46, 37–51. Crawford, Lindsay (2019): “Believing the Best: On Doxastic Partiality in Friendship”, Synthese 196, 1575–1593. Dancy, Jonathan (2004): Ethics without Principles, New York: Oxford University Press. Feldman, Richard (2000): “The Ethics of Belief”, Philosophy and Phenomenological Research 60, 667–695. Goldberg, Sanford (2018): “Against Epistemic Partiality in Friendship: ValueRefecting Reasons”, Philosophical Studies 176, 2221–2242. Hazlett, Allan (2013): A Luxury of the Understanding: On the Value of True Belief, New York: Oxford University Press. Kawall, Jason (2013): “Friendship and Epistemic Norms”, Philosophical Studies 165, 349–370. Keller, Simon (2004): “Friendship and Belief”, Philosophical Papers 33, 329–351. Keller, Simon (2013): Partiality, Princeton: Princeton University Press. Keller, Simon (2018): “Belief for Someone Else’s Sake”, Philosophical Topics 46, 19–35. Kelly, Thomas (2002): “The Rationality of Belief and Other Propositional Attitudes”, Philosophical Studies 110, 163–196. Kolodny, Niko (2005): “Why Be Rational?”, Mind 114, 509–563. Leary, Stephanie (2017): “In Defense of Practical Reasons for Belief”, Australasian Journal of Philosophy 95, 529–542. Lord, Errol (2016): “Justifying Partiality”, Ethical Theory and Moral Practice 19, 569–590. Maguire, Barry/Woods, Jack (forthcoming): “The Game of Belief”, The Philosophical Review. McCormick, Miriam Schleifer (2015): Believing against the Evidence: Agency and the Ethics of Belief, London: Routledge. Parft, Derek (2011): On What Matters, New York: Oxford University Press. Paul, Sarah/Morton, Jennifer (2018): “Believing in Others”, Philosophical Topics 46, 75–95. Raz, Joseph (2011): From Normativity to Responsibility, New York: Oxford University Press. Reisner, Andrew (2018): “Pragmatic Reasons for Belief”, in: The Oxford Handbook of Reasons and Normativity, ed. by Star, New York: Oxford University Press, 705–726.

108

Lindsay Crawford

Rinard, Susanna (2015): “Against the New Evidentialists”, Philosophical Issues 25, 208–223. Rinard, Susanna (2019a): “Equal Treatment for Belief”, Philosophical Studies 176, 1923–1950. Rinard, Susanna (2019b): “Believing for Practical Reasons”, Noûs 53, 763–784. Shah, Nishi (2006): “A New Argument for Evidentialism”, Philosophical Quarterly 56, 481–498. Stroud, Sarah (2006): “Epistemic Partiality in Friendship”, Ethics 116, 498–524.

6

Instrumental Reasons for Belief Elliptical Talk and Elusive Properties Asbjørn Steglich-Petersen and Mattias Skipper

1.

Introduction

There are beliefs you should have and ones you shouldn’t. What makes it so? Epistemic instrumentalists give something like the following answer: you have epistemic reason to believe a proposition insofar as doing so is conducive to certain epistemic goals or aims – say, to believe what is true and avoid believing what is false. According to them, epistemic normativity is just a special kind of instrumental normativity. Among the main attractions of this view are that it unifes epistemic and practical normativity and explains epistemic normativity in terms of a relatively well-understood kind of normativity.1 Perhaps the most prominent challenge for epistemic instrumentalists in recent years has been to explain, or explain away, why beliefs seem normatively constrained by one’s evidence, even when believing in accordance with the evidence does not seem to further any of one’s aims. This challenge can arguably be met.2 But a different challenge looms: instrumental reasons in the practical domain have various properties that epistemic reasons do not seem to share. We will say more about these properties as we proceed. But if epistemic reasons indeed do not share them, that speaks against understanding epistemic reasons as a species of instrumental reasons. The goal of this chapter is to offer a way for epistemic instrumentalists to overcome this challenge. Our main thesis takes the form of a conditional: if we accept an independently plausible transmission principle of instrumental normativity, we can maintain that epistemic reasons in fact do share the relevant properties of practical instrumental reasons. In addition, we can explain why epistemic reasons seem to lack these properties in the frst place: some properties of epistemic reasons are elusive, or easy to overlook, because we tend to think and talk about epistemic reasons in an “elliptical” manner (in a sense to be explained). We will go about defending these claims in three steps. First, we will provide some general background on epistemic instrumentalism and introduce the transmission principle that we will be working with

110

Asbjørn Steglich-Petersen and Mattias Skipper

(Section 2). In the second step, we will explain what we mean by our tendency to think and talk about epistemic reasons in an “elliptical” manner (Section 3). The third step is where most of the action will take place: we describe four properties of practical instrumental reasons that epistemic reasons seem to lack, and we draw on lessons from the previous two steps to explain how epistemic instrumentalists can maintain that epistemic reasons in fact share those properties, despite initial appearances to the contrary (Section 4).

2.

Epistemic Instrumentalism

Why does evidence for a proposition seem to provide normative reason to believe it? And why should we avoid believing what is not supported by our evidence? These are questions that any theory of epistemic normativity must answer. Epistemic instrumentalists propose to answer them by understanding epistemic reasons as a species of instrumental reasons. We often have reason to seek the truth about various matters, and since the best way to promote this aim is to believe in accordance with our evidence, we have good instrumental reason to do so. Although all epistemic instrumentalists share this basic commitment, specifc versions differ in detail as well as motivation. One of the main motivations for epistemic instrumentalism has traditionally been to reconcile a cognitivist view of epistemic normativity with a broadly naturalistic metaphysics. This ambition calls for a purely descriptive version of epistemic instrumentalism, which grounds instrumental reasons in the aims and desires that agents actually hold. On this sort of view, the question of whether an agent has instrumental reason to believe some proposition is settled on the basis of whether the agent in fact aims to form a true belief about the proposition (and, if so, how that aim is best pursued).3 While the metaphysical tidiness of this approach may seem attractive, it is beset with serious diffculties. Perhaps the most well-known problem is that we do not seem to have enough, or suffciently general, epistemic aims or goals to ground all of the epistemic reasons that we intuitively seem to possess.4 For one thing, there are propositions that most people seem to care very little about (“Which percentage of the world population are currently singing in falsetto?”). And to make matters worse, there are even propositions that most people seem to prefer not to form true beliefs about (“How does the crime novel end?”). Yet, even in such cases, many philosophers fnd it intuitive that our beliefs should respect the available evidence. If so, the limited scope of people’s actual epistemic aims appears to make them unft to explain the generality with which our beliefs are normatively constrained by the evidence. One way to avoid this diffculty is to adopt a normatively loaded version of epistemic instrumentalism on which instrumental reasons are

Instrumental Reasons for Belief

111

grounded, not in people’s actual aims or desires, but in the aims or desires they should (or have reason to) have. This move is supposed to make the epistemic instrumentalist better equipped to explain why our epistemic reasons are not sensitive to our actual aims (or lack thereof), since we might fail to have the aims that we should have.5 The obvious cost of this proposal is that it is not easily adaptable by a naturalist metaphysics, since it relies on normative claims about what aims we should have, and not just psychological facts about what aims we do have. Epistemologists of a naturalistic bent will perhaps fnd this cost unacceptable. But it is important to note that there are plenty of attractions left to make the account theoretically desirable: it promises a unifed theory of epistemic and practical normativity, and grounds epistemic norms in a relatively well-understood class of instrumental norms. There are several different ways that one could develop this normatively loaded version of epistemic instrumentalism, and we will not be presupposing any version in particular. But it will be helpful to have a concrete view on the table, so we will briefy sketch an account that we have defended elsewhere.6 To motivate the account, we frst need to get clear on the proper explanandum for a theory of epistemic normativity. We mentioned earlier that one of the central challenges for epistemic instrumentalists is to explain the generality with which our beliefs seem normatively constrained by evidence. But what, exactly, do these general constraints amount to? A natural initial suggestion would be that evidence for p always provides one with a pro tanto reason to believe that p, and that decisive evidence for p always makes it the case that one ought to believe that p. However, we fnd it doubtful that the general normative demands generated by our evidence are as strong as that. As several authors have pointed out, even if we restrict our attention to propositions that are entailed by our evidence, we are left with a set of propositions that vastly outstrips what any human being could ever believe – endless arrays of disjunctions, tautologies, and so on. Thus, if “ought” implies “can,” it cannot be the case that we ought to believe whatever we have decisive evidence for.7 Moreover, even if we abstract away from contingent facts about our cognitive limitations, it seems implausible that we ought to, or even have reason to, believe various trivial propositions that are of no interest to us. What reason is there for me to believe that “either the earth isn’t fat or the number of tables in Shanghai is odd”? Very little, it seems. And yet the proposition is well supported by my evidence. This suggests that if evidence places any general normative constraints on our beliefs, it cannot be by always generating positive reasons to form beliefs that are supported by the evidence. The constraint must be weaker than that. A more plausible suggestion, it seems to us, is that the strongest general evidential constraint on belief is a necessary condition on permissible

112

Asbjørn Steglich-Petersen and Mattias Skipper

belief: roughly, you are permitted to believe p only if p is supported by your evidence (to a suffcient degree).8 This norm says nothing about when you should, or have reason to, believe p. Rather, when you have positive reason to believe some proposition, this must be explained, at least in part, by non-evidential factors of a certain sort. If this is correct, a theory of epistemic normativity must be able to explain two separate normative phenomena. First, it must explain why it is never permissible to form evidentially unsupported beliefs. Second, it must explain why there is sometimes, but not always, positive reason to believe a proposition that is supported by the evidence. On our preferred view, these normative phenomena are explained by two separate principles that feature prominently in the literature on instrumental normativity. The frst principle is a coherence norm that governs which combinations of instrumental attitudes and actions it can be rational to take. Roughly speaking, if you have a certain aim and believe some means to be necessary to achieve that aim, you are instrumentally irrational if you fail to either take the means or give up the aim. This norm says nothing about which particular means you should take, since you may or may not have reason to pursue the relevant aim in the frst place. By contrast, the second norm is a substantive norm, namely a “transmission principle” that governs when you have positive reason to take particular actions or attitudes.9 Roughly speaking, if you have reason to pursue a certain aim and you should expect some means to be conducive to that aim, you thereby have a pro tanto reason to take the means. Together, these two norms provide the resources to explain how our beliefs are normatively constrained by the evidence. As argued elsewhere, the general sense in which beliefs are permissible only if they are evidentially supported can be explained in terms of the coherence norm of instrumental rationality (when suitably adapted to the epistemic context).10 And the transmission principle can be used to explain why we sometimes, but not always, have positive epistemic reason to adopt beliefs that are supported by our evidence. For reasons that will emerge, the ensuing discussion is mainly going to center around the transmission principle. It is therefore worth pausing to give a more refned formulation of the principle. Generally speaking, such a formulation should capture two ideas. First, it should capture the idea that one’s reasons to take some means depend on (are “transmitted” from) one’s reasons to pursue the relevant aim. Second, it should capture the idea that the strength of one’s reasons to take the means depend partly on the strength of one’s reasons to pursue the aim and partly on the likelihood that the means will be effective in achieving the aim. One way of capturing these (and other, more subtle) features has been proposed by Niko Kolodny (2018):11 General Instrumental Transmission: If there is reason for one to pursue an aim, A, and there is positive probability conditional on one’s

Instrumental Reasons for Belief

113

M-ing, that this helps bring about A non-superfuously, then that is a reason for one to M, whose strength depends on the reason for one to pursue A and the probability. How might this principle be applied to understand the situations in which one has positive epistemic reason to believe some proposition? The idea is that evidence for p speaks in favor of believing p only when there is reason to pursue the aim of coming to a true belief as to whether p. With such a “truth-aim” in place, the evidence helps to generate an epistemic reason to believe p by raising the likelihood that believing p will fulfll the aim of coming to a true belief as to whether p. If we make the relevant substitutions, we thus get the following instantiation of the general transmission principle: Instrumental Transmission of Epistemic Reasons for Belief: If there is reason for one to pursue the aim of coming to a true belief as to whether p, and there is positive probability conditional on one’s adopting a belief that p, that this helps bring about the aim nonsuperfuously, then that is a reason for one to adopt a belief that p, whose strength depends on the reason for one to pursue the aim and the probability. To sketch a toy example, suppose that you have good reason to pursue the aim of believing the truth as to whether the toy rubber duck you are considering as a birthday present for your niece contains harmful chemicals. This reason will transmit to various means. One such means might be to seek out further evidence on the matter. But the reason will also transmit to another means that, if true, will constitute the fulfllment of the aim, namely that of believing that the rubber duck contains harmful chemicals. The strength of the epistemic reason you have to adopt this belief depends on two factors. First, it depends on the strength of the reason you have to pursue a true belief on the matter (which in this case is presumably fairly strong). Second, it will depend on how likely it is in light of your evidence that believing that the rubber duck contains harmful chemicals will bring about, or rather constitute, a true belief on the matter. The stronger the evidence, the stronger the transmitted reason, other things being equal. So far, we have focused our attention on the epistemic goal of forming a true belief as to whether p. But there is also another epistemic goal that deserves equal attention: the aim of not forming a false belief as to whether p. Famously, William James pointed out that these aims do not go hand in hand: someone can achieve one aim without achieving the other.12 However, for the most part, it will do us no harm to talk about a single “truth-aim,” which we will take to stand in for the dual aim of believing what is true and not believing what is false. Whenever it becomes important to keep these aims separate, we will do so explicitly (especially in Section 4.2).

114 Asbjørn Steglich-Petersen and Mattias Skipper Before we proceed, let us note that the version of epistemic instrumentalism outlined here is compatible with a number of different views about when and why we have reason to adopt the truth-aim relative to some proposition. Some philosophers hold that there is always some intrinsic value in having true beliefs, no matter the content of those beliefs (see, e.g., Lynch 2004). If so, the present view implies that there is always at least a weak transmitted reason to believe any evidentially supported proposition. On another view, which we are inclined to favor, one has reason to adopt the truth-aim only when doing so is likely to promote the practical values or aims that one has reason to promote or pursue. These need not be restricted to prudential aims or values, but might concern moral, intellectual, or aesthetic matters as well. For example, even if it is not in my own narrow prudential interest to seek true beliefs about chemicals in toys, there might be moral reasons to do so. And even if no prudential advantage would be gained by learning the true nature of black holes, it might still be of purely intellectual value.

3.

Elliptical Talk About Epistemic Reasons

An initial question for the epistemic instrumentalist is what we should make of the common practice of referring to evidence simpliciter as “reasons” for belief. On our view, it is strictly speaking false to say that evidence by itself constitutes a normative reason for belief. Why, then, do we often talk in ways that suggest otherwise? To our mind, the practice of referring to evidence simpliciter as “reasons for belief” is best understood as a way of talking about epistemic reasons in an “elliptical” manner. The idea is that, in most ordinary contexts, we have good pragmatic reason to take the truth-aim for granted and simply focus our attention on whether the available evidence makes it likely that believing p will result in a true belief as to whether p. This makes it appear as if the question of whether there is epistemic reason to believe p is settled by the question of whether the evidence supports p alone.13 The same phenomenon is also familiar from practical contexts where the aim in question “goes without saying.” Consider, for example, the reasons we have to avoid doing things that will get us killed.14 Strictly speaking, these reasons only arise for people who have (or should have) the aim of not getting killed. But most people share this aim; it would, at least in most ordinary contexts, belabor the obvious to make explicit mention of this aim. That is why it typically seems perfectly natural to refer to, say, the risks of falling to one’s death during free solo climbing as a reason not to try it, without mentioning the underlying aim of staying alive. Such talk is elliptical in the sense that it omits reference to the underlying aim, which must instead be inferred from the context, and instead focuses on whether the available means are likely to be effective in achieving the contextually given aim.

Instrumental Reasons for Belief

115

The same goes for our talk about epistemic reasons. With respect to the vast majority of propositions that we ordinarily think and talk about, it seems plausible that we have at least some reason to pursue the aim of forming true beliefs (and not forming false beliefs) about those propositions. Thus, it is typically safe to take this truth-aim for granted and focus solely on whether the evidence at hand makes it likely that believing the relevant proposition will achieve the aim of believing that proposition truly. Of course, it may become relevant to draw attention to the underlying truth-aim in cases where the aim is called into question. For example, if someone starts rehearsing an endless number of trivialities that follow from your evidence, you would be right to object that you simply do not care about those propositions. This reaction, we take it, is just a way for you to convey that you have no reason to form beliefs about those trivialities, even if they are well supported by your evidence.15 It is like rejecting the relevance of a piece of advice about how to achieve some aim on the grounds that you have no reason to pursue the aim in the frst place. Yet, in most ordinary contexts, it will be entirely felicitous to take the truth-aim for granted and speak of evidence simpliciter as a reason for belief. This fact will become important in what follows.

4.

Epistemic Reasons versus Practical Instrumental Reasons

We will now take up the challenge against epistemic instrumentalism that we set out to address in this chapter. The challenge arises when we try to single out various properties of practical instrumental reasons, and ask whether epistemic reasons share those properties. In what follows, we will look at four properties in particular. Two of those properties pertain to contributing or pro tanto reasons (Sections 4.1–4.2), while the other two pertain to how those contributing reasons combine into overall or all-things-considered reasons (Sections 4.3–4.4). In each case, we will argue that epistemic instrumentalists can explain away the seeming contrast between epistemic reasons and practical instrumental reasons by paying close attention to the mechanics of the transmission principle. And we will argue that they can do so in a way that makes it clear why the contrast seems to arise in the frst place. There are no doubt other properties aside from the ones discussed later that might be thought to drive a wedge between epistemic and instrumental normativity. But even so, we hope that the considerations put forth will at least go some way to discharge the worry that epistemic reasons and practical instrumental reasons behave too differently to be of a kind. 4.1

Bounded Versus Unbounded Scales

The frst apparent contrast between epistemic reasons and practical instrumental reasons we want to discuss concerns the scales on which the

116

Asbjørn Steglich-Petersen and Mattias Skipper

weights of these reasons are to be measured. More specifcally, the question of our concern is whether these scales are bounded or unbounded. Let’s begin with the practical domain. It is natural to think that no matter how strong a practical reason you have to perform some action, you can always, at least in principle, get additional reason to perform that action. For example, even if the prospect of winning $1,000,000 gives you an extremely strong reason to φ, the prospect of winning $1,000,001 might give you an even stronger reason to φ. This seems to suggest that we should measure the weights of practical instrumental reasons on an unbounded scale: there is no upper limit beyond which no further strengthening of practical reasons can be gained.16 Things look different in the epistemic domain. As Reisner (ms) has pointed out, we are used to thinking that if you have conclusive or decisive evidence for p – i.e., evidence that makes it certain that p is true – no further evidence could give you additional reason to believe p. For example, if your current evidence includes the facts that “it’s raining” and “if it’s raining, the streets are wet,” any further evidence you might get about the wetness of the streets would seem inert. After all, your current evidence already entails that the streets are wet; evidence doesn’t get any stronger than that! This seems to suggest that we should measure the weights of epistemic reasons on a bounded scale: there is an upper limit beyond which no further strengthening of epistemic reasons can be gained. Intuitive as the contrast may seem, we think the epistemic instrumentalist can explain it away. Doing so will require us to separate the two main inputs to the transmission principle, which jointly determine the weight of the transmitted reason. The frst input consists of the reasons you have to pursue the aim of forming a true belief as to whether p. The scale associated with this input is presumably unbounded: you can always, at least in principle, get additional reason to pursue the aim of forming a true belief as to whether p. The second input is the evidential likelihood that believing p will result in a true belief as to whether p. The scale associated with this input is bounded: no amount of evidence can make it more than certain that believing p will result in a true belief as to whether p. (The point obviously generalizes beyond the epistemic domain: even the most effective means can at most guarantee that the relevant aim will be achieved.) Once we keep these two components of the transmission principle separate, it becomes clear how the instrumentalist can maintain that epistemic reasons, just like practical instrumental reasons, have unbounded weights. Although the evidence can at most make it certain that believing p will result in a true belief as to whether p, the transmitted epistemic reason to believe that p will nevertheless be unbounded in weight, since the reasons to pursue the aim of forming a true belief as to whether p are unbounded in weight. (Again, this point generalizes beyond the epistemic domain: although φ-ing can at most guarantee the achievement

Instrumental Reasons for Belief

117

of some aim, the transmitted instrumental reason to φ will nevertheless be unbounded in weight, since the reasons one might have to pursue the relevant aim are unbounded in weight.) So far, so good for the instrumentalist story. But why does it seem to us that epistemic reasons have bounded weights, if really they do not? That’s a fair question. We think that the epistemic instrumentalist has a good answer, one that centers around the tendency to think and talk about epistemic reasons in an elliptical manner. As explained earlier, we often take the truth-aim for granted and focus solely on whether the evidence at hand makes it likely that believing p will result in a true belief as to whether p. And since the evidential likelihood is measured on a bounded scale, it is thus not surprising that we tend to think of epistemic reasons as having bounded weights, although, strictly speaking, they do not. 4.2 Miniscule Reasons Another seeming contrast between epistemic reasons and practical instrumental reasons can be brought out by refecting on the following consequence of the transmission principle: as long as there is a non-zero evidential probability that a given means will achieve a certain aim that you have reason to pursue, you will have at least a miniscule reason to take the means. How does this feature of the transmission principle fare in the practical and epistemic domains, respectively? In the practical domain, the feature looks appropriate. Suppose you aim to become rich, and suppose that buying a lottery ticket gives you a small, but non-zero, chance of achieving that aim. Do you thereby have a practical reason to buy the ticket? Sure! Of course, the reason will easily be outweighed by other reasons you might have not to buy the ticket (say, reasons transmitted from the aim of not having to pay for the lottery ticket). But this is fully compatible with your having a miniscule reason in favor of buying the ticket. Things look different in the epistemic domain. Suppose you aim to form a true belief as to whether p, and suppose that your evidence leaves a small, but non-zero, probability that p is true. Do you thereby have an epistemic reason in favor of believing p? It seems not. Rather, you seem to have strong epistemic reason against believing p. After all, your evidence speaks strongly against p. When taken at face value, this seems to suggest that the transmission principle has implausible consequences in the epistemic domain. However, we think the face value is misleading. On closer inspection, epistemic instrumentalists can maintain that you do have a miniscule epistemic reason to believe that p whenever the evidential probability of p is non-zero. To see why, note that the aim of coming to a true belief as to whether p is equally fulflled by believing p when p is true, and by disbelieving p when p is false. Thus, whenever there is a non-zero

118

Asbjørn Steglich-Petersen and Mattias Skipper

evidential probability of p being true, there will be two transmitted epistemic reasons in play at once: one in favor of believing p, and one in favor of disbelieving p. What is the relative weight of these opposing reasons? In cases where the evidential probability of p is very low, the latter reason clearly wins the day: you will have a miniscule reason to believe p, but this reason will be outweighed by a much stronger reason to disbelieve p. This is why it’s tempting to think that you have no reason to believe p when your evidence speaks against p, although, strictly speaking, you do. Before we proceed, let’s anticipate a potential worry: how can the very same evidential input to the transmission principle yield both a reason to believe p and a reason to disbelieve p? Doesn’t this amount to saying, absurdly, that the same body of evidence can both raise and lower the probability of p? No. The claim is not that the same evidential input to the transmission principle affects the probability of p in two different ways. Rather, the claim is that the same evidential input to the transmission principle combines with two different means to yield two different reasons: a reason in favor of believing p, and a reason in favor of disbelieving p. And there is nothing absurd about that. 4.3 “Permissive” Versus “Prohibitive” Balancing The next property we want to discuss shows up in situations where you have equally strong reasons to take each of two incompatible actions or attitudes. The question is what you have all-things-considered or overall reason to do or believe in situations like this. Let’s frst consider the practical domain. Suppose you face a choice between cake and ice cream, and that you are equally keen on having cake as you are on having ice cream. Provided that you have no other reasons bearing on the matter, it then seems natural to say that either option is permitted: you may have cake, and you may have ice cream. More generally, if φ and ψ are two incompatible actions such that you have equally strong practical reason to perform each of them, it seems natural to say that either action is permitted: you may φ, and you may ψ. Following Berker (2018), let’s call this property “permissive balancing.” Do epistemic reasons, like practical instrumental reasons, exhibit permissive balancing? It seems not. Suppose you are deliberating about whether to believe or disbelieve p, and suppose that your evidence consists of two equally trustworthy experimental studies: one supporting p, the other supporting ¬p. Here it seems that neither attitude is permitted: you may not believe p, and you may not disbelieve p. Rather, you should suspend judgment about p. Following Berker (2018), let’s call this property “prohibitive balancing.” How might the epistemic instrumentalist explain away this seeming contrast? Once again, we think the key is to pay close attention to the mechanics of the transmission principle. The frst thing to note is that,

Instrumental Reasons for Belief

119

in specifying the epistemic case, we said that the two experimental studies favor p and ¬p equally strongly. But merely specifying the evidential input to the transmission principle is not enough to determine the weights of the transmitted epistemic reasons. We also need to specify the relative importance of the epistemic aims that are operative: the aim of forming a true belief as to whether p, and the aim of not forming a false belief as to whether p. What is the relative weight of these aims? As a number of authors have observed, it seems to be a widely shared commitment that it is more important to avoid false beliefs than to secure true ones. We should, in other words, be epistemically conservative or risk-averse.17 Why? Here is a quick argument that gets the basic idea across: it seems obvious that you shouldn’t believe p unless p is more likely to be true than false given your total evidence. However, if you found it just as important to gain true beliefs as to avoid false ones, it’s hard to see why this would be so. After all, believing p would give you just as good a chance of gaining a true belief as to gain a false one. Thus, it seems that you should consider it more important to avoid false beliefs than to gain true ones. That’s the quick argument. As we argue elsewhere, it is surprisingly diffcult to make the argument hold up under closer scrutiny (Steglich-Petersen/ Skipper ms). But for present purposes, let’s take it for granted that agents should be epistemically risk-averse, at least to some extent. Given this assumption, the transmission principle delivers the result that you have more epistemic reason to suspend judgment about p than to believe or disbelieve p when the evidence is equally balanced for and against p. Believing (or disbelieving) p gives you a 50% likelihood of achieving the aim of forming a true belief as to whether p, and a 50% likelihood of achieving the aim of not forming a false belief as to whether p. Suspending judgment about p gives you a 0% likelihood of achieving the aim of forming a true belief as to whether p, and a 100% likelihood of achieving the aim of not forming a false belief as to whether p. Thus, assuming that the latter aim is more important than the former, the transmitted epistemic reason to suspend judgment about p is stronger than the transmitted epistemic reason to believe (or disbelieve) p. That’s why you are neither permitted to believe p nor permitted to disbelieve p. You might think that this result just reaffrms the apparent contrast between epistemic reasons and practical instrumental reasons. However, the opposite is the case. These considerations reveal a subtle, but crucial, disanalogy between the two cases that was used to motivate the difference between permissive and prohibitive balancing. On the one hand, the practical case is one where the two incompatible options – cake versus ice cream – are associated with a set of aims (say, having to do with hunger, pleasure, etc.) that do not at the same time give rise to a transmitted reason to have neither cake nor ice cream. By contrast, the epistemic case is one where two incompatible attitudes – believing p versus disbelieving

120

Asbjørn Steglich-Petersen and Mattias Skipper

p – are associated with a set of aims (forming a true belief and not forming a false belief) that do at the same time give rise to a transmitted reason to neither believe p nor disbelieve p. The reason, then, why there seems to be a contrast between epistemic reasons and practical instrumental reasons is that we test them against very different cases. In order to test whether the apparent contrast is genuine, we need to consider a practical case that is relevantly similar to the epistemic case. So, consider: Food Boxes You stand before two opaque boxes. You are 50–50 about whether Box 1 contains cake or carrots, and you are 50–50 about whether Box 2 contains ice cream or lettuce. You face a choice between three options: (i) eating the contents of Box 1, whatever it contains; (ii) eating the contents of Box 2, whatever it contains; or (iii) eating nothing at all. There are two aims that you want to attain: to eat something healthy and to not eat something unhealthy. As it happens, the latter aim is more important to you than the former: you’d rather avoid eating something unhealthy than eat something healthy. Which option do you have the most practical instrumental reason to take? Presumably, the third: eating nothing at all. But this means that the practical instrumental reasons that are operative in Food Boxes in fact exhibit prohibitive balancing: you are neither permitted to eat the contents of Box 1, nor are you permitted to eat the contents of Box 2. Thus, the apparent contrast between epistemic reasons and practical instrumental reasons disappears once we test their behavior against cases that are relevantly similar. 4.4

What Does It Take to Have “Suffcient” Reason?

Our fnal focus of discussion concerns the question of what it takes to have suffcient reason for some action or attitude. As usual, let’s begin with the practical case. When do you have suffcient instrumental reason to perform some action? The transmission principle itself is silent on this issue. It merely states that the strength of a transmitted instrumental reason depends on the strength of the reason you have to pursue the relevant aim, together with the evidential probability that the action will achieve that aim. But note that it seems possible to have suffcient instrumental reason to perform an action even when the likelihood that the action will achieve the aim is quite low. Most obviously, this happens in cases where you have very strong reason to pursue the aim in question. Suppose, for example, that you are on a runaway train full of passengers, destined to crash within minutes. The feeble emergency brakes are extremely unlikely to stop the train, but they might just, and

Instrumental Reasons for Belief

121

there is nothing else to be done. Here it seems obvious that you have suffcient instrumental reason to pull the brakes, despite the low probability that doing so will stop the train. Only a little less obviously, there are cases where you do not have particularly strong reason to pursue the aim in question, and yet seem to have suffcient reason to take the means. Suppose, for example, that you are imprisoned with nothing to do but passing time. Looking out the window rarely gives you any pleasure, but there might just be something new to look at this time. Given that you have nothing better to do, it seems that even such an unlikely prospect of mild diversion might give you suffcient instrumental reason to get up and look out the window. Are similar situations possible in the epistemic domain? That is, could you ever have suffcient epistemic reason to believe a proposition that is very unlikely to be true in light of your evidence? It seems not. In fact, when it comes to reasons for belief, it seems that having suffcient reason to believe a proposition requires that the evidential likelihood meets a threshold of at the very least 50%, and according to most epistemologists much higher. How might epistemic instrumentalists explain away this apparent contrast? The frst thing to consider is what explains, on the instrumentalist account, why you cannot have suffcient epistemic reason to believe p when the evidential likelihood of p is 50% or below. The story is familiar by now. There are two epistemic aims at play at once: forming a true belief as to whether p, and not forming a false belief as to whether p. If the evidential likelihood of p is 50% or below, believing p will give you at most a 50% likelihood of achieving the aim of forming a true belief as to whether p, and likewise at most a 50% likelihood of achieving the aim of not forming a false belief as to whether p. By contrast, suspending judgment about p will give you a 0% likelihood of achieving the aim of forming a true belief as to whether p, and a 100% likelihood of achieving the aim of not forming a false belief as to whether p. Thus, given that you have reason to be epistemically risk-averse – i.e., given that you should be more concerned with avoiding false beliefs than with securing true ones – you have stronger epistemic reason to suspend judgment about p than to believe p. This line of reasoning makes vivid an important disanalogy between the practical and epistemic cases considered previously. In the train case, the reason why it is natural to think that you have suffcient reason to pull the emergency brakes is that the aim of not crashing is presumed to be much more important than any other aim you might have in the relevant situation. In the prison case, the reason why it is natural to think that you have suffcient reason to get up and look out the window is that there are no competing options available to you. Neither of these features is shared by the epistemic case. The frst feature does not show up in the epistemic case, since the aim of forming true beliefs is not the only

122

Asbjørn Steglich-Petersen and Mattias Skipper

important aim in the context: the aim of not forming false beliefs is also present and at least as important. The second feature does not show up in the epistemic case, since the option of believing p does compete with the alternative option: namely that of suspending judgment about p. No wonder, then, that the cases seem to reveal a contrast between epistemic reasons and practical instrumental reasons. In order to determine whether the apparent contrast is genuine, we must once again compare the epistemic case to a relevantly similar practical case. Let’s simply recycle Food Boxes from earlier. As we recall, this is a case where you do not have suffcient practical reason to eat the contents of Box 1, nor suffcient practical reason to eat the contents of Box 2. Why? Precisely because the aim of eating something healthy competes against the aim of not eating something unhealthy, and because there is an alternative option available, namely to eat nothing at all. Thus, the apparent contrast between epistemic reasons and practical instrumental reasons again disappears once we compare these reasons against relevantly similar cases.

5.

Conclusion

The goal of this chapter was to defend epistemic instrumentalism against the charge that epistemic reasons lack a number of properties of practical instrumental reasons. Our strategy was not to deny that such differences, if genuine, would count against understanding epistemic reasons as a species of instrumental reasons. Rather, we wanted to argue that the seeming differences can be explained away by paying close attention to the mechanics of the transmission principle. In the course of doing so, we also offered what we take to be a simple and intuitive error-theory on behalf of the epistemic instrumentalist: the reason why epistemic reasons seem to lack various properties of practical instrumental reasons is that we tend to think and talk about epistemic reasons in an “elliptical” manner. That is, we take the truth-aim for granted and focus solely on whether the evidence at hand makes it likely that believing the proposition in question is likely to achieve the aim of forming a true belief about that proposition. If we are right, epistemic instrumentalists might obviously still be wrong for independent reasons. But they will at least have one thing less to worry about.

Notes 1. For defences of epistemic instrumentalism, see Quine (1969), Foley (1987), Laudan (1990), Kitcher (1992), Kornblith (1993), Nozick (1993, ch. 3), Papineau (1999), Cowie (2014), and Steglich-Petersen (2011, 2018). 2. How? One strategy is to deny that we have the epistemic reasons that we are alleged to have in cases where we lack the relevant epistemic aims (Papineau 2013). Another is to adopt a “normatively loaded” version of epistemic

Instrumental Reasons for Belief

3. 4. 5. 6. 7. 8.

9. 10.

11. 12. 13.

14. 15.

123

instrumentalism, which refers to the aims you should have rather than the ones you actually have (Steglich-Petersen 2018). For views of this sort, see, e.g., Quine (1969), Kitcher (1992), Kornblith (1993), and Papineau (1999, 2013). For an infuential development of this objection, see Kelly (2003). For a critical exchange, see Kelly (2007) and Leite (2007). For a helpful review and discussion of various responses, see Côté-Bouchard (2015). For versions of this approach, see e.g. Grimm (2009), Cowie (2014), and Steglich-Petersen (2011, 2018). Steglich-Petersen (2018) and Steglich-Petersen/Skipper (forthcoming). E.g., Daniel Whiting (2013) makes this point. See also Bykvist/Hattiangadi (2007). For a defense of this norm, see Whiting (2013). Even this norm, however, may be too strong. If one can have practical reasons for belief (as we think one can), and such reasons can outweigh one’s epistemic reasons (as we think they can), there will be cases where believing an evidentially unsupported proposition is all-things-considered permissible. For discussion of this possibility, see Steglich-Petersen/Skipper (forthcoming). Still, in some weaker sense, there will be something wrong or defective about these unsupported beliefs, even if they are all-things-considered permissible. See Steglich-Petersen (2018: 274–275) for discussion of this weaker sense of defectiveness. For our present purposes, however, we can ignore this subtlety. That is not to say that the status and content of these principles are uncontroversial. For an overview of the current discussion, see Kolodny/Brunero (2018). See Steglich-Petersen (2018, §4.1). In short, if we suppose that every belief is associated with an aim to believe the truth (either in a personal sense, or in a subpersonal functional sense), and that believing in accordance with evidence is thought to be necessary to achieve this aim, it will be instrumentally incoherent not to intend to believe in accordance with the evidence. For defenses of the required teleological account of belief, see Steglich-Petersen (2006, 2009). The terminology has been adjusted for our purposes. James (1897). For more recent discussions of these “Jamesian goals” and their interaction, see Kelly (2014) and Horowitz (2018). While this elliptical understanding of our common talk about epistemic reasons helps solve a number of problems for the instrumentalist account, it leaves other problems open that we shall not discuss in this paper. One such problem concerns what kinds of considerations we are psychologically able to base our beliefs on. For example, while some have argued that we are only able to base our beliefs on evidential considerations (Shah 2006), others have argued that we are perfectly capable of basing our beliefs on practical considerations (Rinard 2019). For further discussion, see Steglich-Petersen (2008) and Reisner (2009). Kelly (2003) discusses a similar case. Cases like this raise a number of subtle issues that we will have to save for another occasion. For example, wouldn’t there be something awry with you if your friend’s rehearsal of trivialities didn’t make you believe them? And if so, wouldn’t this show that the evidence with which your friend thereby provides you does give you a reason to form the relevant beliefs? We think not. As human beings, we are naturally disposed to automatically form beliefs in response to our evidence. On the whole, this disposition works well in our favor, since we often have reason to want true beliefs. When the disposition fails to deliver this, it is thus regrettable insofar as it indicates that a disposition that is, on the whole, desirable is not entirely infallible. But this does

124

Asbjørn Steglich-Petersen and Mattias Skipper

little to show that whenever we form beliefs in this automatic manner, those beliefs are formed in response to reasons. 16. This assumption is familiar from versions of expected utility theory that use (unbounded) real-valued functions to represent the utility of different outcomes at different states. 17. See, e.g., Dorst (2019), Easwaran (2016), Steinberger (2019), and Pettigrew (2016).

References Berker, Selim (2018): “A Combinatorial Argument against Practical Reasons for Belief”, Analytic Philosophy 59, 427–470. Bykvist, Krister/Hattiangadi, Anandi (2007): “Does Thought Imply Ought?”, Analysis 67, 277–285. Côté-Bouchard, Charles (2015): “Epistemic Instrumentalism and the Too Few Reasons Objection”, International Journal of Philosophical Studies 23, 337–355. Cowie, Christopher (2014): “In Defence of Instrumentalism about Epistemic Normativity”, Synthese 191, 4003–4017. Dorst, Kevin (2019): “Lockeans Maximize Expected Accuracy”, Mind 128, 175–211. Easwaran, Kenny (2016): “Dr. Truthlove or: How I Learned to Stop Worrying and Love Bayesian Probabilities”, Noûs 50, 816–853. Foley, Richard (1987): The Theory of Epistemic Rationality, Cambridge, MA: Harvard University Press. Grimm, Stephen (2009): “Epistemic Normativity”, in: Epistemic Value, ed. by Haddock, et al., New York: Oxford University Press, 243–264. Horowitz, Sophie (2018): “Epistemic Value and the Jamesian Goal”, in: Epistemic Consequentialism, ed. by Ahlstrom-Vij/Dunn, New York: Oxford University Press, 269–289. James, William (1897): “The Will to Believe”, in: The Will to Believe and Other Essays in Popular Philosophy, New York: Longmans, Green, and Co, 1–31. Kelly, Thomas (2003): “Epistemic Rationality as Instrumental Rationality: A Critique”, Philosophy and Phenomenological Research 66, 612–640. Kelly, Thomas (2007): “Evidence and Normativity: Reply to Leite”, Philosophy and Phenomenological Research 75, 465–474. Kelly, Thomas (2014): “Evidence Can Be Permissive”, in: Contemporary Debates in Epistemology, 2nd ed., ed. by Steup, et al., Oxford: Blackwell, 298–311. Kitcher, Philip (1992): “The Naturalist’s Return”, The Philosophical Review 101, 53–114. Kolodny, Niko (2018): “Instrumental Reasons”, in: The Oxford Handbook of Reasons and Normativity, ed. by Star, New York: Oxford University Press, 731–763. Kolodny, Niko/Brunero, John (2018): “Instrumental Rationality”, in: The Stanford Encyclopedia of Philosophy, ed. by Zalta, URL = . Kornblith, Hilary (1993): “Epistemic Normativity”, Synthese 94, 357–376. Laudan, Larry (1990): “Normative Naturalism”, Philosophy of Science 57, 44–59.

Instrumental Reasons for Belief

125

Leite, Adam (2007): “Epistemic Instrumentalism and Reasons for Belief: A Reply to Tom Kelly’s ‘Epistemic Rationality as Instrumental Rationality: A Critique’”, Philosophy and Phenomenological Research 75, 456–464. Lynch, Michael (2004): True to Life: Why Truth Matters, Boston: MIT Press. Nozick, Robert (1993): The Nature of Rationality, Princeton: Princeton University Press. Papineau, David (1999): “Normativity and Judgment”, Proceedings of the Aristotelian Society 22, 17–41. Papineau, David (2013): “There Are No Norms of Belief”, in: The Aim of Belief, ed. by Chan, New York: Oxford University Press, 64–79. Pettigrew, Richard (2016): “Jamesian Epistemology Formalized: An Explication of ‘the Will to Believe’”, Episteme 13, 253–268. Quine, Willard Van Orman (1969): “Epistemology Naturalized”, reprinted in his Ontological Relativity and Other Essays, New York: Columbia University Press, 69–90. Reisner, Andrew (2009): “The Possibility of Pragmatic Reasons for Belief and the Wrong Kind of Reasons Problem”, Philosophical Studies 145, 257–272. Reisner, Andrew (ms): “Comparing Pragmatic and Alethic Reasons”, Chapter II of Reisner: The Pragmatic Foundations of Theoretical Reason, book manuscript. Rinard, Susanna (2019): “Believing for Practical Reasons”, Noûs 53, 763–784. Shah, Nishi (2006): “A New Argument for Evidentialism”, The Philosophical Quarterly 56, 481–498. Steglich-Petersen, Asbjørn (2006): “No Norm Needed: On the Aim of Belief”, The Philosophical Quarterly 56, 499–516. Steglich-Petersen, Asbjørn (2008): “Does Doxastic Transparency Support Evidentialism?”, Dialectica 62, 541–547. Steglich-Petersen, Asbjørn (2009): “Weighing the Aim of Belief”, Philosophical Studies 145, 395–405. Steglich-Petersen, Asbjørn (2011): “How to Be a Teleologist about Epistemic Reasons”, in: Reasons for Belief, ed. by Steglich-Petersen/Reisner, Cambridge: Cambridge University Press, 13–33. Steglich-Petersen, Asbjørn (2018): “Epistemic Instrumentalism, Permission, and Reasons for Belief”, in: Normativity: Epistemic and Practical, ed. by McHugh, et al., New York: Oxford University Press, 260–280. Steglich-Petersen, Asbjørn/Skipper, Mattias (forthcoming): “An Instrumentalist Account of How to Weigh Epistemic and Practical Reasons for Belief”, Mind. Steglich-Petersen, Asbjørn/Skipper, Mattias (ms): “Belief Gambles in Epistemic Decision Theory”, unpublished manuscript. Steinberger, Florian (2019): “Accuracy and Epistemic Conservatism”, Analysis 79, 658–669. Whiting, Daniel (2013): “Truth: The Aim and Norm of Belief”, Teorema 32, 121–136.

7

Suspension of Judgment, Rationality’s Competition, and the Reach of the Epistemic Errol Lord

1.

Introduction

One of the most entrenched commitments in all of philosophy is the idea that, at least for some reactions, rationality should be understood in competitive terms. To understand rationality in this way is to think that there are multiple reactions competing against each other to win rationality’s prize, which is having the status of being rational. This conception is extremely plausible when it comes to the rationality of actions. There is a simple reason for this: At least the vast majority of the time, most of the options in play are incompossible or in some other way in metaphysical tension with one another. In cases where this is so, it looks like one has to evaluate the relative rational merits of each option. This pits the options in a sort of rational competition. The ones that pass the rational bar are the winners. Traditional epistemology has eagerly embraced this competitive conception of rationality. But the participants in the competition have traditionally been seen as a small group. Belief and disbelief are obvious competitors.1 Nearly everyone, at least after some badgering, also admits that there is a third participant, which is suspending judgment about whether p. When we suspend judgment, we adopt a sort of principled neutrality with respect to p and ¬p. On a fatfooted understanding of epistemic rationality’s (coarse grained) competition, for any proposition p, the merits of believing p, disbelieving p, and suspending about p compete against one another. When this competition is won by an option, that option is rational. Many think something even stronger, which is that there will always only be one winner; this is to say that for every proposition p and every agent in some particular rational circumstance, only one of the options will win and thus be rational. While I think this is plausible, I won’t assume it in full generality here. I will, however, assume that most of the time only one option will win.2 This paper has three main tasks. The frst two tasks aim to complicate this traditional picture. The frst complicating factor is that there

Suspension of Judgment

127

are many different reactions one can have about p that intuitively count as a sort of neutrality about p. Given this, it is not clear why we should think that there are only three participants in epistemic rationality’s competition. This frst complicating factor puts the onus on everyone to sort through these different ways of being neutral and take a stand on which ones compete with belief and disbelief. The second complicating factor – discussion of which comprises the second part of the paper – is that there are really four participants in epistemic rationality’s competition. This is fewer than all the ways that one can be neutral but more than the traditional three. I will argue that we should admit two sorts of suspension but shouldn’t allow several other ways of being neutral. The arguments for this conclusion will turn on a conception of what it takes to be a participant in the relevant competition. The fnal task is to argue that investigating the participants in rationality’s competition is a good way of investigating the reach of epistemic rationality. This is important for several reasons. First, it gives us some leverage on just what the epistemic domain is. Second, and more specifcally, it gives us leverage in the debate about whether only the evidence bears on epistemic rationality. A common thought shared by evidentialists – those that think only the evidence matters – is that evidentialism is strongly motivated because it doesn’t make sense to have non-evidential factors bear on epistemic rationality; the common rationale given for this claim is that the epistemic is only concerned with the truth. In the third part of the paper, I will argue that this motivation is illusory. This is because the two sorts of suspension that are participants in epistemic rationality’s competition can be justifed by non-evidential factors. But they are in the epistemic’s reach. So it is just not true that we can only appeal to the evidence when discussing the epistemic.

2. 2.1

On Being Neutral Varieties of Neutrality

Let’s start by thinking about some baseline features of suspension of judgment. The frst feature is that suspension is a reaction. It follows that not taking a stand is insuffcient for suspension. At least before thinking of this example, I had no view about the claim that Trump is President and Mannheim has an odd number of hairs within its limits. That does not mean that I suspended judgment about that claim. I just had no view.3 A second baseline feature is that suspension is rationally evaluable. It is the sort of thing that can be rational and irrational. This is obviously needed if it is going to be a participant in epistemic rationality’s competition. A third baseline feature is that suspension is the sort of reaction that you can have for reasons. When you fnd out that the probability of rain

128

Errol Lord

tomorrow is about the same as the probability of no rain tomorrow, you can suspend judgment about whether it will rain tomorrow for reasons; most naturally, your reason will be the fact that the evidence is balanced. So far these are the features that all participants in the competition will have. Belief and disbelief share them, and I think it’s very plausible that they are necessary for being in the competition. Thus, these three features don’t differentiate suspension from belief and disbelief. What does the differentiation? To help get a grip on this question, it is helpful to think about what I will call one’s intellectual outlook. One’s intellectual outlook is a reckoning of one’s doxastic stance on the world. Epistemic rationality’s competition takes place within one’s intellectual outlook. Different competitors play different roles within the outlook.4 To start, consider the fact that belief and disbelief are both ways to determine whether p within one’s outlook. When you believe p, p is determined within your outlook; functionally, this means that, other things equal, you will proceed as if you inhabit a p-world. Similarly, when you disbelieve p, ¬p is determined within your outlook. When you disbelieve p, you move forward as if you are in a ¬p-world. One difference between suspension and belief and disbelief is that suspension does not involve determining p in either of these ways. To use a nice phrase from Sturgeon (2010), suspension is a committed neutrality. So, in trying to fgure out which state or states suspension is, we need to investigate the ways in which we can be committed to neutrality. We need to investigate the ways in which p can be in your outlook in a neutral way. As it happens, it looks like there are several states that can be reasonably described as commitments to neutrality about whether p. An obvious candidate is middling credence about p.5 Take credence .5 in p. This meets the three minimal requirements for being a competitor. It is also seems to be a way of having a commitment to a sort of neutrality with respect to p. Plausibly, there are many credences near the middle that have the same features. Hence, having a middling credence in p is a sort of neutral stance about p. A second batch of states is a range of related beliefs you might have about your position with respect to p. There are many different beliefs you could have that could naturally be construed as commitments to neutrality. For example, you might believe that the evidence is balanced. You might believe that there is no evidence for p or for ¬p. You might believe that both p and ¬p are unknowable. You might believe that you are neither in a position to know p nor in a position to know ¬p. You might believe that it is neither rational to believe p nor rational to believe ¬p. You might believe that neither p nor ¬p are provable. The contents of each of these beliefs are about ways in which the rational merits of belief and disbelief are relatively neutral. Given this, it is plausible to think that believing those contents is a way of taking a neutral stance.

Suspension of Judgment

129

A third possibility is more practical. Suppose you start investigating p and fnd out that the case for p and the case for ¬p are neutral in one of the previously-mentioned ways. You might come to believe that you are neither in a position to know p nor in a position to know ¬p. But, in an obvious way, this belief is not really a stand about whether p. This belief is about your epistemic situation. So it wouldn’t be crazy to think that forming that belief isn’t suffcient for suspension. One thing the belief might prompt, though, is an intention not to form a belief or a disbelief.6 This intention seems straightforwardly to be a commitment to be neutral about p. These three options are perhaps the most obvious. There are two unobvious neutral states that have been important to recent theorizing about suspension. The frst is what Jane Friedman calls interrogative attitudes (see Friedman 2013a, 2013b, 2017). These are attitudes that take as their object the question whether p. So, for example, when you are curious about whether p, you have an attitude directed at whether p. One of the main functional roles of this attitude is to dispose one to be on the lookout for information that would settle whether p. Often we are in a state like this when we suspend. So, for example, when I’m deliberating about whether to teach at 10 am or at noon and I fnd out that I lack some information about my train’s future schedule, it might be rationally required for me to suspend about which time to teach. But I will keep the issue in my outlook. A natural way of spelling this out is in terms of adopting a questioning attitude about whether to teach at 10 or 12 – this state disposes me to do things that will put me in a position to settle the relevant question. Interrogative attitudes are states like this. Interrogative attitudes are neutral positions about whether p. But they also involve a kind of openness towards answering whether p. In fact, at least paradigmatically, they dispose one to answer whether p. In other words, not only are they attitudes one takes towards the question of whether p; they are also attitudes that orient one towards answering whether p. In contrast, it looks like there are attitudes one can take towards the question whether p that are not attitudes that orient you towards answering whether p. So, consider the question of whether there was an even number of food particles in Karl Marx’s beard at the time of his death. It is very likely that there is no evidence whatsoever that speaks to this question. So neither belief nor disbelief is appropriate. Simply forgetting about it is an option that might be permitted. But there seems to be a stance you could take that is more aggressive than this. Rather than merely dropping the issue, you can bury the issue, which is to say you can adopt an attitude towards whether p that disposes you to ignore the question. This attitude makes you insensitive to information relevant to the question. I’ll call attitudes that are directed towards questions that bury those questions anti-interrogative attitudes. So while adopting an interrogative attitude keeps p in your outlook in a way that orients you towards settling

130

Errol Lord

the question, adopting an anti-interrogative attitude puts whether p in the outlook by, in effect, taking whether p to be a bad question. In some cases this seems to be the right stance to take – e.g., when our epistemic situations look to be permanently impoverished when it comes to some question (later I’ll argue there are some other situations where it makes sense to adopt these attitudes). 2.2 Which Ones Participate? All of the stances just canvassed can be naturally seen as neutral stances towards p. Orthodoxy maintains that there is one neutral stance that is a participant in epistemic rationality’s competition. Given the varieties of neutrality, someone who maintains this view needs to tell us which neutral state is the participant in the competition. Alternatively, one can give up the traditional view and allow for multiple neutral participants. The frst thing to say is that it is unclear what the rules of the game are here. What we need is a method for testing whether or not a stance is a participant in our competition. There are two simple and tempting tests to see if there is competition between our neutral states and belief and disbelief. The frst test examines whether there is a rational tension between a particular neutral state and belief and disbelief. If there is a rational tension between adopting a neutral state and adopting belief or disbelief, then we have some evidence that those states compete. Call this the Tension Test. The second test examines whether suspension is intuitively rational even when some candidate neutral state is not rational or whether suspension is intuitively irrational when some candidate neutral state is rational. If there is either mismatch, then that is evidence that the candidate state is not a participant in the competition. Call this the Rational Incongruity Test.7 As we’ll see in a moment, I think we can use the Tension Test to rule out one of the options. However, it should be said right away that I don’t think this test can rule out all of them. This is because I think it produces false positives. To put some of my cards on the table: Middling credences and beliefs about one’s epistemic position with respect to p will be in tension with believing p or disbelieving p. However, I don’t think this shows that those states are true participants in the competition. This is why we need the Rational Incongruity Test. Before explaining why, let me frst argue against the intention view. 2.2.1

Against the Intention View

Let’s start with the intention view. That view maintains that to suspend is to form an intention to not believe or disbelieve p. I think there are clear cases where this view fails the Tension Test. This is because most of our reasons for intention are straightforwardly practical and thus can be

Suspension of Judgment

131

generated no matter one’s evidential situation. For example, imagine you are a scientist working for big tobacco in the early 1990s. You have been exposed to the overwhelming evidence that smoking cigarettes causes a variety of serious health problems. However, you are also now in a tough legal position because the federal government is poised to punish your company and perhaps even you for decades of deception. Your lawyer informs you that once all the evidence comes out, it will not be credible for you to say that you didn’t believe smoking kills. It would be better, he says, if you didn’t believe one way or the other, especially if this neutrality can be supported by something that looks like sophisticated scientifc reasoning. In this case, I think it is very plausible that the scientist has decisive reasons to intend not to believe either that smoking kills or that it’s not the case that it kills. He has decisive reason to intend to have no stand. This is because this would be the best thing for him practically. If he could get himself in that state, then he’d be able to truthfully testify that he is neutral (of course the explanation that he would give for that is not the true explanation). Nevertheless, this does not change the normative impact of the evidence. The scientist’s legal trouble does not compete with the evidence. The evidence still demands belief. So it doesn’t look like the reasons for intention compete with the reasons for belief in the needed way. Because of this, I think it’s plausible that he has both decisive reasons to believe that smoking kills and decisive reason to intend to neither believe nor disbelieve. Now, fortunately for the lovers of justice, it is also plausible that his beliefs are not under intentional control in the right way for his intention to have neither state to be causally effcacious. So while he has decisive reason to form that intention, he is unlikely to succeed in carrying out the intention. So perjury about his epistemic life is likely his only real option. A similar objection can be leveled at a recent view defended by Masny (forthcoming) for a view that is a hybrid of an intention view and belief view. According to Masny, suspension has three conditions that are all necessary and jointly suffcient: (i) S believes that she neither believes nor disbelieves that p, (ii) S neither believes nor disbelieves that p, (iii) S intends to judge that p or ¬p. The problem with this view is that it distorts the rational profle of suspension. Since this is the section about intention views, we can start there. Masny’s intention condition doesn’t pass either of our tests. First, it fails the Tension Test. It is easy to think of cases where one has very good reasons to intend to come to some judgment about whether p that will not be reasons to suspend. For example, there are very good reasons for me to intend to judge that I was just offered a lifetime yearly income of $1bn (and, thus, I assume, good reasons to intend to judge that I was just offered a lifetime yearly income of $1bn or it is not the case that I was just offered a lifetime yearly income of $1bn). But these don’t seem to be reasons to suspend judgment about whether I was just offered a yearly

132 Errol Lord income of $1bn. What this shows is that Masny’s intention condition does not pass the Tension Test – there is no tension between that intention and belief and disbelief. Masny’s intention condition also fails the Rational Incongruity Test. There are cases where it is irrational to have Masny’s intention even though suspension is rational. It is not rational for me to intend to have an opinion about whether there was an even number of food particles in Marx’s beard when he died, but it might be rational for me to suspend about this. After all, no evidence will ever settle this matter for me. This is reason to suspend; in fact, this sort of reason to suspend takes center stage in many discussions of the epistemology of suspension.8 Now, this might be unfair to Masny’s considered view. After all, there are three conditions. Two of them involve mental states. So perhaps reasons to suspend on Masny’s view will be some combination of reasons to have the belief that I don’t believe p or ¬p and reasons to intend to judge that p or judge that ¬p. In the yearly income case, I have excellent reason to believe that I disbelieve that I was just offered a yearly income of $1bn; so I don’t have strong reasons to believe I don’t believe or disbelieve and intend to believe or disbelieve. But there are cases where I have strong reasons to have both the belief and intention yet lack strong reasons to suspend.9 Suppose that Phillip thinks that he believes that members of a certain marginalized group are morally equal to Anglo-Saxons. But then psychologically observant friends start pointing out patterns of thought and action of his that provide evidence that he doesn’t believe this. He responds by pointing to certain patterns of thought and action that provide evidence that he does believe they are equals. We can imagine that these two bodies of evidence are roughly on a par. Thus, he has good reason to suspend about whether he believes the groups are equal – i.e., good reasons to believe that it is not the case that (he believes they are equal or he disbelieves they are equal). Further, we can imagine that Phillip is an ordinary person in contemporary society who thus has good reason to intend to have an opinion on this matter. Thus, Phillip has strong reasons to believe he neither believes nor disbelieves and has strong reasons to intend to believe or disbelieve. He doesn’t thereby have strong reasons to suspend. We can suppose that he has particularly good evidence for thinking the groups are equal, evidence that is even good enough to put him in a position to know that the groups are equal. This is compatible with him having good reasons to have the relevant belief and intention. To be clear, I am not saying that Phillip is rational. It is plausible that he ought to believe they are equal. So there is a rational tension between the reasons for Masny’s belief and intention and the reasons that bear on whether the groups are equal. So it looks like Masny’s view passes the Tension Test. But it doesn’t pass the Rational Incongruity Test. This is

Suspension of Judgment

133

because this is a case where Masny’s neutral state is rational even though suspension is not. The frst upshot of this subsection is that the prospects for pure intention views are dim because the relevant intentions are not in tension with belief and disbelief in the right way; as we saw, things get better by moving from a pure intention view to a hybrid intention-belief view. But Masny’s particular hybrid view also fails because it fails the Rational Incongruity Test. Its intention-belief bundle can be rational even when suspension is not. 2.2.2. Against the Credal View Now let’s examine credences. The credal view of suspension maintains that to suspend about p is to adopt some credence about p. Intuitively, it won’t just be some single amount of confdence. Rather, there will be a range of credences one can adopt that are suffcient for suspension. We’ll just say that suspension is middling credence. The Tension Test allows us to check this hypothesis. What we are looking for is a rational tension between middling credences and beliefs and disbeliefs. Middling credences seem to pass the test. It is very plausible that having a middling credence about p is in tension with belief or disbelief that p. This is because in order to rationally believe p, your credence in p needs to be high enough, and it will never be high enough if one has a middling credence. That’s the good news for the credal view. The bad news is that it fails the Rational Incongruity Test. To pass this test, it would need to be that non-middling credence (which is non-suspension on this view) is in tension with suspension. This does not appear to be the case, as Friedman (2013c) has demonstrated.10 To see the problem, consider Lack of Evidence and Conjunction: Lack of Evidence: If A lacks evidence for p and for ¬p, then it is permissible for A to withhold about whether p. Conjunction: If A is permitted to withhold about p, q, r . . . n, then A is permitted to withhold about (p ∧ q ∧ . . . ∧ n).11 Both of these principles are very plausible, but the credal view cannot vindicate them given the standard norms on credence. To see why, let’s suppose any credence between .33 and .66 is middling. Now take three probabilistically independent claims p, q, and r. Suppose you lack evidence with respect to all three. Thus, from Lack of Evidence, it follows that you are permitted to suspend about all three. It should follow from Conjunction that you are permitted to suspend about their conjunction. But it doesn’t follow from the credal view. This is because even if you had the absurdly high credence of .66 in all three of them, the standard norms

134

Errol Lord

on credence predict that your credence in their conjunction should be below .33 since your credence in the conjunction should be the product of your credences in the conjuncts. Thus, the credal view maintains that you are not permitted to suspend about their conjunction. That’s false. What’s important for our test is that having non-middling credence is compatible with rational suspension. If suspension is middling credence, then when you are required to have a non-middling credence, you are required not to suspend. But in cases with the relevant form, you are permitted to both suspend about (p ∧ q ∧ r) and have a non-middling credence about (p ∧ q ∧ r). So having middling credence is not suffcient for suspension. Further, the reasons that bear on suspension are not always reasons that bear on whether or not to have middling credence. This is all evidence that middling credence is not a participant in the relevant competition. 2.2.3

Against the Belief View

Now let’s turn to the most complicated case, which is the higher-order belief view. This view maintains that to suspend is to have some belief about your epistemic position with respect to p. To start, I will assume that the relevant belief is that you are neither in a position to know p nor in a position to know ¬p.12 This view, like the credal view, passes the Tension Test. There is a tension between the relevant higher-order beliefs and believing p and believing ¬p. That is, it is plausible that you are generally irrational to both believe or disbelieve p and also believe that you are neither in a position to know p nor in a position to know ¬p. Furthermore, it is plausible that reasons to believe that you are neither in a position to know p nor in a position to know ¬p compete with the reasons to believe p and the reasons to believe ¬p. So the higher-order belief view seems to pass our frst test. As it was with the credal view, it is far from clear that having the relevant higher-order belief is necessary for suspension, and thus it looks like the higher-order belief view fails the Rational Incongruity Test. One reason why is that the higher-order belief requires conceptual sophistication out of reach of many agents who appear to be able to suspend.13 Position to know is, at least seemingly, not an ordinary concept regularly possessed and deployed by ordinary people. But ordinary people regularly suspend judgment. So it seems unlikely that having that higherorder belief is necessary for suspension. The point is even sharper when it comes to immature humans and animals. It is even less plausible that they possess and regularly deploy those concepts. Yet they suspend. Raleigh (forthcoming) anticipates this overintellectualization objection. In response, he maintains that the content of the relevant belief is less sophisticated. Suspending, on Raleigh’s view, involves “a belief or opinion that one cannot yet tell whether or not p, based on one’s

Suspension of Judgment

135

evidence.” This is certainly better, but notice that this still requires one to have the capacity to have beliefs about evidence. Further, one needs to have the capacity to have beliefs about one’s epistemic situation with respect to p in order to suspend. This is a feature that it shares with all higher-order belief views. This is because all higher-order belief views require beliefs about one’s own situation in order to suspend. Suspension, on these views, thus requires a lot more conceptual sophistication than belief and disbelief. To believe or disbelieve p, you just need to have the capacity to entertain p. To suspend, you need to entertain not only p; you also need to entertain claims about your intellectual situation vis-à-vis p. On Raleigh’s view, you need to be able to entertain thoughts about what your evidence indicates. Raleigh is clear-eyed about the added sophistication his view requires. He thinks that it is an important truth that suspension requires more intellectual sophistication than belief. He writes, it is not at all clear that we are prepared to ascribe the attitude of suspending judgement to simple believers such as young children or animals – I might describe my toddler or my dog as hesitant or uncertain or just ignorant whether p, but it would sound pretty strange to describe them as agnostic whether p. Agnosticism is plausibly a more sophisticated, intellectually demanding attitude than belief. Raleigh forthcoming If Raleigh is talking about vanilla suspension – the competitor in our competition – then I think this is seriously mistaken. Young children and animals do suspend. Raleigh and I appear to agree that HOB views threaten this. His modus ponens is my modus tollens. To see why this is seriously mistaken, it helps to have a feshed out alternative on the table. So I will postpone a full explanation of why I think it is wrong to deny young children and animals the capacity to suspend to the next subsection. To warm up to the idea, though, note that it is not surprising for there to be problems for the higher-order belief view, given the way I’ve set things up. After all, what we are looking for are types of mental states that compete against each other. But higher-order beliefs are just beliefs. There is no clash in kind between beliefs and beliefs. So any clash here will simply be a clash in contents. There is a clash in contents here – hence passing the Tension Test. But a simple clash in contents shouldn’t make for a new participant in the competition. After all, any content q that is inconsistent with p will clash with a belief that p. But that doesn’t mean that beliefs that q are participants in the relevant competition. I thus think we should rule out intentions, middling credence, and higher-order beliefs from the competition. While there are rational connections between credences about p, higher-order beliefs about

136

Errol Lord

one’s rational situation vis-á-vis p, and belief and disbelief about p, we shouldn’t think that those states all compete against each other. 2.2.4

For Interrogative and Anti-Interrogative Attitudes

This leaves interrogative attitudes and anti-interrogative attitudes. I do think that these two states are participants in our competition. In order to see why, let’s return to the idea of an intellectual outlook. We already know two ways for p to fgure into one’s outlook. One can believe that p or disbelieve that p. When p fgures into one’s outlook in either of these ways, the status of p is both settled and determined. When you believe that p, p is determined in your outlook in virtue of the fact that you take p to be the case. When you disbelieve p, p is determined in your outlook in virtue of the fact that you take ¬p to be the case. To help contrast this with the sort of determination involved in anti-interrogative attitudes (to be spelled out more later), let’s call this Worldly Determination – so-called because beliefs determine the status of p by taking a stand about the world. The fact that belief and disbelief worldly determine the status of p within an outlook is explained by the functional roles of those states. Belief that p disposes one to use p fat-out in reasoning. So, other things equal, when you believe p, you will move forward in your interactions with the world as if you inhabit a p-world. When we think about things this way, it is easy to see the evidence for thinking that interrogative and anti-interrogative attitudes are participants in the competition. For it looks like their functional roles are in obvious tension with the functional role of belief. Both interrogative and anti-interrogative attitudes are essentially worldly undetermined states. When you adopt either sort of attitude about whether p, you enter a state that essentially lacks a disposition to use p fat-out in reasoning. Further, interrogative attitudes also essentially dispose you to settle the question of whether p. So they dispose you to do things that will induce belief, but essentially lack the function of belief. Given this, it is easy to see why reasons to form interrogative attitudes compete against the reasons to believe. We can say similar things about anti-interrogative attitudes. These attitudes also essentially lack the dispositions involved in worldly determination. Further, these attitudes dispose you to ignore any p-relevant information, at least as it pertains to p. So, unlike the interrogative attitudes, these attitudes seek to prevent you from determining whether p via belief. They seek to prevent you from hearing the siren call of reasons to worldly determine whether p. Anti-interrogative attitudes thus provide a different way of determining p’s status in your outlook. We can call it Agnostic Determination. Given the fact that anti-interrogative attitudes involve agnostic determination, one sort of suffcient reason to adopt an anti-interrogative attitude is information that strongly suggests both that the current information doesn’t suffciently confrm p or ¬p and that future information won’t suffciently confrm p or ¬p.14

Suspension of Judgment

137

Now we are in a position to see why it is plausible to think that these four attitudes are participants in our competition. It is because they are competing ways to settle the status of p in your outlook. They are competing because their respective functional roles are in direct tension with each other. A belief that p is in tension with a disbelief that p because its functional role involves taking the world to be a p world whereas the functional role of disbelief involves taking the world to be a ¬p world; beliefs and disbeliefs are in tension with both interrogative attitudes and anti-interrogative attitudes because beliefs and disbeliefs are essentially worldly determining whereas interrogative and anti-interrogative attitudes are not; and interrogative attitudes are in tension with antiinterrogative attitudes because anti-interrogative attitudes are essentially agnostic determining whereas interrogative attitudes are essentially not.15 Before moving on, it is worth considering an objection to this core thought – the thought that the functional roles of the four states are in direct confict. The objection contends that the functional roles of interrogative and antiinterrogative attitudes are not in fundamental tension with the functional roles of beliefs because one can rationally believe p even though one is disposed to gather more p-related information or is disposed to ignore p-related information. If this is right, then it is not clear that my view passes the Rational Incongruity Test. It looks like we can get cases where one rationally believes and rationally has an interrogative or anti-interrogative attitude. I agree that there are cases where one rationally believes p and is rationally disposed to gather more information (think of a surgeon who rationally believes she needs to take out your liver). I also agree that there are cases where one rationally believes p and is rationally disposed to ignore p-related information (think of the pathological overchecker who has already made sure the oven is off 17 times). But this is not enough to show that there are cases where it is rational to believe p and rational to have interrogative or anti-interrogative attitudes. This is because there is more to interrogative and anti-interrogative attitudes than these interrogative or anti-interrogative dispositions. Remember, interrogative and anti-interrogative attitudes are essentially not worldly determining. That means they are partly individuated by a functional lack – the lack of dispositions to move around as if it is a p or ¬p world. This, I admit, is a bit tricky to get one’s mind around at this level of abstraction. It helps to contrast particular agents who are paradigms of believing and suspending. So, consider Beatrice, Ingrid, and Anjali: Beatrice the Believer Beatrice has just fnished investigating whether smoking causes cancer. It clearly does, and she believes so. She is robustly disposed to treat the actual world as a smoking-causes-cancer world, including being disposed to use that claim as a premise in reasoning and disposed to assert that claim.

138

Errol Lord Ingrid the Inquirer Ingrid has become interested in whether evidentialism about epistemic rationality is true. She has read some journal articles on it and the results are mixed. She has a questioning attitude towards the question of whether evidentialism is true. This disposes her to seek out more information, see evidentialism relevant information as salient etc. Anjali the Agnostic Anjali is prompted by an obnoxious philosophy example to consider whether Karl Marx had an odd or even number of food particles in his beard at the time of his death. This is a ridiculous question to consider, and it is clear to Anjali that she will never be in a position to determine the answer. So she adopts an agnostic attitude towards the question, which disposes her to ignore both the question and any information relevant to answering it (unfortunately the obnoxious philosopher keeps bringing it up, which makes it diffcult to be fully agnostic!).

The key point about Ingrid and Anjali is that they don’t merely have interrogative and anti-interrogative dispositions. Ingrid has interrogative dispositions in virtue of having a questioning attitude, and Anjali has anti-interrogative dispositions in virtue of having an agnostic attitude.16 This is of a piece with believing and disbelieving. One can have belief or disbelief dispositions even if one doesn’t believe or disbelieve. When one does believe or disbelieve, one has those dispositions in virtue of believing or disbelieving. What this shows, at a minimum, is that fnding agents who rationally believe and rationally have questioning or agnostic dispositions is not enough to show that my view fails the Rational Incongruity Test. Our feshed-out intellectual outlook can provide an explanation of why it is that middling credence and certain higher-order beliefs stand in tension with belief and disbelief. It is because reasons to have the higherorder beliefs or middling credence provide reasons against belief and disbelief. Reasons to have middling credence about p will be provided by the evidence, and facts about roughly equitable evidence for p and ¬p will themselves provide reasons to suspend in one of the two ways. Similarly, reasons to hold the higher-order beliefs will be evidence about the defciency of one’s epistemic position, and that evidence itself will provide reasons to suspend in one of our two ways.17 Finally, we are in a position to respond to Raleigh’s challenge laid out earlier. Recall that Raleigh’s higher-order belief view – like all higherorder belief views – maintains that suspension about p is a matter of having beliefs about our epistemic position with respect to p. I objected

Suspension of Judgment

139

earlier to the added sophistication this demands from suspenders. Raleigh embraces this result. He maintains that we hesitate from ascribing suspension to young children and animals. Further, he thinks that suspension only makes sense given certain intellectual sophistication. I hope it is clear now that my picture casts serious doubt over these claims. This much seems right: There is a good design rationale for having interrogative and anti-interrogative attitudes. This is because it is important to be able to place p in your outlook in ways that are not worldly determining. When it comes to interrogative attitudes, it is important to be able to place p in your outlook in a way that aids determination of the truth-value of p. When it comes to anti-interrogative attitudes, it is important to be able to close p from future investigation. It would be surprising if the only ways animals or small children could place p in their outlooks is by believing or disbelieving – i.e., by taking a stance that involves worldly determination. The upshot of this subsection is two-fold. First, there are good reasons for excluding intentions, credences, and higher-order beliefs. Intentions and higher-order beliefs don’t have the right subject matter. The relevant intentions are about believing or disbelieving p, whereas the higher-order beliefs are about one’s epistemic situation vis-à-vis p. Credences do have the right subject matter – viz., p – but they don’t seem to settle whether p when it comes to the relevant outlook. The second upshot is that there is a good case for including interrogative and anti-interrogative attitudes in epistemic rationality’s competition. They are both about the right subject matter and they are ways of settling whether p in the right outlook. If this is right, then orthodoxy is vindicated insofar as it was right to think that suspension is a fullblooded alternative to belief. But, at the same time, if this is right, then orthodoxy was wrong to think that there were only three participants in epistemic rationality’s competition. There are four, including two different varieties of suspension that compete with each other as well as with belief and disbelief.

3.

The Reach of the Epistemic

One of the central debates about epistemic rationality is about whether the evidence provides the only source of epistemic reasons. Evidentialists think that only the evidence can provide epistemic reasons. Pragmatists deny this and hold that sometimes non-evidential considerations provide epistemic reasons.18 Many – including past time slices of me – think that evidentialism has a sort of default status because the evidential is clearly within the reach of the epistemic, whereas it is mysterious how non-evidential factors can be within the reach of the epistemic.19 It’s not entirely clear what this means, but the intuitive idea is that the epistemic is essentially tied to the truth,

140

Errol Lord

and thus in order for some consideration to bear on epistemic rationality, it has to have some connection to the truth. The evidence clearly has this connection; it’s not clear how non-evidential considerations could ft. I don’t doubt the force of this thought when we are at the beginning of our inquiry. But the strategy for thinking about epistemic rationality deployed in this paper gives us the resources to more precisely investigate the ultimate force of this thought. By thinking about the participants of epistemic rationality’s competition, we are thereby thinking about which sort of considerations bear on epistemic rationality – they are the considerations that recommend the reactions that are the participants. The important point is this: If belief and disbelief were the only games in town, then evidentialism would be easily vindicated. But by nearly everyone’s lights they are not. The competition also includes suspension. And the view about suspension defended in this paper has the resources to cause serious trouble for evidentialism. To see this, we need to think about the sorts of reasons one might have for forming interrogative and anti-interrogative attitudes. Some of them will be non-evidential. Here is a pair of famous cases that illustrate this for interrogative attitudes:20 Low Bank It is Friday afternoon and Sarah and Yvette have a paycheck to deposit. They drive by the bank on the way home. There is a long line stretching out the door. Sarah has good reason to think that the bank is open on Saturday – she was there on Saturday two weeks ago. There is no rush to deposit the check. High Bank It is Friday afternoon and Sarah and Yvette have a paycheck to deposit. They drive by the bank on the way home. There is a long line stretching out the door. Sarah has good reason to think that the bank is open on Saturday – she was there on Saturday two weeks ago. Unlike Low Bank, though, it is crucial for them to deposit the check before Monday. Their mortgage payment is due on Sunday and they don’t have enough in the account to pay without the check being deposited. In High Bank, Sarah and Yvette have reasons to suspend that they lack in Low Bank. The fact that their mortgage depends on the deposit is itself a reason to form an interrogative attitude about whether the bank is open on Saturday. It is a reason to form an attitude about whether p that disposes them to do things to settle whether p. And this reason is quite strong. This explains the common intuition that it is not rational for them to believe the bank is open on Saturday in High Bank. It’s not rational

Suspension of Judgment

141

because the reasons to suspend defeat the reasons to believe. They ought to suspend by forming an interrogative attitude about whether p. Things are not so in Low Bank. There isn’t that reason to form an interrogative attitude. So they don’t have that reason to suspend competing with the reasons to believe. Thus, there isn’t that reason to suspend defeating the reasons to believe – i.e., the evidence. This explains the widespread intuition that, in Low Bank, it is rational for them to believe the bank is open on Saturday. Of course, the rub is that by stipulation the evidence is the same in both cases. So if there is a rational difference between the two cases, nonevidential factors can make a difference to epistemic rationality. This, I think, is the conclusion we should draw. The key to demystifying the anti-evidentialist intuitions is understanding the nature and rational profle of suspension. Similar things can be said about anti-interrogative attitudes. Take the following pair of cases as an illustration:21 Gritty PhD Student Jason is a struggling second year PhD student at a top PhD program. Jason is African American and is the frst person in his family to go to university. While he excelled at his small liberal arts college, he is fnding it diffcult to make his place in graduate school. The faculty are busy and distant, and some of them are noticeably nervous interacting with him. As he progresses through his second year, the evidence begins to mount that he will not succeed in the program. He has poured nearly all his efforts as an adult to get to this place, though, and it is entirely unclear what he should do if graduate school does not work out. Trust Fund PhD Student Shep is a struggling second PhD student at a top PhD program. He is a white male student who was a legacy undergraduate at Harvard. He became interested in philosophy and succeeded in it as an undergraduate despite a lack of true passion. He didn’t know what to do after Harvard, so he applied to PhD programs on a whim. He got into several top programs and chose the one he did mainly for its proximity to good bars. As he progresses through his second year, the evidence begins to mount that he will not succeed in the program. This doesn’t bother him too much because if he drops out he can always live off his trust fund. We can stipulate that Jason and Shep’s evidence about whether they will succeed is the same. And we can suppose that it is weighty. How should they react to it? Some have the intuition that it would, at the very least,

142

Errol Lord

be rational for Jason to display grit, which involves sticking to one’s longterm plans despite signifcant evidence that one will not succeed. We can suppose, if it helps, that Jason’s evidence is misleading. Once he pushes through his second year course work, he will fnd his research bearings, connect with a thoughtful and supportive advisor, and go on to have a successful career. The view sketched in this paper can explain what is going on here. Certain non-evidential facts – e.g., the fact that a career in philosophy is what suits Jason the most, the fact that there are no good alternative plans for Jason – provide reasons for Jason to suspend judgment about whether he will succeed. They do this by providing reasons to form an anti-interrogative attitude. This attitude will dispose him to ignore (some of) the mounting evidence, which is what he needs to do in order to push through the second year bottleneck. Now, it should be said, that the dispositions involved in anti-interrogative attitudes can be more or less robust (or more or less wide-ranging). They might dispose him to ignore anecdotal evidence, but they won’t dispose him to ignore evidence provided by a letter offcially removing him from the program. That, though, is all to the good, for it does seem like anti-interrogative attitudes should come in degrees. Perhaps there are other cases where fully robust dispositions are what is called for. Things might be different for Shep. Given his independent resources and lack of serious passion for philosophy, he lacks many of the reasons to suspend that Jason possesses. Given this, the only rational option for Shep might be to believe he will not succeed and seek out something else. If this is right, then non-evidential factors make a difference to epistemic rationality. My point here is that the view of suspension defended in this paper makes perfectly good sense of why these non-evidential factors bear on epistemic rationality. It is because they bear on suspension, and suspension competes against belief and disbelief when it comes to epistemic rationality. So the account gives an independently motivated rationale for allowing some non-evidential factors to be within the epistemic’s reach. Now, of course, this does not settle the issue. Evidentialists will disagree with my verdict of the bank cases and the grit cases. The important point is that the view developed here gives us a new way to adjudicate the debate, and it is not clear why one would think that evidentialism is the default. The evidentialist can disagree that non-evidential facts provide reasons to form interrogative and anti-interrogative attitudes or they could disagree that interrogative and anti-interrogative attitudes are forms of suspension. In order to respond in either way they need to defend precise evidentialist-friendly views about which participants compete in the epistemic competition and precise views about the normative profle of those states. This is a much more challenging task than bunching up their noses in an incredulous stare at the thought that non-evidential considerations

Suspension of Judgment

143

are within the epistemic’s reach. In this way, at least, the framework developed here moves the debate about evidentialism forward.22

Notes 1. The orthodox assumption is that belief and disbelief are fundamentally the same attitude that are differentiated by a difference in content. When it comes to some proposition p, belief involves being settled that p and disbelief involves being settled that ¬p. It is fne for us to assume this view, here. I hasten to note, though, that there is some motivation for thinking that while both involve settling, they are not essentially the same attitude. The basic idea is that belief involves a sort of positive settling, whereas disbelief involves a sort of negative settling. One motivation for going in for this is that you eliminate the requirement that one be competent with negation in order to disbelieve. On this view, disbelief is a relation to p, not ¬p. 2. Those who think that only one option will win hold the uniqueness thesis; those who think sometimes more than one option wins (or there are ties) are so-called permissivists. See Kopec/Titelbaum (2016) for a primer. 3. This isn’t meant to suggest that suspension of judgment is always the conclusion of some bit of reasoning, or is always the result of deliberation, or is in some other way always the result of an occurrent bit of mental activity. We sometimes automatically suspend. The point is just that for some claims we have no view rather than suspend. Thanks to Sebastian Schmidt for discussion here. 4. This is misleading in one crucial way: It implies that there is only one intellectual outlook. This, I think, is false. One’s partial beliefs provide another intellectual outlook. There is, of course, great debate about the relationship between the intellectual outlook composed of one’s coarse grained attitudes (belief, disbelief, suspension) and one’s credences. As we’ll see, I don’t think we can reduce suspension to credences. I don’t think we can do it for belief, either. Thus, I think the two outlooks are constitutively separate, which is not to say there are no necessary connections between the two. 5. Sturgeon (2010) argues persuasively that middling credence actually isn’t neutral in the right way. He thinks a degreed notion – what he calls thick confdence – is neutral in the right way. See footnote 10 later for more. 6. Does this require doxastic voluntarism in order for the intention to be effective? It does require that we can execute intentions not to believe. This doesn’t imply the strongest version of voluntarism, which maintains that we can execute intentions to believe. 7. Just to be clear from the outset: I do not think that failing this test is always a decisive blow against a view. After all, if there are multiple ways of suspending judgment, then you are bound to have cases where one way is rational even though suspension in another way is not rational. Since my view posits two different ways of suspending, each way of suspending, on my view, will fail this test in some cases. That said, I will only use this test against theories that maintain there is only one sort of suspension – i.e., maintain that taking a particular neutral stance is necessary for suspension. The test is easier to wield against these theories. 8. See Rosenkranz (2007) for many of the details. 9. This case is a counterexample to Russell’s (1953) and Crawford’s (2004) belief views too, which is why I won’t discuss them in the subsection about belief views. 10. Sturgeon (2010) points out another reason to doubt suspension can be understood in terms of credence: Suspension is a committed neutrality, yet

144

11.

12. 13. 14.

15.

16.

17. 18. 19. 20. 21. 22.

Errol Lord credences are extremely precise ways of not being neutral. This is evidence that credence is a particularly bad tool for understanding suspension. This claim might be dubious in full generality. If you get enough conjuncts, perhaps one should disbelieve the conjunction while suspending about all the conjuncts. But it is very plausible when restricted to relatively small sets of claims. That is all I will need. This is the view Rosenkranz (2007) ends up adopting after considering several other potential contents. Friedman (2013b) makes a similar point. Anti-interrogative attitudes thus play the role that Rosenkranz (2007) wants agnosticism to do. He tries to do the work with higher-order beliefs but fails for the reasons laid out. Miracchi (forthcoming) also discusses states with the functional role of anti-interrogative attitudes. Some virtue epistemologists – e.g., Sosa and Miracchi – have worried a great deal about how suspension can be accommodated in a virtue theoretic framework. The basic issue is how to explain the epistemology of suspension in terms of the sorts of epistemic aims central to such frameworks (e.g., the aim to know). This has led Miracchi to affrm that the epistemology of suspension is derivative in a certain way from the epistemology of belief. My framework makes this unnecessary. Miracchi’s mistake, it seems to me, is to think that suspension is, at bottom, just a lack of belief (which is sometimes accompanied by some other things). This is what makes it diffcult to show that one can aim at epistemic goods while suspending. While it might appear otherwise, these claims are not incompatible with a purely dispositionalist view of doxastic attitudes. For on plausible dispositionalist views, whether one believes is determined by whether one has a suffcient amount of the relevant dispositions (with suffcient robustness). When one does believe, one might have any particular disposition in virtue of the fact that one believes, even if whether one believes is determined by having a suffcient number of dispositions of the right kind. For more details about this, see Lord/Sylvan (forthcoming). For a primer on this debate, see Lord (forthcoming). See, e.g., Ichikawa et al. (2012) and Gardiner (2018). These are slight variations on cases originally found in Stanley (2005). See Schroeder (2012, 2018), DeRose (2009), and Fantl/McGrath (2010). This case is inspired by Paul/Morton (2018). See also Marušić (2015). Thanks to audiences at the Northeast Normativity Workshop, the Penn Normative Philosophy Group, and a Colloquium at Penn, especially Kate Nolf, Nate Sharadin, Jennifer Morton, Sarah Paul, Daniel Fogal, Max Hayward, Jack Woods, Derek Shiller, Max Lewis, Grace Boey, Michael Vazquez, Paul Musso, Michael Weisberg. Thanks also to Jane Friedman, Jim Pryor, and Kurt Sylvan for crucial conversations and Sebastian Schmidt for very helpful written comments. The research leading to these results has received funding from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme (FP7/2007–2013) under REA grant agreement n° [609305]. Research on this project was also supported by the Freiburg Institute for Advanced Study (FRIAS), University of Freiburg, Germany.

References Crawford, Sean (2004): “A Solution for Russellians to a Puzzle about Belief”, Analysis 64, 223–229. DeRose, Keith (2009): The Case for Contextualism, New York: Oxford University Press.

Suspension of Judgment

145

Fantl, Jeremy/McGrath, Matthew (2010): Knowledge in an Uncertain World, New York: Oxford University Press. Friedman, Jane (2013a): “Question-Directed Attitudes”, Philosophical Perspectives 27, 145–174. Friedman, Jane (2013b): “Suspended Judgment”, Philosophical Studies 162, 165–181. Friedman, Jane (2013c): “Rational Agnosticism and Degrees of Belief”, Oxford Studies in Epistemology, Vol. 4, ed. by Gendler/Hawthorne, New York: Oxford University Press, 57–81. Friedman, Jane (2017): “Why Suspend Judging?”, Noûs 51, 302–326. Gardiner, Georgi (2018): “Evidentialism and Moral Encroachment”, in: Believing in Accordance with the Evidence: New Essays on Evidentialism, ed. by McCain, Basel: Springer, 169–195. Ichikawa, Jonathan J., et al. (2012): “Pragmatic Encroachment and Belief-Desire Psychology”, Analytic Philosophy 53, 327–343. Kopec, Matthew/Titelbaum, Michael G. (2016): “The Uniqueness Thesis”, Philosophy Compass 11, 189–200. Lord, Errol (forthcoming): “Evidence and Epistemic Reasons”, in: The Routledge Handbook of the Philosophy of Evidence, ed. by Lasonen-Aarnio/Littlejohn, London: Routledge. Lord, Errol/Sylvan, Kurt (forthcoming): “Reasons to Withhold, Higher-Order Evidence, and Defeat”, in: Reasons, Justifcation, and Defeat, ed. by Brown/ Simion, New York: Oxford University Press. Marušić, Berislav (2015): Evidence and Agency: Norms of Belief for Promising and Resolving, New York: Oxford University Press. Masny, Michal (forthcoming): “Friedman on Suspended Judgment”, Synthese. Miracchi, Lisa (forthcoming): “When Evidence Isn’t Enough: Suspension, Evidentialism, and Knowledge-First Virtue Epistemology”, Episteme. Paul, Sarah/Morton, Jennifer (2018): “Grit”, Ethics 129, 175–203. Raleigh, Thomas (forthcoming): “Suspending Is Believing”, Synthese. Rosenkranz, Sven (2007): “Agnosticism as a Third Stance”, Mind 116, 55–104. Russell, Bertrand (1953): “What Is an Agnostic?”, The Basic Writings of Bertrand Russell, New York: Routledge 2009, 557–565. Schroeder, Mark (2012): “Stakes, Withholding, and Pragmatic Encroachment on Knowledge”, Philosophical Studies 160, 165–185. Schroeder, Mark (2018): “Rational Stability under Pragmatic Encroachment”, Episteme 15, 297–312. Stanley, Jason (2005): Knowledge and Practical Interests, New York: Oxford University Press. Sturgeon, Scott (2010): “Confdence and Coarse-Grained Attitudes”, in: Oxford Studies in Epistemology, ed. by Gendler/Hawthorne, New York: Oxford University Press, 3–26.

Part II

Facets of an Ethics of Mind

A

Responsibility, Reasons, and Rationality

8

Responsibility for Attitudes, Object-Given Reasons, and Blame Sebastian Schmidt

1.

Introduction

Philosophical thought about responsibility is traditionally structured by taking actions and their consequences as the kinds of things for which we are responsible.1 This way of thinking about responsibility highlights a difference between two modes of being responsible: direct and indirect responsibility. On the one hand, our actions are exercises of voluntary control, and we perform our actions for reasons. Their causal consequences, on the other hand, are not exercises of control, and they are no things performed for reasons.2 Our responsibility for consequences originates in our responsibility for actions which cause them.3 This allows us to say that we are directly responsible for our actions but only indirectly responsible for their consequences. Whenever we are responsible for a consequence, our responsibility can be traced back to our responsibility for prior actions.4 If we try to put attitudes within this traditional framework, we are faced with a puzzle. On the one hand, it seems that we lack direct control over our attitudes. In this respect, attitudes behave similarly to mere consequences of our actions. Indeed, it seems undeniable that they are sometimes consequences of our actions, and that we thus have, to a certain extent, indirect control over them: we can manage our emotions through, say, meditation, and we can form justifed beliefs by, for example, proper investigation. Yet it is hard to see what another, more direct kind of control over attitudes is supposed to look like. If we lack direct control over our attitudes, it seems that we should conclude that we are only indirectly responsible for our attitudes – in the same way as we are responsible for consequences of our actions. On the other hand, our attitudes are within the “space of reasons”: they are subject to evaluations to which brute consequences of actions could never be subject to. Attitudes cannot only be evaluated as better or worse to have (as consequences can be evaluated as better or worse). Rather, we think of our attitudes as rational or irrational. In this respect, it seems that our attitudes do not behave like mere consequences of our actions: if I fall

150

Sebastian Schmidt

from the roof due to my carelessness, my broken leg will be my fault, but my broken leg is not irrational. Given our attitude’s presence in the space of reasons, they seem to be much more like actions themselves rather than their consequences. We are tempted to conclude from this second line of thought that we must be also directly responsible for our attitudes (in addition to sometimes being indirectly responsible for them). Luckily, we need not decide between both lines of thought. We can endorse both. Or so I argue. We can claim that each of them is sound as long as we say that each of them is sound for another face of responsibility for attitudes. To see this, we frst need to get a grip on the problem created by the two lines of thought. Understanding the nature of this problem is half the way to its solution. After sketching my assumptions about reasons and rationality in this paper (Section 2), I will argue that the problem of responsibility for attitudes is ultimately a problem about reasons for attitudes (Section 3). Resolving it requires us, frst and foremost, to understand the status of the norms of rationality as compared to the status of requirements of prudence and morality. In order to motivate my own approach on understanding this status, I frst spell out an argument that calls into question the normative force, and thus the very existence, of reasons for attitudes (Section 4). The argument works with a plausible conceptual claim about how reasons and blameworthiness are related. In reply to the argument, I show how we can be held directly responsible for our attitudes, and how this makes the normative force of reasons for attitudes intelligible (Section 5). This provides the outlines of a solution to the problem of mental responsibility (Section 6).

2.

Rationality and Reasons

Following the recent defenses by Kiesewetter (2017) and Lord (2018), I assume a reasons-based conception of rationality, rather than a conception of rationality as mental coherence (cf. Broome 2013). To be rational is to respond correctly to one’s overall accessible reasons for an attitude. If it turned out that rationality cannot be understood as responding correctly to reasons, then my overall argument is not affected. This is because I am interested in the normative force of reasons for attitudes rather than the normative force of attitudinal coherence. If you do not agree with reasons-based conceptions of rationality, you are free to read my use of “rational” as stipulative: it refers to the property of an attitude of being adequately based on the subject’s accessible reasons for this attitude. To be rational means, according to my terminology, to respond correctly to one’s object-given reasons for the attitude.5 Object-given reasons for an attitude are reasons that indicate (or constitute) facts about the attitude’s object rather than about the attitude itself – facts that support the attitude, or make it rational to have the attitude, by indicating (or constituting) that the attitude (partly) fulflls its constitutive aim. For example,

Responsibility for Attitudes

151

object-given reasons for beliefs are (or are provided by) evidence, because evidence indicates the truth of the object of the belief, i.e., the truth of the belief’s propositional content. Scientifc reports on climate change are thus object-given reasons for belief: they indicate that human-induced climate change takes place. By contrast, that I feel less existential angst if I do not believe in climate change is a state-given reason not to believe in climate change. It is not an object-given reason (i.e., no evidence, or not provided by evidence) against climate change, because this fact does not indicate that the belief fulflls its constitutive aim of truth. Analogously, an object-given reason for a desire shows the object of the desire to be desirable in some respect; object-given reasons for fear indicate the danger of what you fear, and thus make it rational to experience fear;6 object-given reasons for intention are reasons for the object of intention – i.e., the intended action: that I will get poisoned if I drink a toxin is an object-given reason not to intend to drink it (and a reason not to drink it); that I get a lot of money for intending to drink a toxin is a state-given reason to intend to drink it.7 In contrast to object-given reasons, state-given reasons indicate (or constitute) facts about the state of the attitude itself. The most important category of state-given reasons under discussion are practical reasons for attitudes, i.e., reasons that support the attitude by indicating the attitude’s value.8 According to those who believe in such practical reasons – i.e., pragmatists – the fact that it would be benefcial to believe in a guardian angel (e.g., because one would sleep better at night) could be a practical reason to believe in a guardian angel (even in the absence of suffcient evidence).9 Analogously, facts about the value of any other attitude are considered to be reasons for this attitude by pragmatists. It is often argued that practical reasons are no reasons for attitudes but rather mere considerations indicating the attitude’s value.10 They are no reasons because if they were, then we would sometimes be able to form attitudes for practical reasons: we could believe, desire, feel, or intend at will. We could just decide not to fear a dangerous tiger or just decide to intend to drink a poisonous liquid merely because we regard fearing or intending to be benefcial. I will assume here that we cannot adopt attitudes at will, and I thus agree that practical reasons are, strictly speaking, no reasons for the attitude they seem to favour.11

3.

The Problem

McCormick helpfully characterizes the confict between responsibility for belief and the absence of doxastic control as follows: Attributions of responsibility and other deontological judgments in the doxastic realm are puzzling. For much of what we believe is beyond our control; we cannot decide to believe the way we can

152

Sebastian Schmidt decide to act. It seems that such lack of control should excuse us from responsibility and judgment. McCormick 2015: 77

As McCormick notes, the problem is not restricted to responsibility for beliefs. It seems that deontological attribution with respect to belief is puzzling in the face of our lack of doxastic control. After all, in some sense it seems to be true that “Ought implies Can.” But how can sentences like “You should not believe this nonsense” or “There is no reason at all to believe this” be true if there is no doxastic control, that is, no “Can”? The very possibility of an ethics of belief, as well as the project of normative or deontological epistemology, seems to become questionable.12 Are we dealing with one problem here, or with two? Is there a problem about responsibility on the one hand and a problem about reasons, justifcation, and rationality, on the other? In this section, I will argue that the problem of mental responsibility is best understood as the problem of reasons for attitudes. Seeing this will allow us to rule out accounts that want to solve the problem by appeal to indirect control, and it will allow us to see the candidate solutions. 3.1

How We Seem to Lack Mental Control

Let us consider, frst, why it seems that we cannot control our beliefs. Most of our beliefs seem to be passively caused by our environment. We experience the acquisition of a perceptual belief – like the belief that you are reading this text right now – as something that happens to us, rather than something we decide for. You do not stop and refect upon whether you should acquire this belief after the text is in front of your eyes. Rather, belief comes immediately with perception. The same holds in cases where our perception provides us with ambiguous evidence. In these cases, we refrain from judgment quite automatically. To take a classical example (cf. Sextus Empiricus, Outlines of Pyrrhonism, I.32): If a tower seems to be round from distance and I know that I am not close enough for judging the tower’s shape, I cannot just decide to believe whatever strikes me as pleasing to believe. Rather, I refrain from judging quite automatically. If, however, I am close enough to the tower, I will come to believe what shape it really is without any further contribution from me (except for walking towards it). Even in cases where we refect about what to believe, there seems to be no place for genuine doxastic freedom. Imagine a scientist, Lara, who wants to fnd out whether a certain substance is water. Lara conducts some experiments, thinks about them, and comes to believe that it is in fact water. It seems that the only activities Lara did here were the experiments and her intentional deliberating or reasoning about what is true. Lara did not perform another action when she was done with her

Responsibility for Attitudes

153

thinking. She did not have to actively form a belief. Rather, her thinking concluded in the formation of a belief, but this concluding is not an additional activity she had to perform. This is confrmed by the following observation. Lara could engage in the experiments she conducts and in her activity of thinking for practical reasons: She could decide to engage in them or refrain from doing them depending on whether it was important or of interest to know the nature of the substance. By contrast, she could not decide to conclude her investigation for such reasons: considerations about whether it would be good, valuable, important, interesting, or useful to believe that the substance in question is water could not guide the formation of her belief when she is done with her previous activities. Rather, she will form the belief based on her evidence that was uncovered by the preceding investigation.13 Thus, there is a dilemma for the proponent of doxastic control: Either we form our beliefs spontaneously, or we form them refectively. Spontaneous belief formation seems to happen quite automatically, and thus there seems to be no room for genuine freedom. However, even when we form our beliefs refectively, only our refection, or intentional thinking, is active, but not the formation of our belief that results from it. As Richard Moran points out: “[T]here is no further thing the person does in order to acquire the relevant belief once his reasoning has led him to it” (2001: 119). Our rational capacities just work the way they do – we do not exercise direct control over the results of their proper functioning. It seems that it is not us but rather our evidential situation, together with the way our rational capacities function, that determines what we believe in each situation – in which we might happen to be either with or without our contribution (cf. Strawson 2003). This is a dilemma not only for the proponent of direct doxastic control but for the proponent of direct control over attitudes in general. Most of our attitude-forming takes place without us being refectively aware of it. We spontaneously form not only numerous beliefs about our environment, but also other attitudes. We form desires and intentions to make it in time to the meeting, hopes that we will still make it, fears that we won’t, regrets that we did not get up earlier to make it in time, or feelings of anger and resentment directed at the person in the car in front of us who did not drive on when they had green lights. In these cases, we would not even have the time to decide for our attitudes, even if we could. And even if we have the time to form our desires, intentions, and emotions refectively, then we might intentionally engage in the activity of thinking about what is good, right, fearsome, regrettable, or deserving of our anger and blame. But the attitudes that arise from such voluntarily controlled thinking are nothing we choose. Even if we actively think about how to decide, it seems that our resulting intention or decision is nothing that we could ever directly decide for.14

154

Sebastian Schmidt

3.2

How Lack of Voluntary Control Does Not Threaten Mental Responsibility

How could this lack of control over attitudes pose a problem for our practice of holding each other responsible for our attitudes? It is helpful to state the problem as a confict of three claims which, when all true, would result in a contradiction. Thus, to resolve the confict, we have to reject at least one of the three claims.15 A frst try would be to state the problem as follows: (1*) We are responsible for our attitudes; (2*) We are responsible for our attitudes only if we can control them; (3*) We cannot control our attitudes. The problem with this way of formulating the puzzle is that (3*) would be obviously false. As mentioned on various points before, we have at least indirect control over our mental states: we can meditate, investigate, and actively engage in thought and reasoning. We can also control our mind indirectly by engaging in projects of acquiring, say, true beliefs and other correct attitudes (right intentions, ftting emotions) about Saint Petersburg by going there and walking through the city. So, this frst attempt of formulating the problem does not get at its core. Given that it is not any control that seems lacking, but voluntary control, that is, the kind of direct control we have over our actions, we could put (3*) not in terms of control but rather in terms of voluntary control. Still, this would not get to the core of the problem. Note that our exercises of voluntary control – our actions – are subject to requirements of prudence and morality. If we had voluntary control over beliefs, we could evaluate someone’s beliefs in terms of how benefcial they are to the person who has them, or in terms of how harmful they are to others. Yet we would not be justifed in holding people accountable for their epistemic failures: we would not be justifed in regarding someone as criticizable or blameworthy for their epistemic irrationality, i.e., for someone’s failure to have beliefs that are properly based on their evidence. For we would only be justifed to blame or criticize someone for their irrationality if this irrationality is bad either in terms of prudence or morality. Yet it seems that we are sometimes justifed to blame or criticize people for failing to properly base their beliefs on their evidence rather than for failing to have beliefs that are not harmful for themselves or for others. One reply is, of course, to just deny that there is such a thing as epistemic blame: we never blame people merely for being epistemically irrational; rather, we blame them for irrationality only if their irrationality is harmful to themselves or to others. We will return to this option of denying the existence of epistemic blame as well as other forms of blame for holding attitudes in Section 4. For now, it is suffcient to note that denying the criticizability of irrationality would be highly controversial in the context of contemporary philosophical debates. The claim that irrationality is criticizable can be viewed as a starting intuition for defenses of

Responsibility for Attitudes

155

the normativity of rationality (cf. esp. Kiesewetter 2017, ch. 2). If we are criticizable or blameworthy (in what follows, I will stick with the label “blame” instead of “criticism” for the sake of brevity) for such failures in our rationality, then voluntary control over attitudes would provide an extensionally inadequate explanation of our blameworthiness for irrationality: we would be blameworthy only if our irrationality is prudentially or morally bad. Note that the kind of responsibility we have for complying to the norms of rationality cannot be straightforwardly explained by reference to indirect control. Indirect control does a good job in explaining why we are responsible for phenomena that are not responsive to reasons and thus cannot be evaluated as rational or irrational, like brute sensations. The question “Why do you feel pain?” can only be answered by giving a causal explanation (“I fell from the roof”). It cannot be answered by giving justifying reasons why it would be rational to feel pain now. For pain, understood as a mere sensation which can occur at various bodily parts, cannot be rational or irrational. It is because of this difference between mere consequences of our actions and reasons-responsive attitudes that our responsibility for the former (which include some of our sensations, like pain) is exhausted by indirect responsibility, but that our responsibility for attitudes is not – at least not as long as we assume that we are directly responsible for being (ir)rational. If we are directly responsible for being (ir)rational, then there will be cases conceivable in which we lack indirect voluntary control over an attitude, but where we are still responsible for the attitude. For the attitude might be considered as (ir) rational even if we could not exercise indirect control over it. I take this to be an intuitively plausible assumption about the norms of rationality. Thus, any account that explains mental responsibility in terms of voluntary control must deny that we are directly responsible for being (ir)rational. Otherwise, explaining mental responsibility by reference to direct voluntary control would result in an extensionally inadequate account of blameworthiness for attitudes; and explaining it with indirect voluntary control would result in an extensionally inadequate account of responsibility for attitudes.16 Proponents of explanations of mental responsibility in terms of voluntary control could argue that I misconstrued the nature of rationality. They could adopt a pragmatist account of attitudinal rationality: Such a pragmatist account claims, roughly, that to have rational attitudes is to have attitudes that are, on balance, prudentially and morally best to have.17 However, this is just to deny that we are ever held accountable for failures to respond correctly to our object-given reasons for attitudes: it is to deny that we are responsible for being (ir)rational in the way I use the term “(ir)rational” (see Section 2 earlier). Intuitively, however, we are sometimes blameworthy for failing to comply with our object-given reasons.

156

Sebastian Schmidt

3.3

How Lack of the Right Kind of Mental Control Threatens Mental Responsibility

Since the problem of mental responsibility calls for an explanation (at least also) of the kind of responsibility we have for (non-)compliance with the norms of rationality, and since it seems that this responsibility can be made intelligible neither by direct nor by indirect voluntary control, we should now try the following formulation: (1**) We are responsible for (non-)compliance with the norms of rationality; (2**) We can only be responsible for such (non-)compliance if we have direct nonvoluntary control over our attitudes; (3**) We do not have such direct non-voluntary control. I think this is a better way to state the problem. Yet we might wonder why (2**) is true as stated. For what is non-voluntary control supposed to be? Intuitively, the control we exercise when we act – direct voluntary control – is our only paradigm of direct control, and this is the kind of control that (at least partly) explains why we are responsible for our actions and their consequences. So how could a non-voluntary control explain norms? Since it seems that we do not know much about the nature of this control, we also do not see, prima facie, how it could do this trick. It thus seems that, intuitively, there is no form of control – direct or indirect, voluntary or non-voluntary – that could explain how we are responsible for being (ir)rational. The norms of rationality are constitutive standards for an attitude: the more irrational someone is, the less we can be sure what attitudes the person holds. If someone says that they believe that drinking water from the sink at midnight on a certain date has healing effects, but also believes in the epistemic authority of contemporary physics and thus accepts a naturalistic world-view, then we normally do not only regard the person to be irrational, but we might also come to doubt whether the person has one of the beliefs we ascribe to them. The more inconsistent someone’s set of beliefs is, the more doubtful it is what it is that the person believes. More generally: the more a person violates the norms of rationality, the more doubtful it is what attitudes the person holds.18 If we are indeed responsible for complying with these constitutive standards, i.e., the norms of rationality, and if it is not obvious what form of control could explain this, then we have an interesting version of the problem of mental responsibility: 1. We are responsible to the standards of rationality. 2. We can only be responsible to the standards of rationality if there is some form of control that could explain why we are responsible to the standards of rationality. 3. There is no form of control that could explain why we are responsible to the standards of rationality.

Responsibility for Attitudes

157

The plausibility of (1) derives from the intuition that being irrational – failing to respond correctly to one’s object-given reasons – is blameworthy in some sense. This implies that we are responsible for (non-) compliance with the standards of rationality: We are legitimately subject to various reactive attitudes in virtue of our being rational or irrational. As a shorthand formulation, I will say that we are responsible to the standards of rationality. The plausibility of (2) derives from the intuition that responsibility requires control, or “Ought implies Can” – a claim that is, in one version or another, accepted also by philosophers concerned with the standards of rationality (cf. Kiesewetter 2017: 28; Wedgwood 2017, ch. 3). I have already motivated the idea that there seems to be no such thing as direct control over our attitudes (in Section 3.1), and that our familiar paradigm of control – voluntary control (direct or indirect) – cannot do the job of explaining why there is such a thing as responsibility to the standards of rationality (in Section 3.2). This motivates (3). In order to resolve our philosophical perturbation caused by the three claims, we need to fnd reasons for rejecting at least one of them. Thus, all the premises are intuitively plausible. There are at least two other advantages of understanding the problem of mental responsibility in this way. First, we know better what it is that requires an explanation. Merely saying that we need to explain how we can be responsible for our attitudes leaves it unclear what exactly it is that requires an explanation. For it is obvious that we can be said to be sometimes responsible for our attitudes in the same way as we can be said to be sometimes responsible for other consequences of our actions. We are indirectly responsible for our attitudes in this way insofar as we have indirect control over our attitudes: We can intentionally reason to beliefs, manage our emotions by meditation, and determine our intentions and decisions by intentionally thinking about what to do. If we understand the problem as merely requiring an explanation of our responsibility for attitudes without specifying what kind of responsibility we have in mind, then it is unclear why indirect voluntary control does not provide a satisfying explanation. Yet, as I have argued, it is intuitively unclear how our capacity to indirectly control our mind can explain why we are responsible for (non-)compliance to the norms of rationality. Indirect responsibility for attitudes seems thus not to exhaust our responsibility for them. Since we do not see what kind of control could explain the special responsibility to the norms of rationality, we are faced with our problem. Secondly, stating the problem in this way has a metaphilosophical advantage. For it shows us how different debates are relevant for solving the problem. For one, we need to think about the nature of responsibility and control. However, we also need to think about the nature and the status of the norms of rationality. Thinking about their status means to think about how we are responsible, and thus sometimes blameworthy, for (non-)compliance with our reasons for attitudes. Getting clear about

158

Sebastian Schmidt

this kind of responsibility might give us a clue as to what kind of control we are searching for, or about whether this search is futile (because the relevant responsibility might not require control at all). Thus, theorists of responsibility need to think about rationality and reasons, and theorists of rationality and reasons need to think about responsibility. As long as we do not understand in what sense we are responsible for complying with our reasons for attitudes, and thus in what sense we can be legitimately blamed for irrationality, we do not understand the status of these reasons. To see this more clearly, I will spell out how we might call into question the status of the norms of rationality as norms to which we are held responsible (i.e., how we might question (1)), and how this can give rise to three different approaches to our problem, before I sketch my own solution (denying (2)).

4.

Mental Blame and Reasons for Attitudes

A promising line of thought for questioning (1) starts off by considering cases of trivial belief (Section 4.1): why, one might wonder, should we be blameworthy for whether we form irrational trivial beliefs – i.e., beliefs that have no impact on ourselves or others? I show how we can extend this line of thought to conclude in a denial of the very existence of reasons for attitudes (Section 4.2). If there are no such reasons, then there are no norms of rationality to which we could be held responsible. To avoid this position, which I label nihilism,19 we either have to adopt pragmatism about reasons for attitudes, i.e., the idea that some reasons for belief are provided by the value of the attitude itself; or we have to make sense of epistemic blame and other forms of mental blame while assuming that reasons for attitudes are merely object-given. I call this last option, which I endorse, constitutivism.20 4.1

Denying Epistemic Blame

Gilbert Harman (1986: 12) famously points out that the following epistemic requirement would lead to “cluttering one’s mind”: (ER) One ought to believe everything that is suffciently supported by one’s evidence. According to the recent consensus, (ER) is not plausible. To see why, consider that my current belief-stock implies the proposition that [I am sitting in my offce or the moon is made of cheese or Hitler is alive or there is no human-induced climate change or . . .]. There is no doubt that this disjunctive proposition is true right now while I am sitting in my offce, no matter how many odd (but meaningful) claims we add. It is true because the frst claim, that I am sitting in my offce, is true. At the

Responsibility for Attitudes

159

same time, it seems that I do not always believe this disjunctive proposition while I am sitting in my offce. Most importantly, I would neither be irrational nor blameworthy (or criticizable) in any sense for not believing such disjunctive propositions. It follows that there is no unconditional norm to believe everything that is suffciently supported by my evidence. This far, contemporary epistemologists agree.21 Yet many philosophers want to modify (ER) to a more plausible claim by spelling out under which conditions our evidence requires us to believe a proposition. They thus include background conditions for when our suffcient evidence provides us with decisive reasons to believe, i.e., conditions under which suffcient evidence makes it the case that we ought to believe a proposition. In this vein, Benjamin Kiesewetter (2017: 184–185) responds to Harman’s clutter-objection by proposing that the central requirement of theoretical rationality is to believe what one’s evidence suffciently supports if one attends to this evidence. According to this proposal, if I attend to the fact that I have suffcient evidence for a specifc disjunctive proposition, then I would be irrational, and thus criticizable, if I do not come to believe it. Thus, I ought to believe it.22 Yet we can reasonably wonder whether Kiesewetter’s modifcation will do. Steglich-Petersen (2011) argues that epistemic reasons are not suffcient by themselves to determine what we ought to believe. And this, I take it, is independent of whether one attends to them. I may have strong evidence in favour of p to which I attend, and yet it seems that it still might be false that I ought to believe that p: Suppose, for example, that the subject matter is whether there is an even number of dust specks on S’s desk. Let us also suppose that S has excellent evidence, and thus epistemic reason to believe in the sense defned, that there indeed is an even number of dust specks on his desk. In spite of this epistemic reason, it does not seem to be the case that S ought to form the belief that there is an even number of dust specks on his desk. It may be that S as a matter of fact cannot avoid forming that belief, since we are psychologically disposed to form beliefs that are supported by consciously considered evidence. But it is nonetheless not the case that S ought to form that belief. If S failed to form the belief, we wouldn’t fault him or regard him as normatively worse off for that reason. Steglich-Petersen 2011: 23 One could object that two levels of evaluation get confused here: practical and epistemic. Even if S is not necessarily blameworthy from a practical perspective when he does not believe on the basis of suffcient evidence, he is, according to this objection, still blameworthy from an epistemic perspective. However, it is important to see here that certain emotional reactions that are usually conceived of as forms of blame – like

160

Sebastian Schmidt

resentment or indignation – seem to be inappropriate towards the person in Steglich-Petersen’s example. Why should we resent S for believing some utterly unimportant truth nobody cares about? Surely, S should not feel guilty about failing to believe it either. Furthermore, it is unclear what, if any, consequences S’s purported epistemic failure should have for future relationships with other people who know about this failure.23 Being epistemically irrational only once does not seem to have any effect on, for example, whether one is a trustworthy source of information as long as this single case of irrationality is an exception.24 So, in what sense then, is S blameworthy?25 I will call any form of blame that arises merely from the fact that someone fails to properly base one’s doxastic attitudes (belief, disbelief, or suspension of judgment) on one’s evidence epistemic blame. We can reasonably ask whether there is such a thing as epistemic blame. Until now, it seems that merely failing to believe what one’s evidence suffciently supports does not give rise to epistemic blame. Rather, it seems that it at least has to matter in some way that one believes what one’s evidence suffciently supports in order for such a requirement to be placed on us. Whether it matters, however, seems not itself to be determined by one’s evidence, but rather a matter of practical considerations that show it to be important that one believes what one’s evidence suffciently supports. If we end up blameworthy only in cases with such practical stakes – for example, when we fail to believe in human-induced climate change or have racist beliefs – then there does not seem to be much point in a notion of purely epistemic blame.26 4.2

Mental Blame and Reasons

Analogously to the doubts about the existence of epistemic blame arising from our consideration of trivial propositions, we can formulate doubts about the existence of mental blame. I will call any form of blameworthiness that arises merely from the fact that someone fails to respond correctly to one’s reasons for attitudes mental blame. That is, if a subject is blameworthy for their attitude in virtue of the fact that they had indirect control over their attitude, then the resulting blame is not mental blame. Here is the initially plausible argument for a radical claim. As long as being rational does not matter, it seems that we cannot be blameworthy for failing to be rational. Since object-given reasons do not necessarily indicate that having an attitude matters, failing to be rational can never, by itself, make us blameworthy. This means that there is no such thing as mental blame. However, if there is no such thing as mental blame, then it is hard to see how object-given reasons could have any normative force – it becomes doubtful whether (purported) object-given reasons for attitudes are reasons at all. Call the radical conclusion that there are no reasons for attitudes nihilism about reasons for attitudes.27

Responsibility for Attitudes

161

One important assumption in the argument is that there are reasons for attitudes only if there is mental blame. We can spell out this assumption as a general condition that a kind of consideration must satisfy in order to be a kind of reason: Reasons and Blame (RB). Considerations of kind K are reasons for or against φ-ing only if we can be blameworthy for φ-ing just in virtue of the fact that φ-ing is (or seems to be)28 decisively supported by considerations of kind K. (RB) is a very plausible principle. This is because it is a very minimal claim about how reasons and blameworthiness are connected: It must be possible to be blameworthy for non-compliance with these (purported) reasons. If it is not possible, a consideration does not deserve its status as a reason for a response. Harman’s clutter-objection from Section 4.1 would not get any grip on us if we would not assume that failing to comply to an epistemic norm is supposed to make us, at least normally, blameworthy or criticizable (in some sense). If we would not assume this, then we could just accept that by not drawing all the implications from our current beliefs we fail to believe what we ought to believe: we would just constantly violate an epistemic norm, but that would not have any further signifcance. However, the normative force of this “ought” would then be mysterious: why comply with this norm if no one can hold us legitimately accountable for non-compliance? The norm would at best have the force of the norms of etiquette: we can reasonably ask why we have a reason to comply with the norms of etiquette in a given situation, and there are cases conceivable in which we ought not to comply with them (cf. Kiesewetter 2017: 4). Epistemologists propose background conditions on epistemic norms, like Kiesewetter’s attending-condition (cf. Section 4.1 earlier), because they want to make sense of the normative force of these standards: they want to explain why it matters to us whether we comply with the norms, why we can be (at least normally) blamed if we fail to comply with them. The whole project of spelling out a notion of epistemic blame is, I take it, to understand the normative force of evidence: of why it matters to comply with epistemic norms.29 Note that (RB) is not even controversial for objectivists about the meaning of “ought” and “reasons,” who claim that whether a subject ought to or has reasons to φ depends on the actual circumstances the subject is in rather than on what (apparent) facts about the circumstances are epistemically accessible to the subject. (RB) is plausible to objectivists because, even according to them, it is possible to be blameworthy for non-compliance with reasons: namely, in cases in which these reasons are accessible to the subject.30 If (RB) is as uncontroversial as I present it, how can it support such a radical claim as nihilism? It only supports nihilism if we assume that

162 Sebastian Schmidt we are never blameworthy merely in virtue of the fact that we fail to respond correctly to (what seem to be) decisive reasons for attitudes. If we were never blameworthy, then what we deem to be decisive reasons for attitudes would, assuming (RB), turn out to be no reason after all. They would lack the specifc normative force of reasons. However, I have merely presented a case of irrational belief in the last subsection (the dust-speck case) where we do not seem to be blameworthy for our noncompliance with (what seem to be) decisive object-given reasons for this belief (i.e., reasons provided by suffcient evidence to which you attend). How can a single case support such a general conclusion? To see how it might support such a conclusion, note how the dustspeck case shares a general structure with analogous examples: S does not believe some unimportant proposition p even though S has suffcient evidence for p (to which S attends). If all cases with this general structure imply that S is not blameworthy, then it seems that we are never blameworthy in virtue of the fact that we fail to properly base our beliefs on our evidence. For cases in which the proposition is important to know seem, prima facie, not to help us to make sense of epistemic blame. In cases where it is morally or prudentially important to believe what is suffciently supported by the evidence, it might seem that the subject’s blameworthiness is grounded in moral or prudential failure, rather than in failure to base one’s belief on the evidence. In order to support nihilism about reasons for attitudes more generally, we would need cases for other attitudes that are analogous to the dustspeck case in that we judge an attitude intuitively to be decisively supported by object-given reasons, but where the subject is not blameworthy for failing to comply with these (purported) reasons and where it is hard to see how subjects could ever be blameworthy in all structurally analogous cases. My task here is not to spell out these cases, because I will not endorse nihilism but rather present a viable alternative. I think, however, that a strong case for nihilism can be made by spelling out such cases.31 In order to avoid nihilism about reasons for attitudes, we have two options. First, we could endorse pragmatism about reasons for attitudes. That is, we could argue that reasons for attitudes are practical reasons, rather than (merely) object-given reasons. If there were practical reasons, then failing to comply with these reasons would matter in a straightforward sense: in not responding to them correctly, we would fail to promote value. Thus, endorsing (RB), together with an argument that we are never blameworthy merely for failing to comply with our object-given reasons, can motivate pragmatism.32 In the remainder of this paper, I would like to explore how we could avoid nihilism while still regarding reasons for attitudes as purely object-given. My account will provide the material to make it intelligible how we are responsible to the norms of rationality, which I understood as described in Section 2, i.e., as nonpragmatic norms to respond correctly to one’s object-given reasons.

Responsibility for Attitudes

5.

163

Blame and Irrationality

We can motivate the idea that we are sometimes blameworthy for being irrational by focusing on certain emotions (Section 5.1). This will give us a clue towards a general account of mental blame as marking the impairment of relationships (Section 5.2). Understanding mental blame will provide a diagnosis and the outlines of a solution to the problem of mental responsibility (Section 6). 5.1

Blame for Emotions

I have motivated the threat from nihilism by thinking about epistemic blame and reasons for belief. I now suggest that we might get a clue about how to avert this threat by thinking about reasons for emotion and how we blame each other for emotional irrationality. This will suggest a general account of mental blameworthiness that is also applicable in the doxastic realm. Take admiration and love. You can be blameworthy in some sense if and because you do not admire features of the world that are admirable, or if and because you do not love someone who is worthy of your love. Whether you are blameworthy will depend on whether we understand admiration and love as stable dispositions or as occurrent feelings. For we might well imagine a case where I am confronted with something admirable (say, I am standing in front of the ancient pyramids), but where I do not feel admiration (say, due to my being stressed out from travelling), and where I am not blameworthy (I am excused due to stress). Yet one might think that I would be blameworthy in some way if I do no not have a disposition to feel admiration for what is admirable under usual circumstances. Similarly, if I do not feel the love for my partner in a specifc moment (say, because I am stressed out by them right now), this need not make me blameworthy. Yet if I do not have any disposition to feel love for them, I might plausibly be said to be blameworthy for this lack of love for someone who would deserve it. We here have some potential counterexamples against the nihilist’s premise that there is no such thing as being blameworthy in virtue of non-compliance with (purported) object-given reasons for an attitude. How might the nihilist respond to these cases? They could try to deny that we are blameworthy for failing to admire the admirable and love who is worthy of our love in virtue of the fact that we fail to respond correctly to object-given reasons for these emotions. According to this objection, we should give another explanation for the blameworthiness involved in such cases. A nihilist might start out with the following thought: When something is admirable or worthy of our love, this essentially involves practical considerations in some way, i.e., considerations that explain why it matters to have these emotions.

164 Sebastian Schmidt More precisely, they could argue that our intuition that the emotionally irrational are blameworthy arises only because we assume that they could have responded to these practical considerations (rather than to their object-given reasons): what they are blameworthy for is failing to cultivate admiration and love – for cultivating admiration and love is something we can do for practical reasons. In reply, I merely point out that we can very well imagine cases where there was no reasonable opportunity to cultivate admiration and love, and where our intuitions about the criticizability of those who fail to acknowledge admirability and lovability do not fade: there is still something substantially wrong with a person who fails to feel the way we would expect a rational person to feel when it comes to emotions like admiration and love. What exactly is wrong with them will become clearer when I spell out my proposal about the nature of mental blame below. A worry with the way I motivate the idea of mental blame here is that I focus on specifc emotions. One might point out that, even if I am right that we blame one another for certain forms of emotional irrationality, this could at most save the idea that there are reasons for these emotions, but not that there can be reasons for other attitudes. We might conceive of a position that grants that we can be blameworthy for failing to comply with object-given reasons for certain attitudes, but not for others. According to this position, there might be, for example, no epistemic blame: we would never be directly blameworthy for our beliefs merely in virtue of our failure to properly base our belief on our evidence. Rather, we would be blameworthy only if we failed to properly manage our doxastic life. My objection to this position is that it is odd. It grants that there are reasons for certain emotions. But it denies that there are reasons for other attitudes. However, there seem to be too many structural analogies between reasons for attitudes for such a position to be coherent. The reasons in virtue of which we are blameworthy for emotional irrationality are object-given: they are provided by the constitutive aims of admiration (the admirable) and love (the loveable). Also reasons for belief, desire, and intention are object-given reasons provided by the constitutive aim of the respective attitude. Because of this striking structural analogy, I deem it to be more promising to explore how we can be blameworthy also for violations of epistemic norms and of other norms of rationality, rather than accepting a scattered account of reasons for attitudes according to which we can be blameworthy only for some failures of rationality, but not for others. 5.2

The Nature of Mental Blame

My suggestion is that in failing to be rational, we often display blameworthy vices.33 The inability to acknowledge the admirable and lovable is a defect in character that has specifc normative consequences for our

Responsibility for Attitudes

165

relationship with others. These normative consequences result from emotional irrationality independently of whether the person could manage their emotion. Why explore the world with someone who cannot admire its admirable features? Why be in a romantic relationship with someone who cannot appreciate you as being worthy of their love? Even though these defects in character do not, by themselves, warrant responses like resentment, indignation, and guilt, they warrant blame insofar as it can be rational not to get involved in specifc ways with the people who suffer these defects: certain types of relationships with them would necessarily be impaired.34 The key to understanding how we can be responsible to the norms of rationality is, I propose, to acknowledge a broader concept of blame that goes beyond blame as essentially involving such passionate responses as resentment and indignation.35 If we have a broader concept of blame in mind, then our intuition that responsibility for attitudes requires control will fade, and we are free to deny claim (2) of the problem of mental responsibility. Angela Smith argues that some reactions rightly deserve the label “blame” without involving passionate components (2013: 32).36 Especially when blaming loved ones for moral failures, we might do so without hostile emotions. Furthermore, we might mark an impairment in our relationship to someone by “dispassionately ‘unfriending’ someone on one’s Facebook page, for example, or by simply refusing to trust anymore, and these too should qualify as blame” (ibid.). According to such a concept of blame, viewing the relationship to someone as impaired just is the blame (Hieronymi 2004). Thus, here is a sketch of the general account of mental blame I propose.37 If we come to see a person’s irrational vices, we might reconsider our relationship with them. We might want no longer to be friends with the person. We might cease to promote their personal projects or not take pleasure in their successes when we see that they adopt their aims and choose their means only because they are greedy, weak-willed, cowardly, intemperate, ungenerous, unjust, gullible, or dogmatic. According to Owens, when I am gullible, then “I cannot be trusted to think and feel as I ought” (2000: 124). We might also no longer want to please the person whose viciousness we have discovered, and no longer accept credit from them for our own actions. We might doubt their judgments more generally because of a general sense of distrust we develop towards them. These are all reactions we can only show to fully responsible beings, because we can only have the relationships that are presupposed by these reactions towards fully responsible beings. Neither computer nor children can display such defects in character that give rise to the reactions described earlier. Furthermore, I take this form of criticism, insofar as it is legitimate, to be grounded (at least partially) in the vicious person’s irrationality. For a fully rational person is not blameworthy.

166

Sebastian Schmidt

What could epistemic blame be like according to this general account of mental blame? We should grant that there are cases of epistemic blame, i.e. cases where a subject’s blameworthiness is grounded merely in the fact that they fail to believe what their evidence suffciently supports (i.e., cases analogous to the dust-speck case). Kauppinen (2018) proposes that epistemic blame is a form of distrust. To distrust someone epistemically is, in some way, more than just treating them as an unreliable source of information – as we would do with a computer that makes unreliable predictions. I might epistemically distrust someone insofar as I am not ready to provide them with information because I fear that the person will draw the false conclusions from the information. Consider Tom, who always irrationally suspects other people to conspire against him. If I tell Tom that two of his colleagues recently meet more often for lunch than usual, he might interpret this – irrationally – as evidence that they conspire against him. Tom is not trustworthy epistemically because he is paranoid. His epistemic irrationality is a manifestation of some vice – a defect in his epistemic character. Distrusting Tom is compatible with assuming that he could not have avoided the rational defect he suffers. Yet my distrust is grounded in this defect, and only in this defect. I do not want to provide Tom with certain information because I do not trust him that he will understand its implications. Even though legitimate distrust seems to presuppose that one’s epistemic irrationality is not just an occasional lapse, there is nothing more I need to know about Tom than that he regularly fails to properly base his beliefs on his evidence when it comes to certain topics in order to for it to be rational to distrust him epistemically. I can view my epistemic relationship to Tom as impaired in certain respects. Part of this relationship is that we regard each other as trustworthy sources of information, and that we know about each other that we draw the right conclusions from the information we provide each other with. We cannot view our relationship with an unreliable computer as impaired in this way, because we do not have such relationships with computers. I think Tom is a plausible example of someone who is epistemically blameworthy. We might not necessarily be blameworthy for failing to believe what our evidence supports (even if we attend to it). A person is not blameworthy for occasional lapses in their epistemic rationality. Similarly, people are not blameworthy for occasional lapses in their moral conduct (e.g., some lapses due to stress). However, as soon as there is some kind of pattern behind these cases of irrationality – if irrationality cannot be excused as an occasional lapse – we view our relationship to irrational people as impaired in specifc ways, e.g., by partly losing our epistemic trust. If irrationality adds up in such a pattern, it begins to matter for our relationship to the person, and consequently for our actions related to the person. Yet our blame is not grounded in the fact that the

Responsibility for Attitudes

167

person violated any practical requirement. It is grounded in the fact that the person violated (and might violate again) the standards of attitudinal rationality. The existence of purely epistemic blame is what epistemic instrumentalist accounts à la Steglich-Petersen fail to capture when they argue that the normative force of evidence depends on further aims that are promoted by believing what the evidence supports.38 Being irrational, which includes failing to believe what one’s evidence supports, is sometimes suffcient to make us blameworthy. In this sense, evidence has normative force on its own.

6.

The Diagnosis: Responsibility and Reasons Without Control

The problem of mental responsibility originally got a grip on us because we assumed that we are responsible only for what we control (cf. Section 3.1). We here assumed that the control at issue is voluntary control. According to my proposal, having (indirect) voluntary control over an attitude is not a necessary condition for blame if we have the broader concept of blame in mind and do not think about such responses as indignation, resentment, and guilt. Our mistake was to assume that being blameworthy for our irrationality would require that we are, in some way, in control of our attitudes. My diagnosis is that we made this assumption because we still thought of blame as involving the passionate emotions that are essential for moral blame, like indignation, etc. It was a mistake to think of blame only in this narrow way. Without justifcation, we took this narrow concept of blame out of the conceptual space where it is at home – the domain of voluntariness and control – and tried to force it to live in a hostile environment – the domain of reasons and rationality. Yet being subject to the will and under our control is one thing, while being part of the space of reasons is another.39 Our attitudes are within the space of reasons without always being under our control. Consequently, the way we blame one another for non-compliance with the norms of rationality must also be different from the way we blame one another for violating requirements that presuppose voluntary control: mental blame is not moral blame. My proposed solution is thus a hybrid account of mental responsibility: On the one hand, we are indirectly responsible for our attitudes, and this face of responsibility can make us ftting objects of passionate forms of moral blame for our attitudes. On the other hand, if our attitudes are well supported by object-given reasons, we are rational, and if they are poorly supported by object-given reasons, we are irrational. These kinds of evaluation do not presuppose control in any substantial sense, and yet they have specifc normative consequences for our relationships towards one another that can be rationally marked by forms of blame.

168

Sebastian Schmidt

Let us fnally return to the two intuitively plausible lines of thought mentioned in my introduction that pulled us towards two seemingly inconsistent accounts of mental responsibility (Section 1). We can now say that the frst line of thought is plausible if we assume a concept of responsibility that can be understood in terms of the passionate moral blaming-responses, while the second line of thought is plausible if we assume a concept of responsibility that allows for a broader understanding of blame in terms of the impairment of relationships. This solution allows us both to accept that we are, in the narrow sense, responsible for our attitudes as we are responsible for consequences of actions, and that we are, in the broad sense, responsible for anything that is within the space of reasons. Here is a fnal objection. Some authors claim that responsibility for attitudes can only be explained by reference to a form of direct, nonvoluntary control that we exercise in believing, desiring, feeling, or intending.40 I have not argued that we do not exercise such control. I even granted that attitudes are responsive to reasons. Why does this not imply that we exercise non-voluntary control in believing, desiring, feeling, or intending? I cannot, so the objection goes, just endorse the idea that being subject to the norms of rationality does not require control. I would have to argue separately that attitudes are not themselves exercises of control. However, I do not claim that attitudes are not exercises of control in some sense. Rather, I claim that we need not assume that they are exercises of control in order to make it intelligible how we can be directly responsible for our attitudes. In order to understand how there can be such a thing as direct responsibility to the norms of rationality, we merely need to understand the nature of responsibility for our attitudes in terms of the blaming-responses that are made appropriate by non-compliance with rational norms. Our intuition that responsibility requires control fades as soon as we have the nature of these responses clearly in front of us: They are appropriate merely in virtue of the fact that another person has a certain character, no matter how the character came about. We might have to consider concepts of non-voluntary control is in order to solve other problems. But we need not do so in the context of understanding mental responsibility.41,42

Notes 1. First, take one of the central questions of normative ethics: what are the rightmaking features of an action (e.g., its consequences, motives, or properties of the action itself)? Second, take the debate about free will and determinism, where we ask whether we are free with respect to and thus responsible for our actions in a way that is compatible with the causal structure of the universe. Thinking about our responsibility for and control over attitudes (instead of actions and their consequences) might well give us a clue as to

Responsibility for Attitudes

2.

3.

4. 5.

6. 7. 8. 9.

10.

11.

12.

169

how to solve this traditional problem (cf. Hieronymi ms; Wagner 2015). However, I will not be concerned with the connection of my claims to these traditional discussions here. I thus exclude other actions from the consequences of an action. Consequences in my sense are mere consequences which are caused by previous actions but which are not themselves directly controlled. It is important that I include attitudes in the class of potential consequences of actions. I do this by saying that attitudes are no things performed for reasons (although one might say that they are held, formed, or even done for reasons). I do not claim that we are responsible for all the consequences of our actions, of course. The point is rather that if we are responsible for consequences of our actions, then we are so in virtue of the fact that we are responsible for actions which caused them. Cf. Fischer/Tognazzini (2009) on tracing. See Schmidt (forthcoming) on why I prefer understanding rationality as reasons-responsiveness. Some refer to object-given reasons as “reasons of the right kind.” However, this label might be confusing because it has its origin in a specifc debate about ftting-attitude accounts of value (Gertken/ Kiesewetter 2017). It is disputable whether the “state-given”/“object-given” labels are much better. I assume here that, whatever the label, we can make sense of this distinction in some way. See Heuer (2018) for criticism about the distinction when applied to intentions. Some might prefer not to talk about the “rationality” of an emotion. If you do, then you can call the relevant normative category “appropriateness” or “fttingness.” The reason to drink the toxin in Kavka’s toxin puzzle (1983) is state-given in this sense. Kolodny (2005: 551) famously pointed out that facts about an attitude’s coherence would be state-given reasons for the attitude. This would be an example of a state-given reason that would be no practical reason. Pragmatists disagree among each other about when the value of a belief provides a practical reason for this belief. Moderate pragmatists, like McCormick (2015), argue that it is not a reason for this belief if all our evidence speaks against the existence of guarding angels. More radical pragmatists, like Rinard (2015, 2017), argue that every consideration indicating the value of an attitude is a practical reason for this attitude. Cf. esp. Shah (2006). This is what evidentialists claim. State-given considerations can of course be evidence and thus epistemic reasons to believe that a certain mental state M is valuable. But they are not thereby reasons for M but rather for the belief that M is valuable. Recent pragmatists dispute that believing at will is necessary for practical reasons for belief (McCormick 2015: 28–29; Reisner 2009: 268–270; Rinard 2015). My target in this paper is not pragmatism. Rather, my aim is to present the outlines of a viable account of mental responsibility without committing to pragmatism. For voices favourable of pragmatism, see the contributions by Crawford (Chapter 5), Steglich-Petersen/Skipper (Chapter 6), and Lord (Chapter 7). See also Coates’ contribution (Chapter 10) for a rejection of some worries that the fact that practical reasons are “the wrong kind of reasons” undermines the project of an ethics of mind. On how “the deontological conception of epistemic justifcation” might become questionable; cf. Alston (1988). Owens (2000) views epistemic normativity to be threatened by the absence of doxastic control. More recently, McHugh (2014) has presented the problem as a problem about the very existence of reasons for attitudes. See also Gaultier’s contribution in this volume (Chapter 4).

170

Sebastian Schmidt

13. For the Lara-case, cf. also Schmidt (2016). In line with this, Owens (2003) argues that our aims in forming beliefs cannot interact with our other aims we have as agents. Importantly, the infuential critique of Steglich-Petersen (2009) of Owens’ argument is not in confict with what I say here, for Steglich-Petersen claims that certain activities, which conceptually aim at forming a true belief (like inquiry or reasoning), can interact with our wider aims. A belief-formation in the sense I use this term here cannot do so, as Steglich-Petersen accepts. To accept that a belief-formation (in the non-intentional sense of the term) can interact with our wider aims is to commit to pragmatism – a position I cannot discuss in detail here, but which is one of the options of making mental normativity intelligible. Steglich-Petersen now sympathizes with pragmatism (cf. his contribution with Skipper in Chapter 6 of this volume). 14. Cf. esp. Kavka (1983). On how we exercise agency in deciding, cf. Soteriou’s contribution in this volume (Chapter 12). 15. I owe the idea of stating philosophical problems in such a way to Gerhard Ernst, cf. esp. Ernst (2008: 65–71). 16. Accounts that are affected by the latter problem include Chrisman (2008, 2018, Chapter 3 in this volume), Meylan (2013, 2017), and Peels (2017). I do not claim that I have argued here that the problem cannot be met. Rather, given the intuitive signifcance of the norms of rationality, these accounts face an intuitive problem. 17. For an account along these lines, cf. Rinard (2015, 2017). More moderate versions of pragmatism, like the one of McCormick (2015) or Reisner (2008, 2009), argue that the rationality of belief is at least sometimes a matter of responding correctly to practical reasons for belief – i.e., reasons that show that state of believing to be valuable in some way (either prudentially or morally). These more moderate versions could still grant that we are sometimes blameworthy merely for being epistemically irrational. However, they would then not avoid the assumption that we are directly responsible for being epistemically (ir)rational. Only radical pragmatists (like Rinard) and those who deny any form of direct responsibility (like Meylan and Peels) for attitudes can straightforwardly avoid this assumption. 18. Saying that certain norms are constitutive for attitudes also means that attitudes can be individuated by reference to the norms by which they are governed. Anscombe (1957, Section 2) provides such an individuation with her “direction of ft” distinction between beliefs and desires: “beliefs ought to ft the world,” but this is not true for desires. Rather, we “ought to ft the world to our desires.” The Ought in question here is the Ought of rationality – if one realizes that it is false that p, then it is not rational to, instead of giving up the belief, continue believing it and to rather adjust the world to one’s belief. Also the debate about belief’s “aim of truth,” (Chan 2013) – an aim that must be different from the aim of, for example, intending, but also different from the aim of guessing or suspecting, which arguably also aim at coming to a true guess or a true suspicion in some sense (Owens 2003: 290) – is concerned with the constitutive norms of belief. 19. I take the label from Kiesewetter (ms) who mainly discusses (and rejects) the weaker claim that reasons for belief are not normative reasons. On recent sympathies with nihilism or the weaker claim, see Papineau (2013), Rinard (2015), Glüer/Wikforss (2018), Maguire/Woods (forthcoming, ms), and Mantel (2019); for a critical discussion, see Paakkunainen (2018). Maguire (2018) argues that there are no reasons for affective attitudes. 20. I choose this label because, as I mentioned in Section 3.3, I take the norms of rationality, which I understand as norms to comply with one’s object-given reasons, to be constitutive norms. See also note 18 earlier.

Responsibility for Attitudes

171

21. One might doubt that I do not believe the disjunctive proposition. For if I was asked whether I believe it, and I understand the content of the proposition, I will reply that I do believe it. However, even if one thinks that we believe all those disjunctive propositions, one will agree that we do not believe all the implications of our current belief-stock, like certain mathematical or logical implications that are just too hard to fgure out. That we do not believe those implications does not make us blameworthy or criticizable in any sense. Furthermore, we can imagine a case where I fail to believe such weird disjunctive propositions. Why on earth, we might ask, should any reasonable person care about this so much as to regard me as blameworthy for not believing them? 22. In a recent paper, Kiesewetter (ms) acknowledges that there is another, probably more diffcult, problem for the normativity of epistemic rationality than Harman’s clutter-objection. This further problem arises from the question whether we have a reason to care about being rational. While Harman’s clutter-objection can be met, according to Kiesewetter, with the attendingcondition for requirements of theoretical rationality, this condition cannot explain why we have a reason to care about being epistemically rational. Steglich-Petersen’s dust-speck case, which I present later, might be understood as exactly raising this question of why we should care about being rational. Kiesewetter’s solution to this new problem about reasons to care builds on Wedgwood’s (2017) idea that rationality is a virtue. Since we always have a reason to care about being virtuous, we always have a reason to care about being rational. This fts well with my account of why and how we are blameworthy for failing to comply with our object-given reasons (cf. Section 5). 23. I here have in mind Scanlon’s (2008) concept of blame, to which I return in Section 5. 24. It would thus not be appropriate to reduce epistemic trust in the person. Reducing epistemic trust is the essence of epistemic blame, according to Antti Kauppinen (2018). I return to this idea at the end of Section 5. 25. One could argue that S is blameworthy from a practical perspective if one thinks that believing what is not suffciently supported by one’s evidence is always morally blameworthy (Clifford 1877). However, this is not only highly questionable (Schmidt 2017; Lindner in this volume, Chapter 2), it also does not avoid that non-compliance to epistemic reasons does not make you blameworthy. Rather, it would be your non-compliance with moral reasons that makes you blameworthy. 26. A lesson one might draw from Harman’s clutter-objection is epistemic permissivism: we are permitted to draw conclusions from our belief-stock as soon as they are suffciently supported by our evidence, but never required to do so. (More generally, permissivism states that our total set of evidence permits more than one set of doxastic attitudes to take towards each (or at least some) proposition(s)). The denial of permissivism is often discussed as the Uniqueness Thesis (as introduced by Feldman 2007). On epistemic permissivism and some of its problems, cf. White (2005). If we are permitted to believe only what is suffciently supported by our evidence, then this implies the following requirement: (ER*) One ought not [to believe what is not suffciently supported by one’s evidence]. The normative force of this “ought” is as mysterious as the normative force of the “ought” in (ER) when it comes to trivial propositions. 27. I here follow the terminology employed by Kiesewetter (ms). Kiesewetter distinguishes between anti-normativism, which claims that object-given reasons are no normative reasons for attitudes, and nihilism, which claims that object-given reasons are no reasons for attitudes at all. The distinction does not matter for my purposes. I think that denying the normative force of

172

28.

29. 30. 31. 32.

33. 34.

35.

36. 37. 38. 39. 40. 41. 42.

Sebastian Schmidt reasons is to deny their status as reasons. I agree with Kiesewetter that there is no plausible account of object-given reasons as non-normative reasons for attitudes. The reading in parentheses is necessary for the nihilist or the (radical) pragmatist when they want to argue that evidence does not provide us with reasons to believe. For according to them, evidence alone never decisively supports a proposition. It merely seems as if it does. Two such recent accounts of epistemic blame which I think are clearly motivated in this way are Brown (2018) and Kauppinen (2018). On objectivism and subjectivism, cf. the discussions in Kiesewetter (2017, ch. 8) and Lord (2018, ch. 8). I spell out the case for nihilism in more detail in part II of Schmidt (ms). As a response to the problem of mental responsibility, pragmatism can most straightforwardly be understood as presenting an alternative account of the rationality of attitudes (in terms of practical reasons rather than object-given reasons), thus endorsing (1) and, most commonly, also (2), but denying (3). However, pragmatists are not per se committed to either (2) or (3). My point is merely that pragmatist accounts of reasons for attitudes could motivate, or make more plausible, specifc solutions to the problem by being accounts about what kind of norms we are supposed to be held accountable to. This is motivated by Wedgwood’s (2017) account of rationality as a virtue as well as Kiesewetter’s (ms) proposal to make our reasons to care about rationality intelligible by acknowledging rationality as a virtue. It is important to note that I do not argue for some kind of elitism of those who are rational. Saying that relationships with the irrational are impaired does not imply that we are never allowed to enter such relationships, or that we should not enter into other types of relationships with people who are irrational. Arguably, we are all irrational in certain ways. Rather, my claim is much weaker: That someone is irrational is a pro tanto reason not to enter certain types of relationship with the person. The types of relationship will depend on the kind of irrationality at issue. Wallace defends this narrow concept of blame when he argues that “it would indeed be strange to suppose that one might blame another person without feeling an attitude of indignation or resentment toward the person, or that one might blame oneself without feeling guilt” (1994: 75). See also Wallace (2011). Here Smith follows especially Scanlon (2008). The kind of effect such vices can have on our interpersonal relationships is nicely captured by Scanlon’s (2008) account of blame, which is taken up and refned by Hieronymi (2004, 2019) and Smith (2013). See esp. Section 4.1 earlier. Next to Steglich-Petersen (2011), cf. SteglichPetersen/Skipper (this volume, Chapter 6). This is the central point of Owens (2000). Cf. esp. Boyle (2011), Hieronymi (2006, 2008, 2014, ms), and Smith (2005). For some objections against these views, cf. Chrisman (this volume, Chapter 3, Section 5). One such context might be the problem of understanding how there can be agency in a world that is dominated by natural law – i.e., the traditional problem of free will (cf. Hieronymi ms). I am grateful to Dorothee Bleisch, Matthew Chrisman, Gerhard Ernst, Pamela Hieronymi, Christian Kietzmann, Martina Lindner, Steffen Lesle, and Konstantin Weber for comments on previous versions of this paper.

Responsibility for Attitudes

173

References Alston, William P. (1988): “The Deontological Conception of Epistemic Justifcation”, Philosophical Perspectives 2, 257–299. Anscombe, Elizabeth (1957): Intention, Oxford: Blackwell. Boyle, Matthew (2011): “‘Making Up Your Mind’ and the Activity of Reason”, Philosopher’s Imprint 11, 1–24. Broome, John (2013): Rationality Through Reasoning, New York: Oxford University Press. Brown, Jessica (2018): “What Is Epistemic Blame?”, Noûs, 1–19, DOI: 10.1111/ nous.12270. Chan, Timothy (ed.) (2013): The Aim of Belief, New York: Oxford University Press. Chrisman, Matthew (2008): “Ought to Believe”, The Journal of Philosophy 105, 346–370. Chrisman, Matthew (2018): “Epistemic Normativity and Cognitive Agency”, Noûs 52, 508–529. Clifford, William K. (1877): “The Ethics of Belief”, in his The Ethics of Belief: And Other Essays, New York: Prometheus Books 1999, 70–96. Ernst, Gerhard (2008): Die Objektivität der Moral, Paderborn: Mentis. Feldman, Richard (2007): “Reasonable Religious Disagreements”, in: Philosophers without God: Meditations on Atheism and the Secular Life, ed. by Antony, New York: Oxford University Press, 194–214. Fischer, John M./Tognazzini, Neal A. (2009): “The Truth about Tracing”, Noûs 43, 531–556. Gertken, Jan/Kiesewetter, Benjamin (2017): “The Right and the Wrong Kind of Reasons”, Philosophy Compass 12, 1–14. Glüer, Kathrin/Wikforss, Åsa (2018): “Reasons for Belief and Normativity”, in: The Oxford Handbook of Reasons and Normativity, ed. by Star, New York: Oxford University Press, 575–599. Harman, Gilbert (1986): Change in View: Principles of Reasoning, Cambridge, MA: MIT Press. Heuer, Ulrike (2018): “The Relevance of the Wrong Kind of Reasons”, in: Normativity: Epistemic and Practical, ed. by McHugh/Way, New York: Oxford University Press. Hieronymi, Pamela (2004): “The Force and Fairness of Blame”, Philosophical Perspectives 18, 114–148. Hieronymi, Pamela (2006): “Controlling Attitudes”, Pacifc Philosophical Quarterly 87, 45–74. Hieronymi, Pamela (2008): “Responsibility for Believing”, Synthese 161, 357–373. Hieronymi, Pamela (2014): “Refection and Responsibility”, Philosophy and Public Affairs 42, 3–41. Hieronymi, Pamela (2019): “I’ll Bet You Think This Blame Is about You”, in: Oxford Studies in Agency and Responsibility V: Themes from the Philosophy of Gary Watson, ed. by Coates/Tognazzini, New York: Oxford University Press, 60–87. Hieronymi, Pamela (ms): Minds That Matter, unpublished manuscript, spring 2019 version.

174

Sebastian Schmidt

Kauppinen, Antti (2018): “Epistemic Norms and Epistemic Accountability”, Philosopher’s Imprint 18, 1–16. Kavka, Gregory S. (1983): “The Toxin Puzzle”, Analysis 43, 33–36. Kiesewetter, Benjamin (2017): The Normativity of Rationality, New York: Oxford University Press. Kiesewetter, Benjamin (ms): “Are Epistemic Reasons Normative?”, unpublished manuscript, July 2019 version. Kolodny, Niko (2005): “Why Be Rational?”, Mind 114, 509–63. Lord, Errol (2018): The Importance of Being Rational, New York: Oxford University Press. Maguire, Barry (2018): “There Are No Reasons for Affective Attitudes”, Mind 127, 779–805. Maguire, Barry/Woods, Jack (forthcoming): “The Game of Belief”, The Philosophical Review. Maguire, Barry/Woods, Jack (ms): “All Genuine Reasons for Belief Are the Wrong Kind of Reasons for Belief”, unpublished manuscript, April 2018 version. Mantel, Susanne (2019): “Do Epistemic Reasons Bear on the Ought Simpliciter?”, Philosophical Issues, 214–227. McCormick, Miriam S. (2015): Believing against the Evidence: Agency and the Ethics of Belief, London: Routledge. McHugh, Conor (2014): “Attitudinal Control”, Synthese, 1–18. Meylan, Anne (2013): Foundations of an Ethics of Belief, Frankfurt: Ontos. Meylan, Anne (2017): “The Consequential Conception of Doxastic Responsibility”, Theoria 83, 4–28. Moran, Richard (2001): Authority and Estrangement. An Essay on Self-Knowledge, Princeton, NJ: Princeton University Press. Owens, David J. (2000): Reason without Freedom: The Problem of Epistemic Normativity, London: Routledge. Owens, David J. (2003): “Does Belief Have an Aim?”, Philosophical Studies 115, 275–297. Paakkunainen, Hille (2018): “Doubts about ‘Genuinely Normative’ Epistemic Reasons”, in: Metaepistemology, ed. by McHugh, et al., New York: Oxford University Press, 122–140. Papineau, David (2013): “There Are No Norms of Belief”, in: The Aim of Belief, ed. by Chan, New York: Oxford University Press, 64–79. Peels, Rik (2017): Responsible Belief, New York: Oxford University Press. Reisner, Andrew (2008): “Weighing Pragmatic and Evidential Reasons for Belief”, Philosophical Studies 138, 17–27. Reisner, Andrew (2009): “The Possibility of Pragmatic Reasons for Belief and the Wrong Kind of Reasons Problem”, Philosophical Studies 145, 257–272. Rinard, Susanna (2015): “Against the New Evidentialists”, Philosophical Issues 25, 208–223. Rinard, Susanna (2017): “No Exception for Belief”, Philosophy and Phenomenological Research 94, 121–143. Scanlon, Thomas (2008): Moral Dimensions: Permissibility, Meaning, Blame, Cambridge, MA: Belknap Press. Schmidt, Sebastian (2016): “Können wir uns entscheiden, etwas zu glauben? Zur Möglichkeit und Unmöglichkeit eines doxastischen Willens”, Grazer Philosophische Studien 93, 571–582.

Responsibility for Attitudes

175

Schmidt, Sebastian (2017): “Why We Should Promote Irrationality”, Grazer Philosophische Studien 94, 605–615. Schmidt, Sebastian (forthcoming): “Rationality and Responsibility”, Australasian Philosophical Review 5. Schmidt, Sebastian (ms): The Problem of Mental Responsibility: Outlines of an Ethics of Mind, unpublished PhD-thesis at Friedrich-Alexander-University of Erlangen-Nuremberg. Sextus Empiricus (1933): Outlines of Pyrrhonism, Cambridge, MA: Harvard University Press. Shah, Nishi (2006): “A New Argument for Evidentialism”, The Philosophical Quarterly 56, 481–498. Smith, Angela M. (2005): “Responsibility for Attitudes: Activity and Passivity in Mental Life”, Ethics 115, 236–271. Smith, Angela M. (2013): “Moral Blame as Moral Protest”, in: Blame, ed. by Coates/Tognazzini, New York: Oxford University Press, 27–48. Steglich-Petersen, Asbjørn (2009): “Weighing the Aim of Belief”, Philosophical Studies 145, 395–405. Steglich-Petersen, Asbjørn (2011): “How to Be a Teleologist about Epistemic Reasons”, in: Reasons for Belief, ed. by Reisner/Steglich-Petersen, New York: Oxford University Press, 13–33. Strawson, Galen (2003): “Mental Ballistics or the Involuntariness of Spontaneity”, Proceedings of the Aristotelian Society 103, 227–256. Wagner, Verena (2015): “On the Analogy of Free Will and Free Belief”, Synthese 194, 2785–2810. Wallace, R. Jay (1994): Responsibility and the Moral Sentiments, Cambridge, MA: Harvard University Press. Wallace, R. Jay (2011): “Dispassionate Opprobrium: On Blame and the Reactive Sentiments”, in: Reason and Recognition: Essays on the Philosophy of T. M. Scanlon, ed. by Wallace et al., New York: Oxford University Press. Wedgwood, Ralph (2017): The Value of Rationality, New York: Oxford University Press. White, Roger (2005): “Epistemic Permissiveness”, Philosophical Perspectives 19, 445–459.

9

Two Kinds of Rationality Gerhard Ernst

1.

Introductory Remark

In this paper I will try to illuminate the nature of rationality. I will do this by distinguishing and explaining two fundamentally different kinds of rationality. My distinction is related to but not identical with a distinction between two attempts to clarify the nature of rationality which have been widely discussed in the literature: the attempt to understand rationality as consistency (broadly understood),1 and the attempt to understand rationality as correctly responding to reasons.2 I will present a framework in which the strengths and the limitations of both ideas become apparent. In order to be able to do this within the confnes of a short paper, I have to draw on many ideas I can only sketch and cannot argue for in detail here. I hope that in giving a “perspicuous representation” (PI §122), as Wittgenstein calls it, of these ideas, i.e. by integrating them into a coherent whole, they become plausible enough for present purposes. The coherence of the whole picture has to speak for the viability of the parts.

2.

Rational Belief-Formation: Consistency

Our rationality manifests itself (mainly, but not only) in what we believe, feel, and do. Let us start with the rationality of our beliefs. Our ability to form beliefs is part of what I would like to call our rational response system. This system enables us to respond in a rational way to our environment. The entry channels to this system are our senses. We see, hear, smell, feel and taste our environment, and we thereby come to have beliefs about it. This is mainly a causal process. When I look around in the seminar room, I come to believe that I see that there are other people present. If I fail to form this belief under these circumstances, I am blind (or blind-sighted). When is it rational to form perceptual beliefs? I think the answer is: it is rational for me to form a perceptual belief unless there are other (suffciently strong) beliefs I have which are inconsistent with this perceptual belief. For instance, if it looks to me as if I see Angela in

Two Kinds of Rationality

177

Oxford, it would be irrational for me to come (without double-checking) to believe that I see her if I have a very strong belief that she is in London. On the other hand, I don’t need any reasons supporting my perceptual beliefs in order to be rational in forming them.3 It is rather enough for our rational perceptual belief-formation if we don’t have any reason to question the belief. When it comes to perception, rational belief-formation has a default and challenge structure: whenever I form a perceptual belief, I am rational until proven irrational by other beliefs I already have.4 So, what is the job of our rational response system here? Perception causes new beliefs, and new beliefs are then checked by our rational response system internally against the beliefs already present. This check is an internal consistency check. It is a consistency check because the question at issue is: is there anything I believe which contradicts or is in tension with the belief I have newly acquired? I add the clause “or is in tension with” in order to accommodate degrees of belief. For example, if I am pretty sure that Angela is in London, and I am pretty sure that I see her in Oxford, these two beliefs, although not contradicting each other in a strict sense, are in tension with each other. One could elaborate on this point modelling degrees of belief with probabilities (cf. Howson 2000). If there is contradiction or tension, a rational person adjusts her beliefs or her degrees of belief so that contradictions and tensions are avoided. Which beliefs have to be adjusted in what ways is guided by the strength of the beliefs we already have, and, in addition, by beliefs about proper belief-formation.5 The purpose of the consistency check is to ensure rationality. To be irrational here means to have ill-adjusted beliefs, i.e. contradictory beliefs or beliefs which are in tension with each other. The purpose of our rational response system, in this case, is to ensure that our representation of reality is as accurate as possible. Contradictory beliefs cannot be true at the same time (and someone with beliefs that are in tension with each other would accept a Dutch book6). The consistency check is only concerned with my beliefs. It is a totally internal affair, in accordance with the view that rationality supervenes on the mind.7 It is useful to distinguish between a person who is perfectly rational and a person who is rational enough or rational in a more common or garden sense. A perfectly rational person would be a person with a perfect rational response system. And a perfect rational response system would be a system which checked all new beliefs against all old beliefs. Any inconsistency or tension would be detected and removed. Nobody is a perfectly rational person in this sense, and when we say that someone is irrational we don’t mean that the person is not perfectly rational. We rather mean that the person is not rational enough, and a person is not rational enough when she fails to check for more or less obvious inconsistencies and tensions between her beliefs. This is, obviously, not a sharp characterization, since there is no sharp boundary between what is obvious and what is not. In addition, one might ask: Is a person

178 Gerhard Ernst not rational enough when she fails to avoid inconsistencies and tensions which are obvious to her or obvious in a more objective sense? The frst option is more in keeping with the view that rationality supervenes on the mind. But, I think, we often ascribe irrationality on the basis of the second option. For present purposes, we don’t need to resolve this issue. A – roughly formulated – frst principle of rationality might look like this: If someone believes that p and believes that q and if p and q are obviously inconsistent or in tension with each other, the person is irrational.

3.

Rationality as Mental Unity

One might wonder whether irrationality in this sense is even possible: how can someone believe that p and believe that q if it is obvious (to her) that p and q are inconsistent? What does it mean to say of someone that she believes that p, believes that q, and believes that p and q are inconsistent? In order to sketch an answer to this question I have to say a few words about belief. I don’t want to enter the general debate about what beliefs are by their very nature. And I don’t have to, because for present purposes it is enough to highlight one feature of beliefs which is, I think, quite uncontroversial: beliefs are linked to certain dispositions. If a person A has the belief that p she is – at least in favourable circumstances – disposed to ascribe the belief that p to herself, she is disposed to base inferences on the belief that p, she is disposed to act in certain ways and to refrain from certain actions, and so on. If Peter, for example, believes that orange juice contains alcohol, he will – if there is no reason for lying – endorse the belief when asked about the matter, he will reason that excessive consumption of orange juice is bad for the liver, and he won’t drink orange juice and drive. All of these dispositions, of course, are only to be expected in the presence of further beliefs of Peter: the belief that one is supposed to tell the truth, the belief that alcohol is bad for the liver, the belief that it is wrong to drive a car when being drunk, etc. Furthermore, it is possible that none of these dispositions ever results in occurrent behaviour, simply because the circumstances are never favourable. It is not very probable that this actually happens since among the relevant dispositions are not only dispositions to observable behaviour but also dispositions to “inner actions” (like reasoning, thinking, etc.). The observable behaviour is our basis for ascribing beliefs. In particular, the relevant avowals of the person herself immediately provide reasons for ascribing beliefs. If Peter himself utters the sentence “I believe that orange juice contains alcohol” then we have good reason to assume that he really believes in the alcoholic nature of orange juice, at least if there is

Two Kinds of Rationality

179

no evidence suggesting that Peter is insincere. Of course, ascribing beliefs is also possible on the basis of other kinds of overt behaviour. When Peter, who drinks nothing but orange juice day in, day out, joins Alcoholics Anonymous we have reason to ascribe the relevant belief as well. The actualizations of the relevant dispositions have – in Wittgenstein’s terminology (PI §354) – the status of criteria for the presence of the belief. A criterion must not be confused with a necessary (or a suffcient) condition – the belief might be there without the criterion being met (and vice versa). But by necessity, the criteria constitute reasons for ascribing the belief, albeit defeasible ones.8 You only know what a belief is if you know that the sincere self-avowal of a belief constitutes a reason for ascribing the belief. This claim does not imply what traditional dispositionalists (like Braithwaite and maybe Ryle) maintain, namely that beliefs are nothing but certain dispositions – although it does not preclude it either. In my opinion, this form of reductionism is problematic, but I don’t want to enter that debate here. My point is simply that there is a (conceptual) connection between beliefs and certain dispositions, and that granting that connection is compatible with any theory of beliefs one might want to adhere to. The reason why we have leeway here is that no claim about explanatory priority has been made. Theories of belief diverge on the question whether beliefs have to be explained in terms of dispositions or whether having certain dispositions is only the result of having certain beliefs. That there is a connection between dispositions and beliefs is, I take it, uncontested. On the basis of these remarks we can understand the possibility of the kind of irrationality under consideration right now in the following manner: if p and q are obviously inconsistent, the fact that someone honestly claims that p, for example, is a reason to believe that the person does not believe that q. That the person honestly claims that p is a criterion for not ascribing to her the belief that q if there is an obvious inconsistency between the belief that p and the belief that q. Nevertheless, the person can believe that q if she satisfes enough other criteria for ascribing the belief that q. The irrationality, then, consists in a tension between criteria for ascribing beliefs. Someone is irrational in the present sense if she satisfes some criteria for ascribing a belief while not satisfying others. Does the person who is irrational in this way really believe that q? It depends. If she satisfes almost all criteria for ascribing the belief and only fails to satisfy one we would say that the person has the belief (even though there is a reason to doubt it). If someone satisfes some criteria and fails to satisfy some, we are unsure what to say. In any case, the person is in a somewhat “disunifed” state of mind. We cannot, at the same time, ascribe the full-blown belief that p, the full-blown belief that q, and the full-blown belief that p and q are inconsistent. But we can, for example, say that the person more or less believes that p, that she more or less sees that p and q are inconsistent, and that, nevertheless, she seems to

180

Gerhard Ernst

believe that q. Her state of mind is somehow “mixed” or “disunifed.” And the irrationality of the person precisely consists in this “disunifcation.” Another way of putting it would be: irrationality in the present sense is a tension within beliefs.

4.

Rational Belief-Formation: Implications

When we acquire new beliefs through perception our rational response system checks for inconsistencies and tensions. But it does more than that: it checks for implications. When I acquire the belief that there are people present, I also acquire the belief that there are human beings present, since it is obvious (to me and to everybody) that people are human beings. One might even say: part of my acquiring the belief that there are people present just is acquiring the belief that there are human beings present. Again, if I fail to meet criteria for the belief that there are human beings present, this is a reason to think that I haven’t really acquired the belief that there are people present in the frst place. And insofar as it is possible to acquire the latter belief without acquiring the former (i.e. insofar as there is a tension between the relevant criteria) this is irrational. Again, we could distinguish between a perfectly rational person and a person who simply is rational enough. A perfectly rational person would, perhaps, believe all the implications of what she believes. If she acquired a new belief, she would simultaneously acquire all the implicated beliefs as well. That’s not how we use the word “rational” though. A person who is rational enough or rational in the usual sense would acquire only the obviously implicated beliefs. The following principle captures this: If someone believes that p and fails to believe that q even though it is obvious that p implies q, the person is irrational.9 Do we have to add a clause like: the person cares for whether q (cf. Broome 2013: 157) or that the person “takes it to be relevant or at least attends to the question of whether q is the case” (cf. Kiesewetter 2017: 15) in order to take care of the fact that we are certainly not rationally required to believe all the implications of what we believe? No, I don’t think we need a caring- or attention-condition to do justice to this fact. The principle only states that we are rationally required to believe what obviously follows from what we believe. And as soon as it is not obvious what follows from what we believe we are not rationally required to believe what follows from what we believe. For instance, from the belief that there are 23 pupils in each class and 31 classes in the school, it follows that there are 713 pupils in the school. But since this is not obvious, I am not rationally required to believe this even if I have the frst two beliefs. If, on the other hand, I do think that it follows from these two beliefs that there are 713 pupils in the school I would be irrational if I

Two Kinds of Rationality

181

didn’t believe that this is the case (since then it would be obvious for me). Rationality requires us to believe all the obvious implications of what we believe. The rational response system of someone who does not believe the obvious implications of what she believes isn’t working properly. The kind of irrationality characterized here is of the very same kind as the one characterized in the frst principle. Again, a person is irrational if she satisfes some criteria for ascribing a belief while not satisfying others. Again, irrationality consists in a tension between criteria. Again, irrationality can be seen as a tension within belief. We have already seen that talking about “obviousness” is not very precise. But the second principle faces an additional problem: it might be obvious that p implies q and that q implies r even though it is not obvious that p implies r. Hence, our distinction between a perfectly rational person and a person who is simply rational enough is in danger of collapsing since, presumably, all the implications of what we believe can be deduced using only obviously valid inferences. Hence, even the person who is only rational enough needs to believe all the implications of what she believes, i.e. she needs to be perfectly rational. And, of course, the perfectly rational person is also rational enough. Therefore, the distinction would collapse. The problem can be solved, I think, by pointing to the fact that what is obvious in one context may not be obvious in another. When I acquire the belief that p, it might be obvious that p implies q (and so I acquire the belief that q) but it might not be obvious for me at this time that q implies r even though that becomes obvious as soon as I focus my attention on q (and perhaps on r). When a person who is rational enough acquires the belief that p, she acquires all the beliefs obviously implied by p while her attention remains fxed on p. And there is only a limited amount of implications which are obvious for someone whose attention is fxed on one particular belief. Hence, it is possible to be rational enough without being perfectly rational.

5. Adjusting vs. Evaluating The frst function of our rational response system is adjusting. Up till now I have described two kinds of adjustment: (a) adjustment of new beliefs to old ones, and (b) adjustment of more new beliefs to new beliefs. In both cases, the purpose of the adjusting is keeping us rational by (a) avoiding inconsistencies and tensions between new and old beliefs, and by (b) avoiding tensions between new beliefs and the absence of other new beliefs. Both kinds of tension, if they occur, are of the same kind: in both cases, criteria then are met for not ascribing the supposedly newly acquired beliefs. Or, as I said: there is then a tension between the criteria for ascribing beliefs in both cases – and this is what the irrationality consists in.

182

Gerhard Ernst

The rational response system has a second function, though, apart from adjusting: evaluating. As soon as we have formed beliefs about the world, we evaluate what we believe to be the case. We classify what we believe to be the case as good, bad, and neutral (and more or less so). This classifcation together with the same classifcation of possible states of affair are the basis for how we react to what we believe. If, for example, I come to believe that my foot hurts because I have set some heavy object on top of it, I immediately evaluate this state of affairs as bad. We could say: frst of all, I respond to the belief that my foot hurts because of a heavy object on top of it with an evaluative belief, namely the belief that it is bad that my foot hurts because of a heavy object on top of it. That it is bad that my foot hurts because of a heavy object on top of it means that I have reason to change this state of affair if I can. This is an obvious implication, so by adjusting my further beliefs to my belief that it is bad that my foot hurts because of a heavy object on top of it, I come to believe that I have reason to change this state of affairs if I can. Let us assume, I believe that I can change this state of affairs. Hence, I come to believe that I have reason to change it. Let us further assume that I don’t have any other beliefs concerning relevant reasons. Hence, I come to the belief that I have conclusive reason to remove the heavy object.

6.

More Adjusting

I will presently return to the evaluation-function of our rational response system, but let me explain a few further adjustment-functions frst. As soon as I come to believe that I have conclusive reason to do something, a new kind of adjustment takes place if I am rational: I adjust my intentions to my beliefs about reasons. If I believe that I have conclusive reason to do X and do not intend to do X, my intention is in tension with my belief about reasons and my evaluative belief. Again, not satisfying the relevant criteria for having the intention is a reason to think that I really don’t have the belief (and vice versa). If I have the belief and I don’t have the intention there is a confict between criteria of ascribing the belief and ascribing the absence of the intention. The corresponding principle of rationality, roughly, is this: If someone believes that she has conclusive reason to do X and does not intend to do X, the person is irrational.10 Again, rationality is a matter of adjustment, in this case adjustment between evaluative belief and intention, and again this is a matter of coherence between criteria for ascribing evaluative belief and intention. There is another, more direct link between evaluative belief and action: If someone believes that she has conclusive reason to do X and does not do X, then either there are external obstacles for her to do X or she is irrational.

Two Kinds of Rationality

183

The same thoughts apply to this principle as to the other ones mentioned so far. If I don’t do X even though I could, i.e. even though there is no external obstacle, this is reason to think that I don’t believe that I have conclusive reason to do X. If it turns out that I do have this belief (because many criteria for ascribing the belief are met), I am irrational (weak-willed in this case) because there is a tension between criteria, and, therefore, a tension “within” my belief. Let me add a further element to the picture: emotions. I said that our rational response system adjusts beliefs to further beliefs which are obviously implied. It also adjusts our emotions to our evaluative and nonevaluative beliefs. If I believe that Thomas has insulted me, I believe this is bad, and I become angry. If I would become frightened instead, my emotion would be ill-adjusted to my beliefs. And in this sense we could say that it would be irrational for me to become frightened. Rational emotions are emotions which are adjusted to our evaluative and nonevaluative beliefs. If someone is not feeling the appropriate emotion this is a reason to think that she does not have the relevant evaluative and/or non-evaluative beliefs. If I am frightened that is a reason to think that I believe something dangerous is happening. If I don’t have this belief, my emotion is irrational. A corresponding principle of rationality might look like this: If someone feels frightened and does not believe that there is some danger, the person is irrational. There are, of course, many principles of this kind, since there are many different emotions. Our rational response system keeps our emotions in order: if all of our beliefs are adjusted to each other and all our emotions are adjusted to our beliefs, all our emotions are adjusted to each other as well. (Again we could distinguish between a perfectly rational person and one who is just rational enough.) There is even more adjusting going on in our rational response system: beliefs have to be adjusted to each other, intentions and emotions have to be adjusted to beliefs, but also intentions have to be adjusted to each other (and to beliefs). An important principle would be something like this: If someone intends to do X and believes that in order to do X she needs to do Y but does not intend to do Y, the person is irrational.11 In this case the intentions of the person and her means-ends-belief are ill-adjusted. We could add still further elements to the picture: hopes and expectations, for example, but I don’t want to expand on this here. Let me just emphasize that adjustment is the frst basis for ascribing rationality to a person: a person is perfectly rational in this sense if all her beliefs,

184

Gerhard Ernst

intentions, emotions, etc. are well adjusted to each other. In fact, I think often there is no hard and fast boundary between belief, intention, emotion, and other states of mind. Our mental life is, at least in a rational person, pretty much unifed. Emotion might often be seen as an aspect of evaluative belief and of intention. And intention might be seen as an aspect of evaluative belief and emotion. Evaluative beliefs come emotionally favoured as do intentions. Irrationality in the relevant sense, then, can be seen as a disruption of the unity of our mental life. This kind of irrationality consists in some criteria for ascribing a belief, emotion, or intention being met while others aren’t. It can be seen as a disruption within beliefs, emotions, and intentions as well as between different states of mind.

7.

Rationality as Responding With Evaluative Beliefs

But, as I said, the rational response system has two functions: adjusting and evaluating. And the second function is, in my opinion, of a completely different nature than the frst. Responding to some non-evaluative belief with some evaluative belief is not a matter of adjusting (or is a matter of adjusting in very different sense12). If someone fails to form an appropriate evaluative belief, we do not consider him as somehow inconsistent or in tension with himself. We rather consider him to be evaluatively blind, as it were.13 To give an example: if I believe that what I do to you hurts you but I do not believe that this is bad, and, hence, I do not believe that I have reason to stop doing what I do, and, hence do not intend to stop doing what I do – there is no tension between my beliefs or my beliefs and my intentions. I am just blind to the fact that it is bad to hurt you. We could also say: I am blind to the fact that the fact that what I do to you hurts you is a reason for me to stop. Hence, I am blind to the fact that something, which I believe is the case, is (if it is the case) a reason. Therefore, we could say I fail to respond adequately – with an evaluative belief – to a non-evaluative belief about something which is (if it is the case) a reason. And since not responding adequately to a reason (or an apparent reason if my non-evaluative belief turns out to be false) can be called a form of irrationality too, I am, in a sense, irrational. But this is a completely different kind of irrationality from the one considered earlier, namely the kind of irrationality which results from or consists in inner tension. It is worth noting that we respond to some (apparent) reason for action when we form an evaluative belief. The fact that what I do to you hurts you is a reason to stop doing what I do. It is not a reason for the belief that it is bad that what I do to you hurts you!14 In coming to believe that it is bad that what I do to you hurts you, I am not responding to a reason for (this) belief. I rather come to see something as a practical reason, namely the fact that what I do to you hurts you as a reason to

Two Kinds of Rationality

185

stop doing what I do. Are all evaluative beliefs connected to beliefs about reasons for action? And is, therefore, a failure to have an evaluative belief always a failure to respond to a practical reason and, therefore, a defect in rationality? I think, the answer is yes, if we think of potential reasons in a certain sense as well as of actual reasons. If something is good, I have reason to bring it about or preserve it, if I can; if something is bad, I have reason to change it, if I can. A failure to have the relevant evaluative belief is a failure to respond to a reason or a potential reason in this sense.15 The prime purpose of our evaluations is to prepare us for action. And rational action is guided by the reasons for action we discern in our evaluative beliefs. The picture that emerges is this: our rational response system is working well if it keeps our beliefs, intentions, emotions, etc. well adjusted, and if it produces evaluative beliefs which are appropriate on the basis of our non-evaluative beliefs. We are not perfectly rational in one sense, the adjustment-sense, if there is any kind of ill-adjustment or inner tension. We are not rational enough or irrational in the adjustment-sense if there is any kind of obvious ill-adjustment or inner tension. We are irrational in a second sense, the evaluation-sense, if no or inappropriate evaluative beliefs are generated on the basis of our non-evaluative beliefs. The word “irrationality” is, perhaps, not widely used in the second sense.16 We would rather speak of (evaluative or moral) blindness, here. Still, there is reason to speak of irrationality: when we are evaluatively blind we fail to respond to some reason for action. And reason, presumably, is our faculty to respond to reasons. Hence, failing to respond to a reason is a reason-failure. We are “unreasonable.” “Unreasonable” and “irrational” are close enough together, I think, to justify the talk of irrationality in this case (which is wide-spread in philosophy). Still, one must not be misled into thinking that rationality in the adjustment-sense is the same as rationality in the evaluation-sense. There are, in fact, two kinds of rationality.

8.

Rationality and Reasons

On the background of this two-tiered view of rationality, it is interesting to take a brief look at how we talk about reasons in various contexts. Even though there is a common ground – all reasons are facts, reasons can be stated, reasons can be premises in arguments17 – we do speak of “reasons” in several different ways. When we say that the reason why Vesuvius erupted was that there were some disruptions in the ground, we are talking about causes. When we say that Jane has reason to come to the meeting because this makes John feel good, we are talking about what speaks in favour of Jane coming to the meeting and about what makes it good for her to come to the meeting (and what ceteris paribus motivates her to come if she is rational); we are talking about reasons for

186

Gerhard Ernst

action. Reasons for action are at the same time reasons for intentions: As rational beings we form our intentions in the light of our reasons for action. And when we ask whether we have reason to do X we either want to know whether (a) we are right in our relevant non-evaluative beliefs,18 and (b) whether we are rational in the evaluative-sense. In our example: if Jane asks herself whether she has reason to come to the meeting she has to check (a) whether her belief that this makes John feel good is true, and (b) whether her evaluative belief that it is a good thing to make John feel good is true, i.e. whether the fact that John feels good if she comes to the meeting speaks in favour of her coming to the meeting. A person who is rational in the evaluation-sense is a person who responds adequately to (apparent) reasons. So, when it comes to forming intentions a rational person responds adequately to her (apparent) reasons (and is also rational in the adjustment-sense). It is here that “responding to reasons” has its place in characterizing rationality. But there is yet another way of speaking about reasons: when we say that the reason why Jane was frightened was that there was a tiger near her (of which she was aware) we are talking about what makes it rational in the adjustment-sense for Jane to be frightened. We are not talking about what speaks in favour of being frightened or about what makes it good for her to be frightened. It is just that feeling frightened is well adjusted to Jane’s belief that she is in danger. The fact that Jane is aware of there being a tiger close by explains why Jane believes that she is in danger, which is bad – if she did not believe this, given her factual belief that a tiger is present, she would be evaluatively blind – which in turn explains why Jane is frightened, since this is the rational adjustment of her emotions to her evaluative belief. There is an intimate connection between reasons for emotions and rationality in the adjustment-sense: given that we are right about the non-evaluative facts and given that we have the relevant evaluative belief, we have reason to feel something if it would be irrational not to feel it. When I have some emotion and I ponder the question whether I have reason to have the emotion, I have to ask myself three questions: (a) am I right about the facts, (b) am I right about my evaluation, and (c) would a perfectly adjustment-rational person have this emotion, given the answer to (a) and (b) is positive. I am irrational in the adjustment-sense if there is some obvious tension between my beliefs and my emotion. And I am not perfectly rational in the adjustment-sense if there is any kind of tension between my beliefs and my emotion. Is “responding to reasons” relevant for the rationality of emotions, too? Yes, because an evaluation is involved. If Jane believed that a tiger was about to attack her but didn’t believe that this is bad, her not feeling frightened would count as irrational, too. She would be irrational in the evaluation-sense because what goes wrong here is that Jane doesn’t respond to the fact that a tiger is about to attack her with the evaluative believe that this is bad; she doesn’t take the fact that a tiger is about

Two Kinds of Rationality

187

to attack her as a reason (for certain actions, e.g. for trying to escape, and for being frightened19). Her emotion – not feeling frightened – is well adjusted to her beliefs, but she doesn’t have an evaluative belief she should have and is, therefore, not responding adequately to a reason (for emotion and action) that she has. How can we understand reasons for belief? I think we have to understand reasons for belief on the model of reasons for emotions without the component of evaluative beliefs. When I ask myself whether I have reason to believe something this question amounts to the question whether (a) I am right in my relevant other beliefs and, given this, (b) whether a perfectly rational person in the adjustment-sense would have the belief at issue. A perfectly rational person in the adjustment-sense does not believe anything which contradicts or is in tension with what she already believes and she believes everything which is implied by the beliefs she already has. Hence, when I ask myself whether I have reason not to believe something, I ask myself whether the belief at hand contradicts or is in tension with what I believe already (or with what follows from what I believe already); and when I ask myself whether I have reason to believe something, I ask myself whether the belief at hand is implied by what I believe already. What is at issue here is rationality in the adjustment-sense only; rationality in the evaluation-sense plays no role whatsoever when we ask ourselves what we have reason to believe.20 And since rationality in the evaluation-sense is irrelevant here, so is talk about “responding to reasons.” In perception, we respond to our environment with forming beliefs. But it leads to insurmountable problems if we try to understand this response as a response to reasons.21 We can, of course, infer new beliefs from old ones, and this could be described as a response to reasons. But we are not irrational if we fail to make inferences which are not obvious. And if we fail to make inferences which are obvious we are irrational in the adjustment-sense. So talking of “responding to reasons” has a very different sense here from the one we identifed in the case of actions (intentions) and emotions.

9.

Conclusion

I think these considerations show the importance of the distinction between rationality in the adjustment-sense and rationality in the evaluation-sense. We fnd both kinds of rationality in the cases of emotions and intentions. We only fnd one kind of rationality in the case of beliefs.22 Practical reasons, i.e. reasons for action, are only relevant for intentions. But evaluative beliefs are relevant for emotions as well as for intentions. Therefore, rationality in the evaluation-sense is relevant for emotions as well as for intentions – but not for the rationality of beliefs. Rationality in the adjustment-sense, on the other hand, is relevant for beliefs, emotions, and intentions.

188

Gerhard Ernst

The adjustment-sense of rationality is intimately connected to the idea of consistency, while the evaluation-sense of rationality is intimately connected to the idea of “responding adequately to reasons.” The sketch I have drawn, therefore, combines the two pictures of rationality I mentioned in the beginning: the picture of rationality as “responding to reasons,” and the picture of rationality as “being consistent.” In my view, both pictures contain important truths while neither contains the whole truth about rationality. Responding adequately to (apparent) reasons is an important part of our rationality. But it is not everything. One cannot understand the rationality of our beliefs in this way, and one can understand only part of the rationality of our emotions and intentions. Being consistent (broadly understood) is also an important and irreducible part of our rationality. But, again, it is not everything. One can understand only part of the rationality of our emotions and intentions in this way. In the picture I have drawn both views are combined (and a view about what inconsistency consists in, namely in a tension between criteria for mental states, is added). Nothing less, I think, suffces to understand the nature of rationality properly.23

Notes 1. This idea is most notably elaborated by John Broome. Cf. Broome (2013). 2. This idea is often associated with Parft (2011). For detailed accounts cf. Kiesewetter (2017) and Lord (2018). 3. What if someone came to believe that she sees Angela in Oxford even though it wouldn’t look to her as if she were seeing Angela? Isn’t this person irrational? I don’t think that there are any such cases but even if there were such cases, “irrational” would certainly not be an adequate description of a person in this situation. 4. This is one of the contested claims I cannot defend here (but cf. Ernst 2001, 2002). I just want to mention two points which are particularly relevant in this defense: (1) Assuming that perception itself somehow gives us reasons for perceptual beliefs leads directly to the problem of “the myth of the given” (cf. Sellars 1956; Williams 1999, ch. 2). (2) Only the assumption of a defaultand-challenge-structure concerning justifcation gets us out of the Agrippatrilemma (cf. Williams 2001, ch. 13). A third point is mentioned in note 17. 5. In the case of perceptual beliefs almost always the beliefs contradicting them have to be given up, not the perceptual belief. For instance, in the example, it would be irrational to maintain the belief that Angela is in London after a closer look which confrmed that I see her in Oxford. 6. Cf. Howson (2000) for a detailed discussion of a logic for partial beliefs. 7. For a discussion of this idea, cf. Wedgwood (2002). 8. Strictly speaking, I am working with a slightly modifed concept of a criterion in this paper: the fact that Peter says “I believe that orange juice contains alcohol” would only be a criterion in my sense if he was sincere. Hence, I am ruling out certain reasons to defeat the ascription of the belief (in this case: insincerity). 9. Many similar principles have been suggested. Cf. “Rationality requires of N that, if N believes p and N believes that if p then q, then N believes that q” (Broome 2006: 183); “If A believes that p, and A believes that p→q, and A does not believe that q, then A is irrational” (Kiesewetter 2017: 15);

Two Kinds of Rationality

10.

11.

12.

13. 14.

15. 16. 17. 18. 19.

20. 21. 22. 23.

189

“Suppose that you believe p, believe if p then q, but fail to believe q. [. . .] You are closure incoherent, and, very plausibly, irrational” (Lord 2018: 28). Again, similar principles have been proposed. Cf. “Rationality requires people to intend to do what they believe they ought to do” (Broome 2013: 2; cf. also p. 170, where he formulates this requirement more precisely); “Rationality requires one to intend to X, if one believes that there is conclusive reason to X” (Kolodny 2005: 521). “Necessarily, if one is rational, then, if one judges ‘I ought to φ,’ one also intends to φ” (Wedgwood 2007: 25). “If A believes that she (herself) ought to φ, and A does not intend to φ, then A is irrational” (Kiesewetter 2017: 14). Similar principles: cf. “Whoever wills the end also wills (in so far as reason has decisive infuence on his actions) the indispensably necessary means to it that is in his control” (Kant 1785: 63); “Rationality requires of us that, if we intend an end and believe some means is necessary to that end, then we intend the means. This has to be a requirement of rationality. A person is necessarily irrational if she does not intend whatever she believes is a necessary means to an end she intends” (Broome 2005: 2); “If you intend to φ and believe that in order to φ you must intend to ψ, then you are rationally required to intend to ψ” (Lord 2018: 19). I am not going to enter the wide-scope/narrow-scope debate here. Wedgwood (2017: 12) argues that if we understand coherence more broadly, then there is no difference between having coherent mental states and having mental states that are supported by the reasons one has. According to my view, this is understanding “coherence” too broadly. In analogy to the perceptual case: Someone who does not respond to some stimulus with a perceptual belief is literally blind (or blind-sighted). It is important here to distinguish between the belief that it is bad that what I do to you hurts you and the belief that it is bad what I do to you. The fact that what I do to you hurts you is a reason to believe that what I do to you is bad – on the basis of the belief that it is bad to hurt someone. But the fact that what I do to you hurts you is not a reason to believe that it is bad that what I do to you hurts you. Every fact can, of course, be a reason. I don’t mean this when I talk about potential reasons here. Hence, I think, the dispute between Bernard Williams (1981: 110) and John McDowell (1995). This list might not be exhaustive and is disputed. The view that reasons can be stated and that reasons can be premises in arguments rules out perceptions as potential reasons (cf. Ernst 2001). I assume that only facts can be reasons. Does the fact that there is a tiger present make it good for her to be frightened? Under normal circumstances it does: being frightened (but not too frightened) plays an important role in preparing us for action. Again, evaluative belief, emotion, and certain intentions are intimately connected with each other. This is true for non-evaluative beliefs as well as for evaluative beliefs. I have indicated two in note 4, and another one in note 17. I argue for the claim that this is the deep difference between practical and theoretical rationality in detail in Ernst (2017). I have presented versions of this paper on various occasions (in Erlangen, Zurich, Magdeburg, Bielefeld, Toronto, Berlin, Munich, and Regensburg), and I want to thank the audiences for their helpful comments. In particular, I want to thank Anna Wehofsits for detailed comments on a previous version of this paper and most of all Sebastian Schmidt for very helpful suggestions for and comments on the penultimate draft.

190

Gerhard Ernst

References Broome, John (2005): “Have We Reason to Do as Rationality Requires? A Comment on Raz”, Journal of Ethics and Social Philosophy 1, 1–8. Broome, John (2006): “Reasoning with Preferences?”, Royal Institute of Philosophy Supplement 59, 183–208. Broome, John (2013): Rationality through Reasoning, Chichester: Wiley Blackwell. Ernst, Gerhard (2001): “Liefert die Wahnehmung Gründe?”, in: Argument und Analyse – Sektionsvorträge, ed. by Beckermann/Nimtz, Paderborn: Mentis, 171–179. Ernst, Gerhard (2002): Das Problem des Wissens, Paderborn: Mentis. Ernst, Gerhard (2017): “Parft über epistemische Rationalität”, in: Worauf es ankommt. Derek Parfts praktische Philosophie in der Diskussion, ed. by Hoesch, et al., Hamburg: Meiner, 223–243. Howson, Colin (2000): Hume’s Problem: Induction and the Justifcation of Belief, Oxford: Clarendon Press. Kant, Immanuel (1785): Groundwork of the Metaphysics of Morals, ed. by Timmermann, transl. by Gregor, Cambridge: Cambridge University Press 2010. Kiesewetter, Benjamin (2017): The Normativity of Rationality, New York: Oxford University Press. Kolodny, Niko (2005): “Why Be Rational?”, Mind 114, 509–563. Lord, Errol (2018): The Importance of Being Rational, New York: Oxford University Press. McDowell, John (1995): “Might There Be External Reasons?”, in: World, Mind, and Ethics: Essays on the Ethical Philosophy of Bernard Williams, ed. by Altham/Harrison, Cambridge: Cambridge University Press, 68–85. Parft, Derek (2011): On What Matters, New York: Oxford University Press. Sellars, Wilfrid S. (1956): “Empiricism and the Philosophy of Mind”, in: Minnesota Studies in the Philosophy of Science, Vol. 1, ed. by Feigl/Scriven, Minneapolis, MN: University of Minnesota Press, 253–329, reprinted in deVries/ Triplett (2000): Knowledge, Mind, and the Given: A Reading of Sellars’ “Empiricism and the Philosophy of Mind”, Indianapolis, IN: Hackett, 2000. Wedgwood, Ralph (2002): “Internalism Explained”, Philosophy and Phenomenological Research 65, 349–369. Wedgwood, Ralph (2007): The Nature of Normativity, New York: Oxford University Press. Wedgwood, Ralph (2017): The Value of Rationality, New York: Oxford University Press. Williams, Bernard (1981): “Internal and External Reason”, in his Moral Luck: Philosophical Papers 1973–1980, Cambridge: Cambridge University Press, 101–113. Williams, Michael (1999): Groundless Belief: An Essay on the Possibility of Epistemology, Princeton: Princeton University Press. Williams, Michael (2001): Problems of Knowledge: A Critical Introduction to Epistemology, New York: Oxford University Press. Wittgenstein, Ludwig (1953): Philosophical Investigations, Oxford: Blackwell 2001. Abbreviated as PI.

B

The Ethics of Blame, Fear, Decision, Passing Thought, and Phantasy

10 The Ethics of Blame A Primer D. Justin Coates

1.

Introduction

The domain of ethical inquiry is concerned with not only how we act but also with how we engage with the world more generally. And much of our engagement with the world occurs in our own minds. Our commitments, values, hopes, concerns, loves, and dreams are all mental states that affect how we interact with our environments and the people, and each of these states can be more or less rational and as a result, better or worse morally speaking.1 As a result, an “ethics of mind” seeks to understand how moral considerations bear on our own mental states and activities. Since our emotions in particular play such an outsized role in our mental lives, it seems that an adequate ethics of mind must attend to the value of certain emotions, emotional episodes, and the variety of dispositions and tendencies we have to be emotionally exercised by what we encounter in our environments. For example, we might sensibly ask whether we have a distinctively moral reason to be more trusting or more emotionally vulnerable in our interactions with others. Or alternatively, we might be concerned, say, that the schadenfreude we feel in response to a rival’s suffering is unwarranted. This concern might be rooted in the thought that schadenfreude can never be a ftting emotion, but the concern might also be anchored in a sense that independently of whether schadenfreude is ever a ftting response to another’s suffering, it’s often a morally unworthy one. Or perhaps we might think it refects poorly on our characters that we don’t really admire someone who clearly deserves it. These questions loom especially large when we consider blame-manifesting emotions like resentment, indignation, and guilt. How we navigate and police the moral community, how we relate to one another as equals, and how we conceive of ourselves and others as being worthy of respect all depend in some crucial ways on our proneness to these emotions.2 Given their importance in our lives with others, the fact that one of these emotions is ftting seems to, at the very least, license blame (if not telling strongly in favor of blaming). Yet these emotions are often the

The Ethics of Blame

193

source of a great deal of suffering, not only on the part of those who are gripped by feelings of say, resentment, but also on the part of those to whom expressions of these emotions are directed. Moreover, these emotions seem at odds with other moral values that we purport to care about: tolerance, civility, generosity, and mercy to name just a few. How these competing values interact is a delicate matter of no small controversy, but it’s a controversy we cannot ignore. A full ethics of mind must pay careful attention to these emotions and to the normative considerations that bear on them. In other words, no ethics of mind can be complete without an ethics of blame.

2.

On Blame and Its Ethics

The question of what blame is is one that’s proven diffcult to answer satisfactorily.3 But whatever blame is, it seems plausible that certain emotional responses to others’ behavior manifest an agent’s blame: these are the blame-manifesting emotions of resentment, indignation, and guilt.4 Since these emotions so clearly manifest blame (even if they cannot be identifed with what it is to blame), I will focus exclusively on them. As a result, I’ll necessarily be less than comprehensive in what I have to say here, since it’s plausible that other mental states or attitudes and outward expressions of these states or attitudes also manifest blame. Nevertheless, I think that the general taxonomy that I’ll offer will apply mutatis mutandis to other blame-manifesting attitudes and behaviors.5 A central task of an ethics of blame, then, will be investigate what general kinds of considerations bear on the all-things-considered propriety of blame. That is, an ethics of blame will offer an account of the facts that are typically relevant to assessing whether an episode or expression of blame is all-things-considered warranted – i.e., facts that can, in many circumstances, serve as deontic constraints on blame. This set of facts includes, but is not limited to, agents’ culpability for their actions, the standing relationship would-be blamers have to those to whom their blame is directed, turpitude or involvement in the transgression on the part of the would-be blamer, the epistemic status of would-be blamers, the values of mercy, generosity, and forgiveness, prudential considerations, and how well blame (and its expressions) serves to protect, promote, and honor important interpersonal values. Only when we consider each of these issues in turn will we be in a position to offer something approaching an ethics of blame.

3.

Does the Ethics of Blame Rest on a Mistake?

Suppose that you discover that a friend has betrayed your trust. Suppose now that it wasn’t an accident, and that instead, he knowingly and intentionally revealed something that he had previously promised to keep

194

D. Justin Coates

secret. Without some shocking further discovery – that, for example, an assassin had credibly threatened to kill his family members if he didn’t break his promise to keep your secret – it’s natural to think that your friend is blameworthy for his transgression. What this means is that he deserves blame. And he not only deserves blame in the form of your resentment (which is ftting since you were the person he betrayed) but he also deserves blame in the form of feeling guilty for what he’s done and indignation from third parties who witness his disloyalty. But how do the facts concerning what he deserves bear on the all-things-considered propriety of blame? That is, does the fact that he deserves to be blamed mean that you must blame him, or only something weaker, e.g., that it would be permissible if you did so, or (weaker still) that there is some reason to do so? On one view, the question of whether someone deserves blame-manifesting emotions exhausts the list of things one must consider when determining the all-things-considered propriety of blame. If the person in question deserves blame-manifesting emotions, then we should have (and perhaps express) those emotions, and if not, then we shouldn’t. The ground for thinking this is that non-desert-based considerations provide agents with the wrong kind of reasons for blaming (or for refraining from blame). In other words, those who accept this view insist that the only bases for an agent’s being warranted or unwarranted in experiencing an emotion are simply the considerations that go into making that emotion ftting (or, in the case of praise- or blame-manifesting emotions: deserved). Any other basis for or against blaming would be analogous to reasons for or against believing that p that are grounded in how happy or sad believing that p might make you. But such considerations aren’t relevant to belief – one should simply believe according to the evidence. And so too, one should blame according to what’s deserved. On this view, then, taking there to be substantive questions about whether one should be resentful or indignant that go beyond the facts that make actions blameworthy is simply a mistake. The problem with this view is that it’s quite doubtful that all non-desert-based grounds for refraining from blame-manifesting emotions are on a par with refraining from believing that p because believing p to be false would make you happy. For one thing, wishful thinking is often something that is of little value, but the goods that can be secured from minimizing or eliminating your resentment in the face of some acts of wrongdoing are often quite signifcant. To see this, suppose you come to discover that just before your friend betrayed you, he had found out that his partner was cheating on him and was deeply distraught and insecure. In that light, his betrayal makes more sense, since we often gossip about others and betray their confdence as a way of boosting our own self-esteem. However, the fact that his betrayal of you was itself a manifestation of the insecurity he felt in response to his own betrayal doesn’t excuse

The Ethics of Blame

195

him from what he did. So it doesn’t wholly undermine the claim that he deserves your resentment and that he deserves to feel guilty for what he’s done as well.6 But now suppose that upon discovering all this you tell yourself, “he’s been a real jerk, yes, but now’s not the time for anger; now’s the time to be supportive of your friend,” and that in light of this you don’t resent him for what he’s done. Here you’ve refrained from resenting someone who deserves it, and yet it seems that you’re within your rights for having done so. Moreover, you’ve decided not to resent him not because he doesn’t deserve it but because it would be all-thingsconsidered better that you not do so. But in this case, the latter doesn’t look like the wrong kind of reason at all. Indeed in this case, there seems to be no better – no more right – kind of reason for acting than the one that’s supplied by the value of your friendship. Yet even if you don’t accept this rather strong claim, a weaker one will do: facts about what your friend needs seem to bear directly on the question of whether to get angry at him in these circumstances in a way that your happiness doesn’t directly bear on the question of whether to believe that p. If such reasons are “of the wrong kind,” then they are not wrong for the same reasons that one’s anticipated happiness is a reason for belief that is of the wrong kind. Of course, this isn’t yet to say that it would be objectionable for you to decide instead to resent your friend for his betrayal. But the former course of mental action certainly seems to be genuinely admirable, which means it must have something going for it rationally speaking. And even if you’re not willing to concede that it’s the uniquely admirable or preferable way of interacting with your friend, it seems hard to deny that that course of action would be rationally optional for you in the circumstances. Yet even that would mean that non-desert-based reasons can bear on the question of whether you should all-things-considered experience some ftting emotional response to an agent’s act of wrongdoing. There’s also a second reason why worries about the so-called “wrong kind of reasons” are misguided here. Even if we accept that any particular episode of a blame-manifesting emotion is made all-things-considered appropriate only on those grounds that make it ftting or deserved, there are still substantive moral questions about what kinds of dispositions to blame that we should cultivate. And if it’s true that we should be disposed to blame-manifesting emotions to a degree different than that which we actually are disposed to these emotions, then episodes issuing from recalcitrant dispositions will be objectionable. So because the question of whether one will blame depends on one’s dispositions, and there are substantive moral norms governing which dispositions one has (and which dispositions one has reason to cultivate), the question of whether it is all-things-considered appropriate to experience blame-manifesting emotions isn’t exhausted by the question of whether the particular attitude is ftting in the circumstances it arises.

196

D. Justin Coates

And fnally, there is a third reason to doubt that only facts that ground desert bear on the propriety of blame-manifesting emotions. The question of whether you should believe that p often depends on the practical stakes of being wrong about p.7 But if pragmatic considerations bear on what we should believe, even if only in this indirect way, it seems plausible that the all-things-considered propriety of our emotions – including our blame-manifesting emotions – can similarly depend on the practical stakes. Of course, as I’ve stated it here, I haven’t specifed which practical considerations bear on the question of whether to blame (indeed, this is, to a great extent the main task that I will take up in the remainder of this paper). Instead, I only mean to suggesting a parity principle: given that practical considerations (particularly practical stakes) can bear on the question of what to believe, practical considerations can bear on whether to blame. If such a reason holds, then we shouldn’t be concerned that appeals to non-desert-based considerations are of the wrong kind when thinking about the propriety of blame any more than we should be concerned about appeals to practical stakes when asking ourselves what to believe. This all means, of course, that the very idea of an ethics of blame does not rest on a mistake. So having sorted that out, we can now turn to some of the considerations that bear on the question of whether any particular instance of a blame-manifesting emotion is all-things-considered appropriate.

4.

Justice and Blame-Manifesting Emotions

Although I’ve just argued that considerations of desert do not exhaust the set of considerations that are relevant to assessing the all-things-considered propriety of any particular instance of a blame-manifesting emotion, they are surely relevant to any such assessment. After all, it is unjust to blame someone who doesn’t deserve blame. Considerations (or reasons) of justice thus bear on the all-things-considered propriety of blame. Two of the key justice-related considerations that are widely accepted to be constraints on the all-things-considered propriety of blame-manifesting emotions correspond to exempting and excusing conditions. If an agent is exempt from blame – if she is not, in general, a morally responsible agent – then it is because she lacks some general capacity that is required for full participation in our responsibility practices. For example, very young children, adults suffering from the late stages of dementia, and those who suffer from severe developmental disorders are all thought to be exempt from moral responsibility. Such individuals, while undeniably agents, lack the rational capacities (e.g., sensitivity to moral reasons) that are required for fully morally responsible agency. By contrast, if an agent is excused from blame, it isn’t because she is not, in general, a morally responsible agent. Rather, on some particular

The Ethics of Blame

197

occasion she is not, for whatever reason, morally responsible for performing that action (perhaps because, e.g., she was coerced, hypnotized, factually mistaken, etc.).8 In such cases, the agential capacities that ground our status as morally responsible agents are temporarily impaired (and the agent herself is not responsible for their impairment). Yet in either of these cases, the thought goes, there is an especially weighty (perhaps even decisively weighty) reason that’s grounded in the value of justice against blame-manifesting emotions.9 To experience, say, resentment in either of these cases would essentially be to blame the innocent. This suggests two distinct ways in which the considerations of justice bear on the all-things-considered propriety of blame-manifesting emotions. Exemption: If B is not, in general, a morally responsible agent, then there are weighty pro tanto reasons against A’s having blamemanifesting emotions in response to any of B’s actions. Excuse: If B does not deserve to be blamed for x-ing because B has an excuse for x-ing, then there are weighty pro tanto reasons against A’s having blame-manifesting emotions in response to B’s x-ing. Together, Exemption and Excuse circumscribe the circumstances in which exempting and excusing conditions constrain the propriety of blame. And to violate one of these norms is to act unjustly towards the innocent targets of our blame. Of course, as you have no doubt noticed, as I’ve stated them here neither Exemption nor Excuse are hard and fast prohibitions. Instead, these norms identify very weighty pro tanto reasons against blame-manifesting emotions that arise whenever the antecedent conditions are met. This means that although Exemption and Excuse are almost certainly decisive reasons against blame-manifesting emotions in ordinary contexts, they could potentially be outweighed. Yet because such contexts are rare, the pro tanto reasons in question are of suffcient weight that it will almost always follow that A has most all-things-considered reason to refrain from blaming B. However, even in cases in which A’s pro tanto reason is outweighed by countervailing reasons of suffcient weight, Exemption and Excuse still place normative burdens on would-be blamers, since they identify weighty reasons against blame that cannot simply be ignored. For suppose that there really is some circumstance in which it is all-thingsconsidered appropriate to blame someone who does not deserve to be blamed (imagine, for example, a situation of the sort consequentialists sometimes point to). If we blame in such a case, then even though it is all-things-considered permissible, it’s plausible that we nevertheless owe the blamed an explanation of why the reasons grounded in justice are outweighed. In other words, in such a situation we would be obliged to say something like, “hey look, I know this isn’t entirely fair, but in

198

D. Justin Coates

this case . . .” or as we sometimes say to children, “I know you can’t understand this, but it’s for your own good.” What this means is that Exemption and Excuse play an important role not only in determining when blame-manifesting emotions are all-things-considered appropriate, but also what form that blame can legitimately take (since even when they are outweighed, they ground the need for further explanation in such cases). It’s not enough that we attend to the reasons provided by Exemption and Excuse, since attending to these considerations only guarantees that our blame-manifesting emotions don’t run afoul of justice by being directed at an agent who doesn’t deserve blame. More is needed for these emotions (and their expressions) to be all-things-considered appropriate. Angela Smith makes the point nicely: “whether it would be appropriate for any particular person to have and express any particular [blaming] attitude to the agent will depend on many considerations in addition to the agent’s responsibility and culpability” (Smith 2007: 477). To know when it’s all-things-considered reasonable to blame, we must identify further considerations that bear on the propriety of any particular instance of a blame-manifesting emotion.

5.

The Jurisdiction of Blame-Manifesting Emotions

Just as the legal notion of jurisdiction, which concerns “how far the arm of the law can reach,”10 bears on the authority of a government to prescribe and enforce its laws, so too, jurisdictional-like considerations bear on the all-things-considered propriety of blame-manifesting emotions. These are considerations that are relevant to the question of whose place it is, normatively speaking, to blame wrongdoers. In other words, jurisdictional considerations fx the conditions under which agents have the normative power or authority to legitimately blame. Thus, when we respond to blamers (as we so often do) by saying, “It’s not your place to blame me!” or by saying, “You have no right to blame her!” we are implicitly invoking jurisdictional considerations. For just as it isn’t Canada’s place to punish me for failing to pay United States federal taxes, it’s often not our place to blame agents whose wrongdoing falls outside of the scope of our moral jurisdiction. But what facts are relevant to determining the scope of one’s jurisdiction? That is, what facts bear on the all-things-considered propriety of blame in this way? 5.1

The Business Condition

One thing that’s relevant for all-things-considered appropriate blame is that there’s the “right” kind of relationship between would-be blamers and would-be blamees.11 In particular, it seems to matter whether the transgression in question was any of the would-be blamer’s “business.”

The Ethics of Blame

199

Business: If B’s transgression x is none of A’s business, then there is a pro tanto reason against A having blame-manifesting emotions in response to B’s x-ing. The simple idea here is that one way in which it can be “not my place” or “no right of mine” to blame is when the transgression in question occurs outside the sphere of human relations that I am privy to. For example, if I lack a standing relationship to another individual – say, with the people sitting next to me at the coffee shop – then if I were to overhear that one of them was leaving her husband for a coworker, there seems to be a reason that tells against me pushing my way into the conversation in order to express my opprobrium. As I would quickly be told, these matters are “none of my damn business!” Of course, here you might note that this case involves expressed blamemanifesting emotions. But there’s also a reason in this case to refrain from getting indignant towards the person. After all, I might criticize myself for getting hotly indignant in the situation even if I didn’t express my indignation on the grounds that Business identifes. That such selfcriticism is reasonable is remarkably easy to demonstrate. Suppose for reductio that Business is false, such that the fact that another’s transgression is none of your business is no reason at all to refrain from blaming. If so, then you would be no more reasonable for refraining from indignation in these circumstances than you would for refraining from indignation because you found the wrongdoer physically attractive – a fact that pretty clearly does not provide you with a reason to refrain from indignation. But clearly you would be more reasonable for refraining on the basis of the former fact than on the basis of the latter fact. So the former fact – that the wrongdoing is none of your business – must provide you with some reason to refrain from indignation in the circumstances.12 However you might be worried that this lionizes privacy and a nebulous notion of “one’s business” in a way that silences criticism of serious wrongdoing. For example, if Doug hears his neighbor verbally and possibly even physically abusing his partner, then it’s entirely appropriate for Doug to intervene, to call the police, and to cut off his relations with the neighbor. So too, it’s pretty clearly appropriate that Doug feels indignation towards his neighbor. Yet if we suppose that Doug doesn’t know his neighbor, then we might worry that the abuse is, in some sense, none of Doug’s business. Of course, if this is right then that tells against Business, since it entails that Doug has some reason to refrain from blamemanifesting emotions in this case. However, it’s plausible that Doug has no reason against having blame-manifesting emotions in response to his neighbor’s action. So Business must be false. This objection to Business fails. As I stated it, the sense in which some failing – moral or otherwise – is “none of A’s business” is to be understood in terms of whether the failing occurs outside the sphere of

200

D. Justin Coates

human relations that A is privy to. And plausibly, the general wellbeing of the neighbor’s partner is something that falls within the sphere of relations that Doug is privy to, even if Doug doesn’t know his neighbor or his partner. After all, there is a general concern that we owe to everyone in virtue of their standing as persons: we are our neighbors’ keepers. So when a moral fault constitutes an imminent and grave threat to another’s wellbeing, as is the case when Doug hears his neighbor abusing his partner, it becomes Doug’s business; Business does not even apply in such a case. The fact that Doug should, without thinking twice, not only blame his neighbor, but take positive steps towards preventing him from abusing his partner, is no threat to the truth of Business. Moreover, even if it were possible to construe the case in such a way as for the domestic violence not to be Doug’s business, it doesn’t follow that that would give us grounds for rejecting Business. After all, Business is silent as to the strength of the pro tanto reason against blame-manifesting emotions in cases in which the transgression isn’t your business. It’s very doubtful that this reason at all approaches the strength of the reasons identifed in Exemption and Excuse. So given that such a reason is relatively weak even in the best of circumstances, it’s no stretch to think that it would be silenced in the sort of situation described earlier. But though cases like the previously-mentioned do not teach us that Business is false, they do serve to remind us of the rather limited scope of some of considerations that bear on the all-things-considered propriety of blame. Relatively minor faults can be “none of our business,” and in such cases, we’ll have a reason against blame. But serious moral failings are plausibly part of our business, and even if they’re not, their gravity might nevertheless silence the reason against blame-manifesting emotions that Business identifes. Thus, the scope of Business is constrained.13 Yet this doesn’t mean that Business isn’t signifcant. It constrains blame-manifesting emotions in the case of relatively minor foibles and faults, and so, it puts real constraints on our tendency to be judgmental busybodies. And if nothing else, that’s something. 5.2

The Standing to Blame

Business is not the only, or even the most prominent, jurisdictional consideration that bears on the all-things-considered propriety of blame. For example, when we’re critical of Tom Buchanan’s anger towards his wife Daisy and her lover Gatsby, it’s not because their affair was none of Tom’s business (Fitzgerald 1995: 136–137). Indeed, it is his business. Rather, we seem to be thinking that Tom, who was having his own affair with Myrtle Wilson (and who seems generally unconcerned with his marriage), lacks what’s sometimes called “the

The Ethics of Blame

201

moral standing to blame.” That is, we seem to endorse something along the lines of: Standing: If A lacks the moral standing to blame B for B’s transgression x, there is a pro tanto reason against A having blame-manifesting emotions in response to B’s x-ing.14 Among those considerations affecting when it is all-things-considered appropriate for A to have blame-manifesting emotions in response to B’s action(s), Standing has perhaps received the most attention. Despite this, it’s not altogether clear what “the moral standing to blame” is. Nor is it clear why A’s lacking the moral standing to blame might undermine the all-things-considered justifcation of her blame-manifesting emotions. In response to these questions, I’ll simply stipulate that the notion of standing I have in mind is relatively thin. A person has standing to blame in the relevant sense when nothing about that person undermines the force of their blame. This means that “standing” is the presumed or default position of all moral agents; it is not something that, at least in the frst, must be earned or merited. However, when there is something about the person that undermines the force of their blame, then that person will either have less standing to blame or lack the standing to blame altogether, depending on how signifcantly weakened the force of their blame would be. And according to Standing, in these circumstances, the agent with impaired or no standing will have a pro tanto reason against blame-manifesting emotions. Although this idea is somewhat sketchy, it will do for our purposes, since we do not need a full analysis of the moral standing to blame to capture the intuitive idea: Tom Buchanan has fairly weighty reasons against having the blame-manifesting emotions that he does in the circumstances, and plausibly, this is because something about him (viz., his own affair) severely undercuts the force of his blame. To better motivate Standing, then, I now turn to perhaps the two clearest threats to the force of one’s blame: hypocrisy and complicity. 5.2.1

Hypocrisy and Standing

We’re told very early in our moral development that pots shouldn’t call kettles black, that people living in glass houses shouldn’t throw stones, and that we should judge not, lest we be judged ourselves. The simple idea behind these commonplaces is that A’s status as a hypocrite – as being guilty of some transgression t – provides A with a weighty reason to refrain from blame-manifesting emotions in response to B’s t-ing. To illustrate the pull of this familiar idea, consider a fairly straightforward case of hypocrisy:

202

D. Justin Coates Tony’s Hypocrisy Bill and Tony are good friends who have coffee every Thursday. Bill is always there at the time they decide, but Tony is reliably late. Since it’s not too big of a deal, Bill doesn’t really mind, so he never says anything about it. On this particular Thursday, however, it is Bill who is 10 minutes late. Tony is very upset and in an accusatory but somewhat passive-aggressive manner, he says, “Glad you could make it.”

If you’re like me, then you think Tony’s anger, which is expressed in the comment he directs towards Bill, is all-things-considered inappropriate. But of course, Bill’s tardiness is Tony’s business, since they had an agreement to meet, so Business cannot explain the impropriety of Tony’s blame. Better, we can explain the impropriety of Tony’s blame by appealing to his hypocrisy. Because Tony’s a hypocrite who regularly shows up late to his meetings with Bill, something about his own character and actions have undercut his complaint – they reveal that he doesn’t care about tardiness per se, only tardiness that adversely affects himself. This severely weakens the condemnatory force of his blame. Tony is therefore shown to lack the standing to have blame-manifesting emotions in response to Bill’s occasional tardiness. Of course, not everyone is convinced that Standing provides the best explanation of what goes wrong in Tony’s Hypocrisy. Tony is a jerk to Bill, for sure, but unless Bill has an excuse for his tardiness, Tony’s blame might itself be ftting and valuable. For standing-skeptics of this sort, the idea that we can blame only if we’re not ourselves hypocrites puts undue constraints on would-be blamers. After all, who among us can honestly said they’ve never lied, broken a promise, or belittled someone without cause? Surely the fact that I bullied someone as a teenager doesn’t render it wholly inappropriate that I blame bullies now. But though all of us are ourselves sinners, this alone can’t obviate blame-manifesting emotions – particularly in light of how important these emotions are for our lives with others. In response to this objection to Standing I will frst note that as I’ve articulated it, Standing does not entail that it’s all-things-considered inappropriate for Tony to blame Bill – only that he has a pro tanto reason to refrain from doing so. This cuts against the standing-skeptic’s concern that all (or almost all) of us would be normatively barred from blaming in a wide range of cases. The fact that I too was a bully can be a reason for me to refrain from blaming another bully even if it is nevertheless all-things-considered appropriate to blame. Standing thus offers a weaker normative constraint on legitimate blame than some skeptics imagine. However, if a standing-skeptic insists that the fact that I would be a hypocrite in blaming someone is no reason at all to refrain from blaming, then I can only recommend the same argument I

The Ethics of Blame

203

offered in defense of Business earlier in this chapter: surely the fact that he’d be a hypocrite to blame Bill gives Tony a better reason to refrain from blaming Bill than does the fact that Bill is wearing a blue shirt, or the fact that it’s Tuesday, or the fact that . . . (facts such as these, I take it, really do provide Tony no reason to refrain from blaming Bill). But if there is more reason to refrain from blaming Bill on the grounds that doing so would be hypocritical than on other, arbitrary grounds, then Tony’s past behavior (his tendency to be unreliable) must be provide him with some reason to refrain from blaming Bill. This is precisely what Standing posits. Moreover, it’s imminently plausible that even if it is all-things-considered appropriate for Tony to blame Bill, it’s plausible that he’ll have reason to express his blame differently in light of the fact that he too has made Bill wait. And the best explanation for this is simply that even in cases in which the reason identifed by Standing is outweighed, it still puts normative constraints on blame. But it can do this only if facts about would-be blamers matter for the all-thingsconsidered propriety of blame. 5.2.2

Complicity and Standing

Hypocrisy is not the only threat to our moral standing to blame. Another threat is complicity. To see this, consider the following instance of complicity: Giving Orders Otto is a lieutenant who orders a private, Erich, to kill a group of wounded prisoners. Erich carries out the order and guns down the enemy soldiers. Later, a colonel asks Otto about the prisoners. Otto blames their death on Erich, saying to the colonel, “he’s the idiot that killed them.” Here again, it seems that Otto’s blame is inappropriate, even though Erich is morally responsible for murdering the soldiers and plausibly deserves to be blamed. However, when Otto blames Erich for the murders, he is not a hypocrite; after all, he didn’t actually kill the prisoners. Despite this, Otto is involved in the killings in such a way that there is also blood on his hands. And this fact about Otto seems to undercut his blame. It seems, therefore, that it is extremely inappropriate for Otto to blame Erich. And the best explanation for this, it seems to me, is that, like hypocrisy, complicity undermines an agent’s moral standing to blame. Standing thus seems vindicated by our judgments in Tony’s Hypocrisy and Giving Orders: when facts about us undercut the force of our blame, then we have reason to minimize or eliminate blame-manifesting emotions.

204

6.

D. Justin Coates

Transactional Norms of Blame-Manifesting Emotions

Besides justice-based or jurisdictional considerations, there is another distinct class of considerations that affect the all-things-considered propriety of blame. For example, sometimes we object to blame-manifesting emotions by pointing out that in the circumstances, those emotions weren’t suitably responsive to epistemic reasons, or by reminding wouldbe blamers not to get carried away in their opprobrium, or because the blamer has failed to take seriously the response of the wrongdoer. In such cases, we’re suggesting that unless blaming transactions has certain properties, the blame itself is not ftting, and so we have reasons for refraining from it. In other words, we’re recognizing that, in addition to justice and jurisdictional norms of blame, there are also transactional norms of blame. It’s not enough that your blame-manifesting emotions are directed at someone who deserves it and that it’s “your place” to feel such emotions, you must also blame well, and you do so, I contend, only if your blaming transaction is suitably structured. 6.1

An Epistemic Norm for Blame

One transactional norm has a very familiar analog in the domain of the criminal justice system. Reasonable: If A is not reasonable in her belief that B is morally responsible for x-ing (and so, deserving of blame-manifesting emotions), then there is a weighty pro tanto reason against A’s having blame-manifesting emotions in response to B’s x-ing. If Reasonable is true, then even if B is morally responsible for a-ing, it may still be inappropriate for A to hold B responsible – at least until A acquires evidence of the sort that renders her beliefs about B’s blameworthiness reasonable. As a general principle, this seems fairly plausible, but we can further motivate Reasonable by considering the following case: Keyed Car Andrew has been fghting with a coworker about how to best pursue a new client. The fght got pretty heated, and they both decide to take a break for the day. As Andrew is leaving, he notices that someone has keyed hateful words into his car. Getting very angry, he quickly jumps to the conclusion that it was his coworker, even though the coworker has no history of vandalism or pettiness. Without seeking any further evidence, he takes a screwdriver from a toolkit in his trunk and punctures the coworker’s tires. “That’ll teach him,” he thinks to himself.

The Ethics of Blame

205

In this case, Andrew’s blame manifests itself in his anger, and in the actions that his anger motivated. But plausibly, Andrew’s blame-manifesting emotions (and the actions that Andrew took them to rationalize) were inappropriate. Leaving aside legal implications for petty vandalism, Andrew’s blame isn’t obviously sensitive to evidence. And because his action shows a disregard for having a reasonable basis to blame, it reveals that Andrew is insuffciently concerned with only blaming those who deserve to be blamed. There is another reason for thinking that there are reasons against having blame-manifesting emotions in the absence of reasonable belief about the purported wrongdoer’s moral guilt. This argument bootstraps Reasonable from the norms of assertion. To see this, consider that there are reasons for A to refrain from asserting that p if it is not reasonable for A to believe that p. But if assertion requires reasonable belief, then plausibly, there are reasons for A to refrain from implicating that p if it is not reasonable for her to believe that p. However, our blame-manifesting emotions (and particularly their natural expressions) implicate that their target is blameworthy, that they have done wrong, that they have a bad character, etc. Therefore, blame must also be subject to the requirement of reasonable belief. In other words, if A is not reasonable in her belief that B is blameworthy, then there is a pro tanto reason for A to refrain from blaming B. The truth of Reasonable, then, is grounded both in the weighty value of protecting the innocent from blame and in the norms of assertion and implicature.15 6.2

The Right Amount of Blame

But Reasonable is not the only transactional norm of blame. There is also: Proportionality: If A’s blame-manifesting emotions in response to B’s x-ing would not be proportional to the gravity of B’s wrongdoing, then there is a pro tanto reason against A’s blame-manifesting emotions in response to B’s x-ing having either the intensity, duration, or signifcance for B that they would otherwise have. Proportionality, then, simply formalizes the old adage that the punishment must ft the crime. It might be appropriate to fne Vanessa for rolling through a traffc stop, but it would be extremely inappropriate to sentence her to fve years in prison for the same crime. The same is true of blame-manifesting emotions as they occur within our interpersonal relationships: a minor insult might warrant some resentment, but it certainly doesn’t license overwrought expressions of anger, violence, rage, or signifcant revisions to one’s relationships. More concretely, let’s imagine that you’re in a busy parking garage and that you’ve been waiting patiently for a car that’s backing out of its

206

D. Justin Coates

spot. Now imagine another person driving in from the other direction and taking the parking spot that you were clearly waiting for. If this has ever happened to you, you’ll know that it’s an infuriating experience. And while it’s plausibly appropriate to blame the space-thief – maybe by glaring at him or by expressing your indignation to those in the car – it seems inappropriate to pull in behind him, get out of your car, yell at him for 10 minutes, scream, cuss, and cause a huge scene. It also seems inappropriate to go on your way, all the while seething with anger for hours after the fact. These blame-manifesting responses greatly exaggerate the degree to which the space-thief has insulted you. He has, to borrow a phrase from Adam Smith, “made little account of you.” But he hasn’t made that little of an account, since there are other spaces, and he only wasted a few minutes of your time. That we shouldn’t be too exercised in such cases, even though many of us have a natural tendency to do so, is plausibly explained by the truth of Proportionality. 6.3

Blame’s End

A third transactional norm relates the role that the aim or function of blame-manifesting emotions play in determining the conditions under which those emotions are all-things-considered appropriate. Aim: If B’s response to her own wrongdoing has led her to sincerely undertake the reconciliatory responses to wrongdoing that blamemanifesting emotions aim for, then there is a pro tanto reason against A’s having blame-manifesting emotions in response to B’s x-ing as if B hasn’t recognized the signifcance of the wrong. Since that statement of Aim is somewhat obscure on its own, let’s consider the following case: Sergio’s Sabotage Because they once clashed at a conference, Sergio purposefully writes Clark a bad tenure letter even though Sergio recognizes Clark’s work to be good. However, before the tenure decision has been made, Sergio has a change in heart. He realizes that he wasn’t fair to Clark, and that it is petty and underhanded to destroy someone’s career because of one bad experience. So Sergio writes a second letter on Clark’s behalf. On the strength of the second letter, Clark is granted tenure. Now suppose that after he is awarded tenure, Clark learns of Sergio’s multiple letters. I think it’s fair to say that he is within his rights to be very upset with Sergio and to blame him for potentially ruining his career.

The Ethics of Blame

207

But, it also seems that it would be inappropriate for Clark to respond to Sergio as if he had been wholly unrepentant. That Sergio took positive steps to rectify his transgression cannot be discounted. In other words, in Sergio’s Sabotage, it’s plausible that Clark should not respond to Sergio as if Sergio has done nothing to acknowledge his fault. Angela Smith develops this idea when she argues that: When we do need to decide what [blaming] attitude to take toward another on the basis of a moral fault, we should always take into consideration the person’s own response to her failure. If someone has an objectionable attitude toward me, for example, but is already reproaching herself for it and making efforts to change, then I may judge that I have no reason to adopt or express any blaming attitudes toward her at all. Her own self-reproach shows to me that she already recognizes that I have moral standing and deserve better treatment, and therefore I may no longer see her attitude as posing a challenge to me or my status. Smith 2007: 482 That is, given that repentance and reconciliation are among the aims that blame clearly has, the reasons for blame-manifesting emotions are considerably weakened when those aims are met independently of your blame. Additionally, notice that Aim is weaker in its claims than the other structural norms. Reasonable and Proportionality entail that there are pro tanto reasons against blame-manifesting emotions when the antecedent conditions are met. But Aim only claims that there are pro tanto reasons against A’s having blame-manifesting emotions in response to B’s x-ing that are the same as they would be if B hadn’t already recognized her fault and taken steps to atone for it. That means it is consistent with Aim that if Clark’s blame-manifesting emotions for Sergio are suffciently mitigated, they can still be all-things-considered appropriate. So Proportionality tells us that blame must ft the transgression, and Aim tells us that blame must also ft the wrongdoer’s own response to her transgression. What holds Reasonable, Proportionality, and Aim together as a class is that if our transactions with wrongdoers are not structured according to these norms (i.e., if would-be blamers are not sensitive to the reasons identifed here), then our blame-manifesting emotions are being conscripted in the service of lashing out. Like blame, we lash out in response to wrongful, harmful, or unwelcomed treatment. But unlike blame proper, lashing out apparently aims at bringing others down and at inficting vengeful harms not only on those who are responsible for the wrongs but also on those unfortunate enough to be in our orbits. And although lashing-out responses to wrongdoing are natural, they are also damaging

208 D. Justin Coates and disrespectful and prone to excess and self-indulgence. It is only by having our blame-manifesting emotions structured in accordance with Reasonable, Proportionality, and Aim that we avoid this and maintain a healthy and all-things-considered defensible response to wrongdoing.

7.

Blame-Manifesting Emotions, All-Things-Considered

In the previous three sections I have introduced several distinct considerations that bear on the all-things-considered propriety of blame-manifesting emotions: Exemption, Excuse, Business, Standing, Reasonable, Proportionality, and Aim. But an obvious question arises here: how exactly do these considerations combine to yield substantive answers to the question of whether its all-things-considered appropriate to experience blame-manifesting emotions? That is, how do these various norms, which all generate pro tanto reasons of differing weights, interact with one another in such a way as to determine when it is that A should, allthings-considered, blame B for x-ing? The frst thing to note here is that if A were to ignore all of these considerations, the thing which would seem to be most inappropriate about A’s blame-manifesting emotions is not that it was none of A’s business or that B had already taken steps to rectify her wrong. Instead, the thing that would bother us the most is that A is blaming B even though B doesn’t deserve to be blamed (either because she is not, in general, a morally responsible agent or because she is not morally responsible for the action under consideration). This suggests that in an intuitive ordering of these considerations, we should give the most weight to Exemption and Excuse. We can make this point more precisely. To see this, consider that there will be many cases in which the all-things-considered impropriety of blame-manifesting emotions is overdetermined. For example, suppose that B isn’t morally responsible for x-ing and that A is herself guilty of x-ing. It seems then that both Excuse and Standing can provide a normative basis for thinking that there would be weighty pro tanto reasons against A’s having blame-manifesting emotions towards B in this case: B is excused and A is a hypocrite. However, in this case, it seems like the more fundamental normative basis for A refraining from blaming B is that B isn’t morally responsible for x-ing. Once that fact is established, A’s hypocrisy does very little or nothing to tell against the all-things-considered propriety of blame-manifesting emotions. Thus, it seems considerations of justice are normatively prior to jurisdictional and transactional considerations. But even though I think Excuse and Exemption have priority over the other considerations, I do not mean to suggest that the other considerations I’ve discussed aren’t normatively signifcant. For example, Business, which holds that it is inappropriate for us to blame others for moral faults if the fault in question is none of our business, is plausibly tied to

The Ethics of Blame

209

the value of toleration. Since toleration is such an important value, we should expect that considerations that facilitate toleration are of considerable weight. The same is true of Reasonable. Because it is grounded in the value of protecting the innocent, which is quite important, it’s plausible that Reasonable generates weighty pro tanto reasons to refrain from blaming in cases of ignorance. Of course, you might worry that attention to the various jurisdictional and transactional norms of blame invites a kind of pernicious quietism. Given all of these considerations, many of which are putatively of signifcant normative weight, it might seem as if we should, in general, refrain from blaming lest we be guilty of all-things-considered inappropriate blame. However, a world in which wrongdoers deserve to be targeted with blame-manifesting emotions but in which those responses are systematically muted or withheld is not an attractive one. After all, by not blaming those who deserve to be blamed, we encourage further entrenchment of injustices: imagine for example that liberal democracies had not sanctioned apartheid-era South Africa.16 And we allow similar injustices, albeit on a much smaller scale, when we don’t blame bullies and bigots for their treatment of the less well off. Yet if we take seriously jurisdictional and transactional considerations to the degree I am suggesting here, you might think that it’s the world we create for ourselves. This is a legitimate concern. But ultimately, I do not think it is decisive. First, you might think that even if the jurisdictional and structural norms do render it all-things-considered inappropriate to have blame-manifesting emotions in a great number of cases, it doesn’t follow that we must therefore acquiesce to injustice, since blame is just one of the ways in which we can respond to serious wrongdoing.17 It’s possible, after all, to object to wrongdoing without blaming wrongdoers. To take but one very famous example, consider the Buddhist monk Thich Quang Duc’s act of self-immolation in response to Diem’s heinous anti-Buddhist policies. This act is clearly a signifcant response to injustice. Yet in his last words, Thich Quang Duc writes: Before closing my eyes and moving towards the vision of the Buddha, I respectfully plead to President Ngo Dinh Diem to take a mind of compassion towards the people of the nation and implement religious equality to maintain the strength of the homeland eternally. I call the venerables, reverends, members of the sangha and the lay Buddhists to organize in solidarity to make sacrifces to protect Buddhism (emphasis added). Lee et al. 2015: 944 Here Thich Quang Duc is clearly objecting to Diem’s unjust regime. He “respectfully pleads” that Diem change his policies, and he shows his commitment to that plea in his subsequent actions. But it doesn’t seem as

210

D. Justin Coates

if Duc is evincing any blame-manifesting emotions.18 This suggests that it is possible to hold wrongdoers to a high moral standard without blaming them should they fail.19 On a smaller scale, then, we too can respond to injustice without blame, by pleading with wrongdoers to rethink their ends and by calling on others to stand in solidarity with us. Consequently, even if we concede that blaming strictly in accord with the jurisdictional and transactional considerations I’ve adduced here severely limits the range of cases in which it is all-things-considered appropriate to blame, it doesn’t follow that we are impotent in the face of wrongdoing. But second, unless there is a general argument against human persons being morally responsible for their actions, I’m doubtful that adherence to these norms will severely limit the range of cases in which blame is appropriate. As I’ve already pointed out, in cases of signifcant wrongdoing, particularly directed towards those who are less equipped to stop the wrongdoing, it is our business to blame. This means that a proneness to indignation in response to the suffering of the oppressed is morally important. Business identifes a reason not to be judgmental busybodies, not a reason against being prepared to wade into the mess created by serious wrongdoing. Similarly, since Standing only constrains the blame of those who are hypocritical or complicit in any serious wrongdoing, there will be many in the moral community who can legitimately blame wrongdoers. Moreover, in general, when we see agents act wrongly we have suffcient evidence to be reasonable in our belief that they are blameworthy for their action. So in standard cases, Reasonable will also be satisfed. Finally, neither Proportionality nor Aim would, if true, really put pro tanto constraints on blame of the sort that would invite pernicious quietism. Accordingly, while I appreciate the concern that we should not let wrongdoers use these norms to shield themselves from deserved blame, I do not think these norms could actually serve to hide or protect wrongdoers in this way.

8.

Conclusion

Of course, the considerations that I’ve identifed here are surely not the only ones that are ultimately relevant to the all-things-considered propriety of blame-manifesting emotions, though they do seem to be the ones that arise given the very nature of these emotions. Other considerations that arise more generally might matter as well. It might be, for example, that virtue itself requires us to be generous towards others, including our assessment of their actions. If this is true, then maybe we’ll have a pro tanto reason against blame-manifesting emotions whenever generosity is apt. Similar points might be made of other virtues – virtues like mercifulness or forgivingness. But these considerations, while important, are aretaic in nature, whereas the considerations I’ve focused on are deontic ones. This means that the way they bear on blame-manifesting

The Ethics of Blame

211

emotions is importantly different. Deontic considerations, when weighty enough, seem to place real constraints on the all-things-considered propriety of blame. On the other hand, aretaic considerations seem to bear on how we evaluate people qua blamers. We might think that a forgiving person is better than an unforgiving one without also thinking that a blamer who has the trait of being unforgiving is eo ipso allthings-considered unwarranted in having blame-manifesting emotions in response to another agent’s wrongdoing. A full ethics of blame (and so, a full ethics of mind) requires that we have more to say on this point, since we have reason to care not only about deontic considerations but about aretaic ones as well. This chapter doesn’t settle the question of when blame is both right and good, but hopefully by identifying considerations that bear on the all-things-considered propriety of blame-manifesting emotions, it puts us on the right track.

Notes 1. For one especially trenchant challenge to the claim that there are moral limits to what happens in our minds, see Sher (2018). 2. This picture of the role that blame-manifesting emotions play in regulating our lives with others has its source (at least in the contemporary literature) in Strawson (1962). 3. For a more detailed account of this diffculty, see Coates/Tognazzini (2013). 4. There are perhaps other blame-manifesting emotions, including disappointment, contempt, disdain, shame, and rage. 5. One possible exception to this are judgment theories of blame. On these theories, to blame someone is to judge that their behavior speaks poorly of them. If blame can be reduced to a judgment or a belief, then it would seem that it would be governed exclusively by the norms of belief. If such norms are exhausted by epistemic norms, then it’s hard to see how the fact that I would be a hypocrite to blame you for x-ing could be a reason to refrain from blaming. Since such facts seem to many to be such a reason, that puts pressure on either the claim that blame can be reductively analyzed in terms of belief or the claim that the norms governing belief are exhausted by epistemic norms. For reasons to doubt the former claim, see Sher (2006) and Scanlon (2008). For reasons to doubt the latter claim, see McCormick (2015) and Rinard (2017). 6. You might think that if he is suffciently distraught then he is excused from responsibility. I imagine that there might be some degree of distress that is suffcient to fully excuse. Yet I also think that many instances of being in distress do not excuse at all or, at most, only mitigate the degree to which an agent deserves blame. 7. One way to spell this thought out is to say that justifcation or knowledge are “pragmatically encroached,” but one can reject this claim and still think that the practical stakes of being wrong about p give you reason to be more reticent in assenting to p. For more on pragmatic encroachment, see Stanley (2005) and Kim/McGrath (2019). 8. For more on this way of framing the distinction, see Watson (1987). 9. Here I want to highlight an important point: the value of justice is the source of weighty reasons to refrain from blaming in a range of cases, but this does

212

10.

11.

12. 13.

14.

15. 16. 17. 18.

D. Justin Coates not mean that it can never be “appropriate” (whatever exactly that means) or morally permissible to blame in such cases. As we’ll see, this point applies to other values that put rational guardrails on blame. As Antony Duff puts it, “the law’s claim to jurisdiction over conduct committed within the territory of the state whose law it is involves two claims: that it can defne such conduct as criminal; and that its courts have jurisdiction to try the alleged perpetrator of such conduct” (2007: 44). You might think that considerations of this sort are also grounded in the value of justice. One reason for thinking this is that it seems like whether I deserve a certain response at all depends on the person issuing the response. For example, I might deserve criticism from a colleague for failing to get some piece of paperwork done on time and not deserve it from someone with whom I have no relationship. So if a passerby criticized me, I might defend myself by saying, “I don’t deserve that from you.” On this view, the question of what I putatively deserve and who’s giving me what I putatively deserve are inextricably linked. So blaming me when I don’t deserve blame from you seems like a kind of injustice. This conception of desert and our relationship to the arbiter what it is that we deserve is an interesting one. And if it’s ultimately correct, then it suggests that jurisdictional issues have their ultimate source in the value of justice. However, even if that’s right, I still think it’s worth distinguishing them from other ways in which the value of justice bears on the all-things-considered propriety of blame. Consider that in the criminal law, the questions of whether I am criminally guilty and whether the judicial body has the jurisdiction to hold me criminally guilty might both ultimately be questions of justice. And yet, the reasons for the United States not to imprison me for a murder I did not commit seem importantly different than the reasons that Canada has for not imprisoning me for a murder I committed in the United States. Cf. Mark Schroeder’s (2007) argument that we are unreliable detectors of negative reasons existentials. This ties into Angela Smith’s (2007) suggestion that the gravity of the wrongdoer’s fault is a consideration that, in addition to the culpability of the wrongdoer, plays some role in determining whether it is all-things-considered appropriate to blame a wrongdoer. But rather than understanding this as an independent consideration, I am suggesting that we understand the gravity of the fault as being relevant to the propriety of blame in virtue of its role in determining whether some piece of behavior is or isn’t any of my business. Versions of Standing have been defended by Cohen (2006), Smith (2007), Wallace (2010), and Todd (2012). For more discussion (including critical discussion) of this principle, see Bell (2013), Fritz/Miller (2018), and Todd (2019). For a full defense of Reasonable, see Coates (2016). Cf. Bell (2013) on the way a failure to blame risks condonation of serious wrongs. As Derk Pereboom (2013) rightly points out, the practices of moral answerability (e.g., calling on wrongdoers to answer for their transgressions, to repent, to make restitution, etc.) can also serve to respond to wrongdoing. Charles Goodman (2002) has argued that many strands of Buddhism embrace hard determinism, which is the view that blame is never appropriate. If this is right, then Thich Quang Duc might very well have conceived of his protest not as a form of blame, which would be irrational by his own lights, but as a defnitive statement of the value of Buddhism, its adherents, and religious toleration.

The Ethics of Blame

213

19. Gary Watson (1987) makes a related point when he references the ways in which Mahatma Gandhi and Martin Luther King, Jr. held wrongdoers responsible without resentment or indignation.

References Bell, Macalester (2013): “The Standing to Blame: A Critique”, in: Blame: Its Nature and Norms, ed. by Coates/Tognazzini, New York: Oxford University Press, 263–281. Coates, D. Justin (2016): “The Epistemic Norm of Blame”, Ethical Theory and Moral Practice 19, 457–473. Coates, D. Justin/Tognazzini, Neal A. (2013): “The Contours of Blame”, in: Blame: Its Nature and Norms, ed. by Coates/Tognazzini, New York: Oxford University Press. Cohen, Gerald A. (2006): “Casting the First Stone: Who Can, and Who Can’t, Condemn the Terrorists?”, Royal Institute of Philosophy Supplement 58, 113–136. Duff, Antony (2007): Answering for Crime, Oxford: Hart Publishing. Fitzgerald, F. Scott (1995): The Great Gatsby, New York: Scribner. Fritz, Kyle G./Miller, Daniel (2018): “Hypocrisy and the Standing to Blame”, Pacifc Philosophical Quarterly 99.1, 118–139. Goodman, Charles (2002): “Buddhism on Moral Responsibility”, American Philosophical Quarterly 39, 359–372. Kim, Brian/McGrath, Matthew (eds.) (2019): Pragmatic Encroachment in Epistemology, London: Routledge. Lee, Jonathan H. X., et al. (eds.) (2015): Asian American Religious Cultures, Santa Barbara, CA: ABC-CLIO, LLC. McCormick, Miriam Schleifer (2015): Believing against the Evidence: Agency and the Ethics of Belief, London: Routledge. Pereboom, Derk (2013): “Free Will Skepticism and Criminal Punishment”, in: The Future of Punishment, ed. by Nadelhoffer, New York: Oxford University Press, 49–78. Rinard, Susanna (2017): “No Exception for Belief”, Philosophy and Phenomenological Research 94, 121–143. Scanlon, Tim M. (2008): Moral Dimensions: Permissibility, Meaning, Blame, Cambridge, MA: Harvard University Press. Schroeder, Mark (2007): Slaves of the Passions, New York: Oxford University Press. Sher, George (2006): In Praise of Blame, New York: Oxford University Press. Sher, George (2018): “A Wild West of the Mind”, Australasian Journal of Philosophy 97, 483–496. Smith, Angela (2007): “On Being Responsible and Holding Responsible”, Journal of Ethics 11, 465–484. Stanley, Jason (2005): Knowledge and Practical Stakes, New York: Oxford University Press. Strawson, Peter F. (1962): “Freedom and Resentment”, reprinted in: Freedom and Resentment and Other Essays, London: Methuen & Co. Ltd. 1974, 1–25. Todd, Patrick (2012): “Manipulation and Moral Standing: An Argument for Incompatibilism”, Philosophers Imprint 12, 1–18.

214

D. Justin Coates

Todd, Patrick (2019): “A Unifed Account of the Moral Standing to Blame”, Noûs 53, 347–374. Wallace, R. Jay (2010): “Hypocrisy, Moral Address, and the Equal Standing of Persons”, Philosophy and Public Affairs 38, 307–341. Watson, Gary (1987): “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme”, in: Responsibility, Emotions, and Character, ed. by Schoeman, Cambridge: Cambridge University Press, 256–286.

11 How Safe Should We Feel? On the Ethics of Fear in the Public Sphere Sabine A. Döring

1.

Introduction

Although statistics tells us that the number of violent crimes in Germany is declining, people feel fear: objective safety level and subjective feeling of safety seem to depart from each other. Because of this, North Rhine-Westphalia has even launched a study in order to fnd out how safe its citizens feel independently of crime statistics. “Safety,” North Rhine-Westphalia’s Minister of the Interior, Herbert Reul, says, “is not just a fact but also a feeling” (Ministry of the Interior of North RhineWestphalia, Press Portal, January 8, 2019). My concern here is to analyse this so-called “feeling” in order to explain why it is that objective safety level and subjective feeling of safety may come apart. Necessarily, this requires an analysis of the nature of fear in the public sphere since feeling safe means to feel that one avoids the frightening, i.e. the threats or dangers that one perceives in the world. I will not address the empirical question of whether and to what extent there is in fact a gap between objective safety level and subjective feeling of safety. My aim rather is to show that the resistance to facts which fear might display in the public sphere is due to the characteristic function that fear fulflls in this sphere. Like any emotion, fear is basically an experience of something as being in a certain evaluative way. In the case of fear, the thing is presented as frightening in light of the subject’s preference to remain unharmed by a perceived threat. But, in addition to this, fear in the public sphere may itself satisfy subjective preferences and thus be pleasant. Prima facie, this may sound paradoxical: how could a negative emotion like fear be pleasant at the same time? However, pleasant fear actually is an ordinary experience of ordinary people. The prime example is the fear that one feels when watching a scary movie, Alfred Hitchcock’s Psycho or Stanley Kubrick’s Shining, say. While in these examples the target of fear is fctional, pleasant fear may well be felt also at real objects or events. Drawing upon Kant’s analysis of the dynamical sublime, I will argue that a necessary precondition for this is that the subject adopts a

216

Sabine A. Döring

spectator’s point of view: the target of his fear must not pose an immediate threat to him. From a spectator’s point of view, fear may be pleasant if it gives the subject a feeling of superiority relying on that he knows what the threat is and how to respond well to it. We shall see that this feeling is the product of a “fear narrative,” i.e. of a real-life structuring social, cultural, political, economic, religious or otherwise ideological story defning a threat and the appropriate responses to this threat. Fear narratives are the reactions of local groups or communities to feelings of powerless fear and insecurity. By specifying what the threat is and what to do about it, fear narratives reduce felt insecurity and enable a feeling of superiority due to which they may even turn into narratives of enthusiastic heroism. I will establish that it is precisely the preference for reducing insecurity in this way which fear narratives satisfy. By contrast with fear of fctional targets, fear as embodied by real-life structuring fear narratives does not mean to fear something which one knows does not exist. On the contrary, the threat is considered to be a real. However, even though fear narratives in the public sphere thus do not give rise to the notorious paradox of fction, their subjects are liable to irrationality. For, by defning threats and appropriate responses to those threats, fear narratives also defne which facts and arguments have to be taken into account and which could be neglected. This is why objective safety level and subjective feeling of safety may come apart. The question is how we could and should handle this.

2.

Fear: What It Basically Is

Let me begin by explaining what, according to contemporary emotion theorists, fear basically is (for an introduction into the theory of emotion, see, e.g., Döring 2009; Deonna/Teroni 2012). Despite signifcant differences, most emotion theorists agree that, like any emotion proper, fear is an intentional state, i.e. directed at certain targets. Most emotion theorists also accept that fear represents these targets as being a certain way (for criticism, see Shargel/Prinz 2017). The relevant representations must however not assume the form of judgements or beliefs (see Greenspan 1988, especially chapter 2, §1). In any case, even judgementalists concede that emotions differ from typical judgements in that they essentially involve a “what-it-is-like” to experience them (see, e.g., Nussbaum 2001). What is important for my argument here is, frst, that, unlike judgements, emotions may be “recalcitrant,” i.e. persist in the light of better judgement. David Hume (1738/1740: 148) and, before him, Michel de Montaigne (1588, book 2, chapter 12) provide a prime example: your fear of falling may persist although you judge that you are safe. In the same way, you may fnd yourself afraid of an increasing number of violent crimes while knowing that statistics tells the opposite. On

How Safe Should We Feel?

217

a judgementalist account of fear, there is in such cases a rational confict between two beliefs or judgements: you judge that p and that ¬p. However, as Greenspan argues, such incoherence in belief is too strong a diagnosis for the confict in question. The judgementalist would need a special reason to violate what Davidson (1973: 19) has called the “principle of charity”: a methodological principle which requires that we seek to avoid attributing irrationality and interpret a subject’s attitudes so as to fnd her to be rational to the greatest extent possible. The only reason for violating this principle “seems the assumption that judgmentalism is true” (Greenspan 1988: 18). The second important point for my argument here is that, due to having a representational content, emotions are assessable for correctness – or, alternatively, fttingness or appropriateness – and that the subject can be assessed for rationality. For example, although fearing climate change does not mean to judge that climate change is frightening, it does mean to experience climate change as frightening (in a sense to be explained) and is thus subject to correctness conditions. In the same way, fearing an increase of violent crimes by immigrants or by right-wing extremists is to feel threatened by an alleged decrease in internal security which may or may not correspond to the facts. As regards the phenomenon of emotional recalcitrance, this opens up the possibility that it may equally be the judgement, rather than the emotion, which is in error: emotions (their representational content) may be inappropriate, but, sometimes, they may teach our judgements better (see Döring 2010). In order to capture the role of fear in public discourse, we need to understand how such fear is acquired. Drawing upon Ronald de Sousa’s approach, I presume that emotions and their defning properties or concepts are acquired in so-called “paradigm scenarios” (de Sousa 1987: 182). Underlying this approach is a cognitive account of emotion, the Formal Object View. According to this view, different emotion types are defned by certain properties or concepts which restrict and thereby determine the class of things a particular type of emotion can be directed at (see Kenny 1963: 187–194; de Sousa 1987: 122–123; Teroni 2007). These properties or concepts are called the emotions’ “formal objects” (or, alternatively, their “core relational themes”; see Prinz 2004). The Formal Object View starts from the insight that emotions necessarily (conceptually) imply nonarbitrary evaluations (or appraisals) of particular targets. Thus, someone’s fear of something is only intelligible if the subject somehow sees the thing in question as frightening. If the subject were claiming to fear the thing while at the same time denying that there is anything frightening about it, we would be conceptually excluded from understanding. In the same way, experiencing indignation at something is intelligible as indignation only if the subject somehow sees that thing as unfair or unjust (where this does not imply that the relevant “seeing” assumes the form of a judgement). In this way, and only in this way, can we distinguish indignation from mere

218

Sabine A. Döring

anger whose formal object is the property of being annoying. Therefore it has been argued that the evaluative property which has to be ascribed to an emotion’s target in order to render the emotion intelligible – the emotion’s formal object – defnes the emotion in question. De Sousa’s Paradigm Scenario View offers an explanation of how emotions get associated with characteristic targets via their defning properties (formal objects). As the term “scenario” already indicates, on this view, emotions are not directed at particular things as such but only as embodied by complex situations. For example, fear of gorillas is not typically fear of dead or deadly sick gorillas, drugged gorillas, baby gorillas or the like. Even specifying vital properties of particular things will not do – for we do also not fear, say, healthy, adult and aggressive gorillas as such but only in situations where they appear to present a danger to us (or others). Typically, we do not fear a (healthy, adult and aggressive) gorilla whom we believe to be securely caged. The situation changes signifcantly when we suddenly realise that the door to the cage has been left open: only then will we feel fear of the gorilla. Paradigm scenarios thus are always complex situations, and so are the contents of the emotions representing these situations. Emotional content is complex (and not simple; see also Goffn 2018, chapter 2; Döring 2019). Paradigm scenarios as introduced by de Sousa involve two aspects: frst, a situation type providing the characteristic targets of the specifc emotion-type, and second, a set of characteristic responses to the situation. According to de Sousa, beyond the stage of “basic emotions” (or “affect programs”) we acquire and further develop emotions in such paradigm scenarios. Basic emotions are supposed to be inborn traits and as such no emotions proper, but mere dispositions to respond in certain ways. Such response dispositions become emotions proper only due to their association with characteristic targets in paradigm scenarios. The frst stage in this association process is to categorise characteristic objects or events which normally cause a certain emotional response: things which normally cause fear are subsumed under the category of the frightening; things which normally cause anger are subsumed under the annoying; things which normally cause sadness are subsumed under the sad; and so forth. The properties which emotions ascribe to their targets thus are response-dependent properties: things aren’t frightening, annoying or sad independently of our responding to them with fear, anger or sadness. Let us be clear that this does not mean that we just “project” such properties onto things as if those properties actually belong to them. The corresponding evaluative concepts (frightening, annoying, sad etc.) will apply to these things only in case they have properties in virtue of which they would elicit a relevant response in normal subjects in standard conditions. Whether emotional concepts such as frightening, annoying or sad apply to something is determined by the paradigm scenarios of fear, anger and sadness, respectively.

How Safe Should We Feel?

219

The second stage of associating emotions with characteristic paradigm scenarios then goes signifcantly beyond the frst stage in that, as we may say, normality makes way for normativity. On the second stage, society and culture go beyond the original paradigm scenarios. Not only are the original paradigm scenarios extended so as to associate emotional responses with new characteristic objects or events. Furthermore, new emotional responses – such as, according to Adam Smith, admiration – emerge and are associated with characteristic objects or events in order to account for cases which cannot be captured in terms of the original repertoire. How original paradigm scenarios are socially and culturally extended is illustrated by disgust: the paradigm scenarios of the aesthetically or morally disgusting differ signifcantly from that of the pathogenetically disgusting: while pathogenetic disgust is experienced in relation to the sense of taste and is directed at potentially harmful substances, aesthetic disgust is directed, e.g., at kitsch, and moral disgust pertains behaviours such as lying, theft, murder or rape. By stating that, on the second stage, normality makes way for normativity I mean that, from now on, the emotional responses of even normal humans in standard conditions can be mistaken: they can be unftting (or inappropriate or incorrect). For example, a disagreement about whether homosexuality is (morally) disgusting can hardly be settled by appeal to whose responses are more normal: this is not what interests us here. In the same way, in assessing fear of violent crimes by migrants, the question is not whether it is normal to feel threatened by an alleged decrease in internal security. The question is rather whether the fear response is ftting. To say that the number of violent crimes by migrants is frightening means to endorse fear of violent crimes in some way. It is to say that we ought to fear violent crimes by migrants, because they do present an actual threat – while agreeing that evaluative properties such as threatening or frightening cannot be understood except by way of human responses. Normativity in this sense, i.e. in the sense of some Fitting Attitude Theory of Value (see, e.g., D’Arms/Jacobson 2006), is the natural result of original paradigm scenarios becoming subject to social and cultural discourse which necessarily starts from the original repertoire while transcending it and constantly putting it to the test. I have proposed that the Paradigm Scenario View suggests a Complex Content View (as opposed to a Simple Content View): the concepts (or properties) associated with different emotion types typically subsume a whole range of concrete individual situations which might be very different from one to the other. Furthermore, the Paradigm Scenario View makes it seem highly unlikely that, within emotional content, evaluation could be disentangled from nonevaluative description. I shall discuss this point in more detail in the fnal section. The process whereby emotions proper – i.e. felt evaluations with a representational content – are constructed out of genetically fxed response dispositions begins in the frst

220

Sabine A. Döring

few months of life and continues indefnitely. Importantly, beyond basic responses such as fight or attack, this process requires complex linguistic skills because paradigm scenarios are typically learned and educated by means of narratives, be they mythologies, legends, fairy tales, novels or other stories. Since such narratives are, by their very nature, designed to offer orientation in a complex world, by their very nature, they interweave evaluation with nonevaluative description and, perhaps, even fact with fction. Let me give some examples.

3.

Fear Narratives

So far I have argued that, because it has representational content and necessarily involves an evaluation of its target, fear can be assessed to be ftting or unftting, while the inevitable and indispensable reference points of such assessment are the paradigm scenarios of fear of a particular society and culture. Concluding the preceding section, it has been pointed out that, typically, such paradigm scenarios are embodied by narratives. As an example, we may consider the fear narrative of the wolf (canis lupus) and contrast it with its counter-narrative. The wolf plays a central role in the narratives of numerous peoples which illustrate the ambivalent attitude of humans towards the wolf. On the one hand, wolves are seen as vicious killers of livestock and as a threat for human safety. On the other hand, the wolf is admired and even worshiped as a symbol of the wild and untamed nature, mysterious, powerful and impressive. Both narratives are perspective-based constructions of what the wolf means to human beings. They help to structure and defne the reality of a local group, be it in terms of a social, cultural, political, economic, religious or otherwise “ideological” story, i.e. in terms of a story that is based on certain ideas and ideals. Narratives essentially contribute to establishing the identity of local groups and thus affect perception and also potential action. By their very nature they embody paradigm scenarios created to evoke corresponding emotional responses. We must here distinguish between narratives in the public sphere and fctional narratives. By contrast with fctional narratives, narratives in the public sphere – such as, for example, the fear narrative of the wolf and its counter narrative – are supposedly factual narratives. This is not to deny that fction has an indispensable role to play in the education of emotion. On the contrary. According to Iris Murdoch, “the study of literature” even is “the most essential and fundamental aspect of culture” since it is “an education in how to picture and understand human situations” (Murdoch 1970: 134; see also de Sousa 1987: 184). Factual narratives in the public sphere typically draw upon fctional narratives. For example, the story of the Little Red Riding Hood is a cautionary tale that overtly warns about talking to strangers; yet by making use of the “big bad wolf”

How Safe Should We Feel?

221

as the generic archetype of a menacing predatory antagonist it also symbolises wolves in a certain way, namely, as a threat to human safety. Factual narratives in the public sphere pick up on the symbols and employ the paradigm scenarios of fctional narratives. This raises an important question, namely, the question of whether and to what extent narratives could invent paradigm scenarios for emotional responses which could then be authentically applied to real life. To the extent that this is possible, supposedly factual narratives might interweave fact and fction. I shall return to this question in the fnal section. For the present let us note that, as pointed out in the introduction already, fear narratives in the public sphere are (supposedly) factual narratives. As such, they are not susceptible to the paradox of fction. As initially stated by Colin Radford (1975), this paradox is grounded in three true yet incompatible propositions: (1) in order to respond emotionally to a character or event, one must necessarily believe that the character really exists or that the event has really taken place; (2) such existence beliefs are lacking when we knowingly partake of works of fction; (3) works of fction do in fact move us sometimes. While proposition (2) applies to fctional narratives, it does not apply to supposedly factual narratives: narratives supposed to be factual rather are defned by being held true. Therefore, social, cultural, political, economic, religious or otherwise ideological narratives in the public sphere are real-life structuring stories. By their very nature, they are designed to offer orientation in a complex world, which is to say that they are perspective-based and thus typically biased. Let us stick with the example of the wolf. Over the last years, the wolf has become popular within certain ideologically charged milieus. Publications on the comeback of the wolf to Germany peaked and brought along a real internet hype on the wolf. Many of these publications focus on the supposed positive impact the wolf will have on our ecosystem, but typically go beyond that. Furthermore, wolves are pictured as a vulnerable species which belongs to an unspoiled nature and wildlife but has wrongly been hunted to extinction in middle Europe. Within the framing of this narrative, the return of the wolf to Germany is seen as a sign of nature’s being “back in order.” Critics object that the social aspect, namely, that the wolf is making its comeback in one of the most densely populated European countries, cannot be ignored and deserves some serious attention. Such critics point to farmers who struggle to protect their livestock against wolf attacks. Like those they criticise, they typically incorporate their critique into a narrative in which wolf advocates prioritise the protection of vicious killers over the protection even of our children’s safety. The narratives of both sides frame the wolf’s comeback according to the perspective of and in terms of the paradigm scenarios of their respective ideological group and milieu. Since they thereby infuence and shape public opinion on human-wildlife interactions, the question naturally

222

Sabine A. Döring

arises on which factual and scientifc evidence the two parties base their plans, policies, and strategies. How could different narratives and the paradigm scenarios they involve be tested against the facts? The diffculty in doing so is that facts and scientifc data are themselves integrated into the narratives: each narrative is also a story about which facts should be taken into account and which arguments, events or experiences are of interest for understanding the issue (see Buijs et al. 2013). The dispute already begins with competing concepts of nature or territory (see Nie 2001; Figari/Skogen 2011): What type of nature do we talk about? Is that nature in harmony with the wolf? Whose territory is dealt with? This is why fear narratives like that of the wolf, and the feeling of insecurity from which they emerge, are likely to resist even statistics telling that the subjective feeling of safety does not match the objective level of safety: the vast majority of wolves will probably never show any aggressive behaviour towards people (see Linnell et al. 2002). Before fguring out in more detail why and in which way fear narratives tend to become resistant to counter-evidence, let us consider two further examples. What frst comes to one’s mind today most probably is the fear narrative of right-wing populism, within which the greatest threat to states and national communities is not government tyranny or inequality or climate change, but immigration. In its most extreme form, this narrative pushes the vision of a mass invasion of primitive and highly aggressive “knife men” who have come to destroy the Western world and are herein assisted by the current political leaders (Alice Weidel, speech at the German Federal Parliament, May 16, 2018). Another no less infuential fear narrative is told by the climate change advocates whose most prominent fgure, Greta Thunberg, wants us “to panic” in view of the threat of climate apocalypse. Like the fear narrative of the wolf, both the fear-of-immigration narrative and the fear-of-climate-change narrative have their counter-narratives; and like the fear narrative of the wolf, they are by their very nature disposed to immunise themselves to counterevidence. Yet, how could they be so powerful and successful in doing so?

4.

The Spectator’s Feeling of Knowing What Really Matters

In a recent tweet, German journalist Henning Sußebach suggests that “from a spectator’s perspective, we would all be up for a Hard Brexit” (Henning Sußebach on Twitter, October 4, 2019; my translation). Whatever we may think about a Hard Brexit, Sußebach here rightly emphasises the difference it makes for one’s emotional experience whether one is an agent of, or actually caught in, a political event or whether one is a spectator not directly affected by that event. The signifcance of the spectator’s point of view for human emotional life has already been emphasised by Immanuel Kant. Kant regards

How Safe Should We Feel?

223

emotions as at best highly suspect noncognitive sources of motivation, and yet he praises feelings like respect, enthusiasm or the experience of the sublime as necessary for morality (see Sorensen/Williamson 2018). He does so because he attributes to the latter feelings a distinctly moral content, consisting in their internal relatedness to freedom which the subject may strive to make actual. Although Kant’s account thus differs substantially from the cognitive emotion-theoretic account underlying my argument, his analysis of the aesthetic experience of the dynamical sublime contains signifcant insights about the nature of fear. In particular, it reveals how negative fear could nonetheless be pleasurable (see Kant 1790, §28). The “dynamical sublime” names experiences like violent storms or earthquakes in which we are overwhelmed by natural forces. In order to experience something as dynamically sublime, Kant says, it must be represented as frightening. And yet, Kant notes, fear as involved in the dynamical sublime is a pleasurable experience: its pleasantness consists in the subject feeling his superiority over nature. Although the experience of the dynamical sublime presents its target as frightening, namely, as a threat to the subject’s physical existence, according to Kant, it is a pleasurable experience because it makes the subject feel that his freedom remains unaffected even when his physical existence has to submit to the dominion of natural forces. The sublimity thus in truth belongs to the subject’s freedom which Kant holds to be unassailable to the forces of nature. Such a conception of freedom as being outside the order of nature, but demanding action upon that order, is the core of Kant’s moral theory. The point to extract from this theory for my argument here is that, provided that the subject’s viewpoint is that of a safe spectator, fear may be pleasurable due to implying a feeling of sublimity or superiority. However, while in the case of the dynamical sublime pleasure emerges from experiencing one’s freedom as unaffected by natural forces, I hold that in the case of fear narratives, pleasure is due to experiencing that the narrative makes one grasp what the threat really is and how to respond well to it. Let me emphasise that I am not thereby committing myself to Kant’s view that emotions, if at all, could be appropriate (or correct or ftting) only if they somehow are internally connected with, and required for, what is prior defned as morality. Rather, on the emotion-theoretic account I am defending, an emotion’s appropriateness solely depends on whether or not it fts its target. Accordingly, whether fear of wolves or of immigrants or of climate change is appropriate is not a moral question. To condemn fear of immigrants or endorse fear of climate change for moral reasons would be to commit what Justin D’Arms/Daniel Jacobson (2000) have dubbed “the moralistic fallacy.” It would be to confate moralistic presumptions with the appropriateness conditions of emotional

224

Sabine A. Döring

responses. For example, the moralistic fallacy would be committed if someone were to grant that, due to the government’s politics of asylum and refugees since 2015, the number of potential islamist attackers has increased in Germany, while at the same time dismissing fear of islamist attacks as inappropriate on the grounds that the politics of asylum and refugees is humane and thus morally right. “Inappropriate” here means “immoral.” However, emotions considered to be immoral may yet ft their target and be appropriate in this sense: if it were true that the threat of an islamist attack in Germany did increase, a corresponding increase of public fear of such an attack would be appropriate. What makes narratives in general and fear narratives in particular so irresistible is, on my view, that they are real-life structuring stories which do not merely offer orientation in a complex world, but also ground a feeling of superiority relying on that, thanks to the narrative, one grasps what the threat is and how to handle it. As has been shown in the previous section, fear narratives start from facts such as wolf attacks on livestock or violent crimes by immigrants or serious consequences that climate change due to human activities is likely to have. But then they go signifcantly beyond such facts by framing wolves or immigrants or climate change in certain ways. Today, narratives often emerge as a reaction to feelings of powerless fear and insecurity in light of the challenges of globalisation. Fear narratives are reactions to these feelings: they soothe these feelings by defning the danger or threat and by offering instructions against it. This is how they give their subjects a feeling of superiority relying on the fear narrative’s moral. This also explains why narratives may be very effcient in immunising themselves against counter-evidence up to being stories about which evidence has to be taken into account and which could be neglected. Feeling superior due to grasping what the threat really is and how to fght it set in the place of powerless fear and insecurity is what makes fear narratives so powerful and successful. Drawing upon Kant’s analysis of the dynamical sublime, I have argued that a necessary precondition for this is that the subject of such a narrative adopts a spectator’s point of view: the target of his fear must not pose an immediate threat to him. This is in accord with the fact that many people who engage in fear narratives about wolves have never encountered any wolf in the wild and who lead a life that makes such an encounter highly improbable in any case; or that the most toxic fear narratives about immigrants are often found in places where the actual number of immigrants is extremely low; or that fear narratives about climate change seem most prominent in countries which are least affected by it so far (like Germany). However, as is also supported by Kant’s analysis, to adopt a spectator’s viewpoint is not to regard the target of one’s fear as merely fctional. On the contrary, their alleged nonfctional, real character makes the fear these targets evoke even stronger and gives it

How Safe Should We Feel?

225

more depth in the eyes of the subject. In other words, it is constitutive of fear narratives in the public sphere that the target of fear is believed to really exist and is not generated by a belief about what is merely fctionally the case. Therefore, the fear in question does not run the risk of being paradoxical or irrational from the outset (Radford 1975). By clear contrast with the fear we may experience, say, of the White Walkers and Wights in David Benioff’s and Daniel Brett Weiss’s fantasy drama television series Game of Thrones, fear of migrants, even when they do not pose an immediate threat to the subject, is integrated into the subject’s constant real-life experience. This is also why, other than fear of fctional characters of events, fear in the public sphere has motivational force. Like Kantian moral emotions, fear as embodied by social, cultural, political, economic, religious or otherwise ideological narratives can be and quite often is transformed into a motive for action. Its motivational force is further strengthened by its phenomenal dimension which I have characterised as a feeling of superiority which the narrative sets in the place of powerless fear and insecurity. Like all narratives, fear narratives are stories with a moral. They are always stories about what really is a threat. In addition to this, they are always stories about how one ought to respond to the threat – as, e.g., by shooting wolves, closing borders or banning climate-damaging behaviour. Because of this, and this is my central claim here, fear, when embodied by a social, cultural, political, economic, religious or otherwise ideological narrative, may reduce felt insecurity. It does so precisely because the narrative is also a story about what the threat is and how to respond to it and thus grounds a feeling of superiority relying on that, thanks to the narrative, one grasps what really matters. Accordingly, fear in the public sphere differs from basic fear in that it is not merely an evaluation in light of the subject’s preferences: it also satisfes subjective preferences and thus is pleasurable fear. Insofar as the pleasure is grounded in a feeling of superiority, fear narratives may mutate into narratives of enthusiastic heroism. A prime example of a narrative of the latter kind is provided by Sebastian Haffner. In his Geschichte eines Deutschen Haffner depicts how the First World War was experienced by his German contemporaries at home, i.e. by those born between 1900 and 1910 who were too young to actively participate in the war and yet old enough to make up their minds about it (see Haffner 2000). Haffner analyses how his contemporaries, instead of fearing and abhorring the war and its atrocities, got caught in a narrative of enthusiastic heroism culminating in Germany’s undeniable and earned ultimate victory. Necessarily, this narrative requires its subject to be in a spectator position, and not to be agent of, or to be caught in, the actual war events. By grounding a feeling of pleasant superiority due to conveying that it is clear both what the threat is and how to fght it, this

226

Sabine A. Döring

narrative grounds enthusiastic heroism. In this way it enables pleasant fear, such as pleasant shivers when learning about horrible war events. Haffner illustrates how this narrative structured the lives of many of his contemporaries and how, when the First World War came to its, from their perspective, unexpected end, they remained in the grip of the narrative, eager to fnd new ftting targets for their emotions. With subtle irony, Haffner characterises the emotional situation of these men in the aftermath of the First World War as awkward, impoverished, deprived, frustrated and bored. Being used to get their whole purpose in life, all targets for deeper emotions, from the public reporting on the war events, Haffner says, his contemporaries became morose and waited eagerly for the frst disturbance, the frst setback or incident, so that they could put this period of peace behind them and set on some new collective adventure. Haffner in this way identifes the real-life structuring narrative of enthusiastic heroism about the First World which compellingly captured his contemporaries’ minds as a central source of National Socialism. As this shows, the feelings embodied by narratives are feelings not just about something in the world, but feelings of being situated in the world and with other people in a certain way which shapes all experience, thought and activity. As such, these feelings might qualify as what Matthew Radcliffe has characterised as “existential” feelings (see, e.g., Ratcliffe 2005; see also Slaby/Stephan 2008). Fear as embodied by reallife structuring social, cultural, political, economic, religious or otherwise ideological narratives is highly likely to have such an existential phenomenal dimension. I will not discuss this issue any further here, but let me point out that fear as understood here is not a “mood.” Moods are supposed to be nonintentional, whereas it is constitutive of fear narratives that they specify the threat at which fear is (to be) directed. I shall now go back to the question raised of whether and to what extent, by the creation or invention of paradigm scenarios, narratives indissolubly interweave evaluation and nonevaluative description and, perhaps, also fact and fction.

5.

Narratives Create Thick Concepts

I have argued that fear and other emotions, together with the evaluative concepts for their defning properties (formal objects) – such as frightening, annoying or sad – are acquired within paradigm scenarios and that the characteristic source of such paradigm scenarios are narratives. As we have now seen, narratives function so as to provide orientation in a complex world, thereby grounding a (possibly existential) feeling of superiority relying on the “moral of the story,” so to speak. In the case of fear narratives in particular, fear thus becomes pleasurable by reducing felt insecurity through substituting powerless fear by a story about what the danger or threat is and about how to respond to it in the right

How Safe Should We Feel?

227

way. This makes fear narratives powerful and successful up to becoming stories about which facts do count as evidence and which are mere disturbing noises (so to speak). Let me now explain how fear narratives interweave evaluation and nonevaluative description, i.e. how they create “thick” concepts. By providing the relevant paradigm scenarios, fear narratives tell us what we ought to fear because it really is frightening (or poses a danger or threat). Some have argued that emotional (or “affective”) concepts like frightening or admirable are so-called “thick” concepts (see Tappolet 2004). Intuitively, thick concepts differ from two other kinds of concepts. On the one hand, they seem to be evaluative concepts and thus to differ from nonevaluative purely descriptive concepts like scientifc concepts or colour concepts. (I am here presuming that the descriptive is nonevaluative.) On the other hand, the evaluation they contain seems also be different from that of “thin” concepts like good or bad: thick concepts seem to convey more information or to be richer in content than thin concepts by somehow holding together evaluation and nonevaluative description. For example, if one says that someone is generous one says more about him than one would if one only said that he is good, namely, that he shows a readiness to give more of something, especially money, than is strictly necessary or expected. Virtue concepts like generous or courageous are the typical candidates for thick concepts, but thick concepts are also illustrated by concepts like cruel, industrious, lewd or admirable. In accord with my argument here, my primary focus is on emotional concepts and, in particular, on the concept frightening (or threat or danger). Similar to virtue concepts, such concepts are richer in content than thin concepts. The message conveyed by stating that wolves are frightening or that they pose a danger or threat is not just that they are bad; it is that wolves are likely to cause serious harm or damage. To say that emotional concepts are thick is to say that they are inherently evaluative: their meaning contains an evaluation. The most prominent argument for thick concepts being inherently evaluative has been formulated by John McDowell (see McDowell 1981; see also Williams 1985). According to McDowell, if thick concepts were not inherently evaluative, it would be possible to “disentangle” a nonevaluative descriptive and an evaluative “component.” (A simple way of expressing this would be: “x is D and therefore x is E,” where D is a nonevaluative description and E is a thin evaluation grounded on that description.) Since for McDowell it is clear that disentanglement is not possible, on his account, thick concepts are inherently evaluative. In its original version, the Anti-Disentangling Argument relies on the lacking ability of an outsider to master the thick concepts of a community whose evaluative outlook is unfamiliar to him. Such an outsider, so the story goes, would not be able to apply these concepts correctly, i.e. to discriminate correctly between things that do and do not fall under them; he would

228

Sabine A. Döring

also, when presented with a selection of instances, fail to recognise what these instances have in common (or to see the “shape” of the concept). Most people share this intuition, i.e. believe that mastering thick concepts requires to adopt a particular group’s evaluative perspective, and to feel what they feel. But why is this so? (I am here skating over objections, e.g., by Blackburn 1992; Väyrynen 2013). For the concept frightening, to which I confne myself here, the answer has already been partially given. Emotional concepts are, together with their corresponding emotional responses, acquired within paradigm scenarios embodied by real-life structuring stories. The essential function of such stories is to substitute powerless fear and felt insecurity by specifying what the threat is and how to respond well to it. Fear narratives therefore are always evaluative, and the paradigm scenarios they provide are such that they ft the overall evaluation and moral of the story. This is to say that the overall evaluation of the story determines the extension of the relevant emotional concepts, which is why this extension cannot be determined independently of the story. This also explains why an outsider fails to master the relevant concepts: local groups are defned, or get their identity, at least partly through their social, cultural, political, economic, religious or otherwise ideological narratives; outsiders are those unfamiliar with these narratives; therefore, they cannot but fail to grasp the paradigm scenarios embodied by or embedded in these narratives. Mastering thick concepts embedded in narratives is made even more diffcult by the interconnectedness of these concepts. For example, the comparison of the fear narrative of the wolf with its counter-narrative indicated that seeing wolves as a threat goes hand in hand with sympathy for farmers struggling to protect their livestock against wolf attacks. Wolf activists, by contrast, are likely to experience sympathy with the livestock instead and to feel anger and indignation about farming as systemic mistreatment of animals. We have seen, furthermore, that, within narratives, even apparently nonevaluative purely descriptive concepts like nature or territory may become inherently evaluative: typically, a wolf activist’s concept of nature is not only different from that of a wolf opponent, namely, nature in harmony with the wolf; it is also a concept of “real” or “true” nature or of “nature back in order.” The same point could be exemplifed at the fear narrative of immigration or of climate change as compared with their counter narratives. Fear narratives are, in the frst place, stories about what things we ought to fear because they are really frightening, and about what we ought to and could do to gain or regain safety. They are reactions to feelings of powerless fear and insecurity and replace these feelings by a feeling of superiority relying on the narrative’s moral. Wolf opponents feel certain that wolves are about to attack innocent children and feel superior to those who just fail to see the threat. Right-wing populists perceive their opponents as committed to cultural and economic suicide. And

How Safe Should We Feel?

229

climate activists are convinced that immediate action, be it legal or not, is required as all life on earth is on the way to extinction due to humaninduced global warming.

6.

Fear Narratives in Normative Dialogue

Fear narratives in the public sphere indissolubly interweave evaluation and nonevaluative description. Do they also interweave fact and fction? There is reason to suspect this because, frst, they are also stories about which evidence counts and which doesn’t. Secondly, fear narratives, even though they are supposed to be real-life stories that correspond to the facts, draw upon fctional narratives. The question is whether and to what extent narratives could invent paradigm scenarios for emotional responses which could then be authentically applied to real life. My provisionary answer is that objective human experience sets a limit to the invention of paradigm scenarios and thus to factual narratives turning into fction. As pointed out, not just any situation qualifes as frightening or as a paradigm scenario of fear, but only those situations which do have properties in virtue of which they would elicit actual human responses of fear. The invention of paradigm scenarios is thus restricted by the actual responses of actual human subjects. Factual fear narratives may employ paradigm scenarios provided by fction only to the extent that the fctional situations presented qualify as paradigm scenarios of fear which could be transferred to real-life situations. Even if there is such a limit, it follows from what I have said so far that normative discourse about the fttingness and rationality of fear must start out thick, i.e. from where actual agents are, with the actual social, cultural, political, economic, religious or otherwise ideological understandings they have. There is, in other words, no neutral bystander point of view from which we could assess the fttingness and rationality of fear, for we are ourselves always caught up within the paradigm scenarios of fear (and of other emotions) of our own local community (see McDowell 1981). One reason why constructive dialogue about fear’s appropriateness often fails is, precisely, that, instead of conceding this, competing groups or members of local groups insist to be objectively right. This insistence is due not least to the feeling of moral superiority grounded by one’s narrative’s inherent claim to tell what the threat really is and how to respond to it in the right way. Typically, this leads to what has been called “the moralistic fallacy.” For example, while in Europe pro-immigrationists accuse antiimmigrationists of betraying Europe’s multicultural and tolerant ideals, anti-immigrationists declare pro-immigrationists guilty of saddling the European project with impossible expectations. Both sides seem unable to grasp what the other side perceives as a threat. The discussion thereby degenerates into a shouting match in which neither side hears the other. Supporting my argument, the two sides apply different thick concepts or

230

Sabine A. Döring

terms which “signal” the underlying evaluative outlook as a whole. Thus, pro-immigrationists describe activists who engage in picking up migrants in the Mediterranean as “sea rescuers,” whereas anti-immigrationists call them “smugglers (of refugees)” or “facilitators (of illegal migration).” Both sides challenge their opponents’ evaluative outlook by challenging their thick concepts or by challenging how they speak. Those who reject the term “sea rescuer” thereby attack the evaluative outlook of pro-immigration; conversely, those opposing the term “smuggler” or “facilitator” thereby challenge the anti-immigrationist evaluative perspective. If we do not want to leave things at this, but return to a rational discussion deserving this name, we will have to take the trouble to go beyond our own evaluative outlooks and try to understand competing evaluative outlooks and the impact they have on what their proponents perceive as threats or dangers. On the one hand, this requires to accept and to deal with evidence which speaks against or even shakes the narratives in which one lives. On the other hand, this means to respect one’s opponents with their competing narratives, even when one’s own narratives offer felt superiority. After all, the other side feels superior as well. This is not to deny that there are limits: what if, for example, the narratives of some anti-immigrationists do contain fascist ideas? Do thick concepts imply relativism in the sense that the appropriateness of all evaluative and moral claims depends on the point of view of the local group in which these claims are made (see Williams 1985)? I think that there might be a normative discourse which goes beyond not only challenging people’s thick concepts but also beyond making these concepts explicit so as to get the underlying evaluative outlooks clear and to perhaps understand them better. At the beginning of this section, I have suggested that objective human experience sets a limit to the creation of paradigm scenarios and their corresponding thick emotional concepts. Only those situations qualify as possible instances of the frightening which, due to certain nonevaluative descriptive features, would elicit actual human responses of fear. The question is whether it is possible to identify paradigm scenarios of fear that are universally shared. If this is possible, these paradigm scenarios could be the starting point of a universal human dialogue about appropriate human fear responses. Martha Nussbaum, for one, holds that this is possible (see Nussbaum 1987). In her attempt to establish a nonrelative virtue ethics, Nussbaum identifes a number of supposedly objective spheres of human experience, including the fear of important damages such as, in particular, death. These spheres of experience are “objective” in the sense that more or less any human life faces them, and they require an answer to the question of what it means to respond well in that sphere. Identifying these spheres makes up the frst step of Nussbaum’s account; providing the required answer the second. The point of Nussbaum’s argument is that, being the shared experience of all human beings, the relevant spheres invite and ground a

How Safe Should We Feel?

231

universal dialogue about what it means to respond well in these spheres. There might not be a single correct answer, but many potentially correct answers, depending on local conditions. Furthermore, the supposedly shared human experiences could in fact be differently constructed by different local groups. Nonetheless, Nussbaum’s nonrelative virtue-ethicist account might allow for universal normative dialogue. Fictional narratives might contribute to this dialogue by exemplarily presenting possible human responses to objective spheres of human experience which might then fnd their way into factual narratives. Assuming that fear of important damage is an objective sphere of human experience, the question is, frst of all, what, apart from death, counts as an important damage. Different fear narratives have to be compared as regards the import of the potential damage that is feared. The next question then is how to respond well to that potential damage. Insofar as responding well to damages means to respond courageously, this would be to integrate fear and the frightening into virtue ethics. On a virtue-ethicist account, the question of how safe we should feel, would depend on what is a courageous response. For the present, this is only a suggestion – and a hope. It has not yet been my aim to show that grounding virtue ethics in objective spheres of human experience does give it a legitimate claim to universality. What has been my aim, however, is to show that objective safety level and subjective feeling of safety come apart due to the characteristic function that fear in the public sphere fulflls. To sum up: fear in the public sphere is embodied by the fear narratives of local groups which, partly at least, defne themselves via these narratives. Fear narratives originate from feelings of powerless fear and insecurity which today are often caused by the challenges of globalisation. By defning what the threat is and how to respond to it, fear narratives reduce felt insecurity and ground a feeling of superiority which is due, precisely, to the narrative’s defnition of the alleged threat and of the allegedly appropriate response to it. Narratives thereby tend to immunise themselves against counter-evidence by also telling which evidence qualifes as such and which doesn’t. Furthermore, the paradigm scenarios they embody create thick concepts by determining the extension of these concepts in such a way that it fts the moral of the story. Nevertheless, there is a limit to turning fact into fction set by objective human experience. The latter might even ground a universal dialogue about the appropriateness and rationality of fear – or so I hope.

References Blackburn, Simon (1992): “Through Thick and Thin”, Proceedings of the Aristotelian Society, Supplementary Volume 66, 284–299. Buijs, Arjen E., et al. (2013): “Framing: de strijd om het nieuwe natuurbeleidsverhaal”, Landschap: tijdschrift voor landschapsecologie en milieukunde 30, 33–41.

232

Sabine A. Döring

D’Arms, Justin/Jacobson, Daniel (2000): “The Moralistic Fallacy: On the ‘Appropriateness’ of Emotion”, Philosophy and Phenomenological Research 61, 65–90. D’Arms, Justin/Jacobson, Daniel (2006): “Sensibility Theory and Projectivism”, in: The Oxford Handbook of Ethical Theory, ed. by Copp, New York: Oxford University Press, 186–218. Davidson, Donald (1973): “On the Very Idea of a Conceptual Scheme”, Proceedings and Addresses of the American Philosophical Association 47, 5–20. Deonna, Julien/Teroni, Fabrice (2012): The Emotions: A Philosophical Introduction, London: Routledge. de Sousa, Ronald (1987): The Rationality of Emotion, Cambridge, MA: MIT Press. Döring, Sabine (2009): Philosophie der Gefühle, Frankfurt: Suhrkamp. Döring, Sabine (2010): “Why Be Emotional?”, in: The Oxford Handbook of Philosophy of Emotion, ed. by Goldie, New York: Oxford University Press, 283–301. Döring, Sabine (2019): “What Is Expressed When Emotions Are Expressed in Art?”, Grazer Philosophische Studien 96, 361–380. Figari, Helene/Skogen, Ketil (2011): “Social Representations of the Wolf”, Acta Sociologica 54, 317–332. Goffn, Kris (2018): Emotional and Affective Representation: Reliability, Complexity and Aesthetics, n. pag., URL = . Greenspan, Patricia (1988): Emotions and Reasons: An Inquiry into Emotional Justifcation, London: Routledge. Haffner, Sebastian (2000): Geschichte eines Deutschen. Die Erinnerungen 1914– 1933, München: DVA. Hume, David (1738/1740): A Treatise of Human Nature: Being an Attempt to Introduce the Experimental Method of Reasoning into Moral Subjects, ed. by Selby-Bigge/Nidditch, Oxford: Clarendon Press 1978. Kant, Immanuel (1790): Kritik der Urtheilskraft, gesammelte Schriften Bd. 1–22, Preussische Akademie der Wissenschaften, Bd. V (2002). Kenny, Anthony (1963): Action, Emotion and Will, London: Routledge/Kegan Paul. Linnell, John D. C., et al. (2002): The Fear of Wolves: A Review of Wolf Attacks on Humans, Trondheim, Norway: NINA NIKU Stiftelsen for naturforskning og kulturminneforskning. McDowell, John (1981): “Non-Cognitivism and Rule-Following”, in: Wittgenstein: To Follow a Rule, ed. by Holtzmann/Leich, London: Routledge, 141–162, reprinted in his Mind, Value, and Reality, Cambridge, MA: Harvard University Press 1998, 198–219. Montaigne, Michel de (1588): Les Essais, Book 2, chapter 12, Paris 2002. Murdoch, Iris (1970): The Sovereignty of Good, London: Routledge. Nie, Norman H. (2001): “Sociability, Interpersonal Relations, and the Internet: Reconciling Conficting Findings”, American Behavioral Scientist 45, 420–435. Nussbaum, Martha C. (1987): “Non-Relative Virtues: An Aristotelian Approach”, Midwest Studies in Philosophy 13, 32–53. Nussbaum, Martha C. (2001): Upheavals of Thought: The Intelligence of Emotions, Cambridge: Cambridge University Press. Prinz, Jesse (2004): Gut Reactions: A Perceptual Theory of Emotion, New York: Oxford University Press.

How Safe Should We Feel?

233

Radford, Colin (1975): “How Can We Be Moved by the Fate of Anna Karenina?”, Proceedings of the Aristotelian Society, Supplementary Volumes 49, 67–93. Ratcliffe, Matthew (2005): “The Feeling of Being”, Journal of Consciousness Studies 12, 43–60. Shargel, Daniel/Prinz, Jesse (2017): “An Enactivist Theory of Emotional Content”, in: The Ontology of Emotions, ed. by Naar/Teroni, Cambridge: Cambridge University Press, 110–129. Slaby, Jan/Stephan, Achim (2008): “Affective Intentionality and Self-Consciousness”, Consciousness and Cognition 17, 506–513. Sorensen, Kelly/Williamson, Dianne (2018): Kant and the Faculty of Feeling, Cambridge: Cambridge University Press. Tappolet, Christine (2004): “Through Thick and Thin: Good and Its Determinates”, Dialectica 58, 207–221. Teroni, Fabrice (2007): “Emotions and Formal Objects”, Dialectica 61, 395–415. Väyrynen, Pekka (2013): The Lewd, the Rude and the Nasty, New York: Oxford University Press. Williams, Bernard (1985): Ethics and the Limits of Philosophy, Cambridge, MA: Harvard University Press.

12 Determining the Future Matthew Soteriou

Each moment is a leap forwards from the brink of an invisible cliff, where time’s keen edges are constantly renewed. We lift our foot from the solid ground of all our life lived thus far, and take that perilous step out into the empty air. Not because we can claim any particular courage, but because there is no other way. Han Kang, The White Book: 7 Freedom [. . .] is practically necessary – man must therefore act according to an Idea of freedom, and cannot act otherwise. Kant, Lectures on Metaphysics, §29: 898

1 No matter what we sense, think about, or do, all our experiences, thoughts, and actions are recessively framed by a temporal point of view that is centred on the present from which we are oriented to our past and our future. That temporal point of view is associated with signifcant asymmetries in our psychological attitudes to our past and our future – psychological asymmetries that are refected in the fact that we recollect the past but not the future, we anticipate the future but not the past, we regard the future as open in a way that the past is not. This asymmetrical way of being psychologically oriented to our past and future provides each of us with a tensed perspective on reality; so given that the relevant asymmetries in our psychological orientation are not optional, it might be said that as self-conscious agents we cannot help but occupy a perspective on reality that is tensed, whether or not reality is itself tensed. Some have also suggested that as self-conscious agents capable of practical deliberation, we cannot help but act under the idea of freedom.1 One of my central aims in this paper is to explore potential connections between these two suggestions – in particular, connections between the notion that we occupy a tensed temporal perspective from which we regard the future as open, and the notion that we occupy a deliberative standpoint from which we act under the idea of freedom.

Determining the Future

235

A further aim is to suggest that tracing out interdependencies between the psychology of our self-conscious agency and those psychological asymmetries that constitute our tensed temporal perspective can play a role in helping to illuminate how we exercise agency and self-determination in deciding to act. Just as we take ourselves to be responsible for the actions we decide to perform, we also take ourselves to be responsible for our decisions to act. However, it is not straightforward to explain how agency and responsibility manage to attach to the mental act of deciding, as well as the actions decided upon. Although it’s generally accepted that we exercise our agency when we act on decisions we have made, it is more controversial to claim that we also exercise our agency in making those decisions.2 This is because decisions don’t seem to ft the standard accounts of what it is that makes an event an agent’s action. Our decisions are not preceded by decisions or intentions to so decide. Moreover, if anything moves one to decide to Φ on a given occasion, it seems that it is some reason for acting as decided – i.e. some reason for Φ-ing – rather than some reason or desire that concerns the event of deciding itself.3 This might lead one to doubt whether our decisions are mental actions that we are motivated to perform. But if our decisions are not themselves mental actions, in what sense can we be said to be exercising agency in deciding to act, and what makes us responsible for those decisions, given that our decisions are not themselves decided upon? In what follows I suggest that connections between the psychology of self-determination and our temporal psychology should be central to an account of the sort of mental agency we exercise in deciding to act, and hence central to an account of what makes us responsible for our decisions. For I shall be arguing that the key to explaining how the mental act of deciding can amount to an agential mental act that is both selfdetermining and self-determined lies in providing the correct account of what is involved in occupying, over time, the sort of tensed temporal perspective that the self-determining agent adopts towards her past, present, and future when she decides to act. What this approach reveals, I argue, is that when it comes to providing an account of the respect in which the psychological act of deciding is a self-determined act for which we are responsible, it is a mistake to look simply to the psychological causes of the act of deciding.4 For in order to account for the respect in which one’s decisions are self-determined acts that one is responsible for, we need to look to the sort of behaviour (including mental behaviour) that one engages in, and the sort of temporal perspective one thereby occupies, after the act of deciding.5 After having considered ways in which aspects of the psychology of our temporal perspective may contribute to explaining our capacity to exercise agency in deciding to act, at the end of the paper I consider how our capacity to make decisions can contribute to explaining some of the distinctive features of the temporal perspective on the future that

236

Matthew Soteriou

we occupy. There I discuss, and respond to, the suggestion that when it comes to our perspective on our future, there is a potential tension between the standpoint of theoretical reason and the standpoint of practical reason.

2 The fact that one occupies a temporal point of view that is centred on the present can give one the sense that one is perpetually confned to the temporal present. One can of course be preoccupied with one’s future or one’s past. But if one is preoccupied with the future, the conscious activity of being so preoccupied falls within the experienced present, and when one dwells on the past, that conscious activity likewise falls within the experienced present. Even contemplation of the atemporal involves conscious activity shackled by the experienced present. So there is a sense in which one can no more escape the experienced present that one occupies whenever one thinks and acts than one can escape the egocentric spatial frame of reference that one occupies when one moves. However, while our temporal point of view may be centred on the present, we also cannot help but be psychologically oriented to what lies beyond any feeting, present moment. Even when the focus of one’s conscious attention is latched securely onto what is now happening, and one is not consciously thinking about the past or the future, the stream of consciousness, which falls within the experienced present, fows upon a bed of psychological states that are recessively oriented to what falls on either side of the bounds of the experienced present. Arguably, only so can we experience the present as we do – as containing occurrences with temporal parts that have happened, and temporal parts that are yet to unfold. These psychological states give breadth to our temporal perspective on reality – a temporal perspective that cannot be adequately captured by the content of any single conscious thought or experience.6 The asymmetrical way in which these psychological states orient one to what falls on either side of the experienced present can be said to amount to a tensed temporal perspective on reality – a perspective that colours one’s experience of what falls within the experienced present. For what falls within the experienced present is experienced by one as occupying a brief interval of time that intervenes between one’s past and one’s future – a past and future marked out by those asymmetries in one’s psychological attitudes. Aspects of the psychology of our agency are relevant to an account of these psychological asymmetries, and so thereby relevant to an account of the temporal perspective that we occupy. Think, for instance, of the way in which one’s agential perspective on one’s current goal-directed intentional activities involves a distinctive kind of psychological posture to one’s immediate future – the direction of action, as it might be put. That is an aspect of the psychology of our agency that

Determining the Future

237

will need to feature in an account of the asymmetries in our psychological attitudes that give content to our temporal perspective on reality – a content far richer than that captured merely by the notion of times earlier and later than a temporal location referred to by means of some temporal-indexical expression. However, my particular concern in this paper is with connections between the psychology of our temporal perspective and aspects of our psychology that are peculiar to us as self-conscious agents – agents capable of occupying the standpoint of practical reason. As self-conscious agents, we are capable of adopting a practical perspective on our lives that is not limited to the thoroughly immersed hereand-now exercise of bodily agency. We are capable of refecting on our past, and we are also capable of thinking ahead and engaging in practical deliberation and planning in our decision-making. One question to consider is how our capacity to adopt this sort of deliberative standpoint might contribute to explaining some of the distinctive features of the kind of tensed temporal perspective that we occupy – for example, those features relevant to the fact that our temporal perspective is one from which the future is open in a way the past is not. However, we might also look for explanations in the other direction. For example, we might consider how an account of the psychology of our temporal perspective may contribute to explaining our capacity to engage in practical deliberation, planning and decision-making. In order to identify explanations in either direction, we would need to uncover connections between the psychology of our temporal perspective and the psychology of self-determination; and uncovering such connections might in turn help to illuminate our understanding of what self-determination consists in. The notion that the psychology of selfdetermination should be central to an account of what self-determination consists in is related to a Kantian line of thought. Kant proposed that we cannot help but act under the idea of freedom, and he also proposed that this is what it is to be really free in the practical sense.7 On one understanding of that proposal, the idea of a free agent is the idea of a self-determining agent, and so acting under that idea amounts to acting and behaving as though one were a self-determining agent. If the psychology of self-determination can provide an account of what is involved in regarding and treating oneself in that way, then if we assume and apply the Kantian line of thought, this should amount to an account of what it is to be a self-determining agent. So if the psychology of our temporal perspective is relevant to understanding the psychology of self-determination, then the psychology of our temporal perspective should in turn be relevant to understanding what self-determination consists in. That is a line of thought I want to pursue here. In particular, I shall be exploring the following suggestion. Regarding and treating oneself as a self-determining agent involves adopting a certain kind of perspective on oneself. In adopting the relevant perspective, one thereby adopts a

238

Matthew Soteriou

distinctive kind of perspective on one’s past and future, and hence on the past and future more generally (Sections 3–4). By occupying that temporal perspective on oneself over time, one thereby makes true various things about oneself. Since it takes time to occupy the relevant perspective, it can take time to make true certain of those things about oneself, including certain things about one’s past. One of the things that one can make true about one’s immediate past by occupying the relevant temporal perspective is the following: that one exercised agency and self-determination in deciding to act (Section 5). That is why when it comes to understanding how we exercise agency and self-determination in deciding to act, we need to look to what changes in how we regard our past and future once we have decided to act, rather than simply looking to the causal antecedents of the act of deciding. So now my preliminary question is this: What changes in the way one regards one’s past, present and future when one decides to do something?

3 Suppose a self-conscious subject decides that tomorrow she will Φ. It might be said that in the typical case something thereby changes in the way that subject regards her past and her present: she now believes she has decided to Φ (past), and she now believes that she intends to Φ (present). But what changes in the way she regards her future? When a subject decides that she will Φ tomorrow, she settles a question in her own mind by committing herself to a certain course of action, and that commitment is refected in her subsequent behaviour, including the further practical deliberation and planning she subsequently engages in.8 For in committing herself to that course of action by making that decision, the subject will subsequently plan on the background assumption that she will Φ tomorrow. So, when a subject decides to Φ, at least the following changes in the way she regards her future: she adopts an attitude towards her future that serves as a constraint on her further practical reasoning, insofar as she is disposed to assume that she will Φ when she engages in further planning and practical deliberation. If a subject predicts that an event will occur, then she will likewise be disposed to assume that some future event will occur, and that assumption will likewise constrain her practical deliberation and planning. However, when a subject decides that she will Φ tomorrow, the assumption that the subject makes about her future Φ-ing differs from a straightforward prediction in various respects.9 From the subject’s own point of view, the assumption that she will Φ tomorrow is not one that she takes herself to be epistemically obliged to make prior to deciding to Φ, neither is it an assumption that she takes herself to be epistemically obliged to make after deciding to Φ, and indeed the subject doesn’t treat her assumption that she will Φ as an assumption grounded in her evidence. This point

Determining the Future

239

can be brought out by considering the way in which the subject regards and treats the constraint on her practical reasoning that is imposed by her assumption that she will Φ tomorrow. Once a subject has decided that she will Φ tomorrow, even as she continues to assume that she will Φ tomorrow, she does not consider it to be epistemically impermissible for her to give up that assumption, for she does not consider it to be epistemically permissible to change her plans and thereby make an alternative assumption about what she will do – one inconsistent with her current assumption that she will Φ tomorrow.10 Of course, if a subject does change her plans, then there will be a change in her evidential situation, given that what she knows about her plans will have changed. However, the important point is this. Prior to any such change in her plans, from the subject’s own point of view, no change in her evidential situation is required in order for it to be epistemically permissible for her to change her plans, and in consequence no change in her evidential situation is required in order for it to be epistemically permissible for her to make an alternative assumption about what she will do.11 Contrast this with the way in which a subject regards the constraint on her practical reasoning that is imposed by an ordinary, evidentially grounded belief that she has about the future – for example her belief about when the tide will come in. She does take it to be epistemically impermissible for her to relinquish that constraint on her reasoning, unless her evidence changes. That is why, unless her evidence changes, she will continue to regard any change in her plans as subject to that constraint. Note that the contrasting way in which the subject regards the latter constraint on her practical deliberation is relevant to the following, additional point: when a subject decides to Φ tomorrow, although she doesn’t treat her subsequent assumption that she will Φ tomorrow as an assumption that is grounded in her evidence, she does nonetheless take her decision-making to be subject to evidential constraints. For the range of assumptions that a subject is in a position to make about her future by deciding what to do is rationally constrained by her evidence concerning what she is capable of doing and what the future circumstances will allow (including, for example, her evidence about when the tide will come in). The fact that there are such evidential constraints on decision-making can help us to pin down further what is distinctive in an agent’s psychological posture towards her future when she decides to do something. So far, I have said that when a subject decides to Φ, (a) she is subsequently disposed to assume that she will Φ when she engages in further planning and practical deliberation, and (b) she doesn’t treat her assumption that she will Φ as an assumption grounded in her evidence, but nonetheless (c) she takes her assumption that she will Φ to be rationally constrained by her evidence – evidence concerning what she is capable of doing and what the future circumstances will allow. Items (a) – (c) can be explained by

240

Matthew Soteriou

the following proposal: When a subject decides to Φ, she is subsequently disposed to assume that she will Φ, and she regards her assumption as one that is to be discharged by an action that makes it true. This explains (c), because if the subject takes herself to be incapable of performing the relevant action (given what she knows about the future), then she should take herself to be incapable of making an assumption that is to be discharged in that way.12 The fact that the subject treats her assumption as one that is to be discharged by an action that makes it true also accommodates (b). For the sort of psychological posture that goes with the subject regarding her assumption in that way is not equivalent to her predicting that she will do something that makes true what she assumes. For reasons I shall explain, it is not even equivalent to her predicting that she will do something that makes true what she assumes because she so predicts.13 When one predicts that something will happen, the psychological posture towards the future that one adopts is consistent with that of waiting for the predicted events to occur. By contrast, when one regards one’s assumption about a future action as one that is to be discharged by an action that makes it true, one’s psychological posture towards the future does not share the sort of passivity that is associated with merely waiting.14 For in so regarding one’s assumption, one recognises that what is required is activity on one’s own part. One way of putting this is as follows. In deciding to Φ tomorrow, one commits oneself to that course of action, and one thereby regards one’s assumption about that future action as an assumption that is to be discharged by an action that makes it true, insofar as one regards the commitment one made in deciding to act as a commitment that is to be fulflled. That commitment imposes various constraints on one’s conduct. But one recognises that the constraints that come with that commitment are constraints one has imposed on oneself, rather than constraints imposed by the evidence one possesses about how the future will turn out. So one recognises that one is subject to those constraints only as long as one continues to impose those constraints on oneself, and so only as long as one continues to act in recognition of those constraints. That is why one does not consider it to be epistemically impermissible for one to relinquish those constraints at any point – i.e. to change one’s plans and make an alternative assumption about what one will do. This is relevant to understanding the following important point. Although the decisions one makes can determine one’s future conduct, from one’s own point of view there nonetheless remains a respect in which those decisions leave open one’s future until one has completed the actions that one has decided upon.15 That point is relevant to understanding what is distinctive about the sort of temporal perspective on oneself that one adopts once one has decided to act; and that in turn is relevant to understanding how occupying that temporal perspective can make it

Determining the Future

241

the case that one exercised self-determination in deciding to act. I now want to unpack each of those ideas in turn.

4 When you decide what to do, there is a respect in which you thereby determine something about your future that had previously been left undetermined. Here are two possible ways of understanding that idea. (A) Epistemic understanding: Your decision about what to do makes known to you something about your future that you previously didn’t know. (B) Causal understanding: Your decision about what to do is the cause of an event that is yet to occur – i.e. the action you decide to perform. While there is something to be said in favour of each understanding, neither is entirely satisfactory without further supplementation, nor is the conjunction of the two.16 Regarding (A): There is an epistemic understanding of “determining the future” that fails to capture the sort of perspective that a subject has on her future when she decides what to do. One attempts to “determine the future” in the relevant epistemic sense when one attempts to determine, via theoretical reasoning, what will come to pass; and as previously noted, when one fgures out what is likely to happen, and hence what will likely come to pass, the psychological posture towards the future that one adopts is consistent with that of waiting for the predicted events to occur, even if those predicted events are one’s future actions. For reasons I have already given, that is not the sort of psychological posture towards one’s future action that one adopts when that action is an action one has decided to perform. Regarding (B): Your decision to act determines your future insofar as it determines what is to be done, but the causal understanding allows for a reading of “what is to be done” which is simply equivalent to “what is yet to happen.” So the causal understanding of “determining the future” is also not entirely satisfactory without further supplementation, for reasons similar to those which make understanding (A) unsatisfactory. This is because knowing that you have decided to Φ is not equivalent to knowing that some past event (one’s decision to Φ) will eventuate in some occurrence that is yet to happen (one’s Φ-ing). For that is again suggestive of the idea of a psychological posture towards the future that is too close to prediction. As a self-determining agent one regards one’s future as a region containing “things to be done,” and one’s decisions fx what is to be done. But we fail to capture what is distinctive about the perspective on one’s future this provides one with, if we equate “regarding one’s future as a

242

Matthew Soteriou

region containing things to be done” with “regarding one’s future as a region containing things that one will do.” One might be tempted to try to avoid that false equivalence by deploying something like the following normative understanding of “to be done” and hence a normative understanding of the respect in which you determine something about your future by deciding what to do: (C) Normative understanding: Your decision about what to do determines your future insofar as it determines what is to be done, and it determines what is to be done insofar as it determines what you ought to do. A problem with understanding (C), as it stands, is that it fails to capture the distinction between, on the one hand, a decision to do something, and, on the other hand, a judgement that one ought to do something. The latter is not equivalent to the former.17 One’s decision to do something imposes constraints on one’s subsequent conduct in a way that a mere judgement about what one ought to do does not. That is why when one decides to do something one thereby regards that future action as something that one will do, and not simply something that one ought to do. Having decided to Φ, one plans on the assumption that one will act as decided. If one doesn’t plan on that assumption, then the constraints on one’s future conduct that come with the commitment one makes in deciding to act won’t yet have been imposed. And it is the latter notion of commitment – committing oneself to a course of action – and the self-imposed constraints that are incurred in so committing, which provides for the appropriate sense in which one determines something about one’s future that had previously been left undetermined when one decides what to do. When one decides to Φ, one commits oneself to a particular course of action by imposing certain constraints on oneself – e.g., by assuming that one will Φ when one subsequently engages in further planning and practical deliberation. When one imposes those constraints on oneself, one thereby commits oneself to a future that contains that action. But that sort of commitment can be made only if one imposes on oneself constraints which one regards and treats as self-imposed. That is to say, in order to commit oneself (in the relevant sense) to a future that contains that action, one must recognise that the constraints one is subject to, in so committing oneself, are constraints that one has imposed on oneself, rather than, say, constraints that are imposed by the evidence one possesses about how the future will turn out. So, since one must regard and treat such constraints as self-imposed in order to so commit oneself, one cannot take it to be an epistemic requirement to so commit oneself, for epistemic requirements are not self-imposed.18 This means that when one decides to Φ, and one thereby commits oneself to a future that contains that action, one cannot take there to be any

Determining the Future

243

epistemic requirement to commit oneself to that future. And this remains true even after one has decided to Φ.19 That is to say, having decided to Φ, one does not take oneself to be epistemically required to retain the commitment one made in deciding to Φ, and so one does not take oneself to be epistemically required to retain one’s commitment to a future that contains that action. That in turn means that although one’s decision can determine one’s future conduct by imposing constraints on how one goes on to behave, one is subject to those constraints only for as long as one continues to impose them on oneself, and one continues to impose those constraints on oneself only for as long as one acts in recognition of them. So, from one’s own point of view, there remains a respect in which one’s decision leaves open one’s future until one has completed the action that one has decided upon – until one has fulflled the commitment one made in deciding to act. From one’s own point of view, what is left open to one is the perpetual possibility of changing one’s mind, and so the perpetual possibility of committing oneself to an alternative future. So, after you have decided to Φ, and so after you have committed yourself to a future that contains that action, there is a respect in which your future remains open, for the future that contains that action remains yours to determine. It remains yours to determine until you have completed the action and so fulflled the commitment you made in deciding to act. That temporal perspective on one’s future brings with it a distinctive attitude towards, and hence perspective on, one’s past. This is because from that temporal perspective, by engaging in behaviour that manifests one’s recognition of the commitment one made in deciding to act, one is thereby attempting to fulfl that commitment, and such behaviour can be regarded as an instance of remembering to do something – i.e. remembering to fulfl the commitment one made in deciding what to do. That variety of remembering (i.e. remembering to do) is not reducible to remembering that something is the case – e.g., it is not reducible to remembering that one decided to Φ. Behaviour that manifests remembering that one decided to Φ is not suffcient for manifesting the sort of recognition that is of concern here – i.e. the sort of recognition that amounts to attempting to fulfl the commitment one made in deciding what to do. For one can remember that one decided to Φ long after one has completed the action, or after one has changed one’s plans when one no longer intends to Φ, and so when one is no longer attempting to fulfl the commitment one made in deciding to Φ.20 This further brings out the difference between, on the one hand, the psychological attitude towards the future that one adopts when one makes a prediction, and, on the other hand, the kind of perspective that one adopts towards one’s future when one decides what to do. Behaviour that manifests the dispositional state that you acquire when you decide to do something counts as a case of remembering to do, whereas behaviour that manifests ordinary belief and prediction does not.21 That is why

244

Matthew Soteriou

when one has made a prediction, the psychological posture towards the future that one adopts is consistent with that of waiting for the predicted events to occur. In adopting that psychological posture towards the future, one doesn’t need to remember to do anything. We are now in a position to summarise some of the distinctive features of the sort of temporal perspective on yourself that you adopt once you have decided to do something. In making a decision, you determine something about your future that had previously been left undetermined by committing yourself to a certain course of action. However, after you have decided on that course of action, and so after you have committed yourself to a future that contains that action, from your point of view there is a respect in which your future remains open. For the future that contains that action remains yours to determine. And it remains yours to determine until you have completed the action and thereby fulflled the commitment you made in deciding to act. For until you have completed the action, what is left open to you is the perpetual possibility of committing yourself to an alternative future – one that doesn’t contain that action. From that temporal perspective, by engaging in behaviour that manifests your recognition of the commitment you made in deciding to act (behaviour that includes planning on the assumption that you will act as decided), you are thereby attempting to fulfl that commitment, and such behaviour can be regarded as an instance of remembering to do something. Let us now consider what one makes true about oneself by occupying that perspective over time.

5 In deciding what to do, you commit yourself to a certain course of action, and in committing yourself to that course of action, you adopt an attitude towards the future that constrains the way you subsequently behave – for example, you plan on the assumption that you will act as decided. You recognise that the constraints that come with that commitment are constraints that you have imposed on yourself in deciding what to do, rather than constrains that are imposed by the evidence you possess about how the future will turn out, and that is why you do not consider it to be epistemically impermissible for you to relinquish those constraints by changing your plans. However, if you do not change your plans, and you act in recognition of the constraints that you imposed on yourself in deciding what to do, then you thereby engage in behaviour that amounts to remembering to do something – for you are remembering to fulfl the commitment you made in deciding what to do. When you remember to do something, you thereby make true something about your past. For you thereby make it the case that when you decided what to do, you were successful in imposing constraints on your future conduct.22 Note that

Determining the Future

245

this isn’t something you can make true about your past by simply remembering that something is the case – e.g. remembering that you decided to Φ, or remembering that you will Φ (the latter is consistent with merely predicting that you will Φ). Note also that when you do act in recognition of the constraints that you imposed on yourself in deciding what to do, you don’t regard those constraints as ones that it is epistemically impermissible for you to relinquish.23 This is because, from your own point of view, until you have fulflled the commitment you made in deciding what to do, there remains the perpetual possibility of changing your mind and committing to an alternative future. So, even as you act in recognition of the constraints you imposed on yourself in deciding what to do, you continue to regard aspects of your future as open, insofar as you regard aspects of your future as yours to determine; for you don’t take it to be epistemically impermissible for you to relinquish those constraints. In consequence, by acting in recognition of the constraints that you imposed on yourself in deciding what to do, you thereby regard and treat that past decision as an act of your determining your current behaviour – rather than a past event that occurred within you and that is now having its effect on the way you behave. From your point of view, given the respect in which your future is open and yours to determine, your past decision constrains your present conduct only if you continue to grant it the authority to do so – i.e. by now acting in recognition of the commitment you made in deciding to act. And when you do act in recognition of the commitment you made in deciding to act (e.g. by planning on the assumption that you will act as decided), you thereby make it the case that you were successful in determining your present conduct by making that decision. Your past decision about what to do thereby determines how you subsequently behave. In that respect it can be described as a self-determining act. But it only determines how you behave if you subsequently engage in self-determined behaviour which grants it that authority. So by engaging in such behaviour, you make it the case that your past decision is a mental act that is both selfdetermining and self-determined. It is a self-determining act insofar as it determines how you go on to behave. It is a self-determined act insofar as its power to determine your behaviour depends on your subsequent self-determined behaviour that grants it that authority. There is a complexity to the temporal perspective you occupy when you act in recognition of the constraints you imposed on yourself in deciding what to do. By occupying that perspective, you are now doing (progressive, present tense) something that is directed towards fulflling (not yet fulflled, and so directed towards your future) a commitment you made (past). And by occupying that temporal perspective, you thereby regard and treat yourself, and hence your past, present and future selves, as self-determining.24 Once you have decided to act, your present self regards the future as open insofar as your present self recognises the

246

Matthew Soteriou

perpetual possibility of a change in mind, and hence the perpetual possibility of committing to an alternative future. So your present self thereby regards and treats your future self as self-determining. From the point of view of your present self, the completion of the action you have decided upon is dependent on the co-operation of your future self.25 For the same reason, there is a respect in which your past self determines your present conduct, in deciding what to do, only with the co-operation of your selfdetermining present self. But if your present self does act in recognition of the constraints you imposed on yourself in deciding to act, your present self thereby regards and treats the decision made by your past self as a self-determining act of your determining your present behaviour. And by acting in that way, your present self thereby engages in self-determined behaviour that makes true something about your past self – namely that the decision made by your past self was a self-determined act. What this shows, I suggest, is that we fail to capture the distinctive way in which we exercise agency in deciding to act, if our approach to an account of the issue is the one typically adopted by action theorists who attempt to specify what it is that makes an event an action of an agent. According to that familiar approach, if we do exercise agency in deciding to act this will be in virtue of the fact that the mental event of deciding is preceded by (or accompanied by), and appropriately caused by, suitable psychological states and/or events. Whereas I am suggesting that when it comes to accommodating and explaining the respect in which we exercise agency and self-determination in making a decision, we need to look the result of this mental act, rather than its causes. It is what happens after one’s decision to act that makes it the case that one has exercised agency and self-determination in making that decision. There are interdependencies between the psychology of our self-conscious agency and the psychology of our temporal perspective. Such interdependencies make it possible for us to uncover connections between the psychology of self-determination and those psychological asymmetries that constitute our tensed temporal perspective. In this section, I have been suggesting ways in which aspects of the psychology of our temporal perspective may contribute to explaining our capacity to exercise agency in deciding to act. In the fnal section of the paper, I shall make some concluding remarks on forms of explanation in the other direction, and in particular, on how our capacity to make decisions can contribute to explaining some of the distinctive features of the temporal perspective on the future that we occupy.

6 From the deliberative standpoint, my future depends, in part, on the decisions I make and will make. So, my future, and hence the future in general, is no more fxed than those decisions. From that deliberative

Determining the Future

247

standpoint, I regard and treat myself as self-determining, and I thereby regard my future as open, insofar as I recognise the perpetual possibility of a change in mind, and hence the perpetual possibility of committing to an alternative future. In so regarding my future, I thereby regard the future as a region that allows for that degree of openness. Although I have primarily been discussing examples of decisions that concern relatively long-range future actions (e.g. decisions about what one will do tomorrow), many of the points I have been making about the temporal perspective that the self-determining agent adopts towards her future generalise; for one’s possession of this capacity for decisionmaking introduces an asymmetry in one’s psychological orientation that affects one’s perspective on any temporal region lying beyond the futureoriented edge of the experienced present – including one’s immediate future. From that perspective, the future (including the immediate future) is open, insofar as there remains open to one the perpetual possibility of committing to any one of a range of alternative possible futures, and so there remains open to one the perpetual possibility of thereby determining which of those futures unfolds. From that point of view, in fulflling the commitment one makes in deciding what to do, one closes off alternative options, and so one closes off those alternative futures. Earlier, I distinguished the sense in which one determines one’s future by deciding what to do, from an epistemic understanding of “determining the future.” According to that epistemic understanding, one attempts to “determine the future” when one attempts to determine, via theoretical reasoning, what will come to pass. That epistemic understanding of determining the future provides for a different sense in which it is possible to “close off” future options. As I previously noted, there are evidential constraints on one’s decision-making. The range of assumptions one is in a position to make about one’s future by deciding what to do is rationally constrained by one’s evidence concerning what one is capable of doing, and what the future circumstances will allow. We need to make place for the recognition of such evidential constraints on practical deliberation if deliberation is to be successful, or even possible. For one cannot govern one’s own conduct without being guided by the facts, or what one thinks are likely to be the facts, including facts about what will be the case. Practical reasoning, then, depends on the deliverances of theoretical reasoning. One deliberates about what to do in a way that recognises that the deliverances of theoretical reason rationally constrain one’s decisions, and so in a way that recognises that the options that are available to one are constrained by theoretical reason. Hence one deliberates in way that recognises that what one determines about the future in the epistemic sense can close off future options. Recognising such constraints on planning amounts to acknowledging that there are aspects of the future which one cannot oneself determine (in the practical sense), but regarding which one might try to determine more

248

Matthew Soteriou

(in the theoretical sense). So implicit in this stance towards the future is the notion that the future is a region that contains things about which one can gather evidence and make justifed predictions. It is a region about which one has some evidence and can try to gather more evidence. So one might think that the more evidence one gathers, the more future options are closed off, which in turn reduces the available options for one to close off in the other, practical, sense – i.e. by deciding what to do. Given that theoretical reason closes off, in the epistemic sense, the range of options that are available to be closed off in the practical sense, is there then a potential tension between what we might call the standpoint of theoretical reason and the standpoint of practical reason?26 We occupy a temporal perspective from which deliverances of theoretical reason and practical reason are integrated but distinct, for while the deliverances of theoretical reason rationally constrain one’s decision-making, the deliverances of theoretical reason do not, and cannot, determine one’s decision-making. It might be said that the integration of theoretical reason and practical reason provides for a temporal point of view on one’s future that amounts to something like a template – a template that divides the future into what Huw Price (2005) has called “options” and “fxtures.” According to Price, the “options” are “the alternatives among which [. . .] [the deliberator] takes herself to be deliberating” (2005: 275). Whereas “fxtures” “denote everything else – all matters of fact that are not held to be a matter of choice in the deliberation in question” (2005: 275). Price writes, FIXTURES will contain a subset, KNOWNS, comprising those facts the deliberator takes herself to know at the time of deliberation, and also a larger subset, KNOWABLES, comprising matters she regards as either known or knowable, at least in principle, before she makes her choice. Price 2005: 275 That picture may suggest that our temporal point of view, in principle at least, allows for the area of the template that contains “fxtures” to expand indefnitely, which would in turn result in a shrinking of the area of the template that contains “options.” The idea of a potential tension between the standpoint of theoretical reason and the standpoint of practical reason might then be expressed as follows. Implicit in the sort of temporal point of view towards the future that we occupy is the idea that theoretical reason could, in principle at least, keep shrinking the area of the template that contains “options” until it eliminates that area entirely. However, although the generation of this sort of puzzle arises from the temporal perspective we occupy, it is not clear that the sort of potential tension it cites can be actualised from the temporal perspective that we occupy. One’s possession of a capacity for decision-making introduces

Determining the Future

249

an asymmetry in one’s psychological orientation towards time that is inescapable for as long as one possesses that capacity. That asymmetry affects the temporal perspective on the future one actually occupies, and from that perspective there remains open to one the perpetual possibility of changing one’s mind and committing to an alternative future. So there remains open to one the perpetual possibility of committing to any one of a range of alternative possible futures, thereby determining which of those futures unfolds. Since this temporal perspective is inescapable for those of us who are capable of making decisions, the exercise of our theoretical reason is also subject to it. That is to say, when one engages in theoretical reason, one makes decisions about what to discover or attempt to discover, and fnd out; and one also makes decisions about which questions to re-open and which matters to suspend judgement about. One governs one’s epistemic conduct.27 In doing so, one thereby commits oneself to a future that contains that epistemic conduct. From that perspective, the future (including the immediate future) is open, insofar as there remains open to one the perpetual possibility of committing to any one of a range of alternative possible futures, and so there remains open to one the perpetual possibility of thereby determining which of those futures unfolds. In exercising theoretical reasoning, one thereby adopts the practical perspective on one’s future. One cannot determine one’s future in the epistemic sense without thereby determining one’s future in the practical sense. Suppose one takes oneself to establish, via theoretical reason, that there are no alternative futures, there is only one future, but one just happens to be condemned to be ignorant of what it is. From that perspective, the future still remains yours to determine, in the practical sense. For this is not a scenario in which all future options have been epistemically closed off, so there remain options to be closed off in the practical sense; and as long as there are options that have not been epistemically closed off to one, there remain options to be closed off in the practical sense.28 By analogy, consider the fantasy of the time-traveller. As she sets off to travel into the past, aspects of her psychological future are in temporal regions earlier than her current temporal location, and she knows this.29 For temporal regions earlier than her current temporal location contain actions about which she must now deliberate, and she knows this. Although she may think there is only one past, as long she is not omnipotent about that past, from her point of view, there remain options to be closed off in the practical sense. She may try to fnd out more and more about the past with the view that this will improve her decisionmaking, by revealing more and more to her about which options are genuinely available to her; and the more she fnds out, the more options will be epistemically closed off to her. However, although this historical fact-fnding mission may constrain her decision-making, it cannot determine her decision-making. Her decisions, and hence her future (which in

250

Matthew Soteriou

this case falls in temporal regions earlier than her current temporal location), remain hers to determine. For while the deliverance of theoretical reason can close off options, in the epistemic sense, it cannot determine which particular options to close off in the practical sense. Although the deliverance of theoretical reason can constrain one’s decision-making, it cannot determine one’s decision-making. For in committing oneself to a (psychological) future, in the practical sense, one must recognise that the constraints one is subject to, in so committing oneself, are constraints that one has imposed on oneself, rather than constraints that are imposed by the evidence one possesses about how the (psychological) future will turn out. Since one must regard and treat such constraints as self-imposed in order to so commit oneself, one cannot take one’s commitment to be epistemically determined. For epistemic requirements are not self-imposed. I said earlier that since the temporal perspective from which the future is open is inescapable for those of us who are capable of making decisions, the exercise of our theoretical reason is also subject to it. However, it might be thought that despite that fact, we can nonetheless deploy theoretical reason to attempt to attain a perspective-independent (or less perspective-dependent) view of things – one that transcends the tensed temporal perspective from which the future is open. For while the actual exercise of theoretical reason may be subject to the tensed temporal perspective we in fact occupy, the content of the view we attain via theoretical reason need not be. However, if we do manage to attain that less perspective-dependent conception of what there is, it is not clear that we would remain in view from it (which of course isn’t to deny that we would remain in view to ourselves through exercising our agency when engaging in the theoretical reasoning that attains that conception). That point is the fipside of the notion that regarding and treating oneself as a self-determining agent is suffcient for being one. Self-determining agents are perspective-dependent entities. They are entities that adopt a certain kind of perspective on themselves. In adopting the relevant perspective, they adopt a distinctive kind of perspective on their past, present, and future, and hence on the past, present, and future more generally; and it is not clear that a conception that abstracts away from that temporal perspective would be one that keeps in view these perspective-dependent beings. So it is not clear that a conception that abstracts away from the kind of tensed temporal perspective that we occupy could take us any closer to understanding ourselves and our lives. A theoretical conception of ourselves that can add to our understanding of ourselves should be one that keeps us in view. If the freedom we exercise as self-determining agents is perspective-dependent, and if that perspective is intimately bound up with the distinctive kind of tensed temporal perspective that we occupy, then at least one route to attaining

Determining the Future

251

that theoretical understanding of ourselves will be to trace out interdependencies between the psychology of our self-conscious agency, and the psychology of our temporal perspective – interdependencies between the psychology of freedom and the psychology of time.30

Notes 1. The phrase is taken from Kant. In addition to the quote from Kant’s Lectures on Metaphysics, see also his Groundwork of the Metaphysics of Morals (Kant 1785, §4: 448). 2. For arguments against the claim that our decisions are mental actions see Strawson (2003). Pink (1996, 2009) presents an important defence of the claim our decisions are mental actions. Peacocke (2007, 2009) also holds that decisions can be mental actions, but for rather different reasons. Gibbons (2009) and Hieronymi (2009) have also offered accounts that accommodate, in different ways, a role for agency in decision and intention. 3. Some take this to be a moral of Kavka’s “toxin puzzle” (1983). 4. For an account that does attempt to explain the agency in making a decision by looking to the appropriate psychological causes, see Shepherd (2015). 5. This develops a proposal made in Soteriou (2013, ch. 12). 6. Compare Husserl’s (1905) account of time-consciousness, and the discussion of the perception of events in O’Shaughnessy (2000). 7. In his Lectures on Metaphysics, Kant writes, “Freedom is a mere Idea and to act according to this Idea is what it means to be free in the practical sense” (1821, §29: 898). 8. For seminal discussions of this idea, see Bratman (1987). See also his (1999) and (2006). 9. It also differs from a belief about one’s future that one might acquire by testimony. Much has been written on the difference between prediction and decision. See, for example, Hampshire/Hart (1958), Ginet (1962), Taylor (1964), Gauthier (1967), O’Connor (1967), Pears (1968), Levi (1986), and Joyce (2002). A good deal of this literature focuses on the question of whether practical deliberation excludes foreknowledge of one’s decisions. Here I am focusing on a different issue, namely the difference between prediction and the assumption one makes about one’s future having decided to act – i.e. after having decided to act, and in virtue of having decided to act. 10. Compare Velleman’s (1989b) discussion of the “epistemic freedom” associated with intention: “Even if the future is going to turn out a particular way, we don’t have to describe it as turning out that way in order to describe it correctly, since there are several other, incompatible ways in which we would be equally correct to describe it as turning out” (Velleman 2000b: 34). 11. One may of course be subject to non-epistemic requirements to do something, which may impose non-epistemic requirements not to change one’s mind without a change in one’s evidential situation, but the point I am making here concerns what is epistemically permissible for a subject to do. 12. And if a subject decides to do something that she is capable of doing, then when the assumption is false there is a respect in which the fault lies with the action and not the assumption. Compare here Anscombe (1959: 56–57): “The mistake is not one of judgment but of performance.” See also Hampshire/Hart (1958). For discussion of this view, see also Soteriou (2013, ch. 12).

252

Matthew Soteriou

13. Here I depart from Velleman’s account of intention. Velleman argues that intention is a kind of self-fulflling prediction that presents itself as such. See Velleman (1989a, 2000a). 14. Compare Ismael (2013), who writes, “We do not experience our own future as though it were a movie whose outcome we are simply waiting to see. We experience it as something that we actively bring about, something that is no more settled than our decisions, and whose outcome hangs in the balance until those decisions have been rendered” (162). In the quoted passage, Ismael is commenting on our attitude to our future “until our decisions have been rendered.” Here I am remarking on our attitude toward the future after the decision has been rendered, but before the action decided upon has been completed. 15. This is a point that I think is missed in Ismael’s (2013) interesting discussion of the respect in which we regard the future as open. 16. Velleman’s account of intention highlights the signifcance of both. 17. Shah (2008) denies that a normative judgment is identical with an intention, but nonetheless argues that deliberation that aims to conclude in an intention whether to Φ can proceed by settling the question whether one ought to Φ. I think that view captures something of the spirit of understanding (C). 18. On Velleman’s view, an intention is a self-fulflling prediction which presents itself as such. It might be thought that the epistemic requirements imposed by such a prediction are in some sense self-imposed. However, although the epistemic requirements associated with the prediction are imposed by a prediction one may not have been epistemically obliged to make, the epistemic requirements one is subject to, having made the prediction, are not self-imposed. This can be brought out by considering a self-fulling prediction that one makes about the behaviour of another. Consider Velleman’s (1989b) discussion of a doctor saying to a patient, in the presence of a nurse, “Nurse will now take you to the operating theatre.” When the nurse has left the room with the patient, the doctor is epistemically required to continue to assume that the patient will be taken to the operating theatre. 19. This is a point that is often overlooked in much of the literature about decision, prediction and foreknowledge. For references to that literature, see note 9. 20. For further discussion of the connection between decision and memory, see Soteriou (2013, ch. 12.4). 21. This point is one that isn’t accommodated by belief accounts of intention – e.g. those proposed by Velleman (1985, 1989a, 1989b), Joyce (2002), and Ismael (2007, 2013). As the account that I am proposing differs from such belief accounts, it is not susceptible to the sorts of objections that Fernandes (2016) levels against them. 22. This means that whether you have made a decision at a particular time depends on what happens after that time – i.e. whether you subsequently behave as one who has decided. But of course, that does not imply any kind of backwards causation. 23. This includes the constraints imposed by your assumption about what you will do. 24. I intend my use of the notions of present, past, and future “selves” here to be taken as a façon de parler, denoting the tensed perspective you have on yourself. As Korsgaard puts it, “the choice of any action, no matter how trivial, takes you some way into the future. And to the extent that you regulate your choices by identifying yourself as the one who is implementing something like a particular plan of life, you need to identify with your future in order to be what you are even now. When the person is viewed

Determining the Future

25.

26. 27.

28. 29. 30.

253

as an agent no clear sense can be made of a merely present self” (1989: 113–114). I take this point to be relevant to Kavka’s (1983) “toxin puzzle.” You cannot now intend to drink the toxin at a later date, if you believe your selfdetermining future self will have no reason to drink the toxin when the appointed time for action arrives. The idea that there is such a tension is something that Ginet (1962), for example, tries to bring out. See also Goldman’s discussion of the “book of life” (1970, ch. 6). One thereby governs one’s capacity for belief revision; and in revising one’s beliefs, one can re-open what one previously treated as epistemically closed. Compare Price (2005): “We plan under certain assumptions about what the future will be like, which we take as KNOWNS – e.g., normally, that the sun will rise tomorrow. But this seems to be very context-sensitive: if we want to consider an action that involves eliminating the sun, we won’t take the fact that it will rise tomorrow as a given – its rising will be in OPTIONS, not FIXTURES” (276). We can also apply our capacity for belief-revision to our beliefs about our own future actions. Suppose you decide to Φ and you have overwhelming evidence that you are able to act as decided and that you won’t change your mind. Let us assume you take yourself to know that you will act as decided. Can you nonetheless be epistemically entitled to assume that you won’t act as decided? Yes. For you are epistemically entitled to believe there is no purely epistemic obstacle to changing your mind and making alternative plans. In which case you are epistemically entitled to believe that there is no purely epistemic obstacle to your rendering your previous evidence inconclusive – thereby making it the case that you didn’t in fact know what you thought you knew. It might be assumed that if you really did take yourself to know that you would act as decided, then you would not see changing your mind as an option. But that assumption, I suggest, leads to the dogmatism paradox – a paradox that leads to the conclusion that whenever one knows, one should not heed any evidence suggesting one is wrong. (For discussion of the dogmatism paradox, see Kripke (2011), and see also the discussion in Soteriou (2013, ch. 15)). Compare Velleman (1989b): “There being no unique answer to the question ‘What will I do?’ – unlike your mere ignorance of the answer – is easy to mistake for there being no unique thing that you’ll do.” By the “psychological” future, I mean the temporal region containing the actions the agent is deciding to perform – which in the case of the timetraveller may be temporal regions earlier than her current temporal location. I am very grateful to Sebastian Schmidt and Hemdat Lerman for helpful comments on an earlier draft of this paper.

References Anscombe, G. Elizabeth M. (1959): An Introduction to Wittgenstein’s Tractatus: Themes in the Philosophy of Wittgenstein, South Bend, IN: St. Augustine’s Press. Bratman, Michael (1987): Intentions, Plans, and Practical Reason, Cambridge, MA: Harvard University Press. Bratman, Michael (1999): Faces of Intention: Selected Essays on Intention and Agency, Cambridge: Cambridge University Press. Bratman, Michael (2006): Structures of Agency: Essays, New York: Oxford University Press.

254

Matthew Soteriou

Fernandes, Alison (2016): “Varieties of Epistemic Freedom”, Australasian Journal of Philosophy 94, 1–16. Gauthier, David (1967): “How Decisions Are Caused”, Journal of Philosophy 64, 147–151. Gibbons, John (2009): “Reason in Action”, in: Mental Actions, ed. by O’Brien/ Soteriou, New York: Oxford University Press, 72–94. Ginet, Carl (1962): “Can the Will Be Caused?”, Philosophical Review 61, 49–55. Goldman, Alvin (1970): A Theory of Human Action, Princeton, NJ: Princeton University Press. Hampshire, Steward/Hart, Herbert L. A. (1958): “Decision, Intention and Certainty”, Mind 67, 1–12. Hieronymi, Pamela (2009): “Two Kinds of Agency”, in: Mental Actions, ed. by O’Brien/Soteriou, New York: Oxford University Press, 138–162. Husserl, Edmund (1905): The Phenomenology of Internal Time-Consciousness, ed. by Heidegger, transl. by Churchill, Bloomington: Indiana University Press 1964. Ismael, Jennan (2007): “Freedom, Compulsion and Causation”, Psyche 13, 1–11. Ismael, Jennan (2013): “Decision and the Open Future”, in: The Future of the Philosophy of Time, ed. by Bardon, London: Routledge, 149–168. Joyce, James M. (2002): “Levi on Causal Decision Theory and the Possibility of Predicting One’s Own Actions”, Philosophical Studies 110, 69–102. Kang, Han (2016): The White Book, London: Portobello Books 2018. Kant, Immanuel (1785): “Groundwork of The Metaphysics of Morals”, in his Practical Philosophy, transl. and ed. by Mary Gregor, Cambridge: Cambridge University Press 1996, 37–108. Kant, Immanuel (1821): Lectures on Metaphysics, ed. and transl. by Ameriks/ Naragon, Cambridge: Cambridge University Press 1997. Kavka, Gregory S. (1983): “The Toxin Puzzle”, Analysis 43, 33–36. Korsgaard, Christine (1989): “Personal Identity and the Unity of Agency: A Kantian Response to Parft”, Philosophy and Public Affairs 18, 101–132. Kripke, Saul (2011): Philosophical Troubles: Collected Papers, Vol. 1, New York: Oxford University Press. Levi, Isaac (1986): Hard Choices: Decision Making under Unresolved Confict, Cambridge: Cambridge University Press. O’Connor, John (1967): “How Decisions Are Predicted”, Journal of Philosophy 64, 429–430. O’Shaughnessy, Brian (2000): Consciousness and the World, New York: Oxford University Press. Peacocke, Christopher (2007): “Mental Action and Self-Awareness (I)”, in: Contemporary Debates in the Philosophy of Mind, ed. by Cohen/McLaughlin, Oxford: Blackwell, 358–376. Peacocke, Christopher (2009): “Mental Action and Self-Awareness (II)”, in: Mental Actions, ed. by O’Brien/Soteriou, New York: Oxford University Press, 192–214. Pears, David F. (1968): “Predicting and Deciding”, in: Studies in the Philosophy of Thought and Action, ed. by Strawson, London: Oxford University Press, 97–133. Pink, Thomas (1996): The Psychology of Freedom, Cambridge: Cambridge University Press. Pink, Thomas (2009): “Reason, Voluntariness and Moral Responsibility”, in: Mental Actions, ed. by O’Brien/Soteriou, New York: Oxford University Press, 95–120.

Determining the Future

255

Price, Huw (2005): “Causal Perspectivalism”, in: Causation, Physics, and the Constitution of Reality: Russell’s Republic Revisited, ed. by Price/Corry, New York: Oxford University Press, 250–292. Shah, Nishi (2008): “How Action Governs Intention”, Philosophers’ Imprint 8, 1–19. Shepherd, Joshua (2015): “Deciding as Intentional Action: Control over Decisions”, Australasian Journal of Philosophy 93, 335–351. Soteriou, Matthew (2013): The Mind’s Construction: The Ontology of Mind and Mental Action, New York: Oxford University Press. Strawson, Galen (2003): “Mental Ballistics or the Involuntariness of Spontaneity”, Proceedings of the Aristotelian Society 103, 227–256. Taylor, Richard (1964): “Deliberation and Foreknowledge”, American Philosophical Quarterly 1, 73–80. Velleman, J. David (1985): “Practical Refection”, The Philosophical Review 94, 33–61. Velleman, J. David (1989a): Practical Refection, Princeton, NJ: Princeton University Press. Velleman, J. David (1989b): “Epistemic Freedom”, Pacifc Philosophical Quarterly 70: 73–97, reprinted in his The Possibility of Practical Reason, New York: Oxford University Press 2000. Velleman, J. David (2000a): “On the Aim of Belief”, in: The Possibility of Practical Reason, New York: Oxford University Press. Velleman, J. David (2000b): The Possibility of Practical Reason, New York: Oxford University Press.

13 Silence and Salience On Being Judgmental Neal A. Tognazzini

1 Begin with a familiar maxim: just because you think something doesn’t mean you should say it. Failure to adhere to this maxim often constitutes tactlessness and can cause offense, even if you have good reason to think what you think, and indeed, even if what you think is true. Speaking your mind may even, in certain circumstances, be not just imprudent but a moral mistake. To take a much-discussed recent example: there may be certain facts about you that make it morally inappropriate for you to call someone out for a misdeed even if you have good reason to believe they have done it, and indeed, even if they have done it. If, for example, you are yourself guilty of the same (or a similar) transgression, then to call someone else out while ignoring one’s own wrongdoing would seem to involve making a groundless exception for oneself: the Kantian hallmark of moral impropriety. Even if hypocritical blame isn’t a moral mistake, though, there certainly seems to be something “off” about it. At the very least, the person being blamed seems entitled to raise a distinctive sort of objection, often voiced as a question: “Who are you to blame me for this?” The implication is that even if your blame doesn’t involve a factual mistake, in any case it involves some sort of mistake. If you’re no better than I am on this issue, best to keep silent. Identifying this mistake, and making sense of it, is the project of several recent writers who are interested, broadly speaking, in the standing to blame.1 This is not my topic, but it is related, and in any case it is helpful to point out that the best accounts of what goes wrong in cases of standingless blame (like our hypocrite) apply most naturally to blame that has been voiced or otherwise explicitly addressed to the wrongdoer. Since being the target of expressed blame is such a common and uncomfortable feature of our lives, it is certainly worth getting clear about. But unexpressed, or private, blame also seems potentially problematic, though making sense of the mistake seems a bit trickier (since its target may be unaware of being targeted). Whether someone might lack the

Silence and Salience

257

standing even to feel blame is a question even more intimately related to my topic, though my interest is not restricted to the mental state of blame in particular. What I’m interested in exploring is not the maxim with which we began, but a closely related and more controversial one, namely: just because something’s true doesn’t mean you should think it. Cases of standingless unexpressed blame – if such cases exist – would be examples of this maxim at work, but I’m inclined to think there are other examples as well. The one I’d like to explore here is the case of the judgmental person. My questions: what exactly is it to be judgmental, and why is it bad? My suggestion: the judgmental person thinks things that, even if true, they shouldn’t be thinking.

2 Let’s start with a case and a theory. First, the case: a young married couple moves into their frst house and begins the familiar fght against entropy known as home ownership. They manage to keep their house and property in adequate shape, certainly nothing that’s going to get them featured in a home and gardening magazine, but at the same time nothing that will get them in trouble with the city or even their neighborhood association. The yard, in particular, is kept mown but not treated or watered, with the result that it goes partially brown in the summertime, and its spring growth ebbs and fows with the life cycles of the various lawn-like substances (grass, weeds, fowers) that constitute the yard. This doesn’t bother the couple in the slightest – in fact, they rather like the natural look – but the couple’s parents feel differently (pick a set of parents, it doesn’t matter which). In their view every yard should have sod (or at least look that way) and it really ought never to appear brown, even in the summer. Of course, they tolerate the suboptimal yard owned by their children and hardly ever say anything about it, except perhaps the occasional passive-aggressive remark, which everyone knows is a parent’s prerogative in any case. Still, every time they visit the house they are struck with wonder at how anyone can look at the state of the yard and not put re-landscaping toward the top of the to-do list. It’s a good question – one to which we will return – why paradigm cases of judgmentalism so often involve parents and children, but for now let’s simply accept that this is indeed a paradigm case of judgmentalism. What exactly is it about these yard-based judgments that makes them different from, say, the judgment that the house is made of red bricks? Why does making the latter judgment not count as being judgmental, whereas making the former judgments does? And what exactly is wrong with making those judgments? (Or are we wrong in the frst place to suppose that judgmentalism is about what judgments one makes?)

258

Neal A. Tognazzini

Now, the theory. Not my theory: Gary Watson’s. Watson has suggested that judgmentalism is a second-order vice: a vice “pertaining to how we respond to the moral shortcomings of ourselves and others” (Watson 2013: 283). In Watson’s view, judgmentalism manifests in two fundamentally connected ways: as interpretive ungenerosity, on the one hand, and as being too unaccepting of faults, on the other. At bottom, both are about unacceptance leading to interpersonal distance or hostility, and it is this that makes being at the receiving end of judgmental remarks so painful. In light of this account, Watson advocates acceptance as the non-judgmental ideal. Let me say a bit more about each manifestation and about how they are connected. Consider interpretive ungenerosity frst. The thought here is that many cases of judgmentalism involve overlaying a particular interpretation onto the words or actions of another when there is another, more generous, interpretation that is equally consistent with the evidence. Consider an example that Watson invokes, namely the hurt and anger that Beethoven expressed in the Heiligenstadt Testament, after his hearing had begun to fail him and yet he felt obligated to keep it a secret: “Oh you men who think or say that I am malevolent, stubborn or misanthropic, how greatly do you wrong me. You do not know the secret cause which makes me seem that way to you” (Swafford 2014: 302). The interpretively ungenerous person is disposed to jump to unfattering conclusions or perhaps simply fails to see that less damning conclusions are available. Of course, not every false interpretation is necessarily ungenerous – perhaps in the case of Beethoven the men in question were being as generous as the evidence allowed – but what Watson seems to be suggesting here is that one way to fght against becoming a judgmental person is to cultivate an active imagination that allows you to avoid adopting the theory of least resistance and applying it wholesale to your understanding of the person. As Watson says, “the interpretively generous person will be more hesitant to epitomize” another person in the terms of even a generally accurate interpretation (Watson 2013: 291). On Watson’s account, the second way judgmentalism manifests is through being too unaccepting of the faults of others – or, perhaps better, being too unaccepting of the perceived faults of others. In many of our relationships we adopt implicit standards, or expectations, that need to be met in order for us to be on “fully good terms” with the person (Watson 2013: 293), and while this is mostly unproblematic, the vice of judgmentalism shows itself when we adopt standards that are unreasonable or too exacting. As an example, Watson offers the pacifst father who uses his daughter’s decision to join the military as a reason to cease communicating with her. The question of military service may be a deeply important, morally weighty issue for both of them, but we view the father as judgmental to the extent that we think him unreasonable for letting that issue be the litmus test for being on good terms with his

Silence and Salience

259

daughter. (As Watson points out, claims that someone is being judgmental will often be controversial in a way that mirrors controversy about moral questions more generally.) The problem here isn’t that the father overlays onto his daughter’s decision an uncharitable interpretation, but rather that he lets that decision completely color his vision of her character. His judgment that she is doing something morally objectionable may even be correct, but it’s the way that judgment about her decision leads him to a judgment about her that makes his reaction a vice. I said earlier that for Watson, these two manifestations of judgmentalism are connected. The connection is that when an ungenerous interpretation counts as judgmental, it is because the interpretation “serves as a prelude to and a pretext for a dismissal or rejection” of the person whose actions are being interpreted. The interpretation is “in effect [a] brief for nonacceptance, for a stance of rejection if not hostility, or at least for maintaining the distance of superiority” (Watson 2013: 291). So at its heart non-judgmentalism is about acceptance; it’s about the demands we place on our relationships with others and about how we evaluate whether the other person has met those demands (Watson 2013: 294). With Watson’s account in hand, return now to the case of the parents and their yardwork-eschewing children. It’s easy to see this as a case of being interpretively ungenerous: the parents view the state of the yard as resulting perhaps from laziness on the part of their children, when in reality the children just prize a more “natural” aesthetic or perhaps think that using sprinklers is a waste of water or perhaps just don’t have the time to worry about their yard since they have a newborn baby in the house. A differently kempt yard certainly isn’t suffcient evidence of a negligent homeowner and even more certainly isn’t suffcient evidence of any sort of character faw. It’s perhaps more diffcult to see how the case might be embellished to ft with the idea that judgmentalism involves non-acceptance – how the parents could think that their relationship with their children is somehow on less than fully good terms because of their yard – but it helps to recall Watson’s remark that non-acceptance might manifest as “maintaining the distance of superiority.” The passive-aggressive remarks (“Oh did you want me to put out the sprinkler this morning?” “No, mom, we don’t own sprinklers, remember?”) are hurtful precisely because they imply a judgment of inferiority. And to the extent that this causes alienation or emotional distance, it seems as though the parents have not fully accepted the perceived faults of their children. Summarizing his account, Watson says that “we should locate the primary vice of judgmentalism in the faulty ways in which one’s judgment conditions one’s relations with others” (Watson 2013: 287), where the judgment in question is a sweeping overall assessment of the person on the basis of a perceived (perhaps ungenerously interpreted) fault, what

260

Neal A. Tognazzini

Watson calls a “verdictive” judgment (292). This account strikes me as illuminating, helpful, and compelling.

3 Still, Watson’s account doesn’t seem quite right, and refecting a bit further on the case of the parents can help us to see why. Watson’s claim is that judgmentalism involves noticing an alleged fault about someone and then holding that fault against them. Judgmental people are alienated or at an emotional distance from those they judge, on Watson’s account, because that alienation is simply part of what it is to be judgmental. While I don’t deny that alienation is often a consequence of judgmentalism, I’m skeptical that it is required. What if, instead of thinking themselves superior to their children in the arena of homeownership, the parents and the children managed to remain on fully good terms with each other? Even if the parents don’t hold the state of the yard against their children in any verdictive way, wouldn’t they still count as judgmental simply by virtue of always noticing the state of the yard, as though it were something worthy of notice? Watson does draw a distinction between someone who is merely hypercritical and someone who is judgmental. In his view, since criticism may originate from a place of love, and need not imply that the criticizer views the relationship as in any way impaired, the hypercritical parent need not count as judgmental. But it strikes me as a false dichotomy to claim that either criticism comes from a place of love or else it is used as the basis for distancing oneself from the person being criticized. It seems as though there is a variation on the case of the parents and children according to which the parents’ judgments neither originate from a place of love, nor are used as a basis for dismissal or distance. What strikes me as judgmental in this case isn’t that the alleged fault is noticed and then held against the children. The mere fact that it is noticed seems to be enough.

4 To see the point more clearly, let’s talk for a minute about pudding. In a wonderful scene in Charles Dickens’s A Christmas Carol, the Cratchit family is just sitting down for their Christmas feast, the one time each year that the family “splurges,” though their poverty would make their splurge meal look rather more like a snack to Ebeneezer Scrooge (and, let’s face it, to most of us). There’s nothing but effusive praise for the meal all around, and dessert is the pièce de résistance: Oh, a wonderful pudding. Bob Cratchit said, and calmly too, that he regarded it as the greatest success achieved by Mrs. Cratchit since their marriage. Mrs. Cratchit said that now the weight was off her

Silence and Salience

261

mind, she would confess she had had her doubts about the quantity of four. Everybody had something to say about it, but nobody said or thought it was at all a small pudding for a large family. It would have been fat heresy to do so. Any Cratchit would have blushed to hint at such a thing. Dickens 2004: 103 Anyone at that table would have been able to report on what was not said, but only the omniscient narrator of the story can also inform us about what was not thought. And in this scene, the Cratchits don’t just fail to say that the pudding is small, they also fail even to think it. To say or think such a thing would have been heresy. Why heresy? Well, if orthodoxy is a set of beliefs or attitudes prescribed to the members of a group, then a member of that group is heretical when they hold (or espouse) a belief or attitude that is “out of bounds.” The Cratchit family is such a group, and it would be inconsistent with full membership in that group to show oneself ungrateful for rare treats like Christmas pudding. And notably, what would count as ungrateful isn’t just complaining about the size of the pudding; it’s even thinking that the pudding is small. In other words: being a member of the Cratchit family involves a certain orientation, or frame of mind, that draws one’s attention toward, and away from, certain aspects of the world. Watson uses the term “fault-fnders” for those who are “inordinately preoccupied with the putative misdeeds of others” (Watson 2013: 291), and he suggests that fault-fnding has its source in fault-tracking. The problem with fault-tracking, Watson suggests, is that it is often in the service of forming a verdictive judgment – the sort of judgment that may, if the circumstances are right, constitute an expression of judgmentalism. I agree that fault-tracking is problematic because it is often a prelude to a verdictive judgment, but it seems to me that sometimes, regardless of the overall verdict, merely tracking a fault is problematic on its own. This, I suggest, is what gives the pudding scene its warm glow: the Cratchits don’t simply refuse to come to a verdictive judgment about their parents’ ability to provide; they aren’t even tracking the alleged fault that could lead to such a judgment. In other words, not only are the Cratchits obeying the maxim not to say everything they think; they are also obeying the maxim not to think everything that’s true. And as I suggested at the outset, it’s this maxim that takes us to the heart of a more complete account of judgmentalism.

5 I suggested previously that being a member of the Cratchit family involves an orientation, and I suspect that’s generally true for relationships. But what exactly is an orientation? I’m not entirely sure, but I will try to say

262

Neal A. Tognazzini

a few helpful things. At the most basic (and yet metaphorical) level, an orientation is something like a way of seeing the world, a lens through which you see things. And that lens, that orientation, makes certain aspects of your environment salient while at the same time forcing other aspects of your environment to go silent. Consider, for example, Wittgenstein’s response, at Philosophical Investigations II.iv, to the problem of other minds: “My attitude toward him is an attitude toward a soul. I am not of the opinion that he has a soul” (Wittgenstein 1953: 152). One way to understand this cryptic remark is as expressing the view that there’s a difference between the beliefs we form, on the one hand, and the background framework we (already) occupy when we form beliefs, on the other. Anti-solipsism is not a proposition to marshal reasons in favor or against; it is part of the framework, the lens through which we experience the world. When I’m thinking of orientations, I have something like this in mind. Less cryptic is the broadly perceptual theory of emotions favored by Robert C. Roberts (2003), according to which they are concern-based construals. When you fear the spider, for example, you are experiencing an amalgam of a perception-like state (a construal that the spider is dangerous) and a desire-like state (a concern not to be harmed). Notably, the notion of a construal is subjective, in the sense that you and I might construe the very same object in radically different ways without either construal having to be incorrect. Like a duck-rabbit, the world often admits of more than one Gestalt. Also, although Roberts thinks that emotions are associated with characteristic judgments, they are not identical to any judgment. As many of us know well, fear of something may persist despite a wholehearted judgment that the thing feared is not dangerous, and a natural way to make sense of this is to conceptualize emotions as construals, ways the world seems, even if we know that’s not the way the world is. Again, when I think of the praiseworthy Cratchit orientation, I have something like this in mind. Helen Longino (1979) points out that what aspects of our environment we are inclined to count as evidence – and what theories they count in favor of – depends crucially on the paradigms we bring to an investigation. She points out, for example, that if we are working within a geocentric paradigm, the datum that night follows day with predictable regularity will count as evidence for the conclusion that the sun circles the Earth at a regular rate. But if we are working within a heliocentric paradigm, the very same datum will count in favor of the conclusion that the Earth circles the sun at a regular rate (Longino 1979: 42). And of course it’s only if we already have some conception of celestial bodies at all that we will be inclined even to count the regularity of the night/ day pattern as evidence for anything at all. Longino is making a point about how evidence is treated in scientifc inquiry, but the general point holds true further afeld. Background frameworks and implicit general

Silence and Salience

263

understandings are what guide our eye toward certain aspects of our experience and away from others. What I want to suggest is that our interpersonal relationships work in a similar way: our orientation toward others guides our eye toward some aspects of them and away from others. What I’m driving at is similar to ideas found in Margaret Olivia Little’s (1995) work on caring and Troy Jollimore’s (2011) work on love. In developing an account of the epistemic signifcance of emotions, Little makes the point that what we see reveals how we care. She says: What one is attentive to refects one’s interests, desires, in brief, what one cares about. [. . .] More generally put, if one cares about something, one is prepared to respond on its behalf, and preparedness to respond is intimately linked with awareness of opportunities to do so. Little 1995: 122 Jollimore makes a similar point in developing his perceptual model of love. Here’s the way he puts it: Personal relationships [. . .] form part of the background against which practical reasoning, including the perception of one’s reasons, takes place; what counts as a reason is determined largely by the relationships and value commitments one brings to the situation. Jollimore 2011: 115 Instead of focusing on the salience of certain reasons to believe a scientifc theory about the solar system, Little and Jollimore are pointing out that one’s background framework (one’s emotions, one’s standing relationships) can make salient certain practical reasons as well. And it’s not just that you wouldn’t have noticed that you had those reasons for action without the relevant background; rather, it’s that, without the relevant background, certain facts wouldn’t even have counted as reasons for action at all. Both of those things – counting and noticing – are relevant to the point I want to make about the Cratchit family. Whereas the size of the pudding may count as a reason for you and I to complain or feel disappointed (which is part of what gives our view on that scene its poignancy), it certainly doesn’t count in that way for members of the Cratchit family. And because it doesn’t, they don’t even notice the size. It’s not that they notice it and yet manage to avoid being disappointed; it’s that the size is not even on their radar to begin with. I have a suspicion that many aspects of our interpersonal relationships can be understood as a type of orientation: not just love (as Jollimore argues) but also blame and forgiveness, perhaps even faith. But that’s a project for another time. For now I hope these remarks have made

264

Neal A. Tognazzini

tolerably clear what I have in mind when I speak of an orientation. I now want to take this idea back to our discussion of judgmentalism.

6 Recall that according to Watson, a judgmental person is one who is disposed to draw alienating verdictive judgments about others, where those judgments are the result either of unreasonable demands placed on the terms of a relationship or else of unreasonable assessments of whether the others have met those demands. My case of the parents and their yard-neglecting children was meant to suggest that there is a strand of judgmentalism not captured by Watson’s account – in particular, the sort of judgmentalism that consists of taking certain facts about others to be a relevant basis on which to form judgments at all, even of the non-verdictive variety. In fact, I wanted to suggest something even a bit stronger than that, namely that judgmentalism might manifest as the mere noticing of certain facts about others, regardless of their perceived relevance to judgment-making. With the notion of an orientation in hand, we can now accommodate these suggestions: to be judgmental is to make judgments (where this can include explicitly forming beliefs or perhaps just patterns of noticing) whose presence is (or whose consequences are) inconsistent with the orientation that constitutes the type of relationship in question, whether or not those judgments are alienating or verdictive. Judgmentalism, then, stems ultimately from disorientation. I don’t mean to suggest, though, that judgmental people are always disoriented through their own culpable ignorance or incompetence. An orientation is a relation, which means that there are two ways to be (or become) disoriented: you may have gotten yourself lost through your own movements, or you might have become lost due to the movements of the world around you. Relationships are dynamic things, and I suspect that many examples of judgmentalism are simply instances where someone has failed to keep up with the dynamism, perhaps even failed to realize that things have changed. This, I think, is why the most ready-to-hand examples of judgmentalism are examples involving parents and children, as mine was. (Watson gives four examples in his paper; three involve parenthood and the fourth involves an implicit age gap.) As children grow, the nature of the parent/ child relationship changes (should change, anyway) radically, and it can be hard for parents to keep up. It’s hard to think of a compelling case of parental judgmentalism when the child is a newborn baby, for example. Why is that? One reason is that judgmentalism most often manifests itself in judgments about how one lives one’s life, and babies aren’t really “living a life” in the relevant sense. But arguably, it is also because the nature of the parent/child relationship, especially in the early stages, involves an

Silence and Salience

265

orientation where every fact about the child, no matter how “personal,” is fair game for the parental eye to notice and take into consideration. This all-encompassing orientation, in fact, is part of what allows parents to be good parents. But this all-encompassing orientation very often overstays its welcome as children grow and begin making their own decisions about how to live their lives. When children are young, there’s nothing judgmental about a parent noticing a messy room and encouraging their child to clean it. But I hope messy rooms are something that I stop even thinking about once my daughter reaches adulthood. It’s not just that her room becomes her business in a way it didn’t used to be, so that I should keep my mouth shut; it’s also that, in a manner of speaking, I should keep my eyes shut. Of course there is a worry that by adopting an orientation that dampens my awareness of certain facts, I’ll perhaps miss important indications that not all is well with a friend or a child. That messy room, after all, might be a sign of trouble. And of course it is important to be attuned to what counts as “business as usual” for a friend, but this idea of attunement also seems to be a matter of background understanding and not something that necessarily crosses one’s conscious radar, unless something is “off.” It’s not as though I need to pay attention to the state of my friend’s room whenever I’m there: even off-radar facts can rise to consciousness when they present a breakdown in background understanding. On the other hand, it may just be that this is one more item to put on the list of vulnerabilities that we are susceptible to as a result of being involved in intimate relationships.2 Deep trust in the faithfulness of one’s spouse doesn’t just involve refraining from always asking them what they are up to when they are out with friends; it also involves not even considering the possibility that they might be up to no good. (Unless and until one is forced to consider that possibility by something out of the ordinary.) That opens one up to being seriously hurt, true, but it’s hard to see why that should be a reason to look for a different account of trust.

7 I turn, fnally, to the question of what exactly is wrong with judgmentalism. On Watson’s account, judgmentalism is a second-order vice: to be judgmental is to be disposed to respond in faulty ways to the perceived faults of others. On my expanded version of the Watsonian account, too, judgmentalism is a vice, it’s just that it can also manifest by being disposed even to perceive the faults of others, or to perceive them as faults. But I’m not sure that calling it a vice exhausts what’s ethically problematic about being judgmental. In particular, a judgmental person hasn’t just failed to live up to certain ethical ideals; they have also often wronged the person

266

Neal A. Tognazzini

about whom they are making their judgments. And it’s not clear that a mere vice can constitute a directed wrong. Of course, Watson does say that a judgmental disposition will tend to lead to alienation and estrangement, and these ways of holding something against someone can easily ft the mold of directed wrongs, especially when they are due to placing unreasonable standards on acceptance. But I have maintained that someone can be judgmental even without their judgments leading to alienation and estrangement, so how can I make sense of the apparent directedness of the wrong involved in being judgmental? Angela Smith (2011) draws a distinction (the formulation of which she attributes to Laurence BonJour) between “merely behavioral friendship,” on the one hand, and “attitudinal friendship,” on the other. Whereas the former merely has the outward trappings of friendship, the latter involves, in addition, “the presence of attitudes of sincere care and concern” (Smith 2011: 251). In Smith’s view, true friendship is attitudinal: The relation of being a friend, I would contend, is a relation with certain normative demands, expectations, and responsibilities built into it. These demands, expectations, and responsibilities pertain not only to one’s outward behavior, but to one’s attitudes, as well. We reasonably expect our friends to have attitudes of care and concern for us, to respect us, to take pleasure in our accomplishments and feel sadness in our losses. Indeed, when we speak of the “duties of friendship,” we have in mind this whole complex of behavioral and attitudinal demands and responsibilities. Smith 2011: 251 But how exactly can the failure to have certain attitudes constitute a failure of an obligation that we owe to our friends? Set aside the question of whether we have the right sort of control over our attitudes for them to be the proper target of responsibility attributions (though see Smith (2005) to clear up any misconceptions on this score). The question now is how the failure to have a certain attitude (or the having of a certain attitude) can count as a directed wrong. Smith argues that adopting a contractualist moral framework can help to make sense of this. On a contractualist framework, moral principles are principles that no one could reasonably reject being bound by. So the question would be whether anyone could reasonably reject a principle requiring friends or intimates to have certain attitudes toward each other in addition to treating each other well. Plausibly, the answer is no, in which case you could manage to wrong a friend even if you don’t treat them badly, but merely fail to have attitudes that are required by the relationship in question (or, I suppose, have attitudes that are inconsistent with the relationship in question). Of course, the contractualist

Silence and Salience

267

framework is controversial, but it does provide a nice framework with which to make sense of the claim that to be judgmental is not just to exhibit a vice but also sometimes a way of wronging the person you are judging.3

8 One effcient way to come across as condescending is to take up what P. F. Strawson calls “the objective attitude” toward one of your friends or loved ones. This is a stance – or, perhaps, an orientation – from within which you view another person as an object “to be managed or handled or cured or trained” rather than taken seriously as a person.4 The objective attitude isn’t a bad thing in itself: as Strawson points out, it’s often exactly what’s called for in clinical contexts, or as an emotional escape from “the strains of involvement.” But to treat your spouse’s anger as, say, being fully explained by their hunger instead of by their justifable objection to being mistreated, is to use the objective stance as a pedestal on which to stand in superiority. In a way, I think that a judgmental person is often guilty of a similar offense. When we are fully immersed in a relationship – “involved,” as Strawson would put it – we are inside of a framework, operating within an orientation that screens off certain facts about the other person. But those facts become salient as we detach, as we begin to use a more objective eye to look upon the other person, and take the measure of their idiosyncrasies. In the right context – say, the clinical context – taking such measurements need not amount to being judgmental. You are meeting with the therapist precisely so that they can take the measurements and help you make sense of them. But the people in our lives we view as judgmental deserve the label in part because they are not our therapists; they are our friends, parents, loved ones. And in the context of those intimate relationships, being treated like a patient is bound to hurt our feelings. This idea – that being judgmental involves inappropriate detachment and measurement-taking – provides another explanation of how someone might unwittingly, almost innocently, become judgmental. (The frst explanation, offered, was that parents have a hard time keeping up with the way their relationship with their children changes as their children grow.) In brief, the idea is that judgmentalism is one of the hazards of striving for an authentic existence. To see what I have in mind, consider a distinction offered by Heidegger which is similar to Strawson’s. Heidegger distinguishes between a stance we take up toward other people (what he calls “Being-with”) and a stance we take up toward objects (whether those objects are approached as “ready-to-hand” tools or as “present-at-hand” items ft for scientifc inquiry). For Heidegger, this way of encountering other people is built into the framework, into our very nature as the sorts of beings we are,

268

Neal A. Tognazzini

and it’s what dissolves the alleged problem of other minds (Heidegger 1927: 153–163). (Recall from earlier Wittgenstein’s remark in response to the problem of other minds: “My attitude toward him is an attitude toward a soul.”) But on Heidegger’s view, this deep fact about us – that we are “always already” oriented toward other people as people – also brings along with it the threat of inauthenticity, the threat that this involved stance toward other people will leave us with no one in particular to be (we end up being absorbed into what Heidegger calls “the they,” the crowd). So how to avoid absorption and inauthenticity? For Heidegger, the answer is to become keenly aware of one’s own mortality, but that’s not exactly a cheery solution, so here’s an easier way out: remind yourself of all the ways that you differ from others, mark contrasts, develop a sense of your own inner identity over against the identities of others. Discover who you truly are, as they say. Well and good, except that the project of fnding contrasts with others requires taking notice of and measuring the qualities of others, so that you can use those measurements to carve out a distinctive place that’s all yours. Thus embarking on the project of becoming a self of one’s own – an authentic individual – seems to require a measure of detachment and so brings in its wake the risk of judgmentalism. It’s a dilemma I’m not entirely sure we can escape.5

Notes 1. 2. 3. 4. 5.

For details, see Tognazzini/Coates (2018, especially section 2.3). On this theme, see Cocking/Kennett (2000). On the question of how beliefs can wrong people, see also Basu (2019). Strawson (1962), as reprinted in Watson (2003: 79). For helpful comments on this project, thanks very much D. Justin Coates and Sebastian Schmidt. I presented an early draft of these ideas at the University of Puget Sound, so I’d also like to express my gratitude to the philosophers and students there, especially Sara Protasi. Finally, thanks to the editors of this volume for inviting me to contribute.

Bibliography Basu, Rima (2019): “The Wrongs of Racist Beliefs”, Philosophical Studies 176, 2497–2515. Cocking, Dean/Kennett, Jeanette (2000): “Friendship and Moral Danger”, The Journal of Philosophy 97, 278–296. Dickens, Charles (2004): The Annotated Christmas Carol, ed. by Hearn, New York: W. W. Norton & Co, Inc. Heidegger, Martin (1927): Being and Time, transl. by Macquarrie/Robinson, New York: Harper & Row 2008. Jollimore, Troy (2011): Love’s Vision, Princeton, NJ: Princeton University Press. Little, Margaret O. (1995): “Seeing and Caring: The Role of Affect in Feminist Moral Epistemology”, Hypatia 10, 117–137.

Silence and Salience

269

Longino, Helen (1979): “Evidence and Hypothesis: An Analysis of Evidential Relations”, Philosophy of Science 46, 35–56. Roberts, Robert C. (2003): Emotions: An Essay in Aid of Moral Psychology, Cambridge: Cambridge University Press. Smith, Angela (2005): “Responsibility for Attitudes”, Ethics 115, 236–271. Smith, Angela (2011): “Guilty Thoughts”, in: Morality and the Emotions, ed. by Bagnoli, New York: Oxford University Press, 235–256. Strawson, Peter F. (1962): “Freedom and Resentment”, Proceedings of the British Academy 48, 1–25. Swafford, Jan (2014): Beethoven: Anguish and Triumph, New York: Houghton Miffin Harcourt. Tognazzini, Neal A./Coates, D. Justin (2018): “Blame”, in: The Stanford Encyclopedia of Philosophy (Fall 2018 Edition), ed. by Zalta, URL = https://plato. stanford.edu/archives/fall2018/entries/blame/. Watson, Gary (ed.) (2003): Free Will, 2nd ed., New York: Oxford University Press. Watson, Gary (2013): “Standing in Judgment”, in: Blame: Its Nature and Norms, ed. by Coates/Tognazzini, New York: Oxford University Press, 282–301. Wittgenstein, Ludwig (1953): Philosophical Investigations, Oxford: Blackwell 2001.

14 The Value of a Free and Wandering Mind Miriam Schleifer McCormick

1.

Introduction

The targets of our normative judgments go beyond voluntary action and include mental states and attitudes. Many common reactions to each other’s beliefs, for example, exhibit attitudes that display negative judgments and emotions; someone’s belief can elicit what Peter Strawson referred to as “reactive attitudes” such as anger and contempt. We ask in an incredulous tone, “How can you believe that?” or exclaim, “What a ridiculous thing to believe!” We criticize each other for being angry, or fearful when we shouldn’t be, or failing to be angry or fearful when we should. Depending on the state being targeted, the kind of attitude or judgment that is appropriate to have towards it varies. In previous work I have been concerned with widening the scope of agency beyond that which is under our direct voluntary control and have argued that some states of mind are appropriate targets of certain reactive attitudes even if they cannot be directly controlled. I have argued, for example, that we can be responsible for our beliefs and, that some hopes are rational or appropriate and some not; I can fault you for not hoping well, or judge that, in this context, this is a hope you should not have (thus invoking a certain kind of reproach). In characterizing the conditions by which to evaluate hope’s rationality I have argued that different dimensions of assessment track different dimensions of agency and that, in hoping, one’s agency can be augmented or diminished in a number of ways. Many others have been engaged in a similar kind of project, and disputes center on how best to characterize the conditions needed such that one is the appropriate target of such attitudes.1 But in thinking about these matters I have become worried that the scope of agency can be widened too far so that no area of mind is beyond the reach of appropriate assessment and judgement. I begin with something like an intuition, though perhaps more of a conviction, that there is, or ought to be, a domain of the mind that is completely free of normative assessment, where you are safe to let your thoughts and images go wherever they take you without concern that you are doing anything wrong,

The Value of a Free and Wandering Mind

271

where praise and shame do not apply.2 I am not sure how widely shared this intuition is, but it seems that there is wide agreement that this is the case concerning dreams while we sleep. While it is undoubtedly the case that one can wake from a dream and feel some shame for what occurred, I think it is not called for. But on certain conceptions of what matters for being an appropriate target of reactive attitudes (and so responsible in some sense), even dreams fall within the scope. This was ultimately Augustine’s view. For dreams may well be expressive of something deep about you, your character, your desires; it is you who authors them. I will begin (in Section 1) by offering an example of the kind of state I think should be beyond normative judgment; I argue that certain kinds of wakeful fantasies, namely ones which are motivationally inert, are on par with sleeping dreams. If one shares my view that there is a “free” domain of the wakeful mind, then what I am doing can be seen as clarifying why such states exempt them from judgment. If one does not share this intuition, then what I am doing can be seen as specifying what criteria would be needed for a kind of state (or domain) to be free in this sense. And then some may argue that no wakeful fantasies satisfy these criteria. I will address those arguments (in Section 2) and argue that if the fantasies as characterized are appropriate targets of normative assessment then it will be very diffcult to exempt dreams of sleep, as well as other exercises of imagination. Of course, some people (like Augustine and surprisingly many others) will not mind this result.3 I don’t think then that is the end of the discussion, stalemate and parting of intuitions. For I think a case can be made for the value of having a realm of imagination that is beyond the reach of any kind of judgment. I present this case in the paper’s fnal section.

2.

Locating the Space: Pure Fantasy

I am going to offer an example of a kind of mental state with the hallmarks of one where normative assessment (at least of the kind where what you are doing is reproachable in any sense) is inapt, where it would be akin to reproaching you for the size of your nose. It may be that other nearby states share the features required for exemption, but I will focus on what I am calling “pure fantasies,” partly because I think this is a state that many would not want to exempt from the realm of judgment. If we begin with dreams of sleep as our model, we can begin by thinking about what kind of wakeful state most resemble dreams. An obvious place to begin is with “daydreams.” What is a daydream? You let your mind wander, without a clear purpose or intention. But one can do this and end up anxiously obsessing about one’s “to do” list or the potential consequences of the latest of Donald Trump’s tweets. This is not daydreaming. One requirement is that it involves some mental imaging, and the second is that it has a kind of narrative structure. It needn’t be a very

272

Miriam Schleifer McCormick

coherent narrative, no more so than dreams. But when you wake from a dream you can write it down or tell someone what happened. I will call these kinds of wandering narratives (with mental images) “fantasies.” A third feature that seems needed to distinguish fantasies (or daydreams) from other exercise of the imagination is that they have an overall positive valence. It certainly seems constitutive of daydreams that they are pleasant. We think of the student smiling wistfully in class and then being brought back to the present task by the teacher repeating the question that she has not heard numerous times. The dictionary defnes a daydream as “a series of pleasant thoughts that distract one’s attention from the present.” This may seem to distinguish them from dreams of sleep, which are not always pleasant, but perhaps that is why we have the term “nightmares.” And, in general, fantasy also has this positive connotation; it is the pleasurably imagining that distinguishes them from other kinds of imagining. Aaron Smuts, one of the few to discuss the ethics of fantasy says “the notion of a sad fantasy is incoherent” (2015: 385). This may be a bit strong but, in any case, I will include this third feature in what counts as a “fantasy,” partly because it is the fnding the images pleasurable that some may think is what makes it worthy of reproach. I will clarify the kind of fantasizing I have in mind by frst distinguishing it from fantasies that accompany hope, and fantasies that accompany desires. 2.1 Fantasizing Without Hope When trying to become clear about what distinguishes hopes from desires more generally, many point out that when you hope for something as opposed to merely desiring it, there is a focus on the possibility of the hoped-for outcome obtaining that affects you in various ways. It can often affect what you actually do, but even more so it leads you to devote mental and emotional energy toward the hoped-for outcome. This kind of focus and energy that hope elicits can help to explain its potential motivational power. For a mental state to count as a hope it needs to take up space, as it were, in our mental landscape. Its dominating presence is what allows it be such a powerful force. One way that this energy is manifest is in fantasizing about the hopedfor outcome. Luc Bovens argues that this kind of mental imaging is constitutive of hope, that it is this imaging which distinguishes hope from other desires. While I think it is possible to hope without fantasy, it is a very common and central way that hopeful energies are manifest, or, as Adrienne Martin puts it, it is a signifcant way of expressing hope, and further when we encourage others to hope, we often encourage this kind of expression. If a fantasy is expressive of hope, then it is open to normative assessment. In such a context when I ask “Should I have this fantasy?” I am asking “Should I have this hope?” While exactly what it

The Value of a Free and Wandering Mind

273

takes for a mental state to be open to normative assessment (and that may well depend on the kind of normative assessment at stake) is part of what I am here trying to get clear on, at least this much seems needed: that it is the type of state that is, at least in principle, responsive to reasons in a certain way; where it makes sense for me to ask why you have such an attitude, that there could be reasons for you not to have the attitude, which may function in your revising it. As Kate Nolf helpfully puts it, “being in a rationally evaluable mental state (e.g. believing that p) paradigmatically involves being answerable – being responsible, in some normatively signifcant sense of the term – for being in the state” (2015: 45). Hope is such a mental state. It makes sense for me to ask you why you have hope, or for me to ask you to give me reasons for hope. But, while some degree of fantasy may frequently accompany hope, the converse is not true: one can fantasize without hoping. It may be easiest to get a grip on how fantasy can come without hope to think about a child playing make-believe. She does not hope to be an elephant or to be the villain who gets shot by the super-hero. But trying to articulate the difference between these two kinds of fantasizing is surprisingly diffcult. Both the desire-like and the doxastic attitudes contained in fantasy that is not hope-induced are different from fantasy that is expressive of hope. I will begin by discussing the differences in doxastic attitudes, which is more straightforward. One can believe that a hoped-for outcome is very unlikely, but one cannot hope for what one believes is impossible. Depending on what kind of possibility we are concerned with, this may not be quite right. A religious person may hope for a miracle while recognizing that this would mean hoping for a violation of causal laws, and this would mean it is possible to hope for what one believes is causally impossible. But in such a case, the person believes a miracle is possible! Since most non-philosophers do not distinguish metaphysical from causal possibility, I think a better way of making the point is to put it in terms of credences; one cannot hope for something to which one has a credence of 0. But there is no such constraint on fantasy. Even if one does not believe in miracles, has zero credence that one can fy around the room unassisted, or does not believe that someone dead can return to life, one can fantasize about it. Now it may be the case that one cannot fantasize about what is logically or conceptually impossible, but this may be because there is no way to produce mental images of such things. There is some debate about whether one can hope for what one believes is certain (has a credence of 1). When I express hope, it always seems to imply uncertainty, but some argue that it would not be incoherent to hope in the context of certainty. Again, we do not have the same diffculty when thinking about hopeless fantasy. I can be minutes away from home after a long journey, certain I will be lying in bed very soon, and still fantasize about the comfort of my pillow.

274

Miriam Schleifer McCormick

We can see then that just attending to the doxastic attitudes found in the two fantasy states reveals that they differ in how they are assessed. In hope, I can evaluate and criticize your probability estimate, and how low the chances are can detract from hope’s overall rationality. None of that seems relevant when it comes to fantasy. I can ask you “How can you hope for something so unlikely? It will just lead to disappointment!” And you may have a response that explains your reasons, or you may alter your state in response to my reasons. But if I were to say (which I probably could not because hopeless fantasies are rarely expressed) “How can you fantasize about something so unlikely?” such a question seems out of a place; in a sense that’s at least in part what such fantasies are for. The difference between the conative attitudes in the two kinds of fantasies is more diffcult to articulate, and it is ultimately more important for it is this attitude that one may think keeps such fantasies within the scope of judgment. When one engages in hopeful fantasy, one has a desire for the hoped-for outcome. But is this the case in a fantasy without hope? One way of getting at the difference is to think that in the former one is fantasizing about the content of the fantasy really coming to be while in the latter one is not. This idea can be made more vivid if we think about darker fantasies, say like murdering an ex, or about some sexual fantasies; one can fantasize about things while not fantasizing about them really coming to be. Think of a Genie appearing and telling you he will make your fantasy a reality. If the fantasy is of something hoped for, one will welcome the Genie’s gift, but there are some fantasies where one would politely decline the generous offer. But perhaps this is the case with certain kinds of hopes as well. One can hope for something without fully endorsing the hope; Bovens talks about a category of “shameful hopes,” that is ones that connect to partial but ultimately ill-considered desires. Sometimes these can get a grip on our mental imagery, even if it is not what we hope for all-things-considered.4 Perhaps I do have a full-fedged desire to murder my ex (or cause him pain), but I have many other conficting desires that override this one. If shameful hopes are possible then the Genie test doesn’t quite work. Others think the kind of desire contained in hope must be an endorsed desire, and perhaps this would be enough to distinguish hopeful fantasies from fantasies without hope. I worry about this for two reasons: First, it seems to make it harder to explain conficting hopes, and second, in many cases, one will not disavow the desire-like attitude found in hopeless fantasy. Ultimately, for a fantasy to be a pure fantasy, it needs to be void of desire. 2.2

Fantasy Without Desire

What is distinctive about the kinds of fantasies I am thinking about, and what I think insulates them from normative judgment, is that they are motivationally inert (or very close to being so).5 But again, it may seem

The Value of a Free and Wandering Mind

275

that we can hope for many things where what we do will make no difference, and that such hopes must also be motivationally inert just because it is completely out of our hands whether what is hoped for will come to be, and if we cannot act at all, how can we be motivated to do so? A prominent way of defning hope in the psychological literature includes that it must be goal directed. C. R. Snyder, whose work on hope has been extremely infuential – who developed the “hope scale” which is a widely used measurement in social psychology – defnes hope like this: “the perceived capability to derive pathways to desired goals, and motivate oneself via agency thinking to use those pathways” (2002: 249). While this way of thinking about hope is too narrow to capture the range of phenomena that we would want to count as hope (indeed it would seem to render it incoherent to hope to win the lottery once the ticket was bought), it does point to something important. This at least seems to be true of hopes, that part of what it means to hope is that one is disposed, if possible, to seek out pathways to making what is hoped-for come to be. It may be that this disposition never has a chance to manifest, but this conditional seems to be the case: If the conditions were to obtain where one could affect the outcome then one would (again all-things-considered). But what about mere desires that are not hopes, either because one does not believe it possible to obtain what one desires (and so despairs), or it is a desire one does not want to have, perhaps is even actively trying to get rid of? Even such desires seem open to normative assessment. Why? Some think that desires include some kind of representation of the object desired as good, or even more that to desire something entails that one believes one has reason to so desire.6 On such accounts, one can be mistaken in one’s representations. But even on less cognitive views of desires, if you desire something, it has some pull on you. It is quite common to have desires one would prefer not to have, but it takes work to overcome them. And the desires you have (or don’t have) tell me something about who you are. Now if desires are intrinsically motivating states, and I am claiming that pure fantasies are motivationally inert, then it follows that such states are void of desires. And I think this is correct; pure fantasies are pleasurable imaginings but without desire. These fantasies then seem much closer to aesthetic experiences. When you take pleasure in looking at a beautiful artwork, you don’t need to desire it; there are times when you would not want it anywhere other than where it is, or depending on its nature – if it is something temporary like a mural – it may not even make sense to say you desire it. And also there can be aspects of it that are dark and disturbing, but the overall experience is one that is positive. Another way of understanding the conative element of pure fantasy and why it differs from those connecting with hope, or desire, is to think that the desires represented in fantasy are not ours; they are not about the actual world and our actual selves, but instead we represent some

276

Miriam Schleifer McCormick

alternative, or hypothetical self as having certain desires. If we also hope for what we are fantasizing about then we may share these desires or see that we should adopt them. But desires of a possible, but not actual, self in a possible, but not actual, world unless connected or adopted do not motivate me anymore than the desires of a fctional character in a book I am reading or a movie I am watching.7

3.

Concerns and Clarifcations

Now that I have isolated the kind of state I think should be normatively on par with dreams of sleep, I will consider some worries about my proposal. The frst are skeptical concerns about whether pure fantasies as I have described them actually exist. The second concerns the strength of my conclusion; even if I have characterized an actual state of mind correctly, is it really true that no normative assessment of such a state is appropriate? 3.1

Is the Idea of a Pure Fantasy Even Coherent?

Given the potential power of fantasy, one may be dubious that it can really be void of desire. Jerome Neu (2002) says this for example: “[t]o make a thought a fantasy there must be a motivation of a particular sort” and further “[t]he problem with certain fantasies is the desires and the attitudes that they reveal.” But Neu’s own discussion allows a way to think of some fantasies as revealing attitudes that are not clearly motivational. Drawing from Freud, he suggests that both night dreams and daydreams manifest wishes, but often in an indirect and distorted form. One cannot even determine that one wishes for the same things that one dreams or fantasizes about. Moreover,“desires have a tendency to inform inclinations toward action. Wishes may be simply idle” (146). In trying to think about which kinds of mental states may be rationally evaluable, we saw one important element is whether the state is responsive to reasons or whether asking for a reason, or a kind of justifcation, is appropriate. And, as an example of one that is not typically a state of this kind, having a headache is often invoked. It does not seem appropriate for you to ask me to justify my headache in the way it is appropriate for you to ask me to justify my belief, or even my anger. But one may say: well, there is some sense in which I can ask you to justify it or to criticize you for being in this state. What if you have a tendency of drinking too much and then having hang-over induced headaches? Can I not say something akin to “you should not have that headache” the way I would “you should not have that belief”? But here it is clear that what I am criticizing you for is the behavior that led to the state, not the state itself. There is no sense in which the state is an inappropriate response to the circumstances you are in, or that it can be altered in response to

The Value of a Free and Wandering Mind

277

recognizing you have reasons to not have a headache. The inclination to judge pure fantasies I think points to something similar; there is the idea that certain fantasies are linked to problematic behavior, either behavior that led to you having the fantasy or perhaps, more importantly, that it will lead to problematic behavior. These worries often center on fantasies that are repeated or frequent; if you keep fantasizing about that then it begins to seem you must actually desire it. I think it is telling when one is arguing for fantasies being open to judgment, that frequency and repetition get invoked. For example when arguing that even fantasies over which one has little control can be the proper target of certain reactive attitudes, Aaron Smuts asks us to “[c]onsider the case of someone who routinely spontaneously fantasizes about violently raping anyone they meet, man, woman, child, or beast. They aren’t a rapist; they just pleasurably imagine raping, and they do it a lot” (2015: 385). If one is constantly fantasizing about something, it is hard to see how this won’t seep into one’s actual motivational structure, and if and when it does then it has left the “no judgment” zone. 3.2 Is No Kind of Normative Assessment Appropriate? I have been referring rather loosely and interchangeably to states being open to “normative” assessment or rational assessment, open to “rational” norms, open to judgment, open to justifcation, or whether we can ask if a state is appropriate, ftting, as well as whether it is an appropriate target of reactive attitudes. In certain contexts, all of these could be, and perhaps should be, distinguished. And many might agree that pure fantasies should be exempted from certain kinds of judgment but think others are appropriate. Most would agree that we are not responsible for our fantasies the same way we are for our actions, that full-fedged blame may well not be appropriate, certainly not punishment. But, still one may think there is a sense in which I can say that it is bad for you to have certain kind of fantasies, or even more generally, that certain kind of pure fantasies are bad. Those who connect blame to action, and in particular acting in a way that violates a duty, suggest that another kind of negative evaluation, appropriate for desires and many emotions is “disesteem” or “dispraise.” Some might say that those who engage in certain kinds of fantasizing are worthy of this kind of negative evaluation. If someone has a desire to torture the innocent, even if they never will act on it, such an attitude in itself is worthy of disesteem. This is partly because of desire’s close connection to action, as is the case with belief. But, putting instrumental concerns aside, it may be bad in itself to desire the bad. The same may be true of other attitudes. In thinking about the ethics of fantasy, Smuts says it is important to distinguish blame from this other kind of moral evaluation.

278

Miriam Schleifer McCormick

He asks us to consider cases where we fnd someone’s attitude, emotional response, or lack of concern off-putting, such as when someone is callously indifferent to the suffering of a small child, and suggests the kinds of reactions we engage in is a different kind of evaluation from that of blame: When we have these kinds of reactions we are engaged in a different kind of evaluation from that of blame. Most plausibly we are holding the off-putting person in disesteem. [. . .] Plausibly, if a person is a ftting object of disesteem, then it would be appropriate for her to feel shame. Smuts 2015: 382 While David Owens uses the term “blame” a little more widely, he also uses the language of esteem. When thinking about certain epistemic vices (such as gullibility) and why I am responsible for the ill-formed beliefs that result from them, he says “[e]ven before anything bad happens, gullibility means that I cannot be an esteemed human being because I cannot be trusted to think and feel as I ought” (2000: 124). Why then would fantasy without desire be exempt from this form of reproach? Could certain pure fantasies reveal that I am not thinking and feeling as I ought? The attitude being assessed in the case of pure fantasy is that of “fnding pleasurable.” We can ask then: Can such a state be bad in itself? When we think about the badness or even wrongness of fnding pleasure and enjoyment in things, it seems they are activities and, often more importantly, activities involving others. If one takes pleasure in causing pain or subjugating others, this is morally problematic. And again, one might think if I take pleasure in thinking about something this is not too far away from taking pleasure in actually doing this thing. But that is, of course, what I have been urging need not be the case with pure fantasies. And so the badness must be only in having certain thoughts. To elicit the intuition that simply having certain kinds of pleasurable thoughts is bad, and the person having them worthy of disesteem, examples of someone fantasizing about raping and torturing children is invoked. Smuts says, for example, “it is a striking example of morally dubious imaginative experience.” And then to show that our attitude does not arise simply because of the worry about what such a fantasizer might do, a hypothetical scenario is introduced where there is no possibility of action, such as Ross’s thought experiment about two worlds: Imagine two worlds, each having just one inhabitant, a sole survivor of a nuclear holocaust (Ross 1930: 140). In world A, the survivor spends her free time thinking nice thoughts. She often imagines cats playing with rubber bands on sunny window sills. In world B, the

The Value of a Free and Wandering Mind

279

survivor lives a similar life, but rather than imagine cats, he imagines raping and torturing children. Is either world preferable? Smuts 2015: 384 That we would fnd world A preferable is supposed to show that it is “intrinsically bad to enjoy evil.” One immediate response to this kind of scenario is that it is not at all clear that these survivors do not have the desires connected with their imaginings; they both may well want their fantasies to come true if that were indeed possible. And, again a person whose mind is flled with these images, and little else, is very different from the person who lets his mind wander, and his fantasies grow without inhibition, even if this leads one to have a pleasurable reaction to something one would not react to with pleasure in any way if it were real. Smuts’s idea is that the reaction to some thoughts (one which includes pleasure) is itself misguided, inapt, unftting. Most of the time when we feel disesteem toward emotional reactions it is reactions to actual events, something real. Smuts recognizes with fantasy “the question here is whether it is intrinsically bad to enjoy imagined evil” (2015: 382), but in his argument supporting this claim the distinction between enjoying evil and enjoying imagined evil is often blurred. For example, we are asked to consider Charlie, who takes great pleasure in seeing animals in pain though he never actually causes them suffering. In this case Charlie does things; he tours slaughterhouses; he views photographs; he clearly desires and even hopes to see such images. There is much in his mental world that lies within the scope of evaluative judgment. This example is supposed to show that his enjoyment itself is morally bad even if it is not bad for the animals. Smuts introduces the two-world example because he thinks that we might be inclined to think our judgment of Charlie comes from a worry that he might someday torture an animal. But as I said, his emotional life and his choices are already appropriate targets for reactive attitudes; he is very far away from pure fantasy. Maybe a lesson from these thought experiments is that it is morally bad to pleasurably imagine things that it would be bad to actually do in real life. Smuts, who ultimately wants to argue that some fantasies (and dreams) are bad in themselves says this is too severe; it would render too many fantasies worthy of moral disapproval. For example, fantasizing about having sex with someone other than one’s spouse doesn’t seem morally bad even though doing so in real life would likely be bad because it would involve betrayal or violations of certain obligations. He suggests instead that it is morally bad to fantasize about doing something that is bad in itself; there is nothing bad in itself about having sex with one’s tennis instructor; what would make it morally bad depends on the context of the relationship; it is the betrayal that is bad. Why is it bad to fantasize about doing something that is bad in itself? In trying to say more about this, what one fnds is repeated images of

280

Miriam Schleifer McCormick

torturing children to convince us that it must be morally problematic to have certain fantasies. And again it is signifcant that the repeated nature is so often invoked: “Although the person isn’t blameworthy, the chronic rape fantasist is plausibly worthy of disesteem. If shame isn’t appropriate in these cases, I’m not sure it ever is” (Smuts 2015: 385). The gist of it is that one may think that having certain fantasies in itself refects poorly on me, that it displays something wrong with my character. But this is exactly what some people will say about dreams of sleep – that having some kind of violent dream displays something problematic about me, something about my deep self or character. Now doubtless dreams tell us something about ourselves as does our fantasy life, but what they reveal is rather opaque and complex; it cannot be taken that a person who has violent dreams is a more violent person, or that if I dream about having sex with my brother that I want to have sex with my brother. Again remember Neu’s discussion about wishing. Even if one isn’t convinced of all of Freud’s dream theory, it is uncontentious that one cannot directly read off from the content of a dream what its meaning is, or what it represents, or that one has a desire, or even a wish that what was dreamed be the case. I have likened pure fantasies to the enjoyment of fction, as well as to dreams of sleep, but it may seem that a difference which would show why fantasies can be assessed and judged in a way the others cannot has to do with the kind of control we have over them, or that I have an authorship over them. Enjoying horror movies or a novel that has graphic descriptions of violence may seem very different from a fantasy of one’s own creation because one is passive and the other active, in some sense. But the active nature is exaggerated; when engaged in fantasy, a thought, or image can lead to another in a way that involves very little conscious thought. Again, watching kids engaged in play can give us a clue how this is so. There is often no deliberation or planning or scripting; when immersed in the make-believe world, they just fow with what that world seems to demand. Again this kind of fantasy is very dream-like. But, surely, one might insist, there is a difference when it comes to control. Our dreams happen to us in a way fantasies do not, and we can see the difference if we think about the fact that if I choose to stop fantasizing I can in a way I cannot if I am dreaming, largely because I am not aware that I am dreaming in a way that I am aware I am fantasizing. Now here is a difference, and one that has some normative signifcance and reveals where judgment about fantasy may well be appropriate. If I am disturbed by a fantasy I can take efforts to stop having it. As I said before, the worry about fantasies is often a worry about their entrenched or repeated nature. And it may well be that I ought to take steps to avoid or halt such entrenchment. But again, something quite similar can be said about dreams. If I have the same violent or deeply disturbing dream over and over it seems

The Value of a Free and Wandering Mind

281

appropriate for me to take steps to understand why this is happening and take steps to alter it. The lack of control we have over our dream life and the amount of control we have over our fantasy life are both exaggerated. In contrasting enjoying disturbing or violent fction and the engaging in pure fantasy, it may even be that the former is more problematic in the context of control. One actively chooses to watch that movie or read that book; there is an intentionality that fantasy lacks and so it may be that if control is an important factor, enjoying the fantasies created by others is more problematic than the fantasies of your own inner life. Ultimately I have a hard time distinguishing the attitude of disesteem from outright blame, unless one takes blame as always warranting punishment. They both have what Andrea Westlund has called a “verdictive” element, a strong passing of judgment that expresses certain “punishing attitudes.” When one is the target of strong verdictive blame, one may feel shamed, berated, attacked [. . .] as if an implicit sentence has been carried out through the expression of blame itself. One experiencing these reactions experiences herself as (metaphorically) having been convicted, sentenced, and punished all at once: the blamer acts as judge, jury, and executioner. Westlund 2018: 2588 She argues, expanding on some insights developed by Hannah Pickard, that we can hold each other “answerable” without blame in this sense. The following long quotation represents what is essential to holding answerable and how it differs from strongly verdictive blame: Holding answerable involves exercising a moral power to demand a response from another, and puts them under a (defeasible) obligation to respond. It is not, however, punitive. [. . .] [H]olding answerable is akin not to a sentencing (at least as that term is commonly understood), but to a summons to moral dialogue. [. . .] The prospect of dialogue highlights the pro tanto nature of the accompanying verdict, and makes it defeasible in practice and not merely in principle. The meaning or signifcance of an action is typically subject to competing interpretations, particularly in morally complex situations. Currents of answerability run in more than one direction, and may lead to mutual moral insight and growth as well as, in some cases, the sharing of associated moral burdens or costs. Holding another answerable, in other words, is not just a way of checking whether a person is indeed blameworthy by scanning her answers for considerations that might excuse or justify. In holding another answerable, one invites her into an alternative perspective on her action, and opens up the possibility of coming to a constructive shared response. Westlund 2018: 259–260

282

Miriam Schleifer McCormick

This is very similar to how I characterized a possible conversation about hope. But even this weaker kind of normative assessment is out of place in the contexts of pure fantasy because, if I am right, there is nothing of normative signifcance to assess, nothing to answer for. Pure fantasies are “ftless” imaginings akin to supposition for the sake of argument.9 I do want to concede that one kind of evaluation that may be appropriate is a kind of aesthetic evaluation. When we criticize someone for liking bad reality tv shows or video games, our primary criticism is that they have bad taste. Now it could be that if one could inhabit another mind and witness the images that were being found pleasurable, that one would come to a similar view regarding taste. But for aesthetic evaluations to lead to appropriate reactive judgments more is needed. We begin to impugn someone’s character if certain kinds of pleasure and enjoyment begin to take up too much space and time. If you spend the bulk of your life watching tv or playing video games, the kind of criticism that may well be appropriate goes beyond the aesthetic. Even though you may not be blameworthy, some kind of disesteem seems apt. When Susan Wolf asks us to consider “The Blob,” someone who “spends day after day, or night after night, in front of a television set, drinking beer and watching situation comedies” (2015: 92), most of us agree with her that there is something important missing from such a life, even if the person feels good and isn’t hurting anyone. But the problem is not with the activities themselves, or the taking pleasure in them, but with the absence of anything else beyond such activities.

4.

The Value of a Judgment-Free Zone in Mental Life

I said at the outset that there may well be value to having a norm-free domain in our mental life. When looking at summaries of empirical literature the results seem mixed on what the consequences of fantasizing are. Does letting one’s mind go to violent imaginings increase or decrease the tendencies to actual violence? One fnds studies on supporting both sides, but these studies are almost all about consumption of violent or pornographic images, not simply images of one’s own creation. In discussing such fndings Smuts concludes that “2,400 years of debate have failed to establish anything close to a reliable and systematic behavioral impact of fction” and that “most plausibly, enjoying fctional suffering does not lead to anything bad, at least not with signifcant regularity” (2017: 112). He, as well as others, is interested in whether there is intrinsic, not instrumental, disvalue in such fantasizing. The intrinsic badness, as we have seen, is that of fnding “the bad” pleasurable. I have argued against this view, but it might be instrumentally bad to be in such state. Perhaps allowing one’s mind to wander into such places is risky; it can risk a loss of virtue, damage to one’s character.10 I will grant this, but on the other side there is the risk of not allowing

The Value of a Free and Wandering Mind

283

such free wanderings, of one being overly controlled in even one’s private and internal thoughts. It is in the realm of imagination where one fnds creativity and innovation. I will end by considering three ways that such judgment-free mind wandering can be of value. 4.1 Art The frst, and perhaps most obvious, way in which pure fantasies can be valuable is in their connection with artistic creation. Freud suggests that there is little difference between fantasists and creative artists. If we think one is of value, I suggest we should be careful about limiting the other. In his essay “Creative Writers and Day-Dreaming” he argues that creative writing is an extension of childhood play. One gets a sense in reading that essay that Freud might have been suggesting that such writers are emotionally stunted. But ultimately he seems to think of this as a healthy kind of sublimation – an alternative to other ways that fantasies can affect the psyche. I am not concerned with the details of Freud’s theory and to what extent it can explain artistic expression. Freud himself seemed to fnd the psychoanalytic tools wanting, saying of the formula he suggests that creative writing is wish fulfllment stemming from childhood memories, that he suspects “it will prove to be too exiguous a pattern.” But what is important for my purpose is the recognition of the link between free play and creative production. Might we not say that every child at play behaves like the creative writer, in that he creates a world of his own, or rather, rearranges the things in his world in a new way which pleases him? [. . .] The creative writer does the same as a child. He creates a world of fantasy which he takes very seriously – that is, which he invests with large amounts of emotion – while separating it sharply from reality. Freud 1959: 174 Now the kind of emotional seriousness pointed to here may seem to contradict what I have been urging about pure fantasies being insulated in a sense from one’s motivational structure. When the artist is making deliberate choices of how to take the material from both their mind and from the world to create an object, such choices may well be open to different kinds of normative assessment or judgment; this will be even more likely if the object produced is shared with the public. But for there to be the possibility of making such choices, allowing one’s mind to wander to places quite disconnected from reality – like the child at play – is essential. A vivid example of the connection between free mind wandering and artistic creation can be found in the work of David Lynch. The contrast between the darkness of Lynch’s flms and his own positive (even cheerful) outlook on life is frequently noted. In his recent memoir/biography we fnd the following description during the making of Blue Velvet:

284

Miriam Schleifer McCormick Amid all this mayhem, Lynch maintained a sunny disposition, tooling around the set on a pink bicycle with streamers futtering from the handlebars, his pockets full of peanut M&M’s. “David is a genuinely happy person, and this is one of the remarkable things about him – I’ve never met anyone as serene as he is,” said Rossellini. [. . .] “David would say meditation is the source of his happiness,” said Laura Dern, “and I’m sure that’s true. He knows who he was and who he became almost immediately after he started meditating, so he’s the best judge of that. I would add, though, that I think part of his happiness has to do with the fact that he places no limits on himself as a creative person. There’s a lot of self-judgment and shame in our culture, and David doesn’t have any of that. When he makes something he never wonders what people will think of it, or what he should be making, or what the zeitgeist needs. He makes what bubbles up out of his brain, and that is part of his joy.’’ Lynch/McKenna 2018: 213

These “brain bubbles” often come in the form of both night and daydreams. Here is Lynch describing how the fnal scene of Blue Velvet came to be: “An important piece of the Blue Velvet script came to me in a dream, but I didn’t remember the dream until quite a while after I woke up from it. So, imagine me for some reason going over to Universal Studios the day after I had a dream that I didn’t remember. I went there to meet a man and went into the secretary’s room and the man was in the room behind her. In this secretary’s room there was either a couch or a chair near her desk, and because the man wasn’t ready to see me I went and sat down on this chair and waited. Sitting on that chair I remembered my dream, and I asked the secretary for a piece of paper and a pencil, and I wrote down these two things from the dream: a police radio and a gun. That did it for me. I always say I don’t go by nighttime dreams because it’s daydreaming that I like. I love the logic of dreams, though. Anything can happen and it makes sense.’’ Lynch/McKenna 2018: 228 Lynch allows himself the freedom to experience darkness and violence in the realm of the imagination, but none of that penetrates his motivational structure or his character and temperament. 4.2 Innovation In a study of the phenomenon of “mind wandering” described like this – “our minds drifting away from a task toward unrelated inner thoughts, fantasies, feelings, and other musings” – Jonathan Smallwood/Jonathan

The Value of a Free and Wandering Mind

285

Schooler (2006) survey the many ways that such wanderings can inhibit completion of tasks that require a sustained focus, but they also cite fndings that support the idea that such wanderings are importantly linked to problem-solving and creativity. They end their article by summarizing some of these fndings: Klinger (1999) suggested that one advantage to mind wandering is that it could foster creative problem solving. Similarly, Singer (1966) argued that processes associated with fantasy, particularly in early life, facilitate the development of problem-solving skills. In this article, we have argued that mind wandering may be a mode of problem solving. In particular, we have suggested that mind wandering is a situation when controlled processing becomes hijacked in the service of a goal refecting our current concerns. If this is correct, then this process is linked to the pursuit of ideas or problems that have, so far, eluded solution. Considered in this light, mind wandering may share important similarities with incubation processes related to creativity. Accordingly, the experience of sudden “eureka” or “ah-ha” moments (Schooler et al. 1995), which apparently occur out of the blue, may sometimes occur because mind wandering addresses more remote goals (e.g., discerning the solution to a heretofore unsolved problem). Smallwood 2006: 956 4.3

Meditation

Meditative practices may seem to be antithetical to mind wandering, in that they increase focus and attention. Smallwood and Schooler, for example, reference a technique, mindfulness-based cognitive therapy (MBCT), aimed at reducing mind wandering: “In MBCT, participants are trained in meditative techniques that are focused on reducing mind wandering. MBCT shows promise in reducing the likelihood of depressive relapse and, posttherapy, in helping participants form more detailed autobiographical memories” (2006: 955). While it is the case that as one becomes more experienced in meditation, the frequency of mind wandering episodes decreases, it is essential that when one’s mind does wander that the mediator does not either (i) judge themselves for allowing their mind to wander or (ii) pass judgment on the content of the wanderings. The idea is to simply observe and be aware of the thoughts as they occur, and then bring one’s attention back to the focus of meditation such as one’s breath or sensation in the body. The meditative process is thus enhanced if one can feel free to let one’s mind wander without judgment. Here is a representative example of this idea: If you want to obtain perfect calmness [. . .] you should not be bothered by the various images you fnd in your mind. Let them come,

286

Miriam Schleifer McCormick and let them go. [. . .] The true purpose to see things as they are, to observe things as they are, and to let everything go as it goes. Suzuki 2011: 20

5.

Conclusion

I have argued that a certain kind of imaginative state, what I have called “pure fantasy,” has no fttingness conditions, is not rationally evaluable, and has no intrinsic value or disvalue. If we discover the contents of this state it will tell us no more about you than would the contents of your dreams. And so the normative judgments made about such fantasizing should not go beyond what is appropriate for dreams. While mind wandering may be risky, allowing freedom of the thought in the imaginative realm is importantly linked to creativity and need not have any negative consequences.

Notes 1. See, for example: Hieronymi (2008), McHugh (2014), Meylan (2018), Owens (2000), and Smith (2005) among many others. I present arguments about belief in (2015) and about hope in (2017). 2. George Sher (2018) has recently argued (against some of those referenced in the frst footnote) that only actions are the proper targets of moral assessment, “that the realm of the purely mental is best regarded as a moralityfree zone” (484). My argument is more modest than his in some ways and more ambitious in others. I think certain kinds of mental states are open to normative assessment, even of a narrowly moral kind; some kinds of beliefs may well be morally blameworthy. But Sher says at the outset that “a person’s preoccupations and fantasies can refect badly on his character” and that they are open to rational evaluation. I will argue that for a certain domain of the mind even these assessments are also not appropriate. One of his arguments against those who think morality can apply to the mental is that he thinks there is no principled way to delineate some mental states from others in a way that explains why some are up for moral evaluation and not others. Part of what I am attempting to do here is offer such criteria of delineation. 3. See Couenhoven (2010) for a discussion of Augustine’s view, especially p. 116. He is quite sympathetic to his view, as are Matthews (1981) and Cherry (1988). 4. See Bovens (forthcoming). 5. I am not here claiming that any particular state which is motivationally inert is thereby exempt from any kind of normative judgment. Many beliefs of trivial matters may have no effect on me, and yet they can be negatively evaluated if false, or perhaps if based on insuffcient evidence. Elsewhere (forthcoming) I argue that a belief violating a particular norm and so being negatively evaluated does not mean that the believer is subject to blame or any blame-like attitude. Fantasies, unlike beliefs, are not subject to standards of correctness, so there is no evaluation independent of whether reproach of the agent is justifed. What is of concern here is what distinguishes the kind of fantasizing I am arguing is exempt from judgment from that which is not. 6. For discussion of ways to understand desires as representations that are correct if good see Hazlett (2018) and Gregory (2017).

The Value of a Free and Wandering Mind

287

7 Cherry (1988) distinguishes between what he calls “surrogate” and “autonomous” fantasies, and Smuts adopts this language in some of his discussion: “Surrogate fantasies are those that the fantasist would like to take place. They are surrogates for reality. But not all fantasies are like this. Some fantasies are ‘autonomous.’ [. . .] [A]utonomous fantasies are unlike surrogate fantasies in that the fantasist does not desire the fantasy to be actualized” (2015: 387). 8. There is a lot of disagreement among theorists of blame about how to best characterize the attitude, but all agree that blame carries with it some normative force that seems to require that it includes a negative evaluative judgment of the person being blamed. Yet, it seems wrong to think of blame as a purely cognitive attitude because when we blame others, negative emotions are often included in the blaming. While there is dispute among theorists of blame whether any particular emotions are necessary, the challenge for those who deny any are required is to explain the special force that blame has. All agree that to blame someone goes beyond mere description; it is not simply to register that someone has failed to meet a particular standard. I can describe you as fawed in a certain way without blaming you or judging you blameworthy. See Tognazzini/Coates (2018) for a very helpful overview. 9. Langland-Hassan (2015) discusses the view the “ftless view” of imaginings and argues that even if some types of imaginings are ftless, not all are (673–674). 10. Sher (2018) admits that thoughts can be morally hazardous (487–488).

References Bovens, Luc (forthcoming): Coping, Princeton, NJ: Princeton University Press. Cherry, Chris (1988): “When Is Fantasising Morally Bad?”, Philosophical Investigations 11, 112–133. Couenhoven, Jesse (2010): “Dreams of Responsibility”, Augustine and Philosophy 6, 103–123. Freud, Sigmund (1959): “Creative Writers and Day-Dreaming”, in: Sigmund Freud’s Collected Papers, Vol. 4, ed. by Jones, London: The Hogarth Press, 173–183. Gregory, Alex (2017): “Might Desires Be Beliefs about Normative Reasons?”, in: The Nature of Desire, ed. by Deonna/Lauria, New York: Oxford University Press, 201–217. Hazlett, Alan (2018): “The Guise of the Good and the Problem of Partiality”, Canadian Journal of Philosophy, DOI: 10.1080/00455091.2018.1433794. Hieronymi, Pamela (2008): “Responsibility for Believing”, Synthese 161, 357–373. Klinger, Eric C. (1999): “Thought Flow: Properties and Mechanisms Underlying Shifts in Content”, in: At Play in the Fields of Consciousness: Essays in the Honour of Jerome L. Singer, ed. by Singer/Salovey, Mahwah, NJ: Erlbaum, 29–50. Langland-Hassan, Peter (2015): “Imaginative Attitudes”, Philosophical and Phenomenological Research 90, 664–686. Lynch, David/McKenna, Kristine (2018): Room to Dream, Munich: Random House. Matthews, Garreth (1981): “On Being Immoral in a Dream”, Philosophy 56, 47–54. McCormick, Miriam Schleifer (2015): Believing against the Evidence: Agency and the Ethics of Belief, London: Routledge.

288

Miriam Schleifer McCormick

McCormick, Miriam Schleifer (2017): “Rational Hope”, Philosophical Explorations 20, 127–141. McCormick, Miriam Schleifer (forthcoming): “Believing Badly: Doxastic Duties Are Not Epistemic Duties”, in: Epistemic Duties: New Arguments, New Angles, ed. by McCain/Stapleford, London: Routledge. McHugh, Conor (2014): “Exercising Doxastic Freedom”, Philosophy and Phenomenological Research 88, 1–37. Meylan, Anne (2018): “The Reasons-Responsiveness Account of Doxastic Responsibility and the Basing Relation”, Erkenntnis 84, 877–893. Neu, Jerome (2002): “An Ethics of Fantasy?”, Journal of Theoretical and Philosophical Psychology 22, 133–157. Nolf, Kate (2015): “Which Mental States Are Rationally Evaluable, and Why?”, Philosophical Issues 25, 41–63. Owens, David (2000): Reason without Freedom: The Problem of Epistemic Normativity, London: Routledge. Ross, William D. (1930): The Right and the Good, Indianapolis: Hackett Publishing. Schooler, W. Jonathan, et al. (1995): “Putting Insight into Perspective”, in: The Nature of Insight, ed. by Sternberg/Davidson, Cambridge, MA: MIT Press, 589–597. Sher, George (2018): “Wild West of the Mind”, Australasian Journal of Philosophy 97, 483–496. Singer, Jerome L. (1966): Daydreaming, New York: Random House. Smallwood, Jonathan/Schooler, W. Jonathan (2006): “The Restless Mind”, Psychological Bulletin 132, 946–958. Smith, Angela M. (2005): “Responsibility for Attitudes: Activity and Passivity in Mental Life”, Ethics 115, 236–271. Smuts, Aaron (2015): “The Ethics of the Imagination and Fantasy”, in: The Routledge Handbook of Philosophy of Imagination, ed. by Kind, London: Routledge, 380–391. Smuts, Aaron (2017): Welfare, Meaning and Worth, London: Routledge. Snyder, Charles R. (2002): “Hope Theory: Rainbows in the Mind”, Psychological Inquiry 13, 249–275. Suzuki, Shunryu (2011): Zen Mind, Beginner’s Mind, 14th anniversary ed., London: Shambhala Publications. Tognazzini, Neal/Coates, D. Justin (2018): “Blame”, in: The Stanford Encyclopedia of Philosophy (Fall 2018 Edition), ed. by Zalta, URL = . Westlund, Andrea C. (2018): “Answerability without Blame?”, in: Social Dimensions of Moral Responsibility, ed. by Hutchison, et al., New York: Oxford University Press, 253–274. Wolf, Susan (2015): “The Meanings of Lives”, in her The Variety of Values: Essays on Morality, Meaning, and Love, New York: Oxford University Press, 89–106.

Contributors

(a) Editors and Contributors Gerhard Ernst holds a chair for philosophy at the Friedrich-AlexanderUniversität Erlangen-Nürnberg (Germany). Since 2018 he is President of the Deutsche Gesellschaft für Philosophie (German Society for Philosophy). His main research areas are epistemology and metaethics. His works include Das Problem des Wissens (The Problem of Knowledge) (Paderborn 2002) and Die Objektivität der Moral (The Objectivity of Morality) (Paderborn 2008). Sebastian Schmidt is a postdoctoral researcher in philosophy at the University of Zurich. His research in epistemology and metaethics focusses on responsibility and reasons for attitudes.

(b) Contributors Matthew Chrisman is Professor in Philosophy at the University of Edinburgh. His research is focused on ethical theory, political philosophy, the philosophy of language, and epistemology. He is the author of The Meaning of ‘Ought’: Beyond Descriptivism and Expressivism in Metaethics (OUP 2016) and What Is This Thing Called Metaethics (Routledge 2017). D. Justin Coates is Assistant Professor of Philosophy at the University of Houston, Texas. His research focusses on Ethics and the Philosophy of Action. One of his main interests during the last years concerned the ethics of blame (see the book he published with Neal A. Tognazzini on Blame: Its Nature and Norms, OUP (2013)). Lindsay Crawford is Assistant Professor of Philosophy at Connecticut College. She is working in epistemology, especially on normative questions at its intersection with ethics, such as partiality in friendship, epistemic obligations, and whether we can wrong others with our beliefs.

290

Contributors

Sabine A. Döring is Professor of Philosophy at Eberhard-Karls-Universität Tübingen. Her work revolves around the philosophy of emotions and related philosophical disciplines. She is author of Ästhetische Erfahrung als Erkenntnis des Ethischen (Aesthetic Experience as Knowledge of the Ethical) (Paderborn 1999) and editor of volumes on the philosophy of emotion. Right now, she is working on a monograph on Emotions and (Political) Agency (provisional title). Benoit Gaultier is a postdoctoral researcher at the University of Zurich. He is author of Qu’est-ce que le pragmatisme? (What is Pragmatism?) (Vrin, 2015) and edited two volumes in French on knowledge and reasons. His current research is focused on epistemic normativity, epistemic luck, knowledge-how, the nature and norms of belief, the value of knowledge and understanding, and the nature of philosophical knowledge and its epistemological consequences. Martina Lindner is a research assistant at the Philipps-Universität Marburg (Germany). She currently works on moral norms for the acquisition of belief and especially on W. K. Clifford’s ethics of belief. Errol Lord is Associate Professor of Philosophy at the University of Pennsylvania, where he is Chair of the Graduate Program in Philosophy. He is author of The Importance of Being Rational (OUP 2018) and edited a collection on Weighing Reasons with Barry Maguire (OUP 2016). Next to epistemology, ethics, and theory of action, he is also writing on aesthetics. Currently, he is writing a book on the epistemology of ethics and aesthetics (the provisional title is Knowing the Normative World). Miriam Schleifer McCormick is Associate Professor of Philosophy at the University of Richmond, Virginia. Her research focus is in epistemology as well as early modern philosophy. She published her book Believing Against the Evidence with Routledge in 2015. One of her current projects is the ethics of hope. Mattias Skipper is a PhD student in the philosophy department at Aarhus University. Most of his work is in epistemology (including formal and social epistemology), but he has also written on topics in philosophical logic, philosophy of language, and philosophical psychology. He is co-editor (with Asbjørn Steglich-Petersen) of Higher-Order Evidence: New Essays (OUP, forthcoming). Matthew Soteriou is Professor at King’s College London (Chair in Philosophy of Mind). From 2011–2016 he was Editor of the Aristotelian Society. His research interests include philosophy of mind, of action, and epistemology. His monographs are The Mind’s Construction: The Ontology of Mind and Mental Action (OUP 2013), where he also explores connections between the ethics of belief and the mind’s ontology, and The Disjunctive Theory of Perception (Routledge 2016).

Contributors

291

Asbjørn Steglich-Petersen is Associate Professor in Philosophy at Aarhus University (Denmark). He has published widely in epistemology, philosophy of mind, metaphysics, and philosophy of language, with a major strand of work devoted to epistemic normativity and the nature of belief. He has edited the volumes Metaphysics: 5 Questions (Automatic Press 2010), Reasons for Belief (Cambridge University Press 2011), and Higher-Order Evidence: New Essays (OUP 2019). Neal A. Tognazzini is Associate Professor of Philosophy at the Western Washington University. His works revolve around the topic of responsibility. He edited volumes on Blame (with Justin Coates, OUP 2013) and Oxford Studies in Agency and Responsibility (with David Shoemaker, OUP 2014).

Index

Adams, Robert M. 13–14, 16 Alston, William J. 13, 31, 43–45, 49–50, 169 Anscombe, G. E. M. 170, 251 blame see ethics, of blame; irrationality; responsibility Bratman, Michael 251 Broome, John 2, 12, 150, 180, 188–189 Chrisman, Matthew 5, 7, 13–14, 47, 53, 65, 170, 172 Clifford, William K. 1, 4, 6–7, 23–29, 32–39, 41–45, 171 Coates, D. Justin 6, 9, 13–14, 169, 192, 211–212, 268, 287 Crawford, Lindsay 5, 7–8, 13, 15, 87, 107, 143 decision: and determining the future 238–244, 247–250; responsibility for 10, 235; as self-determining and self-determined act 235, 245–247 Descartes, René 1, 58 Döring, Sabine 6, 10, 15, 215–218 epistemic instrumentalism 8, 109–111, 114–118, 121–123 epistemology 1, 3, 5–6, 29, 47–52, 55, 63–64, 66, 152; feminist 5, 15; social 5, 7, 15, 48–49, 64; of suspension 132, 144; see also knowledge Ernst, Gerhard 5, 9, 12, 16, 170, 172, 176, 188–189 ethics 1, 29, 35, 168; of belief 1, 3–7, 13–14, 23–26, 28–30, 32, 34, 37–39, 41–43, 152; of blame

6, 9, 14, 192–193, 196, 211; of fantasy 272, 277; of fear 6, 10, 215; meta- 1, 3; of mind 1, 3, 5–6, 9, 11–12, 14, 169, 192–193, 211; virtue 230–231 evidentialism 6–8, 12, 25–26, 28, 38, 42, 88–92, 104–106, 127, 138–143, 169 fear: and control 33–34; narratives 216, 220–229, 231; and safety 215–216, 220–222, 228; see also ethics, of fear; ought, to fear; rationality, emotional/of emotion Freud, Sigmund 276, 280, 283 Gaultier, Benoit 5, 7, 13–14, 72, 169 Ginet, Carl 13–15, 251, 253 Harman, Gilbert 158–159, 161, 171 Hieronymi, Pamela 13–16, 67, 165, 169, 172, 251, 286 Hobbes, Thomas 60 Hume, David 51, 216 interrogative attitudes 129–130, 136–142, 144 irrationality 178–181, 184–185, 216–217; and blame/criticism 9, 154–155, 160, 165–167; see also rationality James, William 1, 4, 42, 44–45, 113, 123 judgment: free from 270–271, 282–283, 285; -alism 257–261, 264–265, 267–268; see also interrogative attitudes (on suspension of judgment)

Index Kant, Immanuel 63, 189, 215, 222–224, 234, 237, 251, 256 Kavka, Gregory S. 13, 16, 169–170, 251, 253 Kiesewetter, Benjamin 12, 14, 150, 155, 157, 159, 161, 169, 170–172, 180, 188–189 knowledge: analysis of 5 (see also knowledge, theory of); and epistemic communities 48, 62–65; -frst (see Williamson, Timothy); and normative standards 48, 50, 53, 55; and pragmatic encroachment (see pragmatic encroachment; thresholds partialism); theory of 47–50, 52–54, 56, 59–60, 64–65, 68 (see also epistemology); and true belief 53, 65 Korsgaard, Christine 252 Lindner, Martina 5–7, 13, 24, 171–172 Lord, Errol 5, 8, 12, 100, 126, 144, 150, 169, 172, 188–189 McCormick, Miriam S. 6, 11, 12–13, 14–15, 25–26, 42, 51, 67, 106, 151–152, 169–170, 211, 270 McHugh, Conor 13, 43, 52, 169, 286 Meylan, Anne 13–14, 32, 38, 43–44, 66, 82, 84, 170, 286 moralistic fallacy 223 morality 2, 48, 60, 63, 80, 150, 154, 223, 286 Moran, Richard 153 naturalism 110–111, 156 normativity 5–6, 9, 109, 219; of belief/doxastic 6, 8, 66; and control 7 (see also ought); epistemic 5, 8, 65, 109–112, 169; epistemic & evaluative 48, 52; epistemic & moral 25; epistemic & practical 111; instrumental 2, 8, 12, 109–110, 112, 115 (see also epistemic instrumentalism); mental 1, 3–4, 6, 11, 13, 170; moral (see morality; normativity, epistemic & moral); of rationality 155, 171; see also ought; reasons ought 161, 165, 170, 189, 225, 242, 252, 270, 278, 280; to be vs. to do

293

53, 55; to believe 1, 3–5, 12–13, 37, 47–49, 52–58, 63–65, 89–92, 99–100, 111, 132, 158–159, 161, 171; to fear 67, 219, 227–228 (see also fear; reasons, for fear); implies Can 3–4, 13–14, 42, 48–49, 52–58, 63, 72, 78, 111, 152, 157 (see also voluntarism; voluntary control); see also normativity; reasons Owens, David 13–16, 66, 165, 169–170, 172, 278, 286 Papineau, David 14, 122–123, 170 Parft, Derek 12, 105, 188 pragmatic encroachment 5, 12, 15, 106, 211; see also thresholds partialism pragmatism 5, 7–9, 25–26, 88, 92, 100, 105, 139, 151, 155, 158, 162, 169–170, 172, 193–196 prudence/prudential 1–2, 12–13, 49, 55, 105, 114, 150, 154–155, 162, 170, 193 Putnam, Hilary 1 rationality: and adjustment/ coherence/consistency 2, 9, 150, 176, 176–184, 187–188 (see also rationality, and evaluation/ responding to reasons); and akrasia 72; attitudinal/of attitudes 4, 6, 155, 172; competition of 8, 126–128, 130, 139; and control 154–156, 168; emotional/of emotion 6, 163–165, 169, 183, 186–187; epistemic/theoretical 4, 5, 9, 138–142, 189 (see also rationality, requirements of/ principles of); and evaluation/ responding to reasons 2, 9, 150, 176, 182, 184–188; of fear 229, 231; of hope 270, 274; instrumental 112 (see also normativity, instrumental); normativity of 6, 155, 171; norms/standards of 2, 155–158, 162, 164–165, 167, 179; practical 5, 106 (see also rationality, requirements of/principles of); problem 152; and reactive attitudes 157; requirements of/principles of 2, 159, 178, 180, 182–183, 188–189; respect for 80; supervenes

294

Index

on the mind/concerns the mind 2, 177–178; theories of 158; and virtue/vice 165, 171–172; see also irrationality; reasons reason(ing): deliberation/reasoning 49–50, 78, 90, 178, 263; standpoint of theoretical vs. practical reason 10, 236–237, 247–250; theoretical vs. practical reasoning 238–239, 241, 247 reasons: for attitudes/for mental states 2–4, 6, 9, 149–152, 157–158, 160–162, 169–172, 273; for belief 1–2, 4–5, 7–8, 12, 14–15, 26, 44–45, 54, 55–56, 87–90, 92–94, 96–98, 100–105, 114, 134, 163–164, 169–170, 172, 187, 263; for belief-ascription 178–180, 188; to believe and to act/do/manage 2, 54–55, 77, 96–97, 102–103, 106; and blame(worthiness) 150, 160–162; for blame 9, 196–210; for desire 151, 275; for doxastic partiality 8, 88, 93, 96–103, 105, 106; for emotion/to feel 163–164, 186 (see also reasons, for fear); epistemic 2, 8, 14, 93, 109–115, 122–123, 139, 159, 169, 171; epistemic vs. practical instrumental 115–120, 122 (see also normativity, instrumental); for fear 151 (see also fear; ought, to fear); for hope 273; for intention 130–132, 151, 186; object-given/constitutive/right kind of 2, 9, 13–15, 67, 150–151, 155, 157, 162–164, 167, 169–172, 194–196; practical/pragmatic/ non-evidential/extrinsic 4–5, 8–9, 15, 31–32, 35, 38, 42–43, 51, 55, 67, 89, 92, 100–102, 123, 139, 151, 153, 169–172, 187, 263; for a proposition 55–56; and rationality 9, 12, 150–151, 158, 167, 177, 185–188; responding to 2, 9, 49–50, 150, 176, 184–188, 274; responsiveness to 83, 155, 168, 273, 276; state-given/wrong kind of 4, 6, 10, 13–14, 151, 169, 194–196; suffcient 120–122; for suspension 127–128, 131–132, 134, 138, 140–142; see also normativity; ought

Reisner, Andrew 4, 12, 14–15, 106, 116, 123, 170–171 responsibility: for attitudes/mental 4, 6, 9, 13, 149–150, 154–158, 163, 165, 167–170, 172, 266–267; for belief/doxastic 4–5, 7, 13, 16, 72, 76–77, 151–152; and blame 198; and capacity 196; and control 3–4, 44, 78–82, 151–152, 167–168; for decision (see decision); and excuse 211; for hope 14; moral 74, 196; original 75; for thoughts 6 Rinard, Susanna 12–15, 26, 30, 45, 106, 123, 169–170, 211 Rousseau, Jean-Jacques 48, 59–60, 63 Scanlon, Thomas M. 171–172, 211 Schmidt, Sebastian 1, 5, 13–15, 45, 68, 84, 107, 143–144, 149, 170–172, 189, 253, 268 Sellars, Wilfred 188 Sextus Empiricus 152 Shah, Nishi 12, 43, 105–106, 123, 169, 252 Skipper, Mattias 5, 8, 12, 14–15, 110, 119, 123, 169–170, 72 Smith, Adam 206, 219 Smith, Angela 14, 16, 165, 172, 198, 207, 212, 266, 286 Soteriou, Matthew 6, 10, 13, 170, 234, 251–253 Steglich-Petersen, Asbjørn 4–5, 8, 12, 14–15, 109, 119, 122–124, 159–160, 167, 169–172 Strawson, Galen 253 Strawson, Peter F. 12, 211, 251, 267–268, 270 tensed/temporal perspective on reality 234–237, 246, 250, 252 thresholds partialism 92, 106 (see also pragmatic encroachment; reasons, for doxastic partiality) Tognazzini, Neal 6, 10, 14, 169, 211, 268, 287, 289 voluntarism 7, 23–24, 29, 31, 34–35, 39, 41, 43, 64, 71, 143 voluntary control 3–4, 7, 13–14, 23, 29, 36, 52–56, 58, 71–74, 77–78, 82–83, 149, 154–157, 167–168, 270

Index Wallace, R. Jay 172, 212 Wedgwood, Ralph 12–13, 157, 171–172, 188–189 Williams, Bernard 12–13, 15, 30–31, 43, 49–50, 188–189, 227, 230

Williamson, Timothy 66 Wittgenstein, Ludwig 176, 179, 262, 268 Zimmerman, Michael 74

295