Evil: A Philosophical Investigation [1 ed.] 0198712480, 9780198712480

Table of contents :
Cover
Evil
Copyright
Acknowledgements
Contents
Introduction
1 The Secular Moral Concept of Evil
2 Evil Action
3 Evil, Extremity, and Incomprehensibility
4 Thick and Thin Accounts of Evil Action
5 Assessing Thick and Thin Accounts
6 Conceptual Pluralism about Evil Action
7 Aggregative Accounts of Evil Personhood
8 Dispositional Accounts of Evil Personhood
9 Evil Feelings
10 Evil and Explanation
11 Doing Away With Evil?
Bibliography
Index

Citation preview

Evil

Evil A Philosophical Investigation

Luke Russell

3

3 Great Clarendon Street, Oxford, ox2 6dp, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Luke Russell 2014 The moral rights of the author‌have been asserted First Edition published in 2014 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2014930351 ISBN 978–0–19–871248–0 Printed and bound by CPI Group (UK) Ltd, Croydon, cr0 4yy Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

Acknowledgements This book is the product of many years of work, and I am grateful to those who helped me along the way. I started thinking about the concept of evil in the wake of the September 11th terrorist attacks on the United States. In the heated moral debate that followed, I was struck by the fact that many people who condemned the attacks were uncomfortable with any suggestion that they should be denounced as evil. There is some kind of step up between wrong and evil, I realized, but I was unsure how best to characterize this distinction. I began making headway on this project in 2005 while I was in Canada and the United States on research leave, generously granted to me by the University of Sydney. During this period I focused on the question of whether the distinction between evil action and wrong action was qualitative or merely quantitative, and engaged with the work of Eve Garrard, Stephen de Wijze, Hillel Steiner, and Daniel Haybron, all of whom I would like to thank for leading me to think more deeply about these issues. Thanks are also due to Michael Smith, who arranged for me to be Visiting Fellow at Princeton, and to Tom Hurka and Peter Singer, who met with me to discuss my project. While working at Sydney over the next few years I began to think more carefully about the nature of evil personhood. It seems plausible that not every evildoer is an evil person, but here, too, it is hard to say what the difference between an evil person and a merely bad person amounts to. I learned a lot from feedback provided by anonymous referees for several journals, including the Australasian Journal of Philosophy, Philosophical Studies, American Philosophical Quarterly, and Ethical Theory and Moral Practice. My colleagues at the University of Sydney also made many useful suggestions, and I would like to thank Huw Price, Moira Gatens, David Braddon-Mitchell, Kristie Miller, Caroline West, Paul Redding, David Macarthur, Rick Benitez, Stephen Gaukroger, Duncan Ivison, Richard Joyce, Nick Smith, Mark Colyvan, Thomas Besch, Paul Griffiths, Jennan Ismael, Jack Justus, and Anik Waldow. Nothing improves the quality of philosophical work more than presenting papers to a receptive critical audience, and I was lucky enough to have many such opportunities in Australia and elsewhere. Particularly memorable sessions include talks at several AAP conferences, and seminars at the ANU, Macquarie University, University of Western Australia, Charles Sturt University, Boston University, Oxford University, Manchester University, Reading University, Hong Kong University, and the University of Vienna. Many philosophers made helpful suggestions during this period, including David Velleman, Herlinde Pauer-Studer,

vi  Acknowledgements Susan Neiman, Julia Annas, Mike Ridge, David Owens, Brad Hooker, David Liggins, Paul Formosa, Jeanette Kennett, Catriona Mackenzie, Peter Menzies, Ben Fraser, Daniel Cohen, Daniel Friedrich, Nic Southwood, Philippe Chuard, Jenny Louise, Garrett Cullity, Dan Marshall, Jessica Wolfendale, Richard Hamilton, Karen Jones, Glen Pettigrove, Rosalind Hursthouse, Christine Swanton, Michael Brady, Fiona MacPherson, Peter Godfrey-Smith, Patrick Greenough, Matthew Hammerton, and Lucy Allais. What was a mere collection of papers on evil gradually began to take the shape of a book during my research leave in 2009, part of which I again spent as a visitor at Princeton. During this period I wrote some of what turned into the first half of the book, and benefited from giving papers at Boston University and the University of Western Ontario, as well as discussions with Michael Smith, Alexander Nehamas, Daniel Star, Aaron Garrett, Julia Markovits, Ian Blaustein, Amile Rorty, Samantha Brennan, and Heidi Maibom. As I drew towards the final stages of this project, my work was greatly improved by the careful and insightful feedback of anonymous referees for Oxford University Press, who led me to restructure the book, and to clarify many points. Thanks are also due to Stephen de Wijze, who, in organizing the conference “The Idea of Evil” in Manchester in 2012, gave me the chance to meet with many of the philosophers who have been writing on this topic, including Adam Morton, Eve Garrard, Matthew Kramer, and Phillip Cole. This conference provided a wonderful conclusion to my project, and indicated that there is plenty more philosophical work to be done on evil. Material from several of my previously published papers has made its way, in a revised or rewritten form, into this book. Thanks to the editors and publishers of these journals for allowing me to use this material in this book. These papers are: “Evil and Incomprehensibility”, Midwest Studies in Philosophy, 36(1), 62–73 (2012).

Permission granted by the editors on behalf of Wiley. “Dispositional Accounts of Evil Personhood”, Philosophical Studies, 149(2), 231–50 (2010). Permission granted by Springer. “Evil, Monsters and Dualism”, Ethical Theory and Moral Practice, 13(1), 45–58 (2010). Permission granted by Springer. “He Did It Because He Was Evil”, American Philosophical Quarterly, 46(3), 267–82 (2009). Permission granted by the editors on behalf of University of Illinois Press. “Is Evil Action Qualitatively Distinct from Ordinary Wrongdoing?”, Australasian Journal of Philosophy, 85(4), 659–77 (2007). Copyright © Australasian Association of Philosophy, reprinted by permission of Taylor & Francis

Acknowledgements  vii Ltd, on behalf of The Australasian Association of Philosophy. “Evil-Revivalism versus Evil-Skepticism”, Journal of Value Inquiry, 40, 89–105 (2006). Permission granted by Springer. Finally, I want to express my gratitude to my friends and family for supporting me during the writing of this book, and for putting up with my endless attempts to explain the nature of evil. Thanks to Tim Brenstrum, Arthur Montazeri, Kate Sharpe, Daniel Friedrich, and Danielle Lamb for generously providing me with accommodation while I was travelling and writing. I am particularly grateful for the support of several philosophical friends—Catharine Abell, Brent Madison, Stefan Linquist, and Katie Steele—who understand what it is like to do this kind of work, and who have provided me with plenty of encouragement and good humour.

Contents Introduction

1

1. The Secular Moral Concept of Evil

9

2. Evil Action

31

3. Evil, Extremity, and Incomprehensibility

49

4. Thick and Thin Accounts of Evil Action

69

5. Assessing Thick and Thin Accounts

93

6. Conceptual Pluralism about Evil Action

112

7. Aggregative Accounts of Evil Personhood

133

8. Dispositional Accounts of Evil Personhood

154

9. Evil Feelings

176

10. Evil and Explanation

196

11. Doing Away with Evil?

217

Bibliography Index

227 233

Introduction “Evil” is a highly contested word. President George W. Bush polarized opinion in his 2002 State of the Union address, when he called the nations of Iraq, Iran, and North Korea an “axis of evil”, and declared “Evil is real, and it must be opposed”. The claim that evil is real strikes some people as being obviously true, but to others it seems to reflect a naive, chauvinist, or medieval moral outlook. Commentators have offered a range of analyses of this disagreement. Some see it as a clash between moral realists, who take wrongdoing seriously, and moral relativists, who treat morality as nothing more than an arbitrary social convention. Others see it as a clash between right wing conservatives, who want to impose free market capitalism across the world, and left wing liberals, who champion the cause of minority rights and the method of diplomacy. Alternatively, it could be described as a dispute between religious fundamentalists, who think that they possess moral certainty, and atheists, who are tolerant and humble in the face of inter-cultural disagreement. None of these common analyses get at the core of the disagreement over the use of the word “evil”. We cannot understand the dispute by jumping straight into arguments about cultural relativism and moral objectivity, or about the costs and benefits of realist foreign policy, or about the existence of God and supernatural forces. A group of people who agree that a particular action—say, the torture of a prisoner by soldiers—is egregiously morally wrong might disagree over whether this action is evil. This suggests that we need to step back and think carefully about the concept of evil itself. What does it mean to say that evil exists, or that an action is evil, or that a person is evil? What is the difference between an action’s being merely wrong and its being evil, or the difference between a person’s being bad and his being evil? Can the concept of evil be explanatorily useful, and, if so, what can it help to explain? In this book I aim to answer these questions concerning the nature of evil, and to show that the concept of evil has a legitimate place in contemporary secular moral thought. In Chapter One: The Secular Moral Concept of Evil I address the questions of whether the concept of evil is exclusively religious or supernaturally loaded, and

2  Introduction what it means to say that evil exists or that evil is real. It is obviously true that many religious believers speak of evil, and that some atheists are unwilling to use that word because of what they see as its religious connotations. Yet many other atheists use the word “evil” both in condemning and in attempting to explain atrocities such as the Holocaust and the genocide in Rwanda, and the actions of serial killers such as Ted Bundy and Dennis Rader. I argue that the concept of evil, like that of forgiveness, is available for use by theists and atheists alike, in contrast to an exclusively religious concept such as that of sin. Evil exists in the way that courage, malice, or honesty exist; not as a mysterious supernatural force that is capable of taking over a person, but as a character trait and as a moral property of actions. Having situated evil within secular morality, I will move on to apply the philosophical method of conceptual analysis to evil. In Chapter Two: Evil Action I argue that we ought to separate the concepts of evil action and evil person, and first give an account of evil action. A plausible account of evil action should be constrained to some degree by what ordinary people think and say about evil. I offer a list of folk intuitions about evil, and show how recent philosophical accounts of evil action can be classified according to which particular intuitions they accept and which they reject. I propose that all evil actions are extreme culpable moral wrongs. There could be no such things as trivial evils, or excusable evils, or evil actions that we morally ought to perform, and we should reject any philosophical accounts of evil that fail to respect these claims. While it is relatively uncontroversial that these conditions are necessary for evil action, it is not clear exactly what kind of moral extremity is required. In Chapter Three: Evil, Extremity, and Incomprehensibility I assess various competing answers to this question. Some philosophers have claimed that evil actions must be extremely harmful, or that evil actions are extreme in the sense of being incomprehensible. I will reject both of these claims, arguing instead that there can be harmless evil actions, and that many evil actions are explicable and are comprehensible even to some morally decent human beings. Although not all evil actions are extremely harmful, I contend that all evil actions are connected in at least one of a specifiable variety of ways to actual or possible extreme harms. These actual or possible harms need not be maximally harmful, nor so extreme as to be life-wrecking. When it comes to specifying the degree of extremity required for evil action, the best we can do is to be guided by our intuitions concerning specific examples, and give a rough location of the threshold on the spectrum of harms. There is much recent disagreement between philosophers on the question of whether there is a psychological hallmark of evil action. In Chapter Four: Thick and Thin Accounts of Evil Action I catalogue the competing views on this issue. Hannah Arendt and her followers believe that an action is evil only if it is an

Introduction  3 extreme culpable wrong, but they think that evil actions could flow from a very broad range of motives. Arendt suggests that evil actions can be banal, in the sense that some evildoers are psychologically ordinary and unremarkable. An evildoer need not act out of malice or with sadistic pleasure, and need not know that what he is doing is morally wrong. This gives rise to what I call the psychologically thin account of evil action; the view that there is no psychological hallmark of evil action, and that an action is evil if and only if it is an extreme culpable wrong. The psychologically thin account is rejected by several contemporary philosophers. Some advocates of the alternative psychologically thick accounts of evil action directly contradict Arendt’s claims regarding banality, and argue that an action can be evil only if it is malicious, or only if it is accompanied by sadistic pleasure, or only if the agent does what he knows is morally wrong. Other advocates of psychologically thick accounts think that Arendt is right to claim that evils need not be malicious, sadistic, or defiant, but that there is a complex psychological hallmark of evil action nonetheless. In Chapter Five: Assessing Thick and Thin Accounts I move on to consider which, if any, of these competing accounts of evil action is correct. It is not clear that we could decisively favour one account on the grounds that it fits more closely with folk intuitions regarding which specific actions count as evil, because there is significant disagreement amongst the folk over particular cases. However, I contend that accounts which posit a complex, hidden psychological hallmark of evil action diverge substantially from folk thought about evil, and lack other benefits that might have outweighed this cost, and thus should be rejected. While it is true in general that we ought to favour those philosophical accounts that pick out the most useful conception, or the conception that maps onto an important natural kind, these criteria do not give us reason to favour the thin account over several of its thick rivals, or vice versa. In Chapter Six: Conceptual Pluralism about Evil Action I consider two comparatively sophisticated philosophical arguments that might help us decide which out of the remaining viable theories of evil action is best. The first argument is that there is a sharp, qualitative difference between evil actions and ordinary wrongs, and that only certain kinds of thick account properly reflect this qualitative difference. It is not clear exactly what it would mean to say that there is a qualitative difference between evils and non-evil wrongs. I  accept that a weak version of this thesis is true, but the weak version does not give us grounds to favour a thick over a thin account of evil action. The stronger version of the qualitative difference thesis, which would favour some accounts over others, turns out to be indefensible. The second argument that I consider in Chapter Six is that a psychologically thick conception of evil is morally dangerous and leads to mistakes in our moral judgements; mistakes that are easily

4  Introduction avoided by those who employ a thin conception of evil action. This argument, too, is unconvincing. Advocates of a thick account of evil action can avoid falling into these moral errors so long as they reject the claim that any particular evil action is morally worse than any particular non-evil wrong. I maintain that an action is evil only if it is an extreme culpable wrong, and that we should reject any account of evil action that does not fit with this claim. I also believe that we should reject the view that there is a hidden, complex psychological hallmark of evil action. This still leaves several viable accounts of evil action on the table: the thin account, according to which an action is evil if and only if it is an extreme culpable wrong, and various folk versions of the thick account, according to which an action is evil if and only if it is an extreme culpable wrong that is also malicious, or sadistically pleasurable, or defiant. The arguments considered in Chapters Five and Six do not give us a decisive reason to think that any one of these accounts is the single best account of evil action, or that any one of these accounts picks out the actions that really are evil. Rather than arbitrarily favour one of these viable accounts, I advocate a restricted conceptual pluralism with regards to evil action. All of us should agree that if an action is evil, it is an extreme culpable wrong, but it is rationally permissible for a person to stipulate that by “evil action” she will mean “malicious and extreme culpable wrong”, and rationally permissible for others to stipulate that they will mean “defiant extreme culpable wrong”, “malicious and sadistic extreme culpable wrong”, or simply “extreme culpable wrong”. Having given my account of evil action, I will move on to consider what it would take for someone to count as an evil person. Many philosophers agree that not every evildoer is an evil person, but that some war criminals and serial killers, for instance, are evil. It is not clear how we ought to characterize the difference between the merely vicious person and the evil person. Chapter Seven: Aggregative Accounts of Evil Personhood begins with a list of common intuitions about evil personhood, which will help us assess competing accounts of evil personhood. These accounts fall into two broad classes. The first group are the aggregative accounts, according to which an evil person is someone who has done enough evil. The second group are the dispositional accounts, according to which an evil person is someone who is disposed to do evil. In Chapter Seven I aim to develop the strongest possible version of the comparatively neglected aggregative account, before assessing its plausibility. Unfortunately, even the most sophisticated aggregative accounts cannot make sense of the fact that evil personhood can play a role in explaining evil action. Aggregative accounts also fail to allow for some possible kinds of moral conversion, and do not fit with important intuitions concerning luck and evil action. Each of these failings would be avoided by a dispositional account of evil personhood.

Introduction  5 Despite its obvious strengths, a dispositional account of evil personhood threatens to produce two counterintuitive consequences. If an evil person is someone who is disposed to perform evil actions, then it is possible for there to be evil persons who have done nothing wrong because their disposition has not manifested itself in action. It is not clear that such a blameless person would deserve our strongest moral condemnation. Moreover, experiments performed by social psychologists suggest that a very large proportion of us are disposed to perform evil actions when instructed to do so by authority figures. There is also ample evidence from history that in some societies the ordinary citizens have been led to perform horrible atrocities. If evil persons are those who are disposed to do evil, then it seems that a very large proportion of us will turn out to be evil. In Chapter Eight: Dispositional Accounts of Evil Personhood I set out these objections to a basic dispositional account, and explore a variety of more sophisticated versions of the dispositional approach that might preserve the intuition that evil persons are comparatively rare. I reject the view that evil persons are bad in every respect, and propose instead that evil persons are those who are strongly and highly-fixedly disposed to perform evil actions when in conditions that favour the exercise of their autonomy. In other words, an evil person is someone who is markedly likely to do evil when he is allowed to do what he wants to do, and whom we cannot easily change into a good person by using everyday techniques such as moral reasoning. In Chapter Nine:  Evil Feelings I  consider whether the dispositional account of evil personhood should be modified to allow for the possibility that someone could be evil merely in virtue of having evil feelings. For instance, malevolent quadriplegics and sadistic voyeurs take great pleasure in the extreme suffering of others, but they inflict no harm themselves. Yet people of these kinds are disposed to perform evil acts of appreciation, I claim, and hence can be accommodated within the existing dispositional account. It is more difficult to evaluate persons who have involuntary or conflicted evil feelings. I will argue that our answer to the question of whether such persons are evil does not depend on whether their feelings are part of inverted moral values, or whether their feelings can be narratively integrated into the story of their lives, but whether they repudiate their evil feelings. Persons who are strongly and highly fixedly disposed to have unrepudiated evil feelings are evil. Having set out my account of evil action, evil feeling and evil personhood, I will conclude by responding to a range of objections to the use of the concept of evil. In Chapter Ten: Evil and Explanation I try to show how and to what extent the concept of evil can be used in explanations of actions. Some philosophers have argued that the claim “He did it because he was evil” is vacuous or pseudo-explanatory. I defend the view that the concept of evil action and the concept of evil personhood

6  Introduction have roles to play in our explanations of some actions, although they are limited roles, and evil does not provide a complete explanation for any action. Moreover, I argue that we have reason to believe that there really are some evil persons, who are not only strongly disposed to perform the worst kind of wrong actions, but are beyond redemption, for practical purposes, and should be treated as write-offs. In endorsing the concept of evil, I am not suggesting that we should accept each and every folk claim about evil. In Chapter Eleven: Doing Away With Evil? I concede that the concept of evil is often misused, sometimes with disastrous effects. My recommendation is that we continue to use the concept of evil, but that we do so cautiously, and that we disambiguate and clarify our claims that certain actions and persons are evil. Much of our moral discourse, including talk of freedom, virtue, and weakness of will, is equally vague, ambiguous, and in need of clarification. As philosophers, though, we should aim to engage with the folk discourse that occurs outside of philosophy. One significant advantage of analysing and clarifying our talk of evil, rather than attempting to replace it, is that doing so will aid our ability to engage with these everyday moral disagreements. There are many important questions concerning evil that I do not address in this book. I do not set out and defend a general systematic normative ethical theory identifying which actions are right and which are wrong. Rather, I assume that we have a fairly good intuitive grasp of which actions are wrong, and focus instead on the differences between wrong and evil, and between vice and evil. Obviously there remain significant and often recalcitrant disagreements over which specific actions are wrong, but I hope that my account of evil will be compatible with the various views about wrong action that are most popular amongst philosophers. This book also does not contain a metaethical argument for moral realism as opposed to non-cognitivism, error theory, or relativism. Rather, I assume either that moral realism is true, or that, if some other metaethical theory is correct, it will be compatible with our continuing to care deeply about morality, our deploring murder, torture, and rape, and our continued use of first-order moral discourse to condemn wrongdoers and to persuade others to share our opinions. For instance, I think that non-cognitivists and fictionalists should care about the difference between wrong and evil, and might argue about whether Mao Zedong or Jeffery Dahmer were evil. Another important task I do not undertake here is that of identifying the dividing line between those human beings who are so mentally incompetent that they fall below the threshold for moral responsibility, and those who properly are held responsible, despite the depravity of their desires. This is a very difficult question, and answering it in detail would be beyond the scope of this book. Nonetheless, it is plausible that very many of the most extreme wrongdoers are capable of reasoning,

Introduction  7 planning, and carrying out their intentions, are aware of the moral standards to which we hold each other, and hence are morally culpable for their deliberate and voluntary actions, as horrible as those actions might be (cf. Stone 2009, 35–6). Where “illness” is used merely a label for depravity, illness is no excuse. Where compulsion to act comes from within the agent’s own character, compulsion is no excuse. Even if there are some cases in which it is not clear whether a particular serial killer is morally culpable or merely mentally ill, there remain plenty of cases in which extreme wrongs are performed by rationally adept persons who should be held responsible. Some of the recent philosophical, psychological, and historical literature on evil aims to identify the causes of extreme wrongdoing and the causes that contribute to the development of extreme vice. In pursuit of this goal, authors including John Kekes, Claudia Card, David Cesarani, Daniel Goldhagen, Christopher Browning, David Velleman, and Herlinde Pauer-Studer, Michael Stone, and Phillip Zimbardo offer detailed case studies of particular societies or situations in which great wrongs were committed, or of particular human beings who have committed atrocities. These issues too are deeply important, and I draw upon the work of some of these authors at various points in my argument. However, I do not aim to identify the specific causes of evil in this book, nor do I give detailed or elaborate descriptions of specific evil actions and evil persons. I intend to delineate the concept of evil by considering a very broad range of examples, some of them sketched in comparatively minimal detail. While it is true that often we cannot fully understand why a specific evil action occurred without locating that action in its detailed context, it is also true that theories of evil action can be skewed and incomplete if they focus too closely on only one kind of example of extreme wrongdoing. The breadth of my approach is intended to remedy this tendency and provide a comprehensive overview of the main philosophical issues concerning evil. It is plausible that many of the most harmful wrongs are performed by groups of human beings acting together, while each individual member of the group herself remains ignorant of the entire action, and perhaps does nothing that would be particularly harmful if considered in isolation from the activities of the group. In this book I do not attempt to answer the important question of whether a corporation or an institution could be a morally responsible agent, or could be evil. Nor do I explore in any detail the relationship between responsibility for individual actions and responsibility for collective or corporate actions. It is clear that people who contribute to such actions can be culpably negligent and morally responsible for participating in extreme wrongdoing. I simply assume that this is true, but I do not provide any detailed explanation of the nature of corporate action or its moral properties.

8  Introduction Finally, it is important to note that in defending the usefulness of the concept of evil I am not defending every declaration that certain actions or persons are evil. Some recent discussions of evil have been undermined by the assumption that people who disagree over whether evil exists are likely to be diametrically opposed on all significant moral and political questions. By clarifying what we mean when we talk about evil, we will be better able to understand exactly what is at issue in these disputes, and better able to sort merely linguistic disagreements from substantive moral disagreements. It is obviously unrealistic to think that any philosophical account of evil will lead to complete agreement on which particular actions or persons are evil, but my hope is that my account will help us to understand disputes about evil, and will make egregious misuses of the concept of evil stand out in greater relief.

1 The Secular Moral Concept of Evil Some people assume that the concept of evil is exclusively religious, and that those of us who deny the existence of God and other supernatural beings should also deny that anything is evil. In this chapter, I will try to show that this would be a mistake. Atheists and theists alike can believe in the existence of evil. Of course, when we assert that evil exists we cannot expect everyone to understand exactly what we mean, and it turns out to be quite difficult to identify the nature of evil. In part, this is due to the fact that the word “evil” is ambiguous. “Evil” can be used to express the broad concept of bad, or to express a more extreme moral concept. It is the latter that deserves to be called the concept of evil, and it is this concept that I will investigate in this book. Our best hope of forming a clearer view of the concept of evil lies in a method of inquiry that philosophers call conceptual analysis. Since the process of conceptual analysis will be unfamiliar to some readers, I will conclude this chapter by setting out this methodology.

1.1  A Naturalistic Argument for Error Theory Like many other philosophers, I believe that there is no supernatural God, nor any other supernatural agents, such as demons, spirits or angels. I could be described as an atheist and a metaphysical naturalist, and I intend to give a broadly naturalistic account of evil.1 Given that the word “naturalistic” is used by philosophers in many different ways, I ought to specify the sense in which my project is naturalistic. It is not my aim to adopt a naturalistic standpoint and then explain evil away, as a metaethical error theorist might try to show that so-called moral obligations are nothing more than an illusion. Instead, I want to defend a kind of realism about   For the sake of simplicity, I will assume that atheists believe not only that God, as conceived of by theists, does not exist, but also believe that there are no other supernatural agents, such as ghosts, spirits, angels, or demons. 1

10  The Secular Moral Concept of Evil evil; a realism that is compatible with a broadly naturalistic worldview. My goal is not to uphold the kind of narrow or austere naturalism which would demand that, if evil exists, it must be reducible to that which is the subject matter of the physical sciences. My more modest aim is to show that evil is no more metaphysically odd or problematic than wrongness, justice, or generosity. Insofar as atheists can believe that there are such things as injustice and generosity, I contend, we can also believe that there is such a thing as evil. That said, my account of evil is not intended to be unsuitable for people who do believe in the existence of a supernatural God or other supernatural agents. Perhaps it is best to say that I will offer a secular account of evil. Some people think that the concept of evil is supernaturally loaded, or is exclusively religious, and hence that any secular accout of evil must be misguided.2 For certain audiences, the word “evil” brings to mind tales of demonic possession, and horror films populated with ghoulish villains and inhuman monsters.3 To some, “evil” sounds distinctly Biblical, and conjures up thoughts of divine commands and Satan’s disobedience. The journalist Lance Morrow notes that in enlightened political conversation, the word “evil” has been disreputable for a long time— and still is, to a large extent. . . . The word “evil”, in many minds, still smacks of an atavistic, superstitious, and even medieval simplism, of a fundamentalist mindset that might be inclined to burn witches. (Morrow 2003, 12)

As Peter Dews claims, some people believe that the term “evil” has an “inherently antiquated ring about it . . . [that] suggests a vision of the universe as the stage for a battle of supernatural powers” (Dews 2008, 2). When President G.  W. Bush declared that evil is real, Dews claims, it seemed that he was implying that there are “menacing forces at large in the world, working at a level deeper than individual human agency” (Dews 2008, 2). In order to establish the possibility of a secular, naturalistic account of evil, I must address these common concerns about the metaphysical status of evil. As we shall see in subsequent chapters, there are several other possible grounds for scepticism about evil that deserve our attention. Yet the worries about religion and 2   For the purposes of this argument, I will use the labels “exclusively religious” and “supernaturally loaded” interchangeably. Of course, not all people who describe themselves as religious and who participate in religious practices also believe in supernatural agents, and not all people who believe in supernatural agents are religious. Nonetheless, belief in supernatural agents is an important part of religion for very many religious people, and I think that those who claim that atheists cannot believe in evil typically would assume that belief in evil involves belief in such agents. 3   Following a common philosophical convention, when I am mentioning a word rather than using that word, I will enclose the word in inverted commas. Thus, “evil” contains four letters, whereas evil is worthy of condemnation. This use of inverted commas should not be mistaken for the use of so-called scare-quotes.

The Secular Moral Concept of Evil  11 supernaturalist metaphysics can act as an immediate barrier that prevents some atheists and agnostics from even engaging with more complex questions regarding the nature of evil. For this reason I will respond to the metaphysical objections in this first chapter, and, in the process of doing so, will bring the target of this book into sharper focus. Some atheists suppose that they ought not say that anything is evil because evil-talk has religious connotations. What certainly is true is that the word “evil” makes many people think of religion. Yet it would be a mistake to believe that anything that makes people think of religion is out of bounds for those of us who are not religious. When we hear people talk about stained glass windows, mosques, or prayer wheels it is hard not to think of religion, but we do not infer that everyone who speaks of those things is a religious believer, nor that in using such language we would implicitly uphold a religious worldview. Talk of forgiveness also makes many people think of religion, but it would be not only foolish but tragic if atheists felt that we had to deny the existence of forgiveness, or, indeed, if we always refused to forgive. It could also be the case that the word “evil” makes many people think of religion, but that not everyone who believes that evil exists is thereby committed to a religious worldview. Alternatively, some atheists might shy away from judging that anything is evil because they believe that the concept of evil originated in religious belief or practice, and hence remains exclusively religious. It is not clear whether the concept of evil did have these beginnings. The word “evil” derives from the Old English “yfel” meaning “over” or “beyond” (Oxford English Dictionary), but it is at least possible that the concept of evil is much older than this (cf. Jackson 1998, 33; Robinson 2009, 336). According to one genealogical story, it originated in the dualistic theology of Zoroastrianism, entered Judaic thought with the Persian dominance of the Near East, and subsequently spread through Western culture via the theology of the Christian church (cf. Stone 2009, 10). Even if this genealogy of the concept of evil were correct, though, it is not the case that all concepts that originated in religious practices are out of bounds for atheists. The idea of a scapegoat, for instance, is derived from a Jewish religious tradition involving a goat that was part of the observance of Yom Kippur, but it is possible to claim that Lee Harvey Oswald was a scapegoat without thereby expressing commitment to the Jewish faith, or belief in a supernatural realm. It seems possible that atheists could believe in the existence of evil just as easily as in the existence of scapegoats. If atheism really is a significant barrier to believing in evil, then there must be some deeper problem than the mere fact that the word “evil” is used in religious texts, and makes many people think of religion. There would have to be some kind of inconsistency between atheism and evil-realism. This view can be encapsulated

12  The Secular Moral Concept of Evil in what I will call the naturalistic argument for error theory about evil (cf. Garrard 2002, 325): 1. Evil exists only if God or other supernatural agents exist. 2. God and other supernatural agents do not exist. Therefore, 3. Evil does not exist. The plausibility of this argument depends, in part, on the truth of the second premise; namely, that God and other supernatural agents do not exist. Instead of directly assessing this highly contentious second premise, I will try to undermine the conditional claim that evil exists only if God or other supernatural agents exist. If this first premise is false, then theists and atheists can agree that the naturalistic argument for error theory is no good. Let us begin by asking why some of us might be inclined to think that the first premise is true. It is undeniable that the word “evil” is often used when people are talking about a supernatural realm. Evil, some say, is a supernatural force that can enter our bodies and influence our actions. Although it is hard to know exactly what people mean when they speak of evil as a force, many of those people describe it in agential terms, suggesting that the so-called force is malevolent, goal-directed, and controlling. Often the word “evil” is used not as a name of a supernatural agent, but as a description of such an agent. Fire-and-brimstone preachers, for instance, say that Satan is an evil fallen angel who tempts us into wrongdoing. As Phillip Cole points out, it is also common for extreme wrongdoers to be described as evil monsters, and as demonically evil. According to Cole, describing someone as an evil monster implies that he has “demonic and supernatural powers” (Cole 2006, 215). If the word “evil” functioned merely as a name for a supernatural agential force, or as a label for a supernatural property of supernatural beings, then it would be true that atheists ought not believe in the existence of evil. Yet it is common for people who believe in the existence of supernatural agents to apply the word “evil” to natural objects and events as well. For instance, some theists might say that Satan is an evil being, that Charles Manson is an evil man, that atheism is an evil doctrine, and that murder is an evil act. The fact that even theists judge that some things in the natural world are evil is prima facie evidence that atheists can believe in the existence of evil. After all, atheists believe in the existence of Charles Manson, and in the existence of murder. It is possible, though, that the theists who say that Charles Manson is evil believe that Manson was causally influenced by a supernatural being, such as an evil spirit. Similarly, theists might believe that an action such as murder, which is part of the natural world, could be evil only if that action is performed in defiance of the commands issued by God. Maybe supernatural agents need somehow to be involved

The Secular Moral Concept of Evil  13 whenever something in the natural world counts as evil. If this were true, then atheists, who deny that any supernatural beings exist, should not believe in evil. Advocates of a secular account of evil must admit that some people use the word “evil” with an intent to refer to supernatural entities, and that some of those same people would say that any part of the natural world could be evil only in virtue of being connected to a supernatural agent. Yet the same is true of other moral concepts that are accepted as being available for use by atheists. Some theists believe that there are supernatural beings who do wrong, and that human actions count as morally wrong only in so far as they clash with the commands issued by God. This gives us no reason to suppose that the concept of moral wrongness is out of bounds for atheists, or that every person who says, for instance, that sexual harassment is morally wrong thereby expresses commitment to a supernaturalistic worldview. Theists and atheists use the same concept of moral wrongness when they agree that gratuitous torture is morally wrong, even though they might disagree about which other actions are morally wrong, and might give different explanations of why a particular action is morally wrong. As yet, we have been given no reason to suppose that this is not equally true of the concept of evil.

1.2  Supernaturally Loaded Concepts Perhaps the way forward is to identify some examples of other concepts that are supernaturally loaded, and note what leads us to place them in that category. It is plausible that the concept of God, the concept of ghost, and the concept of sin are supernaturally loaded in the relevant sense. Strictly speaking, this is not to say that they are concepts that can be used only by people who believe in supernatural agents. There are ways we can use a concept that do not commit us to the belief that the concept in question picks out something real. For instance, when I say “G. W. Bush believes in God”, or “If God exists then God will be angry with me”, I use the concept of God, but I do not express a belief that there is a God. In contrast, if I assert that God is worthy of worship, or that God commands that we love one another, then, ceteris paribus, my audience would assume that I believe in a supernatural being.4 Similarly, if I say “There is a ghost in the next 4   In relation to the word “God”, sometimes other things are not equal. There is a long-running tradition of arguing that God exists, and that “God” refers not to a supernatural being, but to the entire physical universe, or to a sense of the sacred (e.g. Johnston 2011). Arguably, this tradition is motivated by a desire to save what is good and significant about religious discourse and practice while rejecting contentious supernaturalistic metaphysics. Regardless of whether this project is worthwhile, it remains the case that the everyday sense of the word “God” is that of a supernatural creator of the universe. Deists who wish to talk about God and to be understood by a general audience must preface their claims with a specification of the non-standard sense in which they use the word “God”, and Deists must maintain that many common beliefs about God—that God wants us to be happy, or that God is watching over us, for instance—are incoherent on their understanding of God. We might say that Deists use the word “God”, but do not use the common supernaturalistic concept of God.

14  The Secular Moral Concept of Evil room”, then, ceteris paribus, listeners will ascribe to me a belief in supernatural beings. When atheists are being honest and speaking literally, they simply do not make such claims, and their refusal to make assertions about the nature or the activities of God or ghosts is the clearest indication that the concept of God and the concept of ghost are supernaturally loaded. It is harder to know what we ought to say about sin. In order to see why this is the case we must consider the relationship between the English word “sin” and the concept of sin. The primary definition of the word “sin” offered in the Oxford English Dictionary (OED) is an “act which is regarded as a transgression of the divine law and an offence against God; a violation (esp. wilful or deliberate) of some religious or moral principle”. The first half of this definition suggests that the word “sin” picks out a supernaturally loaded concept, but the second half suggests that “sin” also applies to violations of moral principle, whether they are religious or not. However, as the OED makes clear, “sin” can also be used to mean a violation of some non-moral and non-religious standard of taste or propriety, as when an author admits to being guilty of certain literary sins. This meaning of the word “sin” clearly does not carry any supernatural baggage at all. If an avowed atheist says that combining blue and green in the same outfit is a sartorial sin, we would know exactly what she means, and would not be tempted to ascribe to her any beliefs regarding God’s commands about clothing. In light of these various definitions of the word “sin”, what should we say about the concept of sin? Should we say that all of these definitions get at one and the same concept? Or does each definition pick out a distinct concept? And if so, which counts as the concept of sin? There are many cases in which a single word can be used to express two different concepts. For instance, the word “bear” can be used to express the concept of load-carrying, and also to pick out an animal of the family Ursidae. The case of “sin” is more complex, though, because the various meanings of “sin” are more closely related. The broadest definition, according to which a sin is the violation of a standard of propriety, includes the narrower definitions, given that moral rules and divine laws count as standards of propriety. This overlap might lead us to suppose that all of these definitions pick out a single concept of sin. I think that this supposition would be mistaken. The supernaturally loaded definition of “sin” picks out a concept that plays a role that is significantly different from the role that is played by the broader concept of violation or transgression. People who believe that an action is a sin in the supernaturally loaded sense are comparatively less likely to perform that action, are more likely to feel shame or guilt if they do perform that action, and are likely to warn other people against performing that action. To some degree, this will also be true of people who believe

The Secular Moral Concept of Evil  15 that an action is a sin in the broader sense of being a violation of a standard of propriety. The difference between the two concepts becomes apparent when we note that people who believe that an action is a sin in the supernaturally loaded sense are likely to take the violation more seriously, are likely to provide evidence for the action’s wrongness by pointing to religious texts or other sources of supposed divine revelation, and are likely to believe that people who perform that action thereby incur the risk of divinely administered punishment in this life or in the afterlife. We could expect that someone who asserts that adultery is a sin would have a cluster of these dispositions and beliefs. In contrast, a critic who accuses an author of committing certain literary sins is not at all likely to justify her claims by pointing to commands in religious texts, nor to believe that the author will incur divine disapproval, or burn in Hell for his transgressions. Given this significant difference, we ought to conclude that the word “sin” is ambiguous. It picks out two distinct but related concepts, only one of which is supernaturally loaded. This ambiguity can cause confusion. For example, if we hear a stranger say that combining meat and cheese in the one dish is a sin, we may have to ask whether she means that it is a mere violation of some culinary standard or that it is a violation of a divine law concerning food. Even though the word “sin” picks out two distinct but related concepts, there are several reasons for concluding that the concept of sin is the one that is picked out by the supernaturally loaded definition. The first reason is that there are other terms in English which clearly pick out the alternative concept, and which are more obvious labels for that concept. Thus it makes sense to ask “Do you mean that adultery is a sin, or merely that it is a violation or transgression?” Imagine that we are asked to label the common concept that is in play in the following three thoughts: that being offside is against the rules of football, that eating with your mouth open is rude, and that lying for profit is immoral. It would be not only odd but misleading to say that in these three thoughts the common concept is that of sin, but not at all misleading to say that the common concept is that of violation. A second reason for concluding that the concept of sin is the supernaturally loaded concept is the fact that people who use the word “sin” to mean mere violation tend to do so when speaking or writing in a comparatively light-hearted or playful register. There is a hint of comic exaggeration in the claim that wearing odd socks is a sin. When someone who makes this claim is challenged to explain her meaning, she is likely to say that she does not think that wearing odd socks literally is a sin, but merely that it is a violation of a sartorial rule. A third reason for saying that the concept of sin is supernaturally loaded is that a person who is religious and who describes certain actions as sinful is likely to retract these claims upon abandoning her religious beliefs. For instance, someone who has lost her faith might

16  The Secular Moral Concept of Evil say that she used to think that lying was both morally wrong and sinful, but that now she thinks that lying is merely morally wrong. There are good grounds for concluding that the concept of sin is an exclusively religious or supernaturally loaded concept. Even so, the word “sin” is sometimes used to express a different secular concept: the concept of a violation or transgression. It would be a mistake to suggest that atheists who say that wearing odd socks is a sin are thereby contradicting their own atheistic beliefs, or that they are misapplying an exclusively religious concept. Rather, they are not using the concept of sin at all. They are using the ambiguous word “sin” in this instance to express the concept of violation. A philosophical debate over whether any actions really are sins will be a debate about the application of the supernaturally loaded concept of sin, not merely the concept of violation. This is reflected in the fact that a naturalistic argument for error theory about sin is entirely plausible to atheists. Sin can exist only if God exists. If God does not exist, then there are no sins (although it is not at all obvious that the absence of God would imply that nothing is morally wrong).

1.3  The Word “Evil” and the Concept of Evil In order to defend the view that the concept of evil is not supernaturally loaded, I must show that evil is not like sin in the relevant respects. I will argue that the word “evil” is like the word “sin” in so far as both are ambiguous, and both can express a broad concept and a narrower concept that is a subspecies of the broader concept. Moreover, it is the narrower concept that properly deserves to be called the concept of evil, and that holds special philosophical interest. The crucial difference between evil and sin, though, is that the narrower concept of evil is not supernaturally loaded. When it comes to identifying the meaning of the word “evil”, the place to start is with dictionary definitions. Such definitions cannot settle sceptical disagreements, nor can they do the hard philosophical work of conceptual analysis for us, but they do reveal common ways in which words are and have been used. In this case, dictionary definitions can help us to see which concept, if any, deserves to be called the concept of evil (cf. Kingsbury and McKewon-Green 2009, 170). People who are sceptical of the existence of evil are likely to be surprised by the breadth of current dictionary definitions. The entry in the OED suggests that the primary meaning of “evil” is “the antithesis of good”, or “that which is censurable, mischievous, or undesirable.” The word “evil”, according to the OED, sometimes means “bad in a positive sense”, which includes “[m]‌orally depraved, bad, wicked, vicious”, and sometimes means “bad in a privative sense”, or that which lacks goodness. The fact

The Secular Moral Concept of Evil  17 that the adjective “evil” has been applied to supernatural beings, as in the phrase “evil spirit”, is noted in the OED, but there is no suggestion that any of the meanings of the word “evil” are restricted to the supernatural realm or otherwise tied to supernatural entities. As the entry in the OED makes clear, there are many obsolete meanings of the word “evil”, including “unfortunate, miserable, wretched”, “inferior in quality, condition or appearance”, and “unwholesome” or “wanting in the essentials of healthy nutrition”. The latter sense of the word is used by Gloucester, in Shakespeare’s Richard III, when he says that the sickly King has “kept an evil diet long” (Richard III I.i.139). Another meaning of “evil” listed in the OED is “unpleasant, disagreeable”, or “troublesome, painful”. Arguably, it is this meaning that was intended by John Locke when he declared “We name that Evil, which is apt to produce or increase any Pain, or diminish any Pleasure in us” (Locke 1987, 229). Jeremy Bentham also appears to have had this meaning in mind when he stated that all punishment in itself is evil. Upon the principle of utility, if it ought at all to be admitted, it ought only to be admitted in as far as it promises to exclude some greater evil. (Bentham 1948, 170)

To contemporary ears, the claim that a poor diet is evil, that a painful toothache is evil, or that justified punishment is evil might sound like hyperbolic misuse of the term. Yet this jarring effect is better interpreted as a sign that the meaning of the word “evil” has shifted over time (cf. Robinson 2009, 331). According to the OED, the word “evil” in Old English was “the most comprehensive adjectival expression of disapproval, dislike or disparagement”. In modern colloquial English, the dictionary tells us, this adjective is “commonly superseded by bad”. While it is true that in many cases the word “evil” has been replaced by “bad”, “evil” certainly has not dropped out of the lexicon. There are plenty of contemporary examples of declarations involving “evil”, but usually these are extreme moral claims: that suicide bombing is an evil act, that serial killers are evil persons, and that slavery is an evil institution. This suggests that the word “evil” was once used to express the broader concept of bad but now it means something else; something more precise and more extreme. It would be a mistake, though, to think that the word “evil” once had a single, unified meaning and now has a different single, unified meaning. The seemingly old-fashioned usage of “evil” to mean nothing more than bad persists to this day in particular contexts. In the stock expression “the lesser of two evils”, for example, we take “evil” to mean bad. When an airline passenger says that she chose the chicken rather than the fish because it was the lesser of two evils, we do not assume that she believes that the chicken or the fish are extremely bad, much less that they are morally bad.

18  The Secular Moral Concept of Evil The use of the word “evil” to mean bad also survives in certain pockets of contemporary discourse that remain strongly connected to older texts, including some religious and philosophical discourse. This use of “evil” is evident in contemporary discussion of the theological problem of evil. When theologians and philosophers ask how an all-good, all-knowing, and all-powerful God could have created a world that contains “evil”, the examples of so-called evil that they consider include suffering caused by earthquakes or by cancer, and the predation of non-human animals by other non-human animals (Tooley 1980, 372; Plantinga 1990, 107; Van Inwagen 2008, 113). Clearly, these things are bad in some sense, but they are not morally bad or wrong. As Marcus Singer points out, the word “evil’ is sometimes used even in recent book titles as “a euphonious surrogate for the opposite of ‘good’ ’ (Singer 2004, 188), despite the fact that those books do not focus exclusively on extreme wrongs or on extremely bad people (e.g. Shafer-Landau 2004; Dews 2008). It is apparent that in some settings the word “evil” is still used to express the broad concept of bad, a concept whose extension includes things that are extremely bad as well as things that are trivially bad, and things that are morally bad as well as things that are otherwise deficient. As we have seen, though, there are heated contemporary disputes over whether anything at all ought to be described as evil. We cannot make sense of this unless we suppose that in these disputes the word “evil” is used to express something far more contentious than the concept of bad. For instance, if people are arguing over whether the gratuitous infliction of torture by members of the United States military in the Abu Ghraib prison was evil, it is clear that the disputed issue is not merely whether those acts of torture were bad in some respect. Most of us agree that they were bad, but there is serious disagreement over whether they were evil. It would be absurd for someone in this context to argue that since a toothache is evil, torture is obviously evil as well. Similarly, a dispute as to whether the serial killer John Wayne Gacy was an evil person is not focused on the question of whether he was a bad person. It seems that in most contemporary contexts we restrict the application of the term “evil” to those things that are extremely bad, and there is much dispute over which extremely bad things, if any, ought to be called evil. Moreover, in most contemporary contexts we do not apply the term “evil” to things that are bad in a non-moral way. The drawn out suffering of a young seal that is toyed with by a killer whale is terrible, but it would be jarring to say that this suffering is evil, or that the whale is an evil creature performing an evil action. This view that the concept of evil is morally loaded and extreme is reflected in the accounts of evil that have been offered by contemporary philosophers. Claudia Card maintains that “[n]‌atural events—earthquakes, fires, floods—not brought

The Secular Moral Concept of Evil  19 about by moral agency are not evil. Catastrophes are not the same as atrocities” (Card 2002, 5). Susan Neiman agrees that, according to the modern conception, “evil is the product of will” (Neiman 2003, 268). The distinction between so-called natural evil and moral evil is accepted as standard in the recent philosophical literature, as is the assumption that contemporary debates over the existence of evil are focused exclusively on moral evil (Thomas 1993, 4; Garrard 2002, 320; Singer 2004, 189; Kekes 2005, 1; Cole 2006, 5; Russell 2006, 93; Formosa 2008, 217). Someone who calls a mildly harmful action “evil” when she is in the context of a theological debate over the problem of evil might deny that the very same action is “evil” when she is in the context of an ordinary moral argument. In such a case it is unlikely that the speaker is making inconsistent claims about evil. Rather, the word “evil” is ambiguous, and this speaker is probably using it to express two different concepts: the broad concept of bad, in the first debate, and a comparatively extreme moral concept in the second. It is the latter rather than the former concept that deserves to be called the concept of evil. In part, this is because there is a clearer alternative label for the broader of the two concepts; namely, the concept of bad. If we were asked to provide a single word that described a toothache, an insult, and a murder, it would be jarring to say that the word is “evil”, but not at all odd to say that the word is “bad”. It is the narrower concept that deserves to be called the concept of evil, and this is the concept that is in play in disputes amongst philosophers, historians, psychologists, and journalists about the existence of evil. A clear indication that a person is employing the word “evil” with the intent of expressing the concept of evil, as opposed to the concept of bad, is that she would accept the paraphrase “not merely bad, but evil”, or “not merely wrong, but evil”. What is under dispute is whether terrorist bombings are not merely wrong, but evil, or whether Adolf Eichmann was not merely a bad man, but an evil man. It is this extreme moral concept of evil that will be the focus of this book. Henceforth, I  will refer to this simply as the concept of evil. I  take it that the concept of evil is usually picked out via the word “evil” in English, by “böse” in German, and no doubt by other words or phrases in other languages. However, we should not assume that the concept of evil is always picked out when those specific words are used, nor that it could be picked out only by those specific words (cf. Jackson 1998, 33; Kingsbury and McKewon-Green 2009, 178). We should also acknowledge that there may be some cultures that do not use the concept of evil, and that in some cases it will be hard to tell whether a particular concept that is in play in another culture ought to be referred to as the concept of evil (cf. Robinson 2009, 330, 335). As yet, much remains unclear about the content of the concept of evil, and I hope to fill in the details in the chapters to come. Throughout this investigation it is important that we remain aware that the word “evil” is ambiguous, and that

20  The Secular Moral Concept of Evil sometimes, including in many cases in the Bible and in canonical philosophical texts, “evil” is used to express nothing more than the concept of bad.5 My target is that which is evil, not merely that which is bad.

1.4  Is the Concept of Evil Supernaturally Loaded? Earlier I claimed that the concept of sin is exclusively religious. What distinguishes the concept of sin from the broader concept of violation or wrongdoing is some extra supernaturally loaded content: a sin is a violation of a divine law. For this reason, atheists ought to be error theorists about sin. We have also seen that the concept of evil is somehow narrower and more extreme than the concept of bad. It might be tempting to think that the concept of evil, like the concept of sin, essentially includes reference to a supernatural agent, and that this extra content distinguishes it from the merely bad. I will argue that this thought would be mistaken. The concept of evil is unlike the concept of sin in this respect. The primary obstacle to the claim that the concept of evil is exclusively religious is the fact that many atheists, when speaking honestly, seriously, and literally, assert that some things are evil. For instance, Hannah Arendt describes Adolf Eichmann’s actions as evil (Arendt 2006, 252). Primo Levi implies that some concentration camp guards were evil (Levi 1989, 169) and says that the autobiography of Rudolf Hoess is “filled with evil” (Levi 2000, 19). Holocaust historian Yehuda Bauer, also an atheist, says that Hitler represents “near-ultimate evil” (Rosenbaum 1998, xxi). Some of the best-known recent advocates of atheism also judge that some things are evil. Christopher Hitchens, for instance, thinks that we need a word to describe extreme wrongdoing, and that “evil” is “the best negative superlative that we possess” (Hitchens 2002).6 These atheists are not writing or speaking in a lighthearted register, in which we could suppose that they are using the term “evil” playfully. Nor do they use the word in a way that suggests they intend merely to employ the broader concept of bad.   Kant uses “böse”, translated as “evil”, to express the concept of that which is morally wrong, but not necessarily extreme (Kant 1997, 52). Kant’s account of “radical evil” is an explanation of how a free rational agent could be capable of wrongdoing simpliciter, not of extreme, cruel or horrendous wrongdoing (Kant 1996, 76–85). Nietzsche’s use of the word “böse” also fails to map neatly onto contemporary usage of “evil”. The Nietzchean contrast between the good/bad dichotomy and the good/evil dichotomy suggests that the category of evil is a moral category, and that the person who judges that an action is evil holds the agent morally responsible, and feels a self-righteous sense of indignation. However, Nietzsche does not clearly indicate that those who employ the concept of evil restrict it to only the worst wrongs or the worst persons (Nietzsche 1996, 21–4; Leiter 2002, 209–11). In this respect Nietzsche’s category of evil does not map onto contemporary use of the word “evil”. 6   Richard Dawkins also believes that some things are evil. For instance, he says that the belief that apostates need to be killed “clearly is evil” (Paulson 2006). However, it is harder to tell whether Dawkins merely uses “evil” as a synonym for “bad” or “wrong”. 5

The Secular Moral Concept of Evil  21 These atheists, along with many others, say that there are no supernatural agents, but they also say that some things are evil. It might be suggested that atheists who speak of evil do not intend thereby to express commitment to a supernaturalist worldview, but that they do so unintentionally. Phillip Cole thinks that something like this takes place when professed metaphysical naturalists use the language of evil: What we have here is a mythology of the evil enemy, such that that enemy possesses the demonic, supernatural powers needed to destroy our communities. This, surely, is a step too far—nobody seriously believes that migrants and terrorists have supernatural powers. But this is exactly what happens through the discourse of evil: the migrant and the terrorist, while they are not represented as agents of Satan, are represented as possessing demonic and supernatural powers. (Cole 2006, 215)

Cole’s claim is, at best, an exaggeration. It is possible that some of the people who say that terrorists are evil believe that terrorists possess supernatural powers, but when Arendt, Levi, and Hitchens claim that certain persons or actions are evil, only the most naïve readers would assume that they are representing them as supernatural, or as being connected to the supernatural. When Arendt (2006, 252) pronounces that the Eichmann trial reveals the banality of evil, for instance, who honestly thinks that she is telling us that Eichmann was possessed by some kind of banal demon? One reason for supposing that the concept of sin is exclusively religious is that a religious believer who abandons her faith is likely to give up on claims such as “Lying is a sin”, and replace them with clearly secular alternatives, such as “Lying is wrong”. The same is not true of evil. If the concept of evil really were exclusively religious, then a Holocaust survivor, who ceased believing in God after witnessing such atrocity, could say that she once thought that the Nazis were guilty of evil deeds, but now as an atheist she can think only that their actions were terribly wrong, not evil. In reality, though, there would be nothing unusual in a Holocaust survivor saying that it was the evil deeds perpetrated at Auschwitz that prevent her from returning to her earlier belief in God. What she has given up is her belief in supernatural beings, not her judgement that she has witnessed evil. It is odd that some people think that if evil exists, it must be supernatural. Perhaps this confusion is prompted by the way in which the question about the reality of evil is phrased. If we ask “Does evil exist?”, it might seem that we are asking whether there is a single object, being, or force that is called evil. In reality, though, our question is broader than that. We want to know whether anything is properly described as evil. When we consider whether greed exists we are not tempted to suppose that greed is a separately existing entity that can enter into a human being and make her act in selfish ways. Rather, we assume that greed is a

22  The Secular Moral Concept of Evil character trait, or a motive, or a property of an action. Greed exists in the sense that there are greedy people and greedy acts. Of course, if there are supernatural greedy beings, then greed exists. But if there are only natural greedy beings, then greed still exists. I contend that evil is just like greed in this respect. Evil exists if there are evil things, whether those things are natural or supernatural. As such, the concept of evil is a secular concept that is available for use by theists and atheists alike. While Phillip Cole thinks that the concept of evil is enmeshed in a supernaturalistic worldview, the contrary position is widely accepted in the recent philosophical literature on evil (e.g. McGinn 1997; Card 2002; Garrard 2002; Haybron 2002a; Steiner 2002; Neiman 2003; Morton 2004; Russell 2006; Formosa 2008). I have argued that this widely shared view is correct. We ought not ignore the fact that there are many conspicuous, dramatic, and memorable applications of the concept of evil to supernatural beings in religious texts and in popular fictions such as horror movies. But nor should we overlook the many other conspicuous, dramatic, and memorable cases in which the concept of evil is applied to human actions, such as genocide, torture, and serial killing, and to human beings, including Pol Pot and Ted Bundy. It is possible to be sceptical of the former, supernatural applications without being sceptical of the latter. As Ron Rosenbaum suggests, many of the people who say that Hitler was evil think of evil not as “some supernatural numinous entity but . . . as a name for a capacity of human nature” (Rosenbaum 1998, xx). Atheists who declare that Hitler was evil are obliged to explain in more detail what they mean by that claim, but such naturalistic accounts of evil should be evaluated on their merits, rather than summarily dismissed on metaphysical grounds.

1.5  Defining the Concept of Evil The naturalistic argument for error theory about evil turns out to be unconvincing regardless of whether God or other supernatural agents actually exist, because it is not true that evil is a supernaturally loaded concept. It is worth stressing that the failure of this particular argument does not imply that evil does exist. There are a range of other considerations for and against believing in the existence of evil, and I will address these in the following chapters. At this stage I hope merely to have narrowed in on our target, and cleared the ground for the development of a secular account of evil. As yet, the account looks rather sketchy. The concept of evil is supposed to be more extreme than the concept of bad, but we require a more precise characterization than this. Ordinarily we would seek this precision by trying to define the concept of evil. Some authors who have recently written about evil are explicitly

The Secular Moral Concept of Evil  23 dismissive of this project, but even these authors appear unable to avoid sliding into disputes over competing definitions of evil. For example, Lance Morrow confidently declares that we “know evil by its works. It defines itself. Good is what ought to be; evil is what ought not to be. Do not trouble yourself overmuch about the definition of evil” (Morrow 2003, 142). In four brief sentences Morrow rejects the need for a definition, immediately attempts to provide a definition, then encourages us not to assess it. This is a shame, given that interesting and difficult definitional questions arise even from Morrow’s abortive attempt. If evil is merely “what ought not to be”, do instances of false belief and violations of norms of etiquette count as evil? If evil covers all of “what ought not to be”, then how can evil also be an agent who has “works”? Rather than not saying enough about the nature of evil, Terry Eagleton says too much, apparently with little thought as to whether his claims are consistent. In his book On Evil, Eagleton either wallows in contradiction or switches between meanings of the word “evil” without signalling to the reader that he is doing so. For instance, he says that evil wants to annihilate being as such (Eagleton 2010, 16) and has a “grisly kind of rationality about it” (103), but also that evil is purposeless (103). Eagleton goes on to claim that evil is simultaneously a deficiency or absence and a positive “death drive” (127). He suggests that evil is “rare” (129) but also that anything which is bad, including “dengue fever, Britney Spears [and] tarantulas”, is evil (138). Evil is “dangerous precisely because of its lethal innocence” (124), but evil is “is not something that we should lose too much sleep over” (130). Someone who is seeking to get a firmer grip on the concept of evil is unlikely to be helped by this kind of free-associative riffing. In contrast to such barrages of inconsistent claims, many philosophers have offered more precise accounts of the concept of evil in recent years (e.g. Kekes 1990, 2005; Thomas 1993; McGinn 1997; Garrard 1998, 2002; Calder 2002, 2013; Card 2002; de Wijze 2002; Steiner 2002; Haybron 2002a; Morton 2004; Singer 2004; Cole 2006; Russell 2006, 2007b, 2009a, 2010a, 2010b, 2012; Formosa 2008). As we shall see, these accounts are varied, and significant disagreements on many issues concerning evil remain unresolved in this literature. This persistent disagreement might be disheartening, and could even be taken as evidence that defining the concept of evil is a hopeless task. Susan Neiman, for one, refuses to offer “a definition of evil or criteria for distinguishing evil actions from those that are simply very bad,” arguing instead that we should focus on paradigms of evil (Neiman 2003, 8–9). She elaborates: Evils can be acknowledged as evils without insisting that evil has an essence. Our inability to find something deep that is common to the mass murders committed by terrorists and the starvation furthered by corporate interests does not prevent us from condemning both. Thinking clearly is crucial; finding formulas is not. (Neiman 2003, 286–7)

24  The Secular Moral Concept of Evil As Neiman hints here, the concept of evil could turn out to be what Wittgenstein calls a family resemblance concept, whose content cannot be captured in a set of necessary and sufficient conditions (Wittgenstein 1953, 31–2). We are best able to discover whether this is the case, though, by wholeheartedly attempting to define evil and noting where and why our definitions break down, if indeed they do. If the concept of evil does turn out to be a family resemblance concept, we would benefit from clearly recognizing which resemblances hold between which members of the family. Our becoming aware of these resemblances might be a useful by-product of a failed attempt to define evil (cf. Kingsbury and McKewon-Green 2009, 167). Moreover, Neiman is wrong to imply that every definition of evil must state that there is “something deep” in common between all examples of evil actions. Thinking clearly about evil includes the task of constructing and comparing broader and narrower definitions. Philosophers who reject the view that all evil actions spring from the same kind of motives need not reject the project of defining evil. They might simply offer a broader definition. With these goals and these potential pitfalls in mind, I will make a wholehearted attempt to define the concept of evil.

1.6  A Brief Methodology In this section I will sketch the method that I propose to use in constructing a philosophical account of evil. This method is conceptual analysis, broadly conceived (e.g. Nolan 2009, 286). Conceptual analysis is not the only method available. Some prefer a more historical approach focused on important texts written by the great philosophers, while others aim to build philosophical theories on purely scientific foundations. In relation to the concept of evil, I believe that conceptual analysis is the method that has the most to offer.7 No doubt the practices that constitute conceptual analysis are familiar to trained philosophers, who may prefer to skip this section, but it is worth setting out these methods explicitly to help orient other readers. At various points in the overall argument of this book I will draw conclusions that may strike some philosophers as unusual. Instead of arguing that one particular account of evil action is correct, for instance, I will defend a restricted conceptual pluralism. My reasons for doing so depend on a certain view about the 7   Those who take a primarily historical approach (e.g. Neiman 2003; Dews 2008) run the risk of conflating past uses of the word “evil” with contemporary uses, and of glossing over fine-grained but significant issues concerning the nature of evil, such as the relation between evil actions and evil persons. Those who try to address the question of evil from a scientific perspective (e.g. Stone 2009; Baron-Cohen 2011) typically fail to notice the variety of roles that the concept of evil plays in moral thought, and are insufficiently sensitive to the normative nature of moral judgements. That said, we should not ignore history or science. Appeals to historical developments and to scientific theories crop up as part of the process of conceptual analysis.

The Secular Moral Concept of Evil  25 nature and the limitations of conceptual analysis. This argument may be easier to grasp, when it arrives, if I bring these methodological assumptions out into open at this early stage. The basic technique used in conceptual analysis is that of considering a broad range of actual and possible cases in an attempt to determine the necessary and sufficient conditions for the application of the concept in question (Kingsbury and McKewon-Green 2009, 160). The intended product of this process is an informative definition of the concept in question; what philosophers call an analysis of that concept. For instance, suppose that we want to provide an analysis of the concept of freely chosen action. Our analysis should allow us to see which properties or features make those actions count as freely chosen, and should allow us to explain the difference between actions that are freely chosen and actions or bodily movements that are not freely chosen. An analysis might also help us to see how a particular concept is used, and why we ought to care about that concept. We have seen that the concept of evil is a secular moral concept that goes beyond the merely bad, and the task at hand is to provide an analysis of that concept. When asked to define an obscure concept such as evil, there is a danger that philosophers will resort to mere stipulation. There are many possible ways to carve an extreme category out of the broader category of the morally bad. If we merely set about inventing ways of doing so, then we would end up with a plethora of incompatible accounts of evil. Any number of these accounts might contain useful suggestions for what we could choose to mean by the word “evil” in future, but it is unclear why we would believe that any one of them picks out the concept of evil. Thankfully, the task of conceptual analysis does not consist in merely inventing a definition and then encouraging its future use. There are several techniques that we can use to evaluate competing analyses of the concept of evil. One kind of constraint on philosophical accounts comes from the so-called folk use of the concept in question (Jackson 2001, 618; Sider 2001, 199–201). As we have seen, many claims about evil are made by philosophically untrained people, whom philosophers refer to as the folk. Presidents, journalists, historians, victims of terrible crimes, persecutors of the oppressed, and many other people say that certain things are or are not evil, and say various things about the nature of evil (e.g. Stone 2009, 23–4). While it might sound condescending, the label “folk” should not be taken to imply naïveté or stupidity. Some of the folk are highly intelligent and have various kinds of expertise, including expertise that might be relevant to the question of whether something is evil. Part of our job as philosophers is to offer an account of evil that makes sense of the folk claims and folk beliefs about evil, even though our account will not imply that all of the folk claims and beliefs are justified or true. It is important to note that we cannot get at what the folk mean by “evil”

26  The Secular Moral Concept of Evil simply by asking them directly to define the word (Robinson 2009, 320). Similarly, linguists cannot get at the rules of English grammar by asking English speakers to list those rules, because many competent speakers are utterly incompetent when it comes to giving an explicit account of the rules that they actually follow. What we ought to do instead is to notice the various ways in which people use the concept of evil, see whether these fall into a systematic pattern, and then attempt to construct an account that fits with this pattern of use (Devitt 2012, 27–9). There are two different aspects of the folk use of the concept of evil that might provide some kind of constraint on philosophical accounts. These two aspects are captured by distinct questions: “Which things are evil?”, and “In judging these things to be evil, what are we judging them to be?” In answering the first of these questions we are trying to determine the extension of the concept of evil; that is, the class of things to which the concept applies. In answering the second we are trying to get at the inferential or functional role of the concept of evil. These two aspects of use suggest two different methods by which we can evaluate competing analyses of the concept of evil. I will call the first of these the extensional method (cf. Putnam 1975, 198–202; Kripke 1980, 116–17). I will refer to the second as the inferential method (cf. Searle 1983; Joyce 2001, 156–7). Philosophers should note that I take the inferential method to include consideration of the broader functional role and the practical implications of concept use (Robinson 2009, 337–9). The extensional method will be applicable only if we have a clear grip on which things count as evil. For the sake of argument, let us suppose for a moment that all of the folk agree on which things are evil and which things are not evil. For instance, suppose we agree that Hitler and Stalin are evil persons, and that torture and so-called ethnic cleansing are evil actions, and we agree that Tony Blair and G. W. Bush are bad but not evil persons, and that cheating on your tax and pretending to be sick to avoid work are wrong but not evil actions. If we were firmly committed to these judgements, then we would have grounds for rejecting any philosophical account of evil which implied that Hitler was not evil, or that cheating on your tax was evil. In other words, one way of evaluating a philosophical analysis of the concept of evil is by checking that the analysis in question picks out the actual extension of the concept, as judged by the folk. As we shall see, there are some cases in which the analyses of evil that have been proposed by philosophers clash egregiously with widely accepted beliefs about the extension of the concept of evil, and this gives us a strong reason to reject those analyses. It should be obvious that the extensional method has far less power when applied to concepts whose extension is unclear or is widely disputed. There is deep disagreement over the extension of many normative and evaluative concepts, and this certainly is true in the case of the concept of evil. When looking for examples of

The Secular Moral Concept of Evil  27 evil action, an evangelical Christian might point to cases of abortion or homosexual sex; actions that many atheists do not judge to be wrong, let alone evil. While some people admire Blair and Bush as strong-willed and virtuous, if sometimes misguided, others take them to be downright evil. If one philosophical definition of evil suggests that Bush is an evil person and a competing definition does not, and if the folk do not agree on whether Bush really is evil, then the extensional method will not help us to decide which of these two definitions is correct. An alternative in such cases is to employ the inferential method, in which we ask not which things are evil, but what is it to judge that something is evil (cf. Mackie 1985, 33–4; Joyce 2001, 3). By way of comparison, let us consider another concept with a highly contested extension: the concept of wisdom. We cannot come to understand the concept of wisdom simply by asking which people are judged to be wise, because there is broad and recalcitrant disagreement on this issue. One person’s guru is another’s reactionary bigot. What we ought to do instead is to ask what people typically infer from their judgement that someone is wise, and what role that judgement plays in motivating their actions (Robinson 2009, 337–9). For instance, if I judge that Karen is wise, then I am likely to infer that Karen is a good source of advice, that Karen is not only instrumentally clever, but knows which goals are worth pursuing, that Karen has admirable foresight, and that Karen has a reliable moral compass. This pattern of inferences will be revealed in what I do and say with relation to the people whom I judge to be wise. Even if you and I disagree over whether Karen is wise, we might agree on what it means to say that someone is wise, because we agree that the fact that a particular person is wise would license the same set of inferences, and the judgement that that person is wise typically would motivate the same kind of behaviour. These common inferences can be pulled together into a list of intuitions about wisdom, and one way of assessing the plausibility of various definitions of wisdom is to check that they fit with these intuitions (cf. Nolan 2009, 274). The inferential method promises to be of great use when it comes to assessing competing analyses of the concept of evil. While we may not be able to resolve some disagreements over which particular things are evil, hopefully we can agree on the content of the concept of evil by identifying the inferences that would be licensed by the judgement that something is evil. In order to do this we must look at what people do and say in relation to the things that they judge to be evil, and from this draw up a list of intuitions concerning evil. If a philosophical account of evil does not fit closely enough with these intuitions, we have reason to be suspicious of that account. In subsequent chapters we shall see that there are many such common inferences regarding evil that can be used in our evaluation of competing philosophical

28  The Secular Moral Concept of Evil analyses of evil. However, the inferential method will prove to be useful in resolving disputes over the concept of evil only in so far as the disputants can reach agreement on which inferences flow from the judgement that something is evil. Unfortunately there are several deep and long-running disagreements over which inferences are licensed by the judgements about evil. The fact that not all of the folk use the concept of evil in the same way creates special difficulties for those of us who hope to define evil (cf. Robinson 2009, 321). I have suggested that we can get at the folk use of the concept of evil by employing some combination of the extensional method and the inferential method, and that this folk use of the concept provides one kind of constraint on philosophical accounts of evil. If we propose a particular analysis that strays too far from the folk use of the concept in question, then we will be accused of having changed the subject; that is, of having drifted away from our original target—evil—and begun talking about something else (Jackson 1998, 38). While the folk use of a concept does provide some constraint on what plausibly can be offered as an analysis of that concept, it would be a mistake to suppose that this is the only relevant constraint. One important way in which philosophers go beyond merely delineating folk usage is by thinking more thoroughly and precisely about the extension and inferential role of the concept in question. In some cases we have grounds for believing that the folk are inarticulate, confused, ignorant, or mistaken on various issues (Jackson 1998, 35). Philosophers have a kind of expertise in abstract thinking that helps us to identify, distinguish, and clarify intuitions regarding extension and inferential role (Robinson 2009, 320; Devitt 2012, 19). Philosophers typically consider a diverse range of examples, some of which may have been overlooked by the folk. Philosophers also consider hypothetical and counterfactual cases that are designed to tease apart subtle distinctions, and to isolate the factors that are responsible for particular intuitions (Sider 2001, 199). Philosophers often add detail or structure to an account in an attempt to unify or explain intuitions that otherwise seem disparate. Of course, philosophers can be mistaken in these judgements, and often disagree with each other. Just as a philosophical account should be constrained to some degree by what the folk say about evil, we should give significant weight, but not necessarily decisive weight, to various philosophers’ intuitions about evil. There are other constraints that might lead us to accept an analysis of evil that clashes with some folk intuitions and some of the claims that philosophers have made about evil. As Sider puts it, when doing conceptual analysis we must consider “use plus eligibility”, where the eligibility of a particular candidate analysis is determined by a range of other considerations (Sider 2001, 191). One important factor that bears on the eligibility of a particular analysis is whether that analysis

The Secular Moral Concept of Evil  29 is internally consistent rather than self-contradictory. We might discover that the various folk intuitions concerning a concept are in conflict with one another, and hence that any definition that respected all of those intuitions would be self-contradictory (Joyce 2001, 7). If this turns out to be true with regard to the concept of evil, some people would conclude that we should reject the concept as incoherent. Alternatively, we might decide to abandon one or more of the folk intuitions so as to allow for a coherent definition, and conclude that evil does exist, but that the folk are a little confused about what they suppose evil to be (cf. Jackson 1998, 44–5; Robinson 2009, 322). Another important kind of constraint on philosophical analyses comes from our best theories about the nature of the world, including scientific theories. Philosophers sometimes recommend that we revise a folk concept so that it picks out a so-called natural kind, or at least so that it points to something whose existence is compatible with our best theories about what exists (Dupré 1993, 22; Sider 2001, 191). This is the case with regards to the concept of time. Our folk understanding of time includes certain beliefs about the absolute simultaneity of events, but these beliefs do not fit with the theory of relativity. Einstein taught us that there is nothing in the actual universe that has exactly the characteristics that we once believed belonged to time. In response to this discovery it would be possible for us to adopt an error theory about time; that is, to conclude that there is no such thing as time. What seems far more attractive in this case, though, is to maintain that there is such a thing as time, and that time is not exactly what the folk supposed it to be. Here considerations of eligibility override some facts about folk intuitions. The folk understanding of time needs to be revised and replaced by a theoretically informed concept of time in order to fit with our best understanding of the world (cf. Lycan 1988, 31–2; Finlay 2008, 363). In contrast, consider the concept of witchhood. Throughout history many women were believed to be witches, and the inferential method reveals that witches are supposed to be women with magical powers who can cast spells and who commune with the Devil or with evil spirits. Given that our best theories about the nature of the world lead us to believe that there is no magic, no Devil, and no evil spirits, what should we say about the existence of witches? One option is to revise the concept of witch in a way that would allow us to say that witches really do exist, but that the folk were mistaken about their true nature. We could try to revise the concept of witch so that the existence of witches does not depend on the existence of magic powers or a supernatural realm. For instance, we could claim that a witch is merely any woman who is believed to have magic powers, or any woman who tries to cast spells. Yet these definitions seem so far from the folk understanding of a witch as to be totally implausible. Nothing that we find in the real world, as

30  The Secular Moral Concept of Evil understood through our best theories, is sufficiently close to being what a witch was supposed to be. Thus we ought to conclude that there are no witches (Devitt and Sterelny 1987, 73; Joyce 2001, 156–7). When it comes to assessing various philosophical analyses of the concept of evil, we must ask whether they describe things that fit with our best theories about the nature of the world. It might turn out that our folk understanding of the nature of evil does not fit neatly with our best theories. For example, it is possible that the folk understanding of what an evil person is supposed to be simply does not match up to the nature of any actual person as described by psychologists. As is the case with time and witches, we have two options when faced with such a predicament. We could choose to revise the folk concept of evil so that it turns out that there really are some evil persons, and claim that evil personhood is not quite what the folk took it to be. Alternatively, we might maintain that the folk understanding of evil determines what could count as evil, and conclude that there are no evil persons. Which of these options seems more attractive to us will depend, in part, on how firmly committed we are to our intuitions regarding the extension of the concept of evil (cf. Jackson 1998, 44–5). Competing philosophical analyses of evil can also be assessed in light of various pragmatic considerations. If one account of evil picks out a concept that it would be dangerous for us to employ in our moral thinking, then we might have good reason to favour alternative accounts of evil that allow us to avoid these bad consequences, even if those alternatives clash with some folk intuitions about evil. However, these pragmatic evaluations should be weighed against the other kinds of consideration listed above. The task at hand is to figure out what evil is supposed to be, and whether evil exists. We cannot complete this task by describing some interesting moral concept and simply deciding to call it the concept of evil. An optimal philosophical analysis of evil will be internally consistent, will accord with our best theories about the nature of the world, and will accord fairly closely with folk intuitions and with philosophers’ intuitions concerning evil. It should be clear by now that the method of conceptual analysis, by which I hope to construct such an account, is a fairly messy procedure that requires the balancing of several competing kinds of consideration. Rather than describe it in any more detail, let us put it into practice.

2 Evil Action We have seen that the ambiguous word “evil” is sometimes used to express nothing more than the concept of bad, but on other occasions is used to express the concept of evil. I have argued that the concept of evil is an extreme moral concept that is available for use by theists and atheists alike. My task in this chapter is to begin applying the method of conceptual analysis to the concept of evil. The two main categories of things that are called evil are actions and persons. While some philosophers have suggested that we ought to begin by focusing on the notion of evil personhood, I will argue that the concept of evil action is more fundamental, and hence a more natural starting point for a theory of evil. In this chapter, I will set out eight common intuitions about evil action, then offer a rough categorization of existing philosophical accounts of evil action in light of those eight intuitions, and finally defend the view that all evil actions are extreme culpable wrongs.

2.1  Why Begin With Evil Action? Many different kinds of thing can be rectangular: tabletops, picture frames, windows, gardens, and so on. However, if we want to understand what it means to say that something is rectangular, we do not need to give a separate account of what rectangularity consists in for each of these possible kinds of rectangular object. All we need is an account of the meaning of the word “rectangular”. Just as many different types of thing can be rectangular, it is common for many types of thing to be called evil, including actions, motives, intentions, feelings, persons, organisations, governments, and social institutions. It would be convenient if the predicate “is evil”, like “is rectangular,” made the same semantic contribution to every sentence, regardless of the subject to which it was applied, for then we could give a neat and unified account of the meaning of “evil”. Unfortunately, though, this seems not to be the case. What it is for an action to be evil is not identical to what it is for a person to be evil or for a social institution to be evil. For instance, I will argue that

32  Evil Action part of what it means to say that an action is evil is to say that there are conclusive moral reasons against performing the action, but clearly it makes no sense to say that there are conclusive moral reasons against performing a person. This gives us a reason to break down the task of analysing the concept of evil into sub-tasks: analysing the concept of evil action, the concept of evil person, the concept of evil institution, and so on, and then explaining how these more fine-grained concepts are related. It will be best to start by examining the most fundamental of these concepts. I believe that we ought to begin with the notion of evil action. One reason for beginning with evil action is that it is less contentious to claim that there are evil actions than to claim that there are evil persons or evil organizations (Stone 2009, 21). Another reason is that it is easy to see how we could use the notion of evil action as a foundation for at least some of the other applications of the word “evil”. For example, evil motives and intentions might be understood as the motives and intentions that are directed at the performance of evil actions, and evil persons might be those who perform or are disposed to perform evil actions (Card 2002, 21). Evil organizations and governments might be those that culpably lead people to perform evil actions, or that perform evil actions as groups, or that can be understood as evil corporate persons. Social institutions such as slavery or human sacrifice might be evil insofar as they incorporate, encourage, or perpetuate the performance of evil actions. More detailed argument would be required to defend the view that these various evil things are evil in virtue of their being related to evil actions in these ways, and I will fill in at least some of this detail in later chapters. At this stage, though, it is worth responding briefly to Daniel Haybron and Marcus Singer, who advocate the alternative approach of deriving an account of evil action from a prior account of evil personhood. Haybron suggests that we should reverse the usual order of explication and understand the evil action in terms of its relation to the evil character: e.g., perhaps an action is evil if it manifests profoundly deadened or perverted moral sensibilities—the sensibilities characteristic of an evil person. (Haybron 2002a, 280)

It is hard to know how best to interpret Haybron’s suggestion. In particular, it is not clear what it would be for an action to manifest deadened or perverted sensibilities. If Haybron means that an evil action is any action that is produced by profoundly deadened or perverted moral sensibilities, then his account will be implausible. The problem is that some actions that are produced by such sensibilities are merely trivial wrongs. Petty theft could be the product of profoundly deadened moral sensibilities, and telling a relatively harmless lie to an acquaintance could be the

Evil Action  33 product of profoundly perverted moral sensibilities. It is implausible that such trivially wrong actions are evil, even though they are amongst the various kind of actions that might be the product of an evil person’s character. Perhaps instead we should read Haybron as suggesting that an evil action is one that could be produced only by profoundly deadened or perverted moral sensibilities. This would rule out the possibility of trivial evils, but would also entail that only an evil person, who possesses such sensibilities, could perform an evil action. Yet many of us believe that not every evildoer is an evil person, and Haybron himself claims that “most evil actions are not the product of evil people” (Haybron 2002a, 279). Neither interpretation of Haybron’s suggestion produces a plausible derivation of the concept of evil action from the concept of evil person. Marcus Singer claims that the concept of evil “applies primarily to persons and organizations, secondarily to conduct and practices”. Singer continues, “when we say that someone did something evil, we are saying something about that person, that person’s motives and consequently about that person’s character” (Singer 2004, 190). Singer is right to claim that there are connections between evaluations of action, motive, and person, but this does not imply that all kinds of act evaluation are secondary, or are derived from person evaluation. Moreover, the fact that our evaluation of an action includes an evaluation of the agent’s motives does not imply that our evaluation of the action depends on a prior judgement about the person’s settled character traits, much less on a judgement about the person’s character as a whole. In some cases the occurrent motives that produce an action are surprisingly uncharacteristic of the agent. When a typically cowardly soldier steels himself and does something brave to protect his comrades, for instance, his action strikes us as courageous in virtue of his occurrent motives, but our judgement that he acts courageously is not derived from a prior judgement to the effect that the soldier is a characteristically courageous person (Hurka 2006, 72). Classifying an action as belonging to a particular evaluative type often requires us to make a judgement about the agent’s motives. An action is a lie only if it is performed with an intent to deceive the audience, and an action is compassionate only if the agent is moved by the suffering of others. Nevertheless, it is plausible that we understand the nature of a characteristically deceitful person and the nature of a characteristically compassionate person by first grasping what it is to tell a lie, and what it is to feel and act out of compassion. As we shall see in later chapters, evaluating a person is a much more demanding task than that of evaluating an action in terms of its motives. Perhaps it is true that in saying that an action is evil we are saying something about the agent’s occurrent motives, but this falls short of saying something about the agent’s character, and certainly falls short of saying that the agent is an evil person.

34  Evil Action Since neither Haybron nor Singer point to good reasons for thinking that evil personhood is logically prior to evil action, I will start by focusing on evil action in the hope that later I can build an account of evil personhood on this foundation. Many philosophers who recently have written on evil also take this approach (e.g. Calder 2002, 56; Card 2002, 21; Morton 2004, 66; Kekes 2005, 2), and even Singer, who initially says that evil applies primarily to persons and secondarily to actions, proceeds to define evil person in terms of a prior definition of evil action (Singer 2004, 196).1

2.2  Intuitions Concerning Evil Action I will begin the task of defining evil action by offering a list of eight common intuitions and showing how recent accounts of evil can be grouped according to their response to these various intuitions. Each of these intuitions has been endorsed by at least some of the philosophers who have recently written about evil, and each has some prima facie plausibility. Some are vague and require clarification, which I hope to provide in the following chapters. Even when clarified, it is not obvious that this group of intuitions are mutually consistent, and I will argue that at least two of them are misguided. Nonetheless, a fully developed philosophical account of the concept of evil action must engage with these intuitions and make sense of as many as possible. Here are the eight intuitions: (1) Saying that an action is evil is a means of expressing very strong moral condemnation of that action. (2) Evil actions are morally wrong. (3) The person who performs an evil action is blameworthy and properly held responsible for that action. (4) Evil actions are extreme and never merely trivial. (5) Evil actions are incomprehensible. (6) Evil actions can be banal. (7) There is a psychological hallmark of evil action. (8) Evil actions are qualitatively distinct from ordinary wrongs. In due course I will examine these intuitions in detail, but first it may be helpful to show how these intuitions can be used to categorize the extant philosophical 1   Singer writes: “An evil action is one so bad, so awful, so horrendous that no ordinary decent reasonable human being can conceive of himself (or herself) doing such a thing. And an evil person or organization is one who knowingly performs, wills, or orders such actions, or remains indifferent to them when performed by another in a situation where one could do something to stop or prevent them” (Singer 2004, 196).

Evil Action  35 accounts of evil action. (What follows will be brief but theoretically dense. Readers who are not already familiar with this literature may prefer to skip ahead to Section 2.3, and review this categorization of theories at a later stage.) Accounts of evil action have been offered by Kekes (1990 and 2005), Thomas (1993), McGinn (1997), Garrard (1998 and 2002), de Wijze (2002), Calder (2002, 2003, and 2013), Card (2002), Perrett (2002), Steiner (2002), Neiman (2003), Singer (2004), Morton (2004), Arendt (2006), and Formosa (2008).2 These accounts can be categorized according to which of these eight intuitions are endorsed and which rejected by the philosophers in question. Intuitions (1)–(4) are almost universally accepted in the recent literature, and can be described as forming a common core of intuitions about evil. Admittedly, there are a handful of cases in which intuitions (1)–(4) are either rejected or less-than-wholeheartedly endorsed. For instance, Kekes equivocates on the question of whether all evil actions are morally wrong, and Calder thinks that an action could be evil yet morally right (Calder 2002, 51–2; Kekes 2005, 2, 207). Morton and Steiner agree that all evil actions are wrong, but could be accused of allowing that some merely minor wrongs count as evil (Steiner 2002, 190; Morton 2004, 60; cf. Russell 2007b, 664, 670). Nonetheless, there is a rough consensus that intuition (1) is true, and that intuitions (2)–(4) set out necessary conditions for an action’s being evil. There is also widespread endorsement of intuition (5), that evil actions are incomprehensible, although Singer comes closer to suggesting that incomprehensibility is a necessary condition for an action’s being evil (Singer 2004, 196), whereas others leave open the possibility that incomprehensibility is a common but not an essential feature of evil action (e.g. Neiman 2003, 303; Morton 2004, 14). The major division amongst recent accounts of evil action turns on intuitions (6)–(8). According to Hannah Arendt, the trial of Adolf Eichmann revealed the banality of evil. Card, Calder, Neiman, and Formosa echo Arendt by endorsing intuition (6). Their assertion that evil action can be banal leads them away from intuitions (7) and (8). In denying that evil actions have distinctive motives, Card, Calder, Neiman, and Formosa offer what I will call a psychologically thin account of evil action. In contrast, Kekes, Thomas, McGinn, Garrard, Steiner, Singer, and Morton are committed to the view that evil actions have a distinctive psychological hallmark, and this inclines them towards accepting intuitions (7) and (8) and   Some other recent accounts of evil, notably those offered by Haybron (2002a) and Cole (2006), are focused on evil personhood rather than evil action. As such, I will assess these accounts in later chapters that address the nature of evil personhood. I also leave out of the classificatory scheme some other philosophers who have made certain claims about evil without providing detailed or systematic elaboration, e.g. Rawls (1972), Vargas (2010). 2

36  Evil Action rejecting the claim that evildoing can be banal.3 The main task for advocates of these psychologically thick theories is to identify exactly which psychological feature distinguishes evil actions from other extreme wrongs. Kekes, Thomas, McGinn, Steiner, Singer, and Perrett all claim that the distinctive psychological feature is one that is reasonably familiar from folk psychology: for instance, malice, or sadistic pleasure, or defiance of morality. Garrard and Morton argue instead that evil actions are marked out by a distinctive psychological feature that is comparatively unfamiliar and complex, such as the absence of certain kinds of psychological barriers, or the presence of a certain kind of psychological silencing. The accounts of evil action can grouped as follows: Psychologically thin account

Psychologically thick account: folk versions

Psychologically thick account: non-folk versions

Arendt Card Calder Neiman Formosa

Kekes Thomas McGinn Steiner Perrett Singer

Garrard Morton

This classification of the recent accounts of evil action is fairly coarse-grained, and does not capture every interesting and significant disagreement within this body of literature. For instance, advocates of a psychologically thick folk account of evil action disagree amongst themselves over whether it is malice, sadistic pleasure, defiance, or some combination of these three conditions which marks out evil action. Some accounts of evil, notably that offered by de Wijze, do not fit neatly into one of the three categories that I have listed. Nonetheless, I hope that this basic schema will prove useful as we assess these competing accounts of evil action.

2.3  Evil, Disapproval, and Wrongdoing Although intuitions (1)–(4) are almost universally accepted in the recent philosophical literature on evil, they require clarification and defence. In particular, we

3   Morton explicitly aims to develop a theory that respects the view that evildoing can be banal, while also arguing that evil action is psychologically distinct from ordinary wrongdoing. Thus, Morton probably would not accept that he must choose between intuition (6), on the one hand, and intuitions (7) and (8), on the other.

Evil Action  37 should be explicit about which kinds of action will be ruled out of the category of evil if we endorse intuitions (1)–(4). According to intuition (1), saying that an action is evil is a means of expressing very strong moral condemnation of that action. This intuition is important not merely because it is widely shared, but because it can be called upon as part of a response to those who doubt the usefulness of the concept of evil. The concept of evil is an essential part of contemporary moral thinking, we might argue, because of its unparalleled ability to capture and to convey our wholehearted condemnation of an action. When Christopher Hitchens says that “evil” is “the best negative superlative that we possess”, it is plausible that he believes that it is the best because it is the most emotionally forceful negative evaluation available to us (Hitchens 2002). Some philosophers share this sentiment. Singer claims that “evil” is our “worst term of opprobrium” (Singer 2004, 190), and Garrard agrees: The terrible massacres of the twentieth century, the hideous and endless ingenuity of its tortures, seem to require description in terms of evil . . . because other kinds of moral condemnation do not capture their nightmarish horror. (Garrard 2002, 321)

Given the obvious expressive power of the claim that an action is evil, we might be tempted to think that “evil” is a purely expressive word, and that saying that an action is evil is equivalent to giving a loud and sustained “Boo!” in its direction. According to the psychiatrist Michael Stone, when we say “That was evil!” we are not “putting forth some philosophical comment”. Rather, we are “expressing an emotion”. Stone continues: “Evil, in everyday speech, means something to the effect: ‘I feel a horror beyond my ability to understand, beyond my ability to put what I feel into words.’ ” (Stone 2009, 23) Yet it would be a mistake to believe that we could reduce the meaning of “That was evil!” to a mere expression of horror or strong disapproval of that action. One reason that such a reduction would be mistaken is that there are various non-moral grounds for strong disapproval of an action—that it is impolite, that it is imprudent, that it is disgusting—and we do not express these other kinds of disapproval by calling the action in question evil.4 It is specifically moral disapproval, or disapproval on moral grounds, that commonly is linked to the judgement that an action is evil. It also would be a mistake to think that there is a necessary connection between judging that an action is evil and strongly disapproving of that action, or that there is a necessary connection between saying that an action is evil and communicating disapproval of that action. Some people are amoralists who simply do not care one way or the other about moral considerations (Brink 1989, 48; Morton 2004, 52;   There can be a morally loaded kind of disgust, as when someone finds paedophilia disgusting, but there are also non-moral kinds of disgust, such as when someone find rotting food or sentimental kitsch disgusting. 4

38  Evil Action Nichols 2004, 72; Stone 2009, 24). Even if amoralists can recognize what is morally right and what is morally wrong, they are not moved to act in accordance with their own moral judgements. An amoralist who judges that an action is evil need not disapprove of that action. If we also admit the possibility of principled anti-moralists, who make it a rule to perform actions because those actions are morally wrong, then we can imagine someone saying that torture is evil and thereby conveying approval of torture (Joyce 2001, 20). For instance, consider the character of Satan in Milton’s poem Paradise Lost, or that of Dr Evil from the Austen Powers films. When these villains describe certain actions as evil, we assume that they are in favour of the actions in question (Milton 1993, 123 and 251 (Bk I, 165 and Bk IV, 110); cf. Augustine 1951, 24 (Confessions Bk 2, IV)). The fact that a person calls an action evil does not imply that she is booing and giving it the thumbs down. Any link between my saying that an action is evil and my expressing moral disapproval of that action seems to depend on my commitment to morality. In this respect, the claim that an action is evil is similar to many other moral claims. My saying that an action is harmful, or that an action is a lie, or that an action is a moral atrocity would express my disapproval of that action so long as I want people not to be harmed, and so long as I value honesty and morality. While each of these negative judgements— that an action is harmful, that an action is a lie, and that an action is an atrocity—is typically accompanied by some kind of disapproval of that action, it is important to note that these three moral judgements are not all of the same kind. These types of judgement are worth distinguishing so that we can ascertain exactly which kind of negative moral judgement we make when we think that an action is evil. Ethicists distinguish evaluative judgements from deontic judgements. Evaluative judgements are judgements to the effect that something is good or bad, better or worse. The evaluative domain also includes aretaic judgements, which are expressed in the language of virtue and vice, such as that an action is cowardly or courageous, generous, or selfish.5 Deontic judgements are judgements to the effect that some things are right or wrong, that certain actions ought be performed and that others ought not be performed, or that we have reason to do one thing rather than another. Deontic judgements, broadly speaking, include both judgements about pro tanto reasons for acting in certain ways, and judgements about what we have conclusive reasons to do or to refrain from doing.6 (It is important to note that not every philosopher who makes deontic judgements is an advocate of   Judgements of this form (e.g. that an action is courageous) appear not to be purely evaluative, in that they also include descriptive judgements about motives. 6  A pro tanto reason counts in favour of an action, but might be outweighed by countervailing reasons against performing that action. For example, imagine that you must decide whether to grant an extension to a distressed student who has failed to get her work in on time, but who has no legitimate 5

Evil Action  39 so-called deontological ethics.) In everyday moral discourse there is not always a clear distinction drawn between evaluative and deontic claims. Within philosophical ethics, where such distinctions proliferate, there is much disagreement over the relationship between the evaluative and the deontic. Consequentialists claim that the evaluative is prior to the deontic; roughly, that we ought to do whatever will maximize the good. Deontologists and virtue ethicists maintain that we cannot determine which actions are right merely by finding out which would maximize the good. Further disagreements cut across this broad divide. Many of these disputes are irrelevant or peripheral to the question of the nature of evil action, but it will prove useful to ask where the concept of evil action sits in relation to some of the main categories within this area. Specifically, we ought to ask whether the judgement that an action is evil is an evaluative judgement that leaves open the question of whether that action is wrong, or whether it includes the deontic judgement that it is wrong. Let us return to three other kinds of negative moral judgement: that an action is harmful, that an action is a lie, and that an action is an atrocity. The judgement that an action is harmful includes the judgement that the action is bad in some respect for the person who is harmed, but it does not include the judgement that the action is bad overall, much less the judgement that the action is wrong. For instance, the imposition of a well-justified punishment might be harmful to the person who is punished without thereby counting as a wrong action. In believing that an action harms someone we make a limited kind of negative evaluative judgement, but this does not commit us to a deontic judgement about what morally we have most reason to do or not to do.7 The case of dishonesty is more complex. Some people claim that the act of lying can be captured in purely descriptive terms. For instance, we might define a lie as the act in which S tells an audience something that S believes to be false, where S intends through this act to mislead the audience. If this is what it means to say that an action is a lie, then we should conclude that the judgement that an action is a lie does not include the judgement that the action is morally wrong. This outcome fits neatly with the consequentialist view that it is morally right to tell a lie when doing so would produce sufficiently good consequences, and morally wrong to excuse. The fact that granting an extension would ease the student’s distress is a pro tanto reason for doing so, even if it is outweighed by the fact that it would unfair to grant the extension. In a case in which you have a pro tanto reason to perform a particular action, and this reason is not outweighed by countervailing reasons, then you have a conclusive or decisive reason to perform that action. 7   Some consequentialists believe that if an action is harmful overall (that is, more harmful than beneficial) then it is morally wrong. Nonetheless, the judgement that it is harmful overall still falls short of the deontic judgement that it is wrong. A consequentialist and a non-consequentialist might agree that a particular action is harmful overall but disagree over whether it is wrong.

40  Evil Action tell a lie when doing so would produce sufficiently bad consequences. In contrast, some people believe that the judgement that an action is a lie necessarily is a negative moral judgement. It could be that if an action is a lie, then we have a defeasible or pro tanto reason against performing that action, or even that it is never morally permissible to lie (Kant 1994, 280). Those who think that lying is never permissible might still accept the evaluatively neutral definition of lying, and claim that the evaluatively neutral judgement that an action is a lie always morally ought to be accompanied by a separate judgement that the action is morally wrong. Alternatively, they might argue that the judgement that an action is a lie includes the judgement that the action is morally wrong, and hence that lying ought to be defined as a kind of wrongful deception, or as a kind of deception that violates the moral rights of the deceived party (Kemp and Sullivan 1993, 153). We have seen that the judgement that an action is harmful uncontentiously falls short of being a deontic judgement to the effect that the action is wrong, but that there is philosophical dispute over how close the ties are between the judgement that an action is a lie and the judgement that the action is wrong. In contrast to both of these cases, it is obvious that the judgement that an action is an atrocity includes the judgement that the action is morally wrong (Card 2002, 9). While it is contentious but coherent to claim that it is morally permissible to tell a lie, it would be incoherent to say that an action is an atrocity but is morally permissible. When we judge that an action is evil, we are making some kind of negative moral judgement about that action, but which kind? Is it the evaluative judgement that the action is bad in some respect, or the deontic judgement that we have a pro tanto but not necessarily conclusive reason not to perform the action? Or is it the deontic judgement that the action is morally wrong; that is, that there are conclusive or decisive moral reasons that count against performing that action? I will argue for the conclusion that judging that an action is evil includes the conclusive deontic judgement that it is morally wrong. This conclusion may seem incontrovertible, yet at least two philosophers have suggested that not all evil actions are morally wrong. One way to motivate the view that it could sometimes be morally permissible to do evil is to note that we sometimes describe moral dilemmas as cases in which we are forced to choose between the lesser of two evils. Someone who faces a dilemma of this form should choose the better of the two options, so there is a sense in which this person morally ought do something that is evil. Were we to accept this claim, it seems we should conclude that all evil actions are morally bad in some respect, and perhaps that an action’s being evil indicates that we have a pro tanto reason not to perform that action, but that sometimes it is morally permissible to perform an

Evil Action  41 evil action. John Kekes appears to be following this line of thought when he says that in some cases “exceptional circumstances demand that evil be done in order to prevent even greater evil”, and that evildoers ought not be held morally responsible when their evil actions are morally justified (Kekes 2005, 207). In the documentary film The Fog of War, former United States Secretary of Defense Robert McNamara expresses a similar thought when he asks “How much evil must we do in order to do good?” Suggesting that it could be morally right to do evil is not merely contentious, but confusing. It certainly is coherent to claim that an action is harmful but permissible, and arguably coherent to claim that an action is a lie but is permissible, but to many of us it sounds contradictory to say that an action is evil but morally permissible. Elsewhere Kekes himself declares that evil actions are “worse than morally bad”, and that they lack a “morally acceptable excuse” (Kekes 2005, 2). How could an action be inexcusable and worse than bad, but not wrong? I think that Kekes’ claim that sometimes we ought to do evil to prevent greater evil is best understood in light of the ambiguity in the word “evil”. In Chapter One, I claimed that “evil” can be used to express the broader concept of bad, or the narrower and more extreme moral concept of evil. Kekes’ claim that some evil actions are morally justified could be due to the fact that, at this point in his argument, he has conflated these two uses of the word “evil”, and is merely suggesting that some actions which are bad in significant respects might nonetheless be morally justified in rare circumstances (cf. Card 2002, 17). Todd Calder raises a more significant challenge to the view that an action can be evil only if it is morally wrong. He describes a case that he calls Malicious Hirer, in which a person performs an action that Calder takes to be evil, but that would not count as morally wrong according to some plausible versions of consequentialism. In Calder’s example, person A, who works for a charity, can choose to hire one of two equally qualified and deserving candidates, B and C. A knows that C is a celebrity, and that hiring C rather than B will ultimately increase donations to the charity and produce the best consequences overall, even though it will leave B suffering from extremely intense and sustained depression. If A hires B instead of C, then neither B nor C would suffer depression, but the charity would not secure increased donations, and the overall good would not be maximized. According to many versions of consequentialism, in this situation it is morally right for A to hire C rather than B, because it is clear that doing so would maximize good consequences. But Calder imagines a case in which A hires C without any concern for the good consequences that would flow from this action. Person A “cares only about getting pleasure from the suffering of others”, and the reason A hires C is “to

42  Evil Action take pleasure in the suffering this will cause B” (Calder 2013, 184). Calder thinks that it is plausible that the malicious hirer’s action is morally right but is also evil.8 The malicious hirer could be described as having done the right thing for the wrong reason. In cases of this kind it is very difficult to disentangle our evaluations of act, motive, feeling, and person. I think that consequentialists should respond to Calder’s example by accepting that A’s hiring of C is morally right, but denying that it is an evil action. What gives the illusion that this is an evil action is the fact that A takes pleasure in B’s suffering. Calder himself accepts that in some circumstances the act of taking pleasure in witnessing another person’s intense suffering is an evil action (Calder 2002, 56) and, in Chapter Nine, I will argue for a similar conclusion. While some austere versions of consequentialism might imply that there is nothing intrinsically morally bad in taking pleasure in the undeserved suffering of another, more plausible versions of consequentialism correctly describe this act as morally vicious (Hurka 2001, 16; Russell 2007a, 486). Thus I think Calder is right to say that the malicious hirer does something evil in this situation, and consequentialists should agree with him.9 Calder’s mistake, I contend, is to think that it is A’s hiring of C that is the evil action. Rather, it is A’s taking pleasure in C’s suffering that is the evil act. Consequentialists believe that A should hire C. If A was not around to do the job, consequentialists would recommend that someone else step in and hire C. This is not what we would say about an action that we condemn as evil. Unlike Calder, I think that an action cannot be both morally permissible and evil. If an action is evil, then it is not merely badly motivated or deplorable in some way, but also is morally wrong, and ought not be done. It does not make sense for someone who cares about morality to say “What you are about to do is evil, but please go ahead and do it”, or even “It’s evil, but in this case it is ok for you to do it”. The view that all evil actions are morally wrong is reflected in folk claims about evil. For instance, the psychiatrist Michael Stone, who proposes a scale of evil, claims that justified homicide is not evil at all (Stone 2009, 33). Almost all contemporary philosophical accounts of evil action also rule out the possibility of morally justified evil actions. Claudia Card claims that we ought to restrict our definition of evil action so as to include only wrong actions (Card 2002, 17). Stephen de Wijze   For an earlier and different example of roughly the same form, see Calder (2002, 52). Calder notes that Kantians would judge that the action of the malicious hirer would be morally wrong because the hirer fails to respect the humanity of B, and uses B merely as a means for pleasure. We should also note that some virtue ethicists would judge that the malicious hirer performs a wrong action. 9   For sake of the argument at hand, I set aside the question of the whether the suffering that is connected to the malicious hirer’s action—namely, B’s depression—is sufficiently extreme for the act of taking pleasure in that suffering to count as evil. Calder could easily construct an example in which the suffering was more extreme. I discuss the issue of extremity in Chapter Three, and the issue of sadistic pleasure in Chapter Nine. 8

Evil Action  43 concurs, suggesting that if the infliction of great harm, such as dropping an atomic bomb on a city, were morally justified, then it would not belong in the category of evil action (de Wijze 2002, 224). In general, other contemporary philosophers agree that all evil actions are morally wrong (e.g. Thomas 1993, 74; Joyce 2001, 43; Morton 2004, 9; Singer 2004, 193). This weight of opinion suggests that the judgement that an action is evil, like the judgement that an action is an atrocity, includes the conclusive deontic judgement that the action is morally wrong. Part of what it means to say that an action is evil is that it is morally wrong.10 The fact that all evil actions are moral wrongs helps us to explain why we can use the term “evil” to express moral disapproval even though “evil” is not a purely expressive word like “boo”. Regardless of whether we accept motivational internalism or motivational externalism about moral judgement (Brink 1989, 37–42; Smith 1994, 71–6), the prevalence of moral behaviour and the effectiveness of moral censure indicates that most people have some kind of motivational commitment to doing what they judge to be morally right. In other words, most people are not amoralists or anti-moralists. This allows us to assume that most people who say that an action is evil also disapprove on moral grounds of the performance of that action, and that they believe that there was a morally preferable alternative action available. Unless we have reason to believe that the speaker is an amoralist or an anti-moralist, the speaker’s claim that an action is evil can express her strong disapproval of the action, and will be heard by others as communicating this attitude. The emotional punch of her claim than an action is evil depends on her commitment to morality.

2.4  Extreme Culpable Wrongdoing All evil actions are morally wrong actions, but are all morally wrong actions evil actions? According to the fairly broad meaning of “morally wrong” that I have been using so far, an action is morally wrong if and only if there are conclusive or decisive moral reasons against performing that action. To put the same thought in more prosaic terms, an action is morally wrong if and only if it is something that we morally ought not do. However, there is also a narrower sense in which we might judge that an action is morally wrong; namely, that it is a culpable moral wrong, a moral wrong for which the agent could properly be held responsible.

  This falls short of the claim that the agent who performs an evil action must also judge it to be morally wrong. That is a separate and much more contentious claim, which we will consider in Chapter Four. 10

44  Evil Action There are several kinds of case in which an action can be described as being wrong but not culpably wrong. Sometimes people accidentally perform actions that harm others, as when an unlucky driver skids off the road and hits a pedestrian. In cases in which the person who accidentally performs the harmful action does so with culpable negligence, we hold that person morally responsible for inflicting the harm. In cases in which the agent was not being reckless or unduly careless, we do not hold the agent morally responsible for the harm that has been inflicted. Even though we might say that what that unlucky driver did was wrong, in the broader sense that there are decisive moral reasons against hitting pedestrians with your car, we excuse the unlucky driver who accidentally but non-negligently inflicted this harm. Similarly, sometimes people deliberately perform actions that unforeseeably turn out to have bad consequences. For instance, a doctor might vaccinate a child who subsequently is debilitated by the vaccine. In such cases there is sense in which the doctor has chosen incorrectly and has done the wrong thing, although she was not to know this at the time of choosing. Yet we should not hold people morally responsible for unforeseeably tragic mistakes. Nor should we hold people morally responsible for harming others when they inflict those harms only because they are subject to extreme threats or coercion. In such cases we might even say that the agents did not freely choose to do what they did. While their actions were morally wrong in the broader sense that there are decisive moral reasons against harming innocent people, the coerced wrongdoers have a non-justifying excuse for having acted as they did, and hence they are not morally culpable for having performed those actions. I think the judgement that an action is evil includes the judgement that the action is culpably wrong. If an action is evil then there are decisive moral reasons against performing that action, and the agent can properly be held morally responsible for having performed that action. An action is evil only if it is a moral discredit to the agent, and the agent is blameworthy for having performed that action. If a terribly harmful action is performed by non-negligent accident, then we would say that it was a tragic event, but not that it is an evil action. If someone performed a terribly harmful action but has a good excuse for having done so, then it would be unfair and misleading to say that her action was evil (cf. Calder 2013, 187). This condition is built into many of the recent philosophical accounts of the concept of evil. For instance, Morton believes that harmful actions that result from mere “incompetence or miscalculation” do not count as evil (Morton 2004, 16, 61). Singer claims that through “accident or misadventure one can do something wrong or bad, even terrible, but not something evil” (Singer 2004, 190). Kekes agrees, maintaining that, if an action is evil, it must lack an excuse (Kekes 2005, 1; but cf. Kekes 2005, 207).

Evil Action  45 There are two potential challenges to the view that all evil actions are culpable wrongs. The first concerns cases in which extremely difficult circumstances, such as those faced by prisoners in Nazi concentration camps, have pushed ordinary people to perform terrible actions (Levi 1989, 40–8). In such cases the victims of oppression can themselves, in Card’s carefully qualified phrase, become “implicated in the perpetration of . . . evil on others” by cooperating with their oppressors (Card 2002, 224). These might seem to be cases in which certain actions are evil but are excusable rather than culpable and blameworthy. The second possible challenge to the view that all evil actions are culpable wrongs concerns cases in which the wrongdoer has some kind of illness or incapacity which some people think should count as an excuse for his wrong actions. For instance, Manuel Vargas has argued that psychopaths do evil, but that the nature of the psychopaths’ emotional deficit means that they could not have done otherwise, and hence that they ought not be blamed for what they do (Vargas 2010, 74–6). It is unclear whether we should judge that the actions performed by desperate prisoners or by psychopaths count as evil. Arguably, though, our intuitions about evil action are unclear here precisely because in these situations it is hard to discern the facts about agency and responsibility. This serves to support rather than undermine the claim that an action is evil only if it is a culpable wrong that is a moral discredit to the agent. It is plausible that the concentration camp guards were acting through the coerced prisoners, and it is comparatively uncontroversial to judge that the deliberate exertion of pressure by the guards to co-opt others into extreme wrongdoing was, as Card puts it, “unambiguously evil” (Card 2002, 234). If we accept that the guards are acting through the prisoners, and if we believe that the prisoners are not blameworthy for what they do, then it is reasonable to maintain that the prisoners’ actions were not evil.11 Our judgements about the status of the terribly harmful actions performed by some psychopaths are similarly contentious and unclear. What looks to some to be blameworthy bad character in the psychopath seems to others to be an exculpating inability to act otherwise (Duff 1977, 193; Calder 2002, 58). Yet in these cases, too, it is reasonable to believe that an action can be evil only if it is a culpable wrong that is a moral discredit to the agent. After all, we do not judge that great white sharks or grizzly bears which kill and eat humans are thereby doing evil, or even that the lorry driver who accidentally kills a pedestrian, in Bernard   It is possible that some prisoners in this situation would describe themselves as having being forced to do evil, and might feel terrible about having done these things. According to my view, if these prisoners are actually blameless, then their own judgement that their actions were evil is incorrect; they are being too hard on themselves. But in those cases in which the prisoners are blameworthy, it is possible that their actions were evil. 11

46  Evil Action Williams’ example, is an evildoer (Williams 1981, 28). We might judge that the emotional deficits possessed by psychopaths diminish but do not extinguish their responsibility, acting as a mitigating factor but not a total excuse for their actions. In this case we could maintain both that an action is evil only if it is a culpable wrong and that some psychopaths perform evil actions for which they are culpable. If we were convinced that psychopaths really are not blameworthy for any of their harmful actions, then we should say that psychopaths are dangerous human beings, but are not evildoers. In any case, it seems fair to conclude that there are no clear counterexamples that would undermine the claim that every evil action is a culpable moral wrong. While it is true that all evil actions are culpable wrongs, it is not the case that all culpable wrongs are evil actions. In Chapter One, we saw that, unlike the broader concept of bad, the concept of evil is meant to mark out some kind of extremity. Shoplifting, lying in order to get out of jury duty, and cheating on a school exam are all culpable wrongs, but calling them evil actions would be hyperbolic overstatement. If evil actions were not necessarily extreme wrongs, then it would be hard to make sense of the expression “not merely wrong, but evil”. Similarly, the heated debates over whether terrorist bombings or war crimes are evil actions can be understood only if we agree that evils are more than minor wrongs. Almost all contemporary accounts of evil action build in an extremity condition that is designed to respect these intuitions. For instance, Morton equates evil actions with atrocities (Morton 2004, 13), and Card takes atrocities to be paradigms of evil (Card 2002, 8). Card writes: “Evil” is a heavy judgment. Much that is bad is disappointing, undesirable, inferior, even unjust or unfair, but not evil. Many wrongdoings are trivial. Evils never are, even if their perpetrators are ordinary people and their motives not unusual. (Card 2002, 7)

Kekes maintains that evil goes “beyond badness” (Kekes 2005, 1), Formosa claims that evil actions inflict extreme, life-wrecking harm (Formosa 2008, 228), Singer says that evil actions are “horrendously wrong” (Singer 2004, 193), and Laurence Thomas suggests that an evil act “must have a certain gravity to it” (Thomas 1993, 77). According to Thomas, evils sit at the opposite end of the scale from morally heroic supererogatory acts (Thomas 1993, 77). In accordance with these accounts, I conclude that it is impossible for there to be an evil action that is merely trivial or insignificant. An action can be evil only if it is morally extreme. Some philosophers have offered theories that arguably imply that not all evil actions are morally extreme. For instance, both Hillel Steiner and Colin McGinn have claimed that evil actions are wrong actions that are pleasurable to

Evil Action  47 the wrongdoer (McGinn 1997, 62; Steiner 2002, 189). On Steiner’s and McGinn’s accounts, ordinary, non-evil wrongs are those in which the wrongdoer does not take pleasure. Steiner claims that, because the scale for measuring evil is a “compound scale” that combines degree of wrongness with amount of pleasure taken in the wrong, there can be highly evil acts that are “only marginally” wrong (Steiner 2002, 190). Steiner is right to suggest that pleasure taken in a morally wrong action makes that action morally worse, but it is implausible that even the most intense pleasure taken in certain kinds of minor wrong, such as shoplifting or exaggerating deductions on a tax return, could elevate those actions to the status of evil. Steiner’s and McGinn’s accounts clash with the intuition that all evil actions are extreme and never merely trivial, and this gives us a good reason to reject those accounts as they stand. Morton propounds a more complex theory that arguably also clashes with intuition (4). He claims that A person’s act is evil when it results from a strategy or learned procedure which allows that person’s deliberations . . . not to be inhibited by barriers against considering harming or humiliating others that ought to have been in place. (Morton 2004, 57)

On Morton’s account, ordinary, non-evil culpable wrongs are those that do not result from the breaching of such barriers. Morton points out that his account allows that there could be evil actions that are not very harmful (Morton 2004, 60). I think that Morton’s account further implies that there could be evil actions that are merely minor or trivial wrongs, on the assumption that there ought to be psychological barriers in place against considering inflicting gratuitous but minor harms. I will assess Morton’s account in more detail in Chapter Five. For now, let us note that if Morton’s theory of evil action clashes with intuition (4), then it too ought to be rejected or modified. In this chapter, I have argued that intuitions (2)–(4) set out necessary conditions for evil action. In other words, an action is evil only if it is an extreme and culpable moral wrong. The fact that evils are extreme culpable wrongs allows us to explain the sense in which we can endorse intuition (1). If I say that an action is evil, I am implying that it is an extreme culpable wrong. This claim, when combined with my public commitment to morality, expresses my disapproval and condemnation of the action in question. Given that my judgement is that the action is not only wrong but extreme, the disapproval that I express is strong rather than weak. However, the judgement that an action is evil would not imply condemnation or disapproval on the part of amoralists and anti-moralists, who lack a commitment to morality. Intuitions (1)–(4) are not only endorsed in virtually all recent philosophical accounts of evil action, but also accord with widespread folk judgements

48  Evil Action concerning evil action. Any account of evil action that fails to respect intuitions (1)–(4) will for that reason seem to be implausible. While I have argued that intuitions (1)–(4) constitute a relatively uncontentious common ground in recent philosophical disputes about evil action, further investigation of this domain in required. As we shall see in the next chapter, the claim that evil actions are extreme and never trivial is widely accepted, but it is unclear how best to characterize the extremity that is required for evil action, and also how best to interpret the claim that evil actions are incomprehensible.

3 Evil, Extremity, and Incomprehensibility While there is broad agreement in recent philosophical literature that all evil actions are extreme culpable wrongs, it is not clear in which respects an evil action must be extreme. Nor is it clear exactly where the boundary lies between ordinary wrongs and extreme wrongs. Some philosophers have claimed that an evil action must inflict an extreme amount of harm, but I will argue for the contrary view that there can be harmless evil actions, such as harmless failed attempts to bomb innocent civilians. Nonetheless, these harmless evils must be extreme and never minor or trivial. It might be thought that we could capture the kind of extremity required for evil actions by pointing to their incomprehensibility. After having compared several kinds of incomprehensibility, I will argue that we do not have good reason to believe that incomprehensibility is a necessary or a sufficient condition for evil action. The correct view, I will claim, is that evils are extreme in so far as they are connected in one of an identifiable set of ways to actual or possible extreme harms. This leaves unanswered the question of exactly how extreme these harms have to be for the action to count as evil. While some philosophers have suggested that the relevant harms must be so extreme as to threaten the autonomy of the victim, or to be life-wrecking, I will claim that neither of these conditions help us to locate the boundary between evils and ordinary wrongs.

3.1  What Kind of Extremity? Evil actions cannot be minor or trivial, but instead must be extreme. Yet there are a great many respects in which an action could be judged to be extreme, and it is not obvious exactly which kind of extremity is a necessary condition for an action’s counting as evil. A simple answer to this question is that evil actions must be extremely wrong. While this is an appealing thought, we need to explain what it would mean to say that an action is extremely wrong. I have claimed that an action

50  Evil, Extremity, and Incomprehensibility is morally wrong if and only if there are decisive or conclusive moral reasons against performing that action. This suggests that an action is either wrong or not wrong, but cannot be more wrong or less wrong. Perhaps the claim that one action is more wrong than another is shorthand for a more complex thought: that the first action warrants greater censure or punishment than the second, or that the first action is more vicious or deplorable than the second, or that the reasons for which the first action is wrong are more important than the reasons for which the second action is wrong, and hence, were you forced to choose which of these actions to perform, you should choose the second over the first. Even if we accepted one of these accounts of the degrees of wrongness, we would need to say more about why one action is more wrong—more severely punishable, less choiceworthy—than another. Thus we ought to consider the various dimensions on which we could rank wrong actions, with the hope of identifying the kind of extremity that is relevant to the category of evil action. There are numerous dimensions along which we could evaluate and rank wrong actions. We can distinguish actions that are obviously wrong from actions that are wrong for more subtle and abstruse reasons. We can distinguish wrongs that inflict harm from those that do not, wrongs that were intended to inflict harm from those that were not, and wrongs that foreseeably would inflict harm from those that foreseeably would not. We can sort wrongs that lump their harms, so to speak, on a limited number of victims from wrongs that spread their harms very thinly amongst many victims. We can also distinguish and rank wrongs via the type of harm that they inflict, are intended to inflict, or foreseeably would inflict. Some harms are physical pains. Others are constituted by the frustration of desires. Some harms involve the destruction of life, of mental or physical capabilities, or of property. Others are episodes of humiliation or shame, or the removal of valuable possibilities from the lives of the victims. Within each type of harm we can sort the less extreme from the more extreme cases: irritating pains from searing pains, minor scrapes from maimings, and fleeting blushes from episodes of profound and alienating shame. In explaining and defending intuition (4)—that evil actions are extreme and never trivial—we must specify the relevant dimension or dimensions along which evil actions occupy an extreme position. Some kinds of extremity clearly are not necessary or sufficient for evil action. For instance, many extremely obviously wrong actions are not evil actions, because they remain morally trivial. It is also plausible that some actions might be evil even though they are wrong for reasons that, to some people, at least, are not extremely obvious. A callous bureaucrat who issues murderous orders could plausibly be described as doing evil even if his ideological convictions make it hard for him to see why his actions

Evil, Extremity, and Incomprehensibility  51 are morally wrong. Evils are culpable wrongs, so the wrongness of an evil action cannot be so hard to discern as to render the action excusable, but it is not the case that all evil actions stand out from ordinary wrongs in virtue of being extremely obviously wrong. Another possibility is that evil actions must inflict harm of a narrowly circumscribed extreme type in order to count as evil. Many of the paradigmatic examples of actions that have been called evil are instances of murder, in which the harm inflicted is undeserved and unwanted death. This might lead us to propose that all evil actions are extreme in that they inflict undeserved and unwanted death. The problem with this suggestion is that many of us judge that some actions which do not cause death are evil nonetheless. The murder and the torture and the acts of cruel debasement of prisoners that took place in Abu Ghraib prison seem to be candidates for evil actions. Some actions that deny people valuable freedoms might be described as evil as well. Josef Fritzl’s imprisoning of his daughter in the basement for 24 years, for example, strikes many people as being an evil act, as do the buying and selling of slaves. Admittedly, it is much less common to see cases of destruction of property described as evil, except where such destruction is part of a campaign of terror, or is likely to lead to other kinds of extreme suffering, such as starvation (Stone 2009, 286; Zaibert 2012, 95). Nonetheless, the types of harm that are inflicted by evil actions can be quite varied. It seems that what these examples of evil actions have in common is not that they all inflict one type of harm, but that each inflicts great harm. Perhaps we should say that all evil actions are extreme wrongs in the sense that each inflicts an extreme amount of harm. In developing this line of thought, we must be careful to specify the relevant sense in which an action is extremely harmful. Some actions could be said to inflict an extreme amount of harm in that they inflict harm on a very large number of victims, but inflict no more than a minor harm on any single victim. A credit card scammer who took fifty cents from each of the respective bank accounts of two million people has stolen a million dollars, but has inflicted a very small harm on any one of her victims. Similarly, we can imagine someone who inflicts a very small amount of physical pain or humiliation on a very large number of people, such that the total amount of suffering he has inflicted is, in one sense, enormous. Yet someone who deliberately and maliciously inflicts a small amount of suffering on each of a very large number of people does not thereby do something evil. Thus we might suggest that the kind of extremity that is necessary for evil action is the infliction of an amount of harm that is sufficiently extreme for at least one victim, rather than extreme merely in aggregate.

52  Evil, Extremity, and Incomprehensibility Another case in which an extremely harmful culpable wrong would fall short of being an evil action is when the wrongdoer is properly held responsible for the wrong action, but ought not be held responsible for the extreme harm that results. This occurs in cases of what I will call unforeseeably calamitous culpable wrongdoing. For instance, let us imagine that an angry young man maliciously insults a bus driver, and by chance this insult distracts the driver at a crucial moment and causes the bus to crash, killing many people. The angry young man ought to be blamed and held responsible for insulting the bus driver, so the action is a culpable wrong. Yet the fact that the insult turned out to be an extremely harmful wrong action was neither intended nor reasonably foreseeable, so it would be inappropriate to hold the wrongdoer responsible for the extreme harm that resulted, and certainly inappropriate to call the insult an evil action (cf. Card 2002, 20). In light of these examples it might be suggested that if an action is evil then it is an extremely harmful culpable wrong in which the harm inflicted is extreme for at least one victim, and in which the wrongdoer is properly held responsible not merely for having performed the wrong action, but also for having inflicted this extreme harm. Some philosophers have offered accounts of evil action that endorse this view, or something very close to it. Singer states that “evil acts are acts that are horrendously wrong, that cause immense suffering” (Singer 2004, 193). Card claims that an evil intention is the “culpable intention to do someone intolerable harm, or to do something with that foreseeable result”, and in “an evil deed, the intention succeeds” (Card 2002, 20). Formosa offers what perhaps is the most detailed version of this kind of account. He claims that “an evil act is one where an agent perpetrates a moral wrong that makes him at least partly responsible for the harm others suffer” (Formosa 2008, 228), and specifies that the level of harm is sufficiently extreme so as to “normally have a life-wrecking impact on the victim” (Formosa 2008, 230). These claims are highly contentious, not because they imply that all evil actions are extreme wrongs, but because they imply that all evil actions actually inflict harm. Thomas seems to share this view, stating that an evil act, at minimum, is a wrong act that is harmful to another creature (Thomas 1993, 74). The problem with such claims is that some wrong actions that do not inflict any harm seem to be as morally deplorable as actions that do inflict extreme harm. Many carefully planned culpably wrong actions that are intended to inflict extreme harm actually turn out to be relatively harmless because the wrongdoers are foiled by the authorities, or because there is a malfunction that prevents a bomb from detonating, or because their intended victims are much better equipped to defend themselves than the wrongdoer had realised. Real-life

Evil, Extremity, and Incomprehensibility  53 examples of such failed attempts to inflict extreme harm, from Guy Fawkes to the so-called shoe bomber Richard Reid, are familiar enough.1 The failure to inflict harm seems not to provide any kind of excuse to the wrongdoer in these cases, nor to mitigate the wrong. Even if we were to concede that such comparatively harmless wrongs are never morally equivalent to extremely harmful wrongs, these harmless wrongs still are sufficiently grave and extreme to be labelled evil actions. The most malicious people sometimes try to wreak havoc but fail, and their failed attempts to do so deserve our strongest moral condemnation. For this reason I favour the view that there can be harmless evil actions. We ought to consider in more detail what can be said in favour of the alternative position advocated by Singer, Card, and Formosa; namely, that actually inflicting extreme harm is a necessary condition for evil action. We certainly could choose to speak that way, and perhaps doing so would fit with the way in which some people actually describe failed attempts. For instance, if an attempted suicide bombing was thwarted at the last second by a police officer who stopped the bomber from detonating the device, it could be said that the officer prevented the would-be suicide bomber from doing something evil. Yet it is plausible that we ought to punish people only for what they do, not for what they would have done were they in different circumstances. If we want to explain why the unlucky suicide bomber deserved extreme punishment, many of us would say that it was because he actually did something evil: he attempted to detonate a bomb that would have killed a great number of innocent people. This suggests that the attempt to inflict extreme harm itself could be an evil action, whether it is a successful attempt or not. A second argument for the claim that harmless failed attempts can count as evil actions concerns cases in which a wrongful attempt is made to inflict extreme harm, but there is a large temporal gap between the attempt and the moment at which the action either inflicts or fails to inflict harm. For instance, let us imagine a misanthropic mailer, who sends letters containing lethal anthrax spores to his numerous intended victims, all of whom are innocent. Further, suppose that the crime is discovered and the misanthrope is apprehended after the letters have been mailed, but before any have been received. If it were true that a harmless failed attempt could not be an evil action, then we would have to wait until we were sure   Many failed attempts to inflict massive amounts of harm nonetheless inflict some harm. Richard Reid, for instance, did not succeed in killing anybody, but he did frighten both the passengers in the plane that he tried to destroy, and many other airline passengers who subsequently feared similar attacks. This might lead some people to say that what I am calling failed attempts are not in fact harmless evil actions, because they do inflict harm. Yet we can imagine similar cases in which a failed attempt to inflict massive amounts of harm went undetected, and hence never even produced fear in the general population. I contend that these cases could also be described as evil actions. 1

54  Evil, Extremity, and Incomprehensibility that significant harm—in this case the death of at least some of the intended victims—had been inflicted before we could judge whether the misanthropic mailer’s action was evil. I contend that we would not have to wait. We know that his action was evil even as we wait to see whether it succeeds in inflicting the intended harm. In addition to such harmless failed attempts, there are two other kinds of action that can be offered as examples of harmless evil actions; namely, culpably reckless close shaves and harmless acts of sadistic voyeurism. These cases are more complicated and contentious than harmless failed attempts, so I will delay a full discussion of them until later chapters, and merely note here that these examples might give us further grounds to conclude that not all evil actions are extremely harmful.2 While there is some disagreement regarding the moral status of harmless failed attempts, the intuition that there can be harmless evils is at least fairly widely shared, and is endorsed in some recent philosophical accounts of evil action (e.g. Calder 2002, 56; de Wijze 2002, 234 ft. 38; Kekes 2005, 193). Singer, Card, Formosa, and Thomas say that all evil actions are actually harmful, but they do not provide arguments in favour of that view, and they do not explain why harmless failed attempts could not count as evil actions.3 In light of the examples of the unlucky suicide bomber and the misanthropic mailer, we should conclude that there can be harmless evil actions.

3.2  Incomprehensibility and Evil The challenge at hand is to specify the kind of extremity that is possessed by all evil actions. I  maintain that all evil actions are extreme culpable wrongs, but also that some evil actions are harmless. If I am right about this, the extremity   Culpably reckless close shaves are actions that foreseeably would cause an extreme harm that the wrongdoer does not intend, but that, when performed, happen not to inflict that foreseeable harm. It is uncontroversial that knowingly placing many human lives at risk through culpable recklessness is an extreme wrong, regardless of whether the wrongdoer wants those people to die or not, and regardless of whether the potential victims are actually harmed or not. Some people will judge that such actions can count as harmless evils. As we shall see in later chapters, though, advocates of certain kinds of psychologically thick accounts of evil action judge that all evil actions are malicious or sadistic. Since culpably reckless close shaves are not malicious or sadistic, advocates of thick accounts will conclude that culpably reckless close shaves are extreme wrongs but are not evil actions. Another possible category of harmless evil actions are cases of harmless sadistic voyeurism, in which a voyeur takes great pleasure in witnessing extreme suffering but does not himself inflict or contribute to the infliction of any harm. I will discuss the possibility of evil feelings in Chapter Nine. 3   Calder also says that harmless failed attempts are not evil (2002, 56). “Evil has not been done until someone gets hurt”, he claims (Calder 2003, 367). I have explained why I do not share this view. However, Calder does allow that some harmless acts of sadistic voyeurism are evil (2002, 56), so he cannot think that the evils are extreme in virtue of causing extreme harm. On this issue I agree with Calder. 2

Evil, Extremity, and Incomprehensibility  55 that characterizes evil action cannot be reduced to the extremity of harm that is inflicted by those actions. Perhaps the requisite extremity lies in our psychological reaction when contemplating the wrong, rather than in the harmful effect of the wrong on the victim. The kind of psychological reaction most often mentioned in connection with evil action is a failure to comprehend the action in question. When contemplating the grisly deeds of a serial killer we might shake our heads and say “I just don’t understand”, or “I can’t see how anyone could do such a thing”. Thoughts such as these give rise to intuition (5): that evil actions are incomprehensible. The idea that evil actions are incomprehensible is common amongst the non-philosophical folk. Journalist Lance Morrow declares that “ultimately it is not possible to understand evil” (Morrow 2003, 3). Psychiatrist Michael Stone says that evil actions “have the quality of horrifying or shocking whoever witnesses or hears about these acts”, and that in the case of evil actions the deviation from the moral norm must be “over and beyond what the ordinary people in the community could even envision as something another human being could ever do” (Stone 2009, 21). Psychologist Simon Baron-Cohen thinks that people who judge that an action is evil take it to be incomprehensible: The standard explanation is that the Holocaust . . . is an example of the “evil” that humans are capable of inflicting on one another. Evil is treated as incomprehensible, a topic that cannot be dealt with because the scale of the horror is so great that nothing can convey its enormity. (Baron-Cohen 2011, 4)

Many of the contemporary philosophers who write about evil appear to share this folk intuition, and subsequently draw a connection between evil actions and incomprehensible wrongs. Morton says that the “point of view of the victims of evil is usually that of incomprehension” (Morton 2004, 14). Unlike ordinary wrongdoing, evils evoke “horror, disgust and incomprehension”, according to de Wijze (2002, 213). Singer claims that “an evil action is one so bad, so awful, so horrendous that no ordinary decent reasonable human being can conceive of himself (or herself) doing such a thing” (Singer 2004, 196). Comprehensibility is one dimension along which we can evaluate wrongs, so incomprehensibility counts as one possible kind of extremity in wrong actions. Given that we need to clarify and make sense of intuition (4), that all evils are extreme, perhaps we should claim that all evil actions are incomprehensible culpable wrongs, and that incomprehensibility is the kind of extremity that is a necessary condition for evil actions. If this turns out to be the case, then we will have explained intuition (4) via intuition (5). This account would also make sense of the fact that evil actions must be extreme but can be harmless. Failed attempts at

56  Evil, Extremity, and Incomprehensibility suicide bombing might strike us as extreme and non-trivial in the sense of being incomprehensible, even when they are not extremely harmful. The thought that evil actions are incomprehensible requires clarification. For something to be incomprehensible it must be incapable of being understood, yet this could mean a number of quite different things. I will consider several of these possibilities in turn, before focusing on what I think is the kind of incomprehensibility that is the best candidate for philosophers who wish to connect incomprehensibility to evil action. One possibility is that evil actions are incomprehensible in the sense of being shocking, of failing to fit into our existing understanding of the world, or into our current conceptual scheme. This view appears to be endorsed by Stone, who says that an evil action “will appear incomprehensible, bewildering, beyond the imagination of ordinary people” (Stone 2009, 22). Susan Neiman also endorses something like this view when she claims that “designating something as evil is a way of marking the fact that it shatters our trust in the world” (Neiman 2003, 9). The problem with this suggestion, though, is that it implies that an action could be judged to be evil only while it was still new and shocking. Since it is possible for a jaded pessimist to think that the same old evil actions will keep on being performed, thus confirming her settled view that the world is hopelessly corrupt, we should not say that all evil actions are incomprehensible in the sense of being shocking or hitherto unimagined. Another possibility is that evil actions are incomprehensible in the sense that we are unable to dwell on them or contemplate them for any length of time. Evils, we might say, are wrongs that are so horrific that we are forced to turn our thoughts away from them. While it is true that many people find it too uncomfortable to contemplate evil actions for any length of time, it would be a mistake to suppose that this is a necessary feature of evil action. Victims of evil, including survivors of the Holocaust and of torture at the hands of the military, often think long and hard about the actions that they take to be evil. Scholars of evil similarly dwell on cases of evil action for long periods. Thus we should not say that evils are incomprehensible in the sense of being too terrible to contemplate or examine. A more plausible suggestion is that evils are wrong actions that are incomprehensible in the sense that we are unable to identify their causes or the motives from which they were performed. When we are in the heady context of philosophical reflection we might be tempted to say that we do not fully grasp the causes and motives of any actions, given the complexity of the human brain and the fallibility of introspection. Yet if we set aside these general sceptical worries, there is a clear distinction to be drawn between wrong actions whose causes and motives we can identify and those that we cannot explain. For example, it is obvious that most

Evil, Extremity, and Incomprehensibility  57 cases of financial fraud are motivated by greed, and that some murders of wives by husbands are motivated by sexual jealousy. In contrast, we struggle to explain why the Rwandan genocide occurred, or why the English doctor Harold Shipman murdered more than 200 of his patients. Perhaps evils are those wrongs whose motives and causes are beyond our ken (cf. Arendt 1967, 459). This view is superficially attractive, but when we consider a broader range of examples it is revealed to be indefensible. We can point to the historical and psychological causes of many of the actions that strike us as being evil. I am not denying that we often have trouble explaining atrocities. Rather, I maintain that in some cases in which we see what caused the perpetrators to act as they did, their actions still strike us as evil. When news was first broadcast that a lone gunman, Anders Breivik, had brutally killed scores of innocent teenagers on the Norwegian island of Utøya in 2011, we had good reason to believe that the gunman had done something evil, and we had no idea why he did it. His evil action was, in the sense being discussed, incomprehensible. Yet Breivik’s actions did not remain inexplicable. In his online manifesto and during his subsequent trial Breivik detailed his motives and laid out his twisted rationale for all to see. If judging that an action is evil really did include the judgement that it is inexplicable, then we ought to revise our initial belief that Breivik did something evil. Yet this is not what we do. We still judge that Breivik did evil, and now we see what motivated him to do it (Russell 2012, 66).

3.3  I Can’t Imagine Doing That Many people would say that they can see why Breivik did what he did, but that they find his actions incomprehensible nonetheless. Perhaps the most plausible way to understand this claim is to take incomprehensible actions to be those that we cannot imagine performing ourselves, even if we were in the same situation as the agent (Morton 2004, 21). This version of the claim that all evil actions are incomprehensible is attractive in so far as it fits with many folk claims about evil (e.g. Stone 2009, 21). The murderous actions of war criminals and serial killers might strike us as being incomprehensible in this sense, even if we can identify their causes and the reasons for which they were performed. Moreover, some harmless but extreme wrongs, such as failed attempts at suicide bombing, are incomprehensible in this sense. Perhaps the requisite extremity in evil actions could be explained by the fact that all evils are wrongs that we cannot imagine performing ourselves. There are several distinct phenomena that could be described as an inability to imagine performing an action. One is not being able to visualise or mentally rehearse the performance of an action. In some cases we might lack this ability

58  Evil, Extremity, and Incomprehensibility because of simple ignorance about the nature of the action. For instance, I cannot imagine myself recalibrating a Gieger counter because I know nothing about how Gieger counters work. Presumably this kind of imaginative failure is not what people have in mind when they call evil actions incomprehensible. Mass murder and torture are not tasks that require complex practical knowledge. In other cases I might be unable to picture myself performing an action because the imaginative act is so distressing or horrific that I cannot even get it started. It is more plausible to suggest that this is the kind of incomprehensibility that is the hallmark of evil action. However, it is possible that there are some actions that I can visualise myself performing, but that I would still say that I could not imagine doing myself. I can imagine just what it would feel like to jump out of a plane, for example, and the vivid horror that I experience during this imaginative act might be exactly what leads me to say that I could never imagine jumping out of a plane myself. In this case it seems that an action is incomprehensible to me in the sense that I believe that I would never freely choose to perform that action under any circumstances, whether I can visualise doing it or not. I think that this is probably what the majority of people mean when they say that evil actions are incomprehensible. Yet if we accepted that all evil actions are incomprehensible in this sense we would have built a surprising kind of relativism into the concept of evil. We must ask to whom evils are incomprehensible. This question matters, because what one person can imagine herself doing differs greatly from what another person can imagine herself doing. We differ in our desires and inclinations, in our value judgements, in our levels of squeamishness, and in our respective abilities to picture ourselves in scenarios that diverge from our own. Given this disparate set of abilities to imagine choosing to perform actions, we might be led to the unlikely conclusion that a particular extremely wrong action is both evil and not evil, because it is comprehensible to some but incomprehensible to others. Alternatively, we could say that judgements that actions are evil can never be anything more than observer-relative judgements. If this were the case, it would be misleading to say that a particular sadistic murder was evil, full stop. Instead, we would have to say that it was evil-relative-to-me, but maybe not evil-relative-to-you. Some people say that this kind of relativism runs through all of morality, but most philosophers want to avoid it. When we condemn an action as evil, we do not intend to add the relativistic qualifier, “Evil to me, that is” (cf. Joyce 2011, 527). There is another problem with the idea that the judgement that an action is evil is relativised to what the person who makes that judgement can imagine herself doing. Via literature, movies, and television we have become familiar with the figure of the anti-moralist, such as Satan in Paradise Lost, who does what is morally

Evil, Extremity, and Incomprehensibility  59 wrong because it is wrong (Milton 1993, 123 and 251 (Bk I, 165 and Bk IV, 110)). There are also real-life examples of serial killers who believe that they are evil (Stone 2009, 24). The BTK killer Dennis Rader, who got a sexual thrill from murdering his victims and who wrote taunting letters to the authorities seeking media notoriety, seemed self-consciously to aim to do what is evil, and to glory in its achievement (Stone 2009, 237–8). If judging that an action is evil implied judging that it is incomprehensible, it would be incoherent for any person, fictional or real, to plan self-consciously to do evil or to glory in having done what he takes to be evil. How could Rader, for example, plan to do what he cannot imagine himself doing under any circumstances, and how could he gloat over having done something that he cannot imagine himself doing under any circumstances? Since self-conscious evildoing seems to be a coherent possibility, we have good reason to reject the claim that judging an action to be evil implies judging that it is incomprehensible. Perhaps we could avoid these problems by refining our account of the connection between evil and incomprehensibility, such that evils are culpable wrongs that are incomprehensible to a certain kind of person. For example, Singer defines an evil action as “one so bad, so awful, so horrendous that no ordinary decent reasonable human being can conceive of himself (or herself) doing such a thing” (Singer 2004, 196). If all evil actions are incomprehensible to morally decent human beings, then we would be able to say that the murders committed by Dennis Rader were evil actions, full stop. The fact that the actions were not only comprehensible but enjoyable to Rader himself would be compatible with the fact that those actions are evil, because Rader himself is not an ordinary, decent, and reasonable human being. Beyond this, it might be suggested that this kind of incomprehensibility constitutes the kind of extremity that is a necessary condition for evil action. If both extremely harmful atrocities and harmless failed attempts at terrorism are culpably wrong actions that decent and reasonable people could not imagine themselves performing under any circumstances, then both could count as evil on this account.

3.4  Separating Evil and Incomprehensibility Even if we were to accept that all evil actions are incomprehensible to morally decent human beings, I think that the notion of incomprehensibility cannot be used to specify the kind of extremity that is possessed by all evil actions. In order to see why this is the case, we should first notice that there are a variety of non-moral reasons for which an action would be incomprehensible to an ordinary, decent, reasonable human being. Some actions are so incredibly pointless and time-consuming that a reasonable person could never imagine freely choosing to

60  Evil, Extremity, and Incomprehensibility do them. Some such actions are also culpable moral wrongs. For example, writing and then publicly reciting a 100,000 line poem that repeatedly insults your innocent landlord, against whom you hold a grudge, would be morally wrong and also would be incomprehensible to an ordinary, reasonable person. Clearly, though, this incomprehensible culpable wrong is not sufficiently morally extreme to count as evil. In response to this example, Singer presumably would point out that his definition specified that evil actions are incomprehensible in virtue of being “so bad, so awful, so horrendous”, rather than incomprehensible in virtue of being pointless or otherwise irrational (Singer 2004, 196). Yet there are other possible cases of wrong actions that are so awful and horrendous as to be incomprehensible, but lack the requisite moral gravity to count as evil. The act of licking the toes of randomly selected, incapacitated, and unconsenting hospital patients is morally wrong and is so viscerally disgusting as to be incomprehensible to ordinary people, but it is a merely perverse wrong that falls short of evil. If Singer wants his definition of evil to map onto common intuitions about which particular actions are evil, he will have to say that evils are those wrong acts that are incomprehensible to ordinary, decent, reasonable people in virtue of being extremely morally wrong. Once this is specified, though, there is a danger that the incomprehensibility condition will become superfluous to the definition of evil action. Suppose that we committed ourselves to the claim that no ordinary, decent, reasonable person can imagine freely choosing to do an evil action, but then we were asked how we could tell whether a particular person is ordinary, decent, and reasonable. If our answer to that question included the claim that someone is an ordinary, decent, and reasonable person only if she could never imagine herself freely choosing to do that we independently regard as an evil action, then the account will be circular. Alternatively, we could offer an independent characterization of an ordinary, decent reasonable person, and use this person’s imaginative capacity as a genuine criterion for evil action. However, it is not clear that incomprehensibility to an ordinary, decent, reasonable person really is a necessary condition for evil action. As we shall see in later chapters, there are a range of social and environmental conditions, such as those present in the Milgram experiment, that would cause many of us to commit extreme culpable wrongs (Milgram 1974; Doris 2002, 39). When ordinary people like ourselves become acquainted with these experimental results, some of us might suddenly be able to imagine ourselves performing the same horrible actions—perhaps performing them against our better judgement—were we placed in similar conditions. I contend that after this imaginative act it would be possible for us to maintain the judgement that those actions count as evil, even

Evil, Extremity, and Incomprehensibility  61 though now they are comprehensible in the relevant sense. Being able to imagine ourselves performing those actions might lead us to despair about the weakness and lack of character of ordinary, decent people, rather than to judge that the actions in question are less wrong or less worthy of condemnation than we had realized. If we judge an action to be evil, then we judge it to be extremely immoral and unjustified, even if we can also imagine that ordinary people like ourselves sometimes would perform extremely immoral and unjustified actions. There is a possible gap between what we think that we morally ought never do, and what we think that we would never choose to do. For this reason we ought to deny that incomprehensibility is a necessary condition for evil action, and also deny that incomprehensible culpable wrongdoing is sufficient for evil action.

3.5  Appropriate Connection to Extreme Harm I have argued that intuition (5), which links evil action and incomprehensibility, is superficially attractive but ultimately should be rejected. While many evil actions are incomprehensible to many people, incomprehensibility is not a necessary condition for evil action. Since extreme wrongness cannot be explained in terms of incomprehensibility, we need to locate an alternative, more informative account of the nature and degree of the wrongness that is a necessary condition for evil action. Let us turn to a list of examples of prima facie evil actions and ask what those actions have in common. Here, as elsewhere, I will be relying on intuitions that are not universally shared, but my hope is to set out a conception of evil action that fits with widely shared intuitions about particular actions, and that is defensible on other grounds. The list of prima facie evil actions can be divided into the relatively uncontentious examples and the more contentious examples, about which there is disagreement even amongst the philosophical defenders of the concept of evil. The uncontentious examples include culpable wrongs in which the inflicted harm is extreme for at least one victim and the wrongdoer is culpable for inflicting that harm, as well as harmless culpable wrongs that were intended to be extremely harmful for at least one victim. Acts of genocide, acts of torture and murder by serial killers, attempted terrorist bombings, and gratuitous torture by members of the military could fit into this category (Stone 2009, 287). The more contentious examples might include culpably reckless close shaves, as well as voyeuristic acts of appreciation of extreme harm. In both of these latter kinds of action, it is possible that the actual infliction of extreme harm is neither intended nor rationally foreseeable.4   In Chapters Four and Five, we shall see why some philosophers would judge that culpably reckless close shaves are not evil actions, and, in Chapter Nine, we shall see why some cases of sadistic voyeurism might not count as evil actions. 4

62  Evil, Extremity, and Incomprehensibility While this list is fairly diverse, all of its members do have a complex property in common: all are culpably wrong actions that are connected in at least one of an identifiable set of ways to an actual or possible harm that is sufficiently extreme for at least one victim, and for which the agent would properly be held responsible were the harm inflicted. I contend that being connected to extreme harm in one of these ways is what is required to give these actions their moral gravity, what renders them non-trivial in the sense required for evil. For want of better terminology, I will describe an action that is connected to an actual or possible extreme harm in at least one of these ways as being appropriately connected to extreme harm. (Clearly this does not mean that it is morally good for an action to be connected to harm in these ways.) It is important to note that my account of the extremity that is required for evil action is not so open-ended as to allow any and all kinds of moral extremity to push a culpable wrong into the category of evil. There are many properties of wrong actions that would make them more wrong, in the sense of being more vicious, more worthy of punishment and condemnation, and less worthy of choice. A particular wrong action might be made more wrong by also being gratuitous, or by also being perversely pleasurable to the wrongdoer (where this pleasure does not make the action more harmful), or by also being disrespectful (where the disrespect does not make the action more harmful). My claim is that all evil actions are extreme culpable wrongs, and that the extremity of the wrong cannot be constituted merely by the fact that the wrong is extremely gratuitous, or extremely perverse, or extremely disrespectful. Rather, a culpable wrong counts as evil only if it is extremely wrong in virtue of being appropriately connected to an actual or possible harm that is sufficiently extreme. I propose that all evil actions are extreme culpable wrongs, where “extreme” means appropriately connected to an actual or possible harm that is extreme for at least one victim, and “appropriately connected” means that the action culpably produces or was intended to produce such a harm, or (more contentiously) that the action foreseeably would have produced such a harm if it was successful or if it had its typical effects, or (even more contentiously) that the action is an appreciation of such harm. I will discuss these more contentious possible additions in more detail in later chapters. Admittedly, this account of the extremity that is required for evil action is less neatly unified than some of the alternatives that we have considered, such as the view that all evil actions are extremely harmful, or the view that all evil actions are extreme in the sense that they are incomprehensible. However, I have argued that those more unified accounts fail to capture the full range of actions that can count as evil.

Evil, Extremity, and Incomprehensibility  63 In claiming that all evil actions are culpable wrongs that are extreme in this respect, I am not aiming to set out necessary and sufficient conditions for evil action. Rather, my claim that all evil actions are extreme culpable wrongs is intended to capture the common core of the concept of evil action. I think that these conditions are implicitly taken to be necessary for evil action by the majority of people who use the concept of evil in the context of contemporary moral discourse. Given that these conditions are internally consistent, and are not objectionable on other grounds, we ought to agree that if an action is evil, then it is an extreme culpable wrong. As we shall see in the following chapters, there is significant disagreement over whether the fact that an action is an extreme culpable wrong in the specified sense is not only necessary but sufficient for its being an evil action. I hope that it is comparatively uncontroversial to claim that all evil actions are culpable wrongs that are appropriately connected to actual or possible extreme harms. If extreme culpable wrongdoing really is the common core of the concept of evil action, we need to explain why some philosophers have defended accounts of evil action which imply that some evil actions are not appropriately connected to actual or possible extreme harms, and why others have proposed that all evil actions actually inflict extreme harm. One possibility is that the philosophers in question simply have not considered a sufficiently broad range of examples of evil actions. For instance, philosophers who claim that all evil actions must be extremely harmful, including Card, Singer, Thomas, and Formosa, do not explain why harmless failed attempts, culpably negligent close shaves, or harmless sadistic voyeurism could not count as cases of evil action. In all such cases, the accounts that these philosophers have offered could easily be modified so as to include some or all of these examples as harmless evil actions. Of course, the proponents of these accounts might prefer to defend the view that all evil actions are actually harmful, and in this case they would need to argue explicitly for this conclusion. It is more difficult to explain why McGinn and Steiner maintain that actions can be evil despite being connected to merely minor harms. I think the correct explanation is that that these philosophers are led astray by the belief that evil actions must be qualitatively rather than quantitatively distinct from ordinary wrong action (Russell 2007b, 661, 670). By claiming that evil actions have a property that is not possessed to any degree by non-evil culpable wrongs—namely, the property of being sadistically pleasurable to the wrongdoer—Steiner and McGinn can draw a sharp distinction that leaves no actions in a grey area between evil and non-evil. The cost of drawing a sharp boundary in this way, though, is that McGinn and Steiner are led to the implausible view that all sadistically pleasurable culpable

64  Evil, Extremity, and Incomprehensibility wrongs are evil, even if they are only minor wrongs. I will return to discuss these issues in more detail in Chapter Six.

3.6  Where is the Boundary of Extremity? My account of the kind of extremity that is required for evil action leaves a significant question unanswered. Given that all evil actions are appropriately connected to extreme harms, we must ask what it takes for a harm to be extreme, and where the boundary lies between extreme and non-extreme harms. In many cases it is fairly easy to rank a number of harms of the same kind on a scale of less harmful to more harmful. An action that inflicts an annoying jabbing pain is less harmful, cet­ eris paribus, than an action that inflicts an excruciating throbbing pain. An action that kills one innocent person is less harmful, ceteris paribus, than an action that kills ten innocent people. Different kinds of harms might seem to be incommensurable, but in many cases we can rank them by asking which we would prefer to suffer, if given a choice, or which we would choose to prevent from being inflicted on our loved ones, if we had the power to ensure this outcome. In those cases in which we cannot choose, it seems reasonable to say that the harms are equally bad. Although it is tricky in many particular cases, there is no insurmountable practical difficulty in ranking actions in virtue of their connection to more or less extreme harms. But exactly how extreme does the connected harm have to be in order for the action to count as evil? One possibility worth considering is that the harm must be maximally extreme; that it must occupy the end point on the scale of harms. It is often suggested that evil action is the worst kind of moral wrong, and this might lead us to conclude that there could not be a worse action than any specific evil action. This view seems to be endorsed by Neiman, who declares that “[d]‌ividing evils into greater and lesser, and trying to weigh them, is not just pointless but impermissible” (Neiman 2003, 286). Singer also hints at something like this when he states that evil “lies at the end point of a scale of badness” (Singer 2004, 196). It is conceivable, though, that an action could be maximally wrong without being connected to a maximal harm, so it is not clear that Singer actually believes that an action can be evil only if it is connected to a maximally extreme harm. A similar proposition is tabled by Daniel Haybron in his discussion of evil personhood. Haybron considers the view that an evil person is maximally bad in every respect, but then rejects it on the grounds that it “defines evil out of existence” (Haybron 2002a, 261). Haybron’s concern is warranted. If evil actions are those culpable wrongs appropriately connected to a maximal harm—a harm than which none greater can be conceived—then it seems guaranteed that there are no actual evil actions. We can always conceive

Evil, Extremity, and Incomprehensibility  65 of a possible wrong action that takes one more innocent life than is taken by any actual wrong action, and hence which is appropriately connected to a harm that is more extreme than any of the harms to which actual actions are connected. We ought not aim to construct an account of evil action which implies there could be no actual evil actions, and hence we have good reason to reject the assumption that evil actions must be tied to harms that occupy the very end point of a spectrum. If we believe that there really are some evil actions, and if we think that those actions are connected to the worst kind of harms, we must mean that they are connected to harms that fall somewhere beyond a certain threshold on the spectrum of harms, rather than harms that occupy the end point of that spectrum. A similar claim holds true with regard to the degree of wrongness of evil actions. If we say that evils are the worst kind of wrongs, and we think that there are some actual evil actions, then we must mean that evils are wrong actions that fall somewhere beyond a certain threshold on the scale of wrongness. Thus it would be a mistake to say that an evil action is an action that is maximally wrong, where that implies that there could not be a more deplorable or less choiceworthy action. There are many features of an atrocious action that might make it morally more deplorable than other atrocities. Even though it is crass and disrespectful in some circumstances to rank evils, we should admit that some evil actions are morally worse than others. The extremity that is required for evil action consists in being beyond a threshold on a spectrum, rather than occupying an end point. It would be helpful if we could specify the location of that threshold. Calder suggests that the harm must be “significant” and that we come closest to capturing the amount of harm required for evil if we think of the harm of evil as a harm that a normal rational human being would take considerable pains to avoid. (Calder 2013, 188)

This strikes me as being too low a threshold, although it is not clear what counts as “considerable pains”. Kekes locates the threshold at the point beyond which the harm “interferes with the functioning of a person as a fully-fledged agent” (Kekes 1998, 217). Yet not every wrong action that interferes with a person’s capacity to function in this way is evil, because some such interferences might be comparatively trivial, harmless, and easily reversible, as when a dentist administers nitrous oxide to her distressed but unconsenting patient, thereby rendering the patient docile and incapable of rational deliberation. Formosa initially suggests that the harms caused by evil must be “life-wrecking”. He claims that such harms are ineradicably burned into the victim via wrong actions including “genocide, gang-rape, prolonged torture and murder”, and adds that we are “fairly good at recognising” which harms are life-wrecking (Formosa 2008, 229). As Formosa goes

66  Evil, Extremity, and Incomprehensibility on to note, though, such extreme harms sometimes are not actually life-wrecking, because some victims of such harms are extraordinarily resilient. For an act to be evil, Formosa concludes, “it need only be foreseeable that such harm would normally have a life-wrecking impact on the victim” (Formosa 2008, 230). I have already disagreed with one aspect of Formosa’s account; namely, his claim that all evil actions are actually extremely harmful. I maintain that harmless failed attempts can count as evil actions as well. It would be easy to modify Formosa’s claim about the connection between evil action and what normally are life-wrecking harms to meet this condition. What we need to consider is whether all evil actions are appropriately connected to actual or possible harms that foreseeably and normally would be life-wrecking. It is certainly true that many evil actions are life-wrecking. Even so, it is not clear that all evil actions, were they successful in inflicting harm, would normally have a life-wrecking impact on the victim. We can tell at the moment of infliction that some harms, like murder, are life-wrecking. In many other cases it is not at all obvious what the long-term effect of a particular harm will be, or even what the long-term effects of that type of harm normally would be. Formosa’s account suggests that, were we to witness a case of gratuitous torture, we should withhold our judgement as to whether it was an evil action until we had gathered evidence about the long-term happiness and success of people who at some point have been tortured. If we discovered that people generally are emotionally far more resilient than we had suspected, and hence that torture was not normally life-wrecking, Formosa’s account suggests that we should conclude that gratuitous torture is not evil after all. Similarly, if torture normally was life-wrecking, but then we developed a new method of counselling or drug treatment that for the first time allowed most victims of torture to recover and lead happy and untroubled lives, Formosa’s account would lead us to judge that gratuitous torture once was evil, but no longer is so. These implications of Formosa’s account are implausible. This is not to deny that the long-term impact of a violent act on the well-being of the victim should be considered in our evaluation of the extremity of the harm that is inflicted by that wrong action. A worse and more sustained long-term impact does render an action more harmful, ceteris paribus. However, I contend that some evil actions might be connected to harms that are extreme but that do not have a life-wrecking long-term impact. Gratuitous torture could be an evil action even if it turned out that victims of torture usually get over it. Gratuitous torture is extremely wrong in virtue of the perpetrator’s wilful and unjustified infliction of intense and horrible pain and suffering on the victim, regardless of whether the victim eventually recovers and manages to live a happy life. It remains true that many evil actions

Evil, Extremity, and Incomprehensibility  67 do inflict life-wrecking harms on the victims, and it is clear that torture often is life-wrecking. Nonetheless, the judgement that an evil action is connected to an extreme harm is not reducible to the judgement that it is connected to a harm that is (or usually is) life-wrecking, nor does it depend on the judgement that it is connected to harm that is (or usually is) life-wrecking. I have suggested that the common core of the concept of evil action is that all evil actions are extreme culpable wrongs. This extremity is not explained by the incomprehensibility of the wrong, if indeed the action in question is incomprehensible. Rather, evils are wrongs that are extreme in virtue of being appropriately connected to actual or possible extreme harms. What we want to know is just how extreme these harms must be in order to satisfy this condition. We have seen that such harms need not be at the end point of a spectrum, and need not be life-wrecking. Perhaps the best we can do in these circumstances is to say that the harms must be extreme, and to give examples of cases that clearly fall on either side of the dividing line. A mild jabbing pain is clearly not a sufficiently extreme harm, nor is the loss of a handbag, or the denial of the opportunity to go on a rollercoaster, or even a hard slap in the face. Multiple murder and gratuitous torture do strike many people as being sufficiently extreme to count as evil. When people are asked to list actions that they think are evil, they point to genocides, serial killings, and sadistic torture, and do not point to actions that are connected only to mild harms. It is very likely that people have and will continue to have conflicting intuitions as to which of the actions that fall in this grey area between these two extremes should be called evil. Of course, we could stipulate that the boundary of extremity lies at a certain point on the scale of harm, but I think that doing so would result in a false precision that would fail to reflect the actual vagueness of the concept of evil action. It might seem objectionable to build this kind of vague boundary into an analysis of the concept of evil. Yet this kind of vagueness is to be expected in moral concepts, as in most other concepts. For example, for an action to be morally heroic it must be morally right and also difficult, dangerous, or costly, but the concept of heroism itself does not specify exactly how difficult, dangerous, or costly an action must be in order to qualify as heroic, and we should not expect that as philosophers we could eliminate disagreement over cases that are in the grey area. The fact that we cannot specify exactly the extremity of the harm that is linked to evil action does not indicate that there is any special problem with the concept of evil. In this chapter, I have tried to clarify and assess intuition (4), that evil actions are extreme and never trivial, and intuition (5), that evil actions are incomprehensible. I have argued that the kind of extremity that is a necessary condition for evil action is complex, and cannot be reduced to the claim that all evil

68  Evil, Extremity, and Incomprehensibility actions are extremely harmful or that all evil actions are incomprehensible. My view is that all evil actions are culpable wrongs that are appropriately connected to actual or possible extreme harms. An action is appropriately connected to an extreme harm if the agent intends to inflict that extreme harm, or (more contentiously) if the action foreseeably would inflict that extreme harm, or (even more contentiously) if the action is an act of appreciation of that extreme harm. This is what underlies the claim that an evil action is an extreme wrong. In the next chapter we will move on from what I have called the common core of conceptions of evil action—namely, the view that all evil actions are extreme culpable wrongs—to the deeper disagreements over whether all evil actions are the product of a narrow set of motives.

4 Thick and Thin Accounts of Evil Action The claim that all evil actions are extreme culpable wrongs forms the common core of most recent philosophical accounts of evil action. When we move beyond this core, we encounter significant disagreement over what it would mean to call an action evil. Some philosophers follow Arendt in offering a psychologically thin account of evil action, according to which being an extreme culpable wrong is both necessary and sufficient for being an evil action. On this view evil actions could come from a very broad range of motives, stretching from malicious defiance of morality, through misguided ideological conviction, to banal and thoughtless self-interest. In contrast, other philosophers endorse various psychologically thick accounts of evil action, according to which an action is evil if and only if it is an extreme culpable wrong that also meets some extra psychological conditions. Some of these extra conditions—malice, sadistic pleasure, and knowing defiance—are familiar from folk psychology, and are fairly easily recognizable. Other possible extra conditions are more complex and unfamiliar psychological properties from outside the realm of folk psychology. In this chapter, I will lay out the competing thick and thin accounts of evil actions and show how they align with various currents of thought beyond philosophy. In the following chapters, I will go on to consider which, if any, of these theories gets at the real or the best concept of evil action.

4.1  Eichmann, and Arendt’s Eichmann Let us remind ourselves of the eight common intuitions about evil action that I set out in Chapter Two. (1) Saying that an action is evil is a means of expressing very strong moral condemnation of that action.

70  Thick and Thin Accounts of Evil Action (2) Evil actions are morally wrong. (3) The person who performs an evil action is blameworthy and properly held responsible for that action. (4) Evil actions are extreme and never merely trivial. (5) Evil actions are incomprehensible. (6) Evil actions can be banal. (7) There is a psychological hallmark of evil action. (8) Evil actions are qualitatively distinct from ordinary wrongs. The time has come to explore intuitions (6)–(8) in more detail. Some of the most arresting examples of evil actions are those performed by serial killers such as Ted Bundy, John Wayne Gacy, and the BTK killer Dennis Rader. These men murdered victim after victim. Their actions certainly were extreme culpable wrongs, and it also seems that these men despised their victims, took immense pleasure in their killing, and knew that what they were doing was morally wrong. These serial killers seem to be monstrous, not in the sense that they are supernatural beings, or are biologically inhuman, but in the sense that they are significantly psychologically different from ordinary people. It is likely that some people fix on these examples of extreme wrongdoing as archetypes of evil action, and judge that other actions are evil in so far as they resemble these examples. People who think in this way will be strongly inclined to endorse intuition (7), because they believe that there are distinctive sadistic and perverted motives that lie behind evil actions. According to Hannah Arendt and numerous philosophers who have followed in her footsteps, it would be a terrible mistake to think about evil in this way. It was a mistake, Arendt believed, that she herself had made in 1951, when she first published The Origins of Totalitarianism. In this book she claimed that the mass atrocities of the twentieth century had shown us a “radical evil” that could not be explained by the everyday motives of “self-interest, greed, covetousness, resentment, lust for power, and cowardice” (Arendt [1951] 1967, 459) Ten years later, in 1961, Arendt attended some of the trial of the Nazi war criminal Adolf Eichmann in Jerusalem, which she discussed in her book Eichmann in Jerusalem, first published in 1963. Here she reports being surprised to discover that Eichmann was not a “perverted sadist” or an “abnormal monster” motivated by anti-Semitic malice (Arendt [1963] 2006, 276). It would have been very comforting indeed to believe that Eichmann was a monster. . . . The trouble with Eichmann was precisely that so many were like him, and that the many were neither perverted nor sadistic, that they were, and still are, terribly and terrifyingly normal. (Arendt 2006, 276)

Thick and Thin Accounts of Evil Action  71 Not only did Eichmann act without malice and without sadistic pleasure, Arendt claims, but there is no evidence that he knowingly defied morality (Arendt 2006, 276–7). Eichmann was “not Iago and not Macbeth, and nothing would have been farther from his mind than to determine with Richard III ‘to prove a villain.’ ” (Arendt 2006, 287) Arendt continues: Except for an extraordinary diligence in looking out for his personal advancement, he had no motives at all. . . . He merely, to put the matter colloquially, never realized what he was doing. . . . He was not stupid. It was sheer thoughtlessness—something by no means identical with stupidity—that predisposed him to become one of the greatest criminals of that period. (Arendt 2006, 287–8)

Nonetheless, Arendt does not shy away from judging that Eichmann’s actions in coordinating the deportation of Jews to the death camps were evil. For instance, she claims that there is a “strange interdependence of thoughtlessness and evil” (Arendt 2006, 288), and that the Eichmann trial taught us the lesson of the “banality of evil” (Arendt 2006, 252). Years later Arendt claimed: It is indeed my opinion now that evil is never “radical”, that it is only extreme, and that it possesses neither depth nor any demonic dimension. It can overgrow and lay waste to the whole world precisely because it spreads like a fungus on the surface. (Arendt 1978, 251)

Arendt’s analysis of Eichmann and her subsequent claims about the nature of evil have been deeply influential, and are a common source of intuition (6), that evil actions can be banal. While intuition (6) is endorsed by many people who have written about evil subsequent to Arendt, it remains highly contentious. The journalist Ron Rosenbaum rejects outright the phrase “the banality of evil”, which he describes as “the most overused, misused, abused pseudo-intellectual phrase in our language” (Rosenbaum 2009). Rosenbaum continues: Adolf Eichmann was, of course, in no way a banal bureaucrat: He just portrayed himself as one while on trial for his life. Eichmann was a vicious and loathsome Jew-hater and -hunter who, among other things, personally intervened after the war was effectively lost, to insist on and ensure the mass murder of the last intact Jewish group in Europe, those of Hungary. So the phrase was wrong in its origin, as applied to Eichmann, and wrong in almost all subsequent cases when applied generally. Wrong and self-contradictory, linguistically, philosophically, and metaphorically. (Rosenbaum 2009)

There are three strands in Rosenbaum’s objection that ought to be disentangled: that the phrase “the banality of evil” is linguistically confused or misleading, that Eichmann actually was not “banal” as Arendt used that term, and that no or almost no other extreme wrongdoers were “banal” as Arendt used that term. Let us consider each of these objections in turn.

72  Thick and Thin Accounts of Evil Action It is plausible that people who lack first-hand familiarity with Arendt’s introduction of the phrase “the banality of evil” in Eichmann in Jerusalem could take it to be an oxymoron. One reason that the phrase is so striking and memorable is its hint of self-contradiction. Evil actions, by definition, are extreme wrongs, and hence deserve our attention. Calling evils “banal” might suggest that they are somehow also boring, trivial, or unremarkable, and to this extent the use of this label is potentially misleading. But it is clear that Arendt herself intends “banality” to denote ordinariness of type of motive and type of agent, rather than ordinariness of action. It is quite possible that Arendt could have avoided this confusion by employing different terminology; for instance, by saying that evil is non-demonic, or non-monstrous, or simply that evil comes from common motives. Such speculation is beside the point, though, because the phrase “the banality of evil” is entrenched in the minds of the general reading public, as well as in academic literature concerning evil. The appropriate response in such circumstances is to make it clear what Arendt and others mean when they say that evil is banal, rather than attempt to eliminate the phrase. The second strand of Rosenbaum’s criticism is that Eichmann’s wrongdoing itself was not banal, even as Arendt intended that term to be understood. On this issue, Rosenbaum stands on solid ground. David Cesarani has convincingly argued that Arendt misread the psychology of Eichmann. According to Cesarani, Eichmann was not motivated merely by a banal desire for career advancement, but was deeply anti-Semitic (Cesarani 2007, 58, 157, 360), was committed to the goal of murdering Jews, and pursued that goal even when it became clear that Germany would lose the war (Cesarani 2007, 183–4). Perhaps there is room for debate over whether Eichmann took sadistic pleasure in his actions (Cesarani 2007, 180, 360), or whether he judged that his own actions were morally wrong (Cesarani 2007, 300, 312), but his actions certainly were deliberate and malicious rather than merely obedient (Cesarani 2007, 353–6). When seen in this light, the so-called thoughtlessness that Arendt reported seeing in Eichmann during the trial is revealed as a dissembling screen designed to cover his true motives and intentions. However, even if we judge that Arendt’s portrayal of Eichmann was fundamentally inaccurate, we might accept the claim that if Eichmann actually had been as he was described by Arendt, his actions still would have been evil. Perhaps we can treat Arendt’s Eichmann as something like a fictional character, and can agree with Arendt in judging that this character is an evildoer despite the absence in his actions of malice, sadistic pleasure, and knowing defiance of morality. In doing so we run the risk of misrepresenting the actual motives of the perpetrator of horrific crimes, and misrepresenting the nature of the Holocaust. Yet if we simply dropped the example of Eichmann from our discussion of evil, we would have difficulty

Thick and Thin Accounts of Evil Action  73 engaging with the various claims that philosophers have made about Arendt’s Eichmann in this context. In evaluating accounts of evil, I will be careful to distinguish between the real Eichmann and Arendt’s Eichmann, regardless of whether the proponents of these accounts do so themselves. By this means I hope to engage with the existing literature without misrepresenting the real Eichmann. Even if we agreed that Arendt’s Eichmann would count as an evildoer, some of us might think that real, non-fictional evildoers are nothing like Arendt’s Eichmann. We must respond to the third strand of Rosenbaum’s objection: that the phrase “the banality of evil” applies to no actual cases, or almost no actual cases, of extreme wrongdoing, and hence is of no use to us. Rosenbaum’s third claim is not defensible. There are many other examples of extremely wrong actions that were not motivated by malice, that were not sadistic, and that were not knowingly defiant. In some cases it appears that, in the absence of malice, cold self-interest can be sufficient to motivate horrendous crimes. For instance, in Truman Capote’s account of the murder of the Clutter family, Capote reports that the murderer, Perry Smith, bore no malice towards his victims. Apparently Smith said of the father of the family, Herb Clutter, “I didn’t want to harm the man. I thought he was a very nice gentleman. Soft spoken. I thought so right up to the moment I cut his throat” (Capote 2000, 237). There are also real examples of extreme wrongdoers who, like Arendt’s Eichmann, do not take sadistic pleasure in their harmful wrong actions. Some extreme culpable wrongs are difficult and positively unpleasant actions that are performed out of a misguided sense of duty. Jonathan Bennett claims that Himmler saw his own murderous treatment of the Jews as fitting this mould. For instance, Himmler “described himself as caught in ‘the old tragic conflict between will and obligation’ ” (Bennett 1994, 300), and had to steel himself to overcome squeamishness when he came face to face with actual executions that he believed must be performed for the good of Germany (Bennett 1994, 300). According to Primo Levi, Rudolph Hoess, the commandant of Auschwitz, “never enjoyed inflicting pain or killing: he was no sadist” (Levi 2000, 20). A Chilean ex-torturer also suggests that extreme wrongs can be done without sadistic pleasure, when he says “there comes a moment when you feel nothing about what you are doing” (Wolfendale 2007, 175). Finally, even if Eichmann, at the time of acting, judged that his own actions were morally wrong, and hence knowingly defied morality, there are other cases of extreme wrongdoing in which the wrongdoers mistakenly believed that their own actions were morally permissible. The historian Hugh Trevor-Roper, for instance, is confident that “Hitler was convinced of his own rectitude” (Rosenbaum 1998, xxii). Terrorists who kill innocent people in pursuit of misguided religious or

74  Thick and Thin Accounts of Evil Action political goals typically believe that their actions are morally right, even heroic, and some leave behind notes or video messages in which they make sustained attempts to explain why their actions are justified. Similarly, some military personnel who administer gratuitous torture falsely believe that their actions are morally justified in that their victims are criminals who deserve to suffer, and that torture is the only way to secure confessions and other information that will help to save lives in future. Unjustified suicide bombings and gratuitous torture do violate morality, but many people who carry out those actions do not knowingly defy morality. Even if the real Eichmann was nothing like Arendt’s Eichmann, there are real examples of extreme wrongdoers who do not defy morality, or act from malice, or take sadistic pleasure in the harm that they inflict. These actions are banal, in Arendt’s sense of “banal”. In saying that these actions are banal we are not saying that the actions themselves are ordinary, unremarkable, or excusable. Rather, we are saying that the motives that lie behind the actions are of a comparatively common, ordinary and unremarkable type, in that they do not include extreme malice, sadistic pleasure or knowing defiance of morality. In the debate about the nature of evil, we ought to accept that the word “banal” is a term of art, just as the word “intentional” is a term of art in debates about the nature of mental states.1 In both cases it would have been better if the founders of these debates had chosen labels that were less prone to give rise to misunderstandings, but the labels are firmly entrenched, so the best we can do is to explain them and use them carefully.

4.2  The Psychologically Thin Account of Evil Action Arendt overstates her case when she claims that evil action is never radical, and never possesses any depth. For instance, many serial killers are driven by perverted sexual desires, by a thrill that they get from overpowering their victims and from knowingly defying morality (Stone 2009, 216). There is nothing banal or ordinary about the motives that lie behind these actions, and they are squarely within the domain of actions that are labelled evil. The most charitable way in which to interpret Arendt is to read her as saying that a great number of evil actions are banal. Intuition (6), that evil action can be banal, has been accepted by many people. It gives rise to what I will call the psychologically thin account of evil action.

1   An intentional mental state, in the context of philosophy of mind, is a mental state that has representational content; that is, a mental state that is about something. This label creates confusion, given that in ordinary language “intentional” means something like “on purpose” or “deliberate”.

Thick and Thin Accounts of Evil Action  75 Bernard Williams drew a distinction between what he called thick ethical concepts, such as “coward, lie, brutality and gratitude” (Williams 1993, 140) and thin ethical concepts such as right and good. There are a range of differences between these two loose collections of moral concepts. One of those differences is what I will call the difference between psychologically thick and psychologically thin concepts. Psychologically thick concepts, such as the concepts of cowardice and deceitfulness, carry a comparatively large amount of information about the specific psychological states of the agents to which they apply, and about the specific motives or intentions that lie behind the actions to which they apply. A person is a coward only if she is disposed to fail to overcome some of her fears, and an action is cowardly only if it flows from such a failure. Similarly, an action is a lie only if the agent intended to deceive his audience by making a claim that he knew to be false, and a person is dishonest only if he is markedly strongly disposed to tell lies or otherwise to mislead. In contrast, psychologically thin ethical concepts, such as the concepts of right, good, wrong, and bad carry comparatively little information about the motives and intentions that lie behind actions. This is not to say that all psychologically thin ethical concepts carry no such information, though. Amongst the thin concepts, some are thinner than others. The concept of a harmful action carries absolutely no information about the psychology of the agent, and hence is completely thin. A harmful action could have been performed for any reason, from whatever motive, and with any intent. However, when we apply the concept of a culpably morally wrong action, we convey some information about the agent’s psychological states. For instance, if an action is culpably morally wrong, then we can infer that the agent possesses at least the base level of rationality that is required for moral responsibility, that the action was not the product of excusable negligence, and perhaps that the agent freely chose to perform the action. Nonetheless, a morally wrong action could have been performed from a great variety of motives and with a great variety of intentions. Compared to concepts like cowardice, weakness of will, and compassion, the concepts of right action and wrong action convey little information about the psychological states of the agent or the specific motives that lie behind the action. Hence, the concepts of right action and wrong action are comparatively psychologically thin. There is another kind of thickness and thinness of moral concepts that ought to be distinguished from what I am labelling psychological thickness and thinness. Moral concepts can also be expressively thick or expressively thin. In normal cases, calling an action wrong functions in part to express an attitude of disapproval towards the performance of that action, but not a more specific attitude

76  Thick and Thin Accounts of Evil Action than that.2 In contrast, when we call an action repugnant, or pathetic, ceteris pari­ bus, we are expressing a richer set of emotions or attitudes towards that action. The concepts of the repugnant and the pathetic are comparatively expressively thick. A concept that is expressively thick need not be psychologically thick. For instance, if I declare that President Nixon’s actions were disgusting, I expresses a specific kind of disapproval, but I have not, in that declaration itself, conveyed any specific information about the personality of Nixon or the specific motives that lay behind his actions. In contrast, some moral concepts, such as those of arrogance and deceitfulness, are psychologically thick but expressively thin. When a person applies these concepts to an action, she gives us fairly specific information about the psychology of the agent, but does not reveal much detail about her own attitudes towards that action. If I were to say that Nixon’s actions were arrogant and deceitful, I would be ascribing specific kinds of motive to Nixon, but not giving us much information about my own attitudes. Henceforth, when I speak of thick or thin accounts of evil I intend to refer to psychological thickness or thinness, rather than expressive thickness or thinness. In claiming that the Eichmann trial revealed the banality of evil, Arendt appears to be committed to the view that the concept of evil action is psychologically thin. Arendt seems to believe that all evil actions are extreme culpable wrongs, but that these actions could flow from a very broad range of motives that need not include malice, sadistic pleasure, or defiance of morality. Nor does Arendt posit any other psychological hallmark of evil action. Psychologically thin accounts of evil action have been defended more explicitly by some of the philosophers who have written on evil since Arendt, many of whom echo her claims about banality. Claudia Card describes evils as foreseeable intolerable harms produced by culpable wrongdoing. On my theory, the nature and severity of the harms, rather than the perpetrators’ psychological states, distinguish evils from ordinary wrongs. (Card 2002, 3)

Evils are, according to Card, “the worst wrongs people do”, (Card 2002, 28) and evil “is not defined by motive, although it implies culpability” (Card 2002, 9). While some evildoers may knowingly perform extreme wrongs, Card stresses that evildoers “commonly do not understand their deeds as atrocities” (Card 2002, 9). Formosa follows suit, claiming that people are “motivated to perform acts that are evil for all sorts of reasons” (Formosa 2008, 220). Haybron maintains that malice 2   As we have seen in Chapter One, this expressive role relies on the fact that most people are affectively and motivationally committed to morality. In the unusual cases in which an amoralist or an anti-moralist says that an action is morally wrong, he does not thereby express disapproval of the action.

Thick and Thin Accounts of Evil Action  77 and malevolence are not necessary conditions for evil (Haybron 2002b, 76), and both Kekes and Calder agree that some evil actions are purely self-interested or instrumental rather than malicious, and that some evildoers are morally deluded rather than defiant (Kekes 1990, 87; Calder 2002, 366–8). Neiman suggests that certain features of moral psychology once seemed to be essential to evil action, but that, after being confronted with Auschwitz, we saw that this was not so (Neiman 2003, 252). Neiman does not conclude that all evil actions are banal, though. She describes the September 11 terrorist attacks as “an instance of evil that was old-fashioned in structure”, insofar as it was “awesomely intentional” and was the product of “malice and forethought”, unlike Arendt’s modern banal evil (Neiman 2003, 283–4). In accordance with the thin account, Neiman believes that some but not all evil actions flow from banal motives. At first glance it seems that intuition (6) is incompatible with intuition (7). If evil actions can be banal, in Arendt’s sense of that term, then they need not be accompanied by malice, sadistic pleasure, or defiance, and we might conclude that there is no psychological hallmark of evil that distinguishes evils from ordinary wrongs. Card, Formosa, and Neiman clearly accept that evil actions can be banal, and go on to conclude that there is no psychological hallmark of evil. However, there are some philosophers who think that evil actions can be banal, but that there is an important psychological difference between evils and ordinary wrongs nonetheless. Garrard and Morton, for instance, believe that evil actions need not be malicious, or sadistic, or knowingly defiant, but they claim that there is another psychological feature that distinguishes evils from ordinary wrongs. I will consider Garrard’s and Morton’s respective accounts of evil action in more depth later in this chapter. For now, let us note that endorsing the claim that evil actions can be banal is not sufficient for endorsing a psychologically thin account of evil action. Advocates of the psychologically thin account believe that some evil actions are not malicious, or sadistic, or defiant, and that there is no other distinctive psychological hallmark of evil action. If my arguments in the preceeding chapters are correct, advocates of the psychologically thin account of evil should say that an action is evil if and only if it is an extreme culpable wrong. The psychologically thin account includes as evil actions the malicious, sadistic, and defiant murders performed by serial killers, but it allows that many evildoers are psychologically very different to serial killers. According to the philosophers who take this approach, we need a thin account of evil action in order to make correct moral judgements about the Holocaust and other twentieth-century state-sponsored atrocities such as the Soviet gulags and the Argentinian Dirty War. These clearly are cases that elicit our strongest moral despair, disgust, and condemnation, and hence seem to warrant being labelled with an extreme term like evil. Yet

78  Thick and Thin Accounts of Evil Action the ideologically driven and socially coordinated nature of the mass murder saw some participants adopt roles that are very dissimilar to that of a sadistic serial killer. The leading Nazis were not seeking attention and notoriety by self-consciously violating moral norms. Rather, they believed that what they were doing was morally right. Many of those who were eventually tried for war crimes continued to believe this, and went to their deaths with their consciences clear (Zangwill 2003, 92). The socially coordinated nature of the mass murder helps to explain why some (but only some) of the people who carried out the killings appeared not to have borne strong malice towards their victims, and why some of the murderers saw the killing as a thoroughly unpleasant and difficult duty that they were required to fulfil in order to protect their own society (Browning 1998, 174–5). Arendt maintains that many of these actions were not defiant or pleasurable: Evil in the Third Reich had lost the quality by which most people recognise it—the quality of temptation. Many Germans and many Nazis, probably an overwhelming majority of them, must have been tempted not to murder, not to rob, not to let their neighbours go off to their doom. . . . But, God knows, they had learned how to resist temptation. (Arendt 2006, 150)

There is no single archetype of evil action around which the psychologically thin conception of evil revolves. Rather, there are a diverse range of evil actions, all of them extreme culpable wrongs, but some malicious and others self-interested, some sadistic and others cold, some defiant and others horrendously misguided. Perhaps it is this inclusiveness of the psychologically thin account of evil that explains why it is fairly popular amongst contemporary philosophers. However, advocates of the thin account face a problem. If all there is to evil action is extreme culpable wrongdoing, how can we explain the fact that many people are very reluctant to judge that actions are evil? It is not only neo-Nazis who baulk at the claim that Nazi war crimes were evil, nor only anti-Americans who deny that the 9/11 attack was evil. Rosenbaum reports that many Holocaust scholars exhibit a “pronounced reluctance” to call Hitler evil (Rosenbaum 1998, xxi). Some people claim that evil does not exist, and that we should “abandon the idea of evil” (Cole 2006, 23). Others might think that there are evil actions, but that such actions are much less common than Arendt would have us believe. In order to understand why this is so, we must set out the competing psychologically thick conceptions of evil.

4.3  A Psychological Hallmark of Evil Action Intuition (7) is that there is a psychological hallmark of evil action. People who endorse this intuition think that even if it is true that only extreme culpable wrongs

Thick and Thin Accounts of Evil Action  79 could count as evil actions, not every extreme culpable wrong is evil. Certain actions fall short of being evil, on this view, because the relevant agents lack the distinctive psychological state or states that are necessary for the performance of evil action, and that mark evils out from other extreme culpable wrongs. It is uncontentious that some other moral concepts are psychologically thick in just this way. For instance, let us consider how we would evaluate a public donation to the needy that was performed purely out of the donor’s desire to enhance his own reputation in the community. We might conclude that such an action was beneficial, and perhaps even that it was charitable, but it would be misleading to say that the action was compassionate. The concept of compassionate action is psychologically thick. To call an action compassionate is to say that the agent was moved to act, at least in part, via his feeling sorry for those who were suffering. (Note that this falls short of saying that this agent is a compassionate person, for it is possible that someone could perform an uncharacteristically compassionate act.) Advocates of psychologically thick accounts of evil action think that, in this respect, evil is like compassion. They judge that some extreme culpable wrongs fall short of being evil actions because they were not performed from the distinctive motives or with the distinctive accompanying psychological states that are required for evil action. Unfortunately, though, advocates of psychologically thick accounts of evil action disagree amongst themselves as to which motives or psychological states count as the psychological hallmark of evil action. Advocates of psychologically thick accounts of evil actions face a choice between two options. The first option is to claim that the psychological hallmark of evil action is a motive or emotion that is a familiar part of folk psychology. For instance, we might judge that all evil actions require some combination of perverted sadism, malice, or knowing defiance of morality. These are the kind of psychological states that Arendt originally had assumed that she would discern in Eichmann, when she supposed that he was a monstrous or demonic evildoer. If, contrary to Arendt, we decide that any of these conditions are necessary for evil action, then we should reject intuition (6), that evil actions can be banal. For want of a better label, I will call such theories folk versions of the psychologically thick account of evil action. The second option is to claim that evil actions can be banal, in the sense of not being malicious, sadistic, or defiant, but that evil actions nonetheless are distinct from ordinary wrongs in virtue of possessing some more complex and less familiar psychological property; a property for which we do not possess a ready-made folk-psychological label, and a property that may be harder to detect. I will call such theories non-folk versions of the psychologically thick account of evil action. Several distinct psychologically thick accounts of evil action have been advocated in recent philosophical literature. However, the advocates of these accounts

80  Thick and Thin Accounts of Evil Action typically do not contrast their own favoured version of a thick account with the alternatives, nor do they adequately explain why a thick account is preferable in the first place. In the remainder of this chapter, I will attempt to set out in detail the possible versions of thick accounts of evil action, before moving on to evaluate them in the following chapters.

4.4  Folk Thick Accounts of Evil Action: Malice Let us proceed by considering the three common psychological conditions that are put to use in folk versions of psychologically thick accounts of evil action. The first of these is the motive of malice. In everyday contexts “malice” can mean ill will, or the intent to do wrong or to inflict harm. I wish to give the word “malice” a more precise meaning in the context of this discussion. I will take “malice” to mean the desire that another suffer harm, where this is desired for its own sake, and not merely as a means to another separate end (cf. McGinn 1997, 63). A caring doctor who jabs you with a much-needed immunization shot thereby deliberately inflicts upon you the harm of a minor pain, but she does not do so maliciously. She harms you only as a means to securing some other end; namely, the end of protecting your health. In contrast, a envious rival who embarrasses you in public does so maliciously, in that he desires your suffering regardless of whether it helps him to achieve some other separate end. What I am calling malicious infliction of harm is often described as gratuitous infliction of harm, or harm inflicted for its own sake. It might be tempting to think that malicious actions are those in which the agent intends to do wrong under that description, or knows that what she is doing is wrong, but this thought would be mistaken.3 Acting out of malice is compatible with knowingly defying morality, but it does not imply knowing defiance. A person who maliciously inflicts harm on another might judge that the victim deserves to suffer that harm, or that the victim does not possess the moral status that would give the agent a reason not to inflict such a harm. For instance, someone could maliciously attack a prostitute without knowingly defying morality, if he falsely believes that prostitution warrants such harm as punishment. Similarly, someone could maliciously kick a cat without knowingly defying morality, if he falsely believes that cats do not have a moral status that gives us a reason not to harm them

3  For instance, Rosenbaum contrasts the view that Hitler was a “vicious cold-blooded hater who counterfeited a mask of rectitude” with the view that Hitler believed his actions were right (Rosenbaum 1998, 212). Of course, it is possible for someone to be a vicious cold-blooded hater who also believes that his actions are right.

Thick and Thin Accounts of Evil Action  81 gratuitously. According to my definition of the term, some malicious acts might even be morally right. In those cases in which someone deserves to be harmed as part of retributive punishment, the person who inflicts the harm might do so maliciously, but do what is right nonetheless (cf. Calder 2013, 183–4). Moreover, an action can be malicious, in this sense, even though it is not pleasurable to the agent. The infliction of what is believed to be justly deserved punishment can be performed out of a sense of duty, and might elicit nothing more than grim satisfaction on the part of the punishing agent. Some types of wrong action, such as slander, are usually performed maliciously. Yet it is crucial to notice that malice, as I am defining it, is absent from the many actions in which agents wrongly inflict harm for purely instrumental reasons, or as a foreseen but unintended by-product of the pursuit of another end. A thief who steals a car thereby inflicts some harm on the owner of the car, but it is rare for such thefts to be motivated by malice towards the victim. What the thief really wants is to possess the car, not to harm the owner of the car. The thief harms the owner knowingly and wrongly, but not maliciously. Someone who gives a sedative to an agitated child merely in order to quiet her, despite knowing that the sedative is harmful to children, has also acted wrongly but not maliciously. It is plausible that the presence of malice in an action renders that action morally worse than it otherwise would have been, ceteris paribus. In other words, malice is an exacerbator of a wrong. There are many real-life cases in which it is difficult to determine whether an extreme wrong was done purely out of malice, or purely instrumentally, or both out of malice and for instrumental reasons. The dispute between the historians Browning and Goldhagen is, in part, a disagreement over whether the huge majority of the so-called ordinary Germans who carried out the mass killings of Jews during the Second World War were motivated primarily by a deep-seated anti-Semitism, or whether many of the killers in fact did not bear strong ill-will towards Jews, but were motivated instead by situational factors or for instrumental reasons (Goldhagen 1997; 414–18; Browning 1998, 202–16) Regardless of how we resolve that particular dispute, it is obvious that there are plenty of examples of extreme culpable wrongs that are not malicious. Many of the murders that take place in the illegal drug trade are cold-blooded, instrumental crimes rather than malicious actions motivated by hatred of the victims. The perpetrators kill their rivals merely to secure a larger piece of the action, and to build a fearsome reputation. Similarly, a bank robber who shoots a teller merely as a means to get the money has not acted maliciously. Nor has an extortionist who tortures his victim merely in order to obtain useful information. Of course, the fact that their actions are instrumentally motivated does not imply that they are not extremely wrong.

82  Thick and Thin Accounts of Evil Action We have seen that malice is a motive that is distinct from sadistic pleasure and from knowing defiance, and that is present in some but not all cases of extreme wrongdoing. How does this bear on the nature of evil action? Arendt and her followers believe that the purely instrumental, non-malicious infliction of harm could be evil. In contrast, some people refuse to call an action evil if they judge that it was not performed maliciously. The view that malice is a necessary condition for evil action has been endorsed by some of the philosophers who recently have written about evil. Kekes, for instance, claims that evildoers cause more serious harm than is needed for achieving their [other] ends. They are not just unscrupulous in their choice of means, but motivated by malevolence to gratuitous excess. They treat their victims with ill-will, rage, or hatred. (Kekes 2005, 2)

Similarly, Laurence Thomas maintains that an evil action must stem from the evildoers’ hostile feelings towards the victim (Thomas 1993, 76). Manuel Vargas claims that in performing an evil action the agent “desires to see other people harmed for no reason beyond the desire itself ” (Vargas 2010, 75). Peter Dews also identifies this as a common view: To do evil, as the term is often understood these days, is to do more than pursue one’s self-interest, even by morally unacceptable means. It is to be involved in some wilfully pain-inflicting, destructive, and—often—self-destructive enterprise, to be driven by forces that lie deeper than the familiar repertoire of unappealing human motives, such as greed, lust, or naked ambition. (Dews 2008, 4)

The view that malice is a necessary condition for evil action clearly has some support.

4.5  Folk Thick Accounts of Evil Action: Sadistic Pleasure The second of the folk psychological states that some people take to be necessary for evil action is sadistic pleasure. Sadistic pleasure, as I will use the term, is pleasure taken in the infliction of suffering on another, but also pleasure taken in the contemplation of harm suffered by another, no matter how the harm was inflicted, or by whom or what it was inflicted. Sadistic pleasure is experienced by a workplace bully who enjoys harassing his colleagues, and by a silent voyeur who does none of the harassing himself, but enjoys watching others suffer. To take pleasure in causing or observing an event is not merely to desire that the event occur, but to experience a qualitative enjoyment in contemplating or witnessing its occurrence. Someone might feel grim satisfaction but no pleasure in dutifully performing a desired but difficult or repulsive task. It is clear that a wrongdoer could take

Thick and Thin Accounts of Evil Action  83 sadistic pleasure in the malicious infliction of harm, but it is also possible for a wrongdoer to take sadistic pleasure in the non-malicious, purely instrumental or unintended infliction of harm. For instance, an extortionist who is motivated by a purely instrumental desire to extract valuable information from his victims might get an unexpected thrill from dominating and frightening his victim. Sadistic pleasure is separable not only from malice, but also from defiance. A wrongdoer could take sadistic pleasure in performing actions that he mistakenly believes to be morally right, as well as in performing actions that he knows to be morally wrong (cf. Zangwill 2003, 96). In either case, sadistic pleasure, like malice, seems to be an exacerbator of a wrong. Contrary to the claims of some utilitarians, the fact that the torturer enjoys inflicting suffering makes his action morally worse, not morally better (Hurka 2001, 138). Some of the most prominent examples of purportedly evil figures in our cultural history are portrayed as taking sadistic pleasure in wrongdoing. In Milton’s Paradise Lost, Satan not only tempts humans to their doom, but delights in having done so, rushing back to Hell “ With joy / And tidings fraught” (Milton 1993, 546 (Bk X, 345–6)) to boast gleefully to his minions that he has corrupted Adam (Milton 1993, 552 (Bk X, 485–93)). God subsequently punishes Satan and his minions in order to “dash thir pride, and joy for Man seduc’t” (Milton 1993, 556 (Bk X, 577)). If Satan is our paradigmatic evildoer, then it is no surprise that stereotypical evil villains—The Wicked Witch of the West in The Wizard of Oz, Mr Burns in The Simpsons, Dr Evil from the Austen Powers films—are portrayed cackling maniacally when contemplating their evil deeds. Not all such examples of pleasurable extreme wrongdoing are fictional. We should not forget that some serial killers and torturers report getting a buzz or a sexual thrill from their crimes (Stone 2009, 219; Waller and Deal 2010, 5), and many atrocities that are perceived by their perpetrators to be acts of revenge are pleasurable for that reason (Stone 2009, 124–5). No doubt struck by examples like these, some people think that pleasure in extreme wrongdoing is the hallmark of evil action. The psychologist Alford, in a study of prison inmates, reported that most of the prisoners he interviewed said that “[e]‌vil is pleasure in hurting and lack of remorse” (Perrett 2002, 306). Several philosophers propose that sadistic pleasure is a necessary condition for evil action. For instance, Perrett claims that when an action is evil “the agent will take pleasure in the wrongness of the action” (Perrett 2002, 304). McGinn suggests that an evil person is one who derives pleasure from others’ pain and pain from others’ pleasure, and he appears to believe that sadistic pleasure is the hallmark of evil action (McGinn 1997, 62, 69). Hillel Steiner maintains that “[e]‌vil acts are wrong acts that are pleasurable for their doers” (Steiner 2002, 189), and suggests

84  Thick and Thin Accounts of Evil Action that this fact explains why we jokingly say that morally benign indulgent pleasures, such as eating rich chocolate, are evil (Steiner 2002, 183). Similarly, Laurence Thomas states that an act is evil only if the agent “delights in performing a harmful act that has a certain moral gravity to it” (Thomas 1993, 77).

4.6  Folk Thick Accounts of Evil Action: Defiance The third of the folk psychological conditions that might be judged to be necessary for evil action is defiance of morality. A wrongdoer defies morality when she not only does what is actually morally wrong, but when she does what she knows is morally wrong.4 It might be argued that it is impossible to do what one judges to be wrong. According to this view, if a person S deliberately performs an action ø, then S at the time of acting judged that, out of the available options, ø-ing was the right thing to do. If S had judged that it was right to ψ rather than to ø, then S would have chosen to ψ instead (Hare 1952, 169). According to this view, all evildoers must have believed, at the time of acting, that what they were doing was right. This view is highly contentious, though, if we consider the cases of what philosophers call akrasia, or weakness of will. On some occasions it seems that we do things that at the time of acting we know to be wrong. It could be that we are overcome by emotion, lose control, and hence behave irrationally (Aristotle 1985 NE, Bk VII), but perhaps there are times when we calmly and rationally choose to do what we judge to be clearly selfish and wrong (Austin 1979, 198). However, the possibility of this kind of akrasia remains under dispute. It could be argued that the cases in which we seem to have acted against our best judgement are merely cases in which, with the benefit of hindsight, we regret and retrospectively disown the short-lived normative judgements that guided our actions. Even if it were true that akrasia never occurs, and hence that S’s deliberate action always aligns with S’s all-things-considered judgement at the time of acting as to what ought to be done, that would not imply that every one of S’s deliberate actions was judged by S at the time of acting to be morally right. It is common for people to judge that it would be morally wrong to ø but that, all things considered, it would be right to ø. It seems that some sociopaths are able to make correct moral judgements, but just do not care about morality at all, and act contrary to their moral judgements whenever it suits their purposes (Morton 2004, 52; Nichols 2004, 72).

4   In defining defiance of morality, I do not mean to place special weight on moral knowledge as opposed to correct moral judgement. Someone who correctly believes that øing is wrong, but lacks justification for believing that øing is wrong, might be said to lack moral knowledge that øing is wrong. Nonetheless, I intend to allow that such a person would knowingly defy morality if she performed ø.

Thick and Thin Accounts of Evil Action  85 Such people are comparatively rare, and there is some dispute about whether their claims about which actions are morally right or wrong are expressions of genuine moral judgements (Smith 1994, 67; Nichols 2004, 73). However, all of us are familiar with specific cases in which non-psychopathic people who do care about morality to some degree judge that moral reasons are trumped by other reasons: by self-interest, by nepotism, or by aesthetic values. Regardless of whether akrasia is possible, defiance of morality is possible. Some of the most striking and most repulsive cases of extreme wrongdoing are those in which the wrongdoer knowingly defies morality. Some people are moved to do what is wrong precisely because it is wrong. They have a de dicto desire to do what is morally wrong; that is, a desire to perform the action that is wrong, whichever action that happens to be. I will refer to this kind of motive as delinquent defiance. Delinquent defiance is memorably embodied in the biker played by Marlon Brando in The Wild One, who, when asked what he is rebelling against, replies “Whaddaya got?”. An earlier fictional delinquent can be found in Milton’s Paradise Lost. Milton depicts Satan lashing out against the constraints of a higher power, and proclaiming that punishment is preferable to subservience: . . . and in my choyce To reign is worth ambition though in Hell: Better to reign in Hell, than serve in Heav’n (Milton 1993, 128 (Bk I, 261–3))

Given his commitment to this inverted set of values, Satan resolves to do wrong because it is forbidden: To do ought good never will be our task, But ever to do ill our sole delight, As being the contrary to his high will Whom we resist. (Milton 1993, 123 (Bk I, 159–62))

The principled nature of Satan’s defiance is captured is his slogan “Evil be thou my good” (Milton 1993, 251 (Bk IV, 110)). Shakespeare presents us with another moral delinquent in the character of Aaron in Titus Andronicus: lucius:  Art thou not sorry for these heinous deeds? aaron:  Ay, that I had not done a thousand more. Even now I curse the day, and yet I think Few come within the compass of my curse, Wherein I did not some notorious ill: As kill a man, or else devise his death;

86  Thick and Thin Accounts of Evil Action Ravish a maid, or plot the way to do it; Accuse some innocent, and forswear myself; Set deadly enmity between two friends; Make poor men’s cattle break their necks; Set fire to barns and haystacks in the night And bid the owners quench them with their tears. Oft have I digged up dead men from their graves And set them upright at their dear friends’ door Even when their sorrows almost was forgot, And on their skins, as on the bark of trees, Have with my knife carved in Roman letters “Let not your sorrow die, though I am dead.” But I have done a thousand dreadful things As willingly as one would kill a fly, And nothing grieves me heartily indeed But that I cannot do ten thousand more. (Titus Andronicus V.i.123–144) Not all examples of delinquent defiance are fictional. In his Confessions, Augustine categorizes his own sinful actions as delinquent: Yet I chose to steal, and not because want drove me to it—unless a want of justice and contempt for it. . . . For I stole things which I already had in plenty and of better quality. Nor had I any desire to enjoy the things I stole, but only the stealing of them and the sin. (Augustine 1951, 24 (Confessions, Bk 2, IV))

Augustine elaborates: Let that heart now tell You what it sought when I was thus evil for no object, having no cause for wrongdoing save my wrongness. The malice of the act was base and I loved it— that is to say I loved my own undoing, I loved the evil in me—not the thing for which I did the evil, simply the evil: my soul was depraved, and hurled itself down from security in You into utter destruction, seeking no profit from wickedness but only to be wicked. (Augustine 1951, 24 (Confessions, Bk 2, IV))

A similar kind of delinquent defiance appears to be one of the motives lying behind the behaviour of those real-life serial killers, including the Beltway Snipers and Dennis Rader, who taunt the police with anonymous notes describing their actions, glorying in their own violation of morality (Morrow 2003, 36–7). There are plenty of cases of defiance of morality that are not delinquent. Many people perform actions that they know to be wrong not because those actions are wrong, but for some other reason. For instance, wrongdoing that is motivated purely by a desire for financial gain counts as defiant when the wrongdoer also

Thick and Thin Accounts of Evil Action  87 knows that his action is morally wrong. When such wrongdoing is extreme we would describe it as ruthless or callous defiance of morality, rather than delinquent defiance. Defiance of morality, whether delinquent or not, is separable from malice and from sadistic pleasure, although sometimes it is accompanied by either or by both. The fact that well-known examples of extreme wrongdoers, including Milton’s Satan, not only violate but defy morality helps to explain why some people think that defiance of morality is a necessary condition for evil action. According to this view, the fact that the agent believed his action to be morally right implies that the action was not evil. An action’s being performed because of moral ignorance or moral error on the part of the agent is, on this account, incompatible with that action being evil. Claudia Card claims that the view that evildoers “do wrong wilfully” and not out of ignorance is a “popular modern stereotype of evildoing.” (Card 2002, 52). There is some evidence that Card is right about this. For instance, the talk-show host David Letterman appeared to express this view during a discussion about the United States invasion of Iraq with Bill O’Rielly on The Late Show on 30 October 2006: o’rielly:  Do you actually think that we are a bad country for doing what we did, and Bush is an evil man? Is that what you are putting forward? letterman:  I don’t know that I think he is evil. I think he is misguided. Letterman seems to be implying that only wrongdoers who know that they are doing wrong, and hence who are not merely misguided, can be evil. If this claim about evil persons carries through to evil actions, then defiance is necessary for evil action. This appears to be accepted by Christopher Hitchens, who claims that evil is “the surplus value of the psychopathic—an irrational delight in flouting every customary norm of civilization” (Hitchens 2002). Ron Rosenbaum also suggests that evil “in the strict sense [is] doing wrong knowingly” (Rosenbaum 1998, xxii, 214). While Card rejects this “popular modern stereotype”, some other philosophers have included defiance as a necessary condition in their accounts of evil action. In A Theory of Justice John Rawls compares bad and unjust men, who do wrong in order to secure wealth or social power, with the evil man: By contrast, the evil man aspires to unjust rule precisely because it violates what independent persons would consent to in an original position of equality, and therefore its possession and display manifest his superiority and affront the self-respect of others. . . . What moves the evil man is love of injustice: he delights in the impotence and humiliation of those subject to him and he relishes being recognized by them as the willful author of their degradation. (Rawls 1972, 439)

88  Thick and Thin Accounts of Evil Action From Rawls’ claims about evil personhood, it is a short step to the idea that every evil action is a case of defiant extreme wrongdoing. More recently, Marcus Singer includes defiance as a necessary condition in the kinds of action that he thinks are evil, and maintains that horribly wrong actions that were mistakenly believed by the wrongdoers to be right are not really evil (Singer 2004, 205). Roy Perrett goes so far as to suggest that delinquent defiance is necessary for evil action, claiming that an action is not evil unless the action is “wrongdoing which is done because it is wrong” (Perrett 2002, 304). The thought expressed by these philosophers is not that an agent’s believing that his action is right counts as an excuse, and hence that defiance of morality is a necessary condition in any case of culpable wrongdoing (pace Rosenbaum 1998, 210). Rather, it is that defiance is an exacerbator of culpably wrong actions, and is a necessary condition for the subclass of culpable wrongs that count as evil. We have seen that there are three distinct folk psychological states—malice, sadistic pleasure, and defiance—that some philosophers have added as necessary conditions for evil action. Of course, it would be possible to propose that other folk psychological states, such as lack of remorse, are necessary conditions for evil action. As far as I am aware none of these alternatives have philosophical advocates, and it is not clear that any deserve further consideration.5 It is important to note that advocates of the folk psychologically thick accounts disagree amongst themselves as to which out of the three psychological states are necessary for evil action. Perrett maintains that all three are necessary, and Rawls’ claims about the evil man also suggest that malice, sadistic pleasure and defiance are characteristic features of evil actions. In contrast, Thomas and Steiner hold that out of these three psychological conditions, only sadistic pleasure is necessary for evil action. McGinn’s account of evil personhood suggests that malice and pleasure are necessary for evil action. Hitchens says that sadistic pleasure and defiance are necessary, Kekes claims that only malice is necessary,6 and Singer thinks that only defiance is necessary.

  Alford’s survey of prison inmates might suggest that lack of remorse by the agent is also a necessary condition for evil action. However, such a claim would have counterintuitive consequences, given the retrospective nature of remorse. For instance, if feeling remorse about having done ø were sufficient for ø’s not being an evil action, then it could be the case that ø now counts as an evil action but in future might no longer be an evil action, because the agent might come to regret it. As we shall see, remorse does play an important role in discussions about the nature of evil personhood, but it is not plausible that lack of remorse it is a necessary condition for evil action. 6   Note that Kekes earlier claimed that malice is not necessary for evil action (Kekes 1990, 87). 5

Thick and Thin Accounts of Evil Action  89

4.7  Thick Accounts of Evil Action: Non-folk Versions Arendt’s followers claim that evil actions can be banal; that is, they can be performed without malice or sadistic pleasure, and without knowing defiance of morality. The psychologically thick accounts of evil action that we have considered so far suggest that Arendt was wrong to claim that evil is banal. It is possible, though, to construct a theory of evil action which allows that evil actions could be banal in Arendt’s sense, but which implies that evils are psychologically distinct from ordinary wrongs nonetheless. According to such non-folk thick accounts, the psychological hallmark of evil action is some motivational state that is not part of Arendt’s rough list of monstrous motives, but is another, more complex and less familiar state. Adam Morton appears to offer one such account. He says that a basic desideratum of a theory of evil is that it allows for banality of motive, claiming that “many participants in evil are not moved by extraordinary hatred or sadism” (Morton 2004, 8). Yet Morton also maintains that “not all seriously wrong acts should be counted as evil”, and that the label “evil” tells us something about the distinctive kind of motive that produced the action (Morton 2004, 30, 64). Morton suggests that only culpable wrongs could count as evil actions,7 and then adds the following condition: A person’s act is evil when it results from a strategy or learned procedure which allows that person’s deliberations over the choice of actions not to be inhibited by barriers against considering harming or humiliating others that ought to have been in place. (Morton 2004, 57)

It is not entirely clear what would count as a “strategy or learned procedure”, but Morton’s posited barrier is supposed to be a violence-inhibiting mechanism that is an “inbuilt piece of mental machinery” which responds to signs of distress by suppressing impulses to perform violent actions (Morton 2004, 42). If some but not all extreme culpable wrongs are the product of this kind of motivational structure, then it looks like Morton’s account will fit with intuition (6), that evil actions can be banal, and with intuition (7), that there is a psychological hallmark of evil action. Another theory of this kind is offered by Eve Garrard, who locates the psychological hallmark of evil action not in malice, sadistic pleasure, or knowing defiance, but in a certain feature of the evildoer’s deliberation. According to Garrard’s 7   Morton claims that wrong actions performed out of negligence are not evil, so long as the negligence was not reckless and persistent. He also claims that wrong actions performed out of a combination of good motives and ignorance are not evil (Morton 2004, 10–11). These are standard conditions for sorting culpable from non-culpable wrongs.

90  Thick and Thin Accounts of Evil Action first formulation of her account, an evil action is a wrong action in which the agent silences any moral considerations against performing that action. Garrard claims that a moral consideration has been silenced by an agent when the agent is aware of the consideration, but does not take it to have any moral weight. Thus a consideration of which the agent remains ignorant is not a silenced consideration, although both of these kinds of consideration fail to feature in that agent’s deliberation (Garrard 1998, 49–53). Non-evil wrong acts, Garrard claims, are those in which the moral considerations that count conclusively against the action are recognized as having some weight, but are undervalued by the deliberating agent, and hence outweighed by other considerations in the process of deliberation. In contrast, the “evil act is done by one for whom (at least some of) the considerations that tell against his committing this wrongful act are silenced altogether” (Garrard 1998, 53–4). However, Garrard notes that this first formulation of her account would implausibly imply that trivial wrong actions are evil if the agent silenced the considerations that count against them. I think that Garrard could solve this problem by claiming that an action can be evil only if it is appropriately connected to sufficiently extreme harms. Yet Garrard is also committed to intuition (8), that evil actions are qualitatively distinct from ordinary wrongs (Garrard 2002, 321), and if an evil action were simply an extreme culpable wrong in which the agent silences considerations against performing the action, it would turn out that evil actions could be merely quantitatively rather than qualitatively distinct from ordinary wrongs (Russell 2007b, 672–3). Instead of tacking on an extremity condition, Garrard adds two further conditions to her initial account of evil action, producing a more elaborate account of what she sees as the psychological hallmark of evil. The first additional condition concerns what Garrard calls “metaphysical silencers”, by which she means considerations for action which should silence competing considerations for action. In this version of her account, Garrard claims that evil actions are those in which the moral considerations against the action which are silenced by the agent are themselves considerations that should silence the considerations for which the agent acts (Garrard 1998, 54). When performing an evil action the agent psychologically silences reasons that are themselves metaphysical silencers. Since it does not seem merely trivially wrong to ignore these powerful metaphysical silencers, it looks like this move will allow Garrard to rule out the possibility of trivial evils. As Garrard points out, though, considerations that in some contexts should be silenced provide good reasons for action in other contexts. Thus it cannot be that evil actions are those performed on the basis of

Thick and Thin Accounts of Evil Action  91 reasons that always should be silenced. Garrard responds to this problem by adding to her account a second condition, which recognizes the context-sensitivity of metaphysical silencers. According to the final version of Garrard’s account, an act is evil if it is wrongful, and if the agent silences the reasons against doing the act, which reasons are themselves metaphysical silencers, and where the agent’s reasons for doing the act are members of the class of considerations which are in this case metaphysically silenced. (Garrard 1998, 55)

Since it is possible for an action to meet Garrard’s psychological condition without being malicious, sadistically pleasurable, or knowingly defiant, Garrard’s account is also compatible with intuition (6), that evil actions can be banal, and with intuition (7), that there is a psychological hallmark of evil action. This brings us to the end of our survey of the extant theories of evil action, which can be grouped as follows: Psychologically thin account

Psychologically thick folk account

Psychologically thick non-folk account

Arendt Card Calder Neiman Formosa

Kekes Thomas McGinn Steiner Perrett Singer

Garrard Morton

Stephen de Wijze offers a disjunctive theory of evil action that does not fit neatly into this table. According to de Wijze, an act is evil if it is a wrongful action that meets at least one of the following three conditions: A. There is a deliberate violation of persons with the intention to dehumanise (that is, deny basic respect and dignity to) those powerless to retaliate. B. The action or project will gratuitously inflict, or bring about, one or more of ‘the Great Harms’ to sentient beings with the relevant moral standing. C. The action or project (or professed morality) seeks to annihilate the ‘moral landscape’ (de Wijze 2002, 218).

This theory is difficult to classify because Conditions A and B are psychologically thick, whereas Condition C appears to be psychologically thin. It is tempting to read de Wijze’s Condition C such that it renders Conditions A and B superfluous: any extreme culpable wrong counts as an evil action, regardless

92  Thick and Thin Accounts of Evil Action of whether it was performed with an intention to dehumanize the victims, and regardless of whether it was gratuitous, because every extreme culpable wrong seeks to undermine the moral landscape (Russell 2007b, 666–70). In this case, de Wijze would belong in the first column of the table above, along with Arendt and the advocates of the other psychologically thin accounts. Alternatively, if “seeking to undermine the moral landscape” is a distinctive psychological state, then de Wijze belongs in the third column with Morton and Garrard.8 We have seen that beyond the widely shared view that all evil actions are extreme culpable wrongs there is much disagreement over the nature of evil action. In the next chapter, I will begin to evaluate these rival theories.

  By the “moral landscape”, de Wijze means

8

those prerequisite values needed for any civilised attempt to manage conflict and to establish a minimal level of respect and dignity between persons . . . [and those values] which seek to protect the weak from the strong and prevent a world where needless pain, suffering and death are seen as preferable to joy, happiness and life. (de Wijze 2002, 221) I t is not entirely clear what de Wijze means by an action that “seeks to annihilate” the moral landscape, but his model for such actions are those performed by participants in the Nazi regime, which can be banal in Arendt’s sense (de Wijze 2002, 221).

5 Assessing Thick and Thin Accounts Advocates of psychologically thin accounts of evil action maintain that all evil actions are extreme culpable wrongs, and that only evil actions are extreme culpable wrongs. In contrast, advocates of psychologically thick accounts of evil action claim that there is a psychological hallmark of evil action, and that some extreme culpable wrongs fall short of being evil because they lack this distinctive psychological feature. There is further disagreement as to what this extra necessary condition for evil action could be. Some say that it is a motive or psychological state that is a familiar part of folk psychology, such as malice, sadistic pleasure, or defiance. Others propose that it is a more complex psychological condition that is harder to detect. In this chapter, I hope to explain why there is such disagreement on this issue, and to identify which, if any, of these competing accounts of evil action gets at the true nature of evil. I will argue that the non-folk versions of the psychologically thick account are not plausible, but that both the psychologically thin account and the folk versions of the thick account fit reasonably well with folk usage of the concept of evil. Moreover, appeals to natural kinds do not give us grounds to favour any one of these accounts over its rivals.

5.1  Evil and Folk Talk about Evil One important constraint on philosophical analyses is the folk use of the term being analysed, or the folk intuitions concerning that term (Jackson 1998, 118; Sider 2001, 10). For instance, a philosopher who gives a theory of free will is obliged to explain how that account connects with the things that a broad range of non-philosophers say about freedom, and the role that thoughts about freedom play outside of philosophy. Freedom is something that people value, freedom is something that people feel is undermined when they are coerced or threatened,

94  Assessing Thick and Thin Accounts freedom is something that people think that they exercise when they choose from a range of good options, and so on. A philosopher who offers an account of free will according to which, for instance, coercion has no impact on free choice will be accused of changing the subject; that is, of purporting to give an account of free will, but actually giving an account of something else (Jackson 1998, 31). This is not the only kind of constraint on philosophical theories, and there may be cases in which there are decisive reasons for overriding some common intuitions and some aspects of folk usage. Nevertheless, if compatibilist theories fit better than libertarian theories with folk talk about freedom, then we have a pro tanto reason to favour compatibilism. Perhaps we can find similar grounds for favouring one of the thick or the thin accounts of evil action by showing that, compared to its competitors, it fits more neatly with folk talk about evil. As we saw in Chapter One, there are two aspects of folk thought about evil that could constrain philosophical analyses of the concept of evil. The first is a set of intuitions and linguistic practices concerning the extension of evil. We can get at these intuitions and practices by asking which particular actions the folk think are evil, and which actions they call “evil”. The second are a set of intuitions and linguistic practices concerning the inferential and functional role of evil. We can get at these intuitions and practices by asking what the folk conclude from the fact that an action is evil, and what they do subsequent to having judged that an action is evil. I have already used a combination of the extensional and the inferential methods to argue against the suggestion that evil actions are, by definition, incomprehensible wrongs or life-wrecking wrongs. I have also used the extensional method to rule out some theories of evil action; namely, McGinn’s and Steiner’s sadistic accounts, which imply that there could be evil actions that are merely trivially wrong. It is true that McGinn and Steiner pick out an interesting and important category of actions, but it is going too far to say that this category deserves the label “evil”. Even if the folk agree that sadistic pleasure makes a wrong action even more wrong, they would not say that sadistically pleasurable shoplifting is evil. The question at hand is whether we could use these methods to argue for the psychologically thin account of evil action, or for one of its thick rivals. Could we test theories of evil action by identifying counterexamples that are backed by widely shared folk intuitions? For example, could we find examples of actions that are not malicious, sadistic, or defiant, but intuitively are evil nonetheless, or actions that are extreme culpable wrongs in which the agents mistakenly believed that they were doing right, and that intuitively are not evil? The first task for those who employ this method is to determine what the folk actually think about evil. We have already seen many examples in which non-philosophers make public pronouncements about which actions are evil,

Assessing Thick and Thin Accounts  95 and what evil actions are supposed to be. It is possible that the journalists, psychologists, historians and politicians who make these pronouncements are an unrepresentative sample of the population as a whole. It is also possible that these public pronouncements do not reflect the detailed intuitions of the people who make them. An appeal to folk usage ideally would be backed by empirical research designed to tease out the intuitions of a broad range of people in this area. As far as I am aware, such research has not yet been undertaken. However, even if this research had been carried out, it is likely that a more significant difficulty would remain. When we look at the public pronouncements that non-philosophers have made concerning evil we discover very little agreement beyond the assumption that all evil actions are extreme culpable wrongs. This can be seen not only in the rough definitions of evil action that the folk offer, but also in their claims about the extension of the concept. Even on the question of whether Hitler’s actions were evil—perhaps the most obvious place to look for shared intuitions—there are plenty of people on either side, as Ron Rosenbaum discovered in the interviews he conducted for Explaining Hitler (Rosenbaum 1998, xxi). It is not that some of Rosenbaum’s interviewees judge that Hitler’s actions are wrong but others see them as being right or excusable. All of them think that Hitler’s actions are extreme culpable wrongs. They disagree over whether they are evil (cf. Cole 2006, 236). It is possible that future empirical research into intuitions about evil could find a hidden unity in folk thought about evil actions which would strongly favour either a thick or a thin account, but at this stage it seems more likely that folk intuitions will turn out to be just as scattered as the existing public pronouncements. If the folk don’t agree on which particular actions are evil, and if they don’t agree as to whether there is a psychological hallmark of evil action, then it is not clear how we could assess competing thick and thin accounts by appealing to folk intuitions. What strikes one person as a case of an evil action that is not defiant could easily strike another as being an extreme wrong that falls short of evil precisely because it is not defiant. When folk intuitions are scattered in this way, argument by counterexample loses its force. Even if there is much disagreement amongst non-philosophers in their usage of the word “evil”, it will be advantageous for an advocate of any particular philosophical account to show that at least some of the folk deploy “evil” in something like the way that her favoured account would recommend. As I have pointed out in the previous chapter, some non-philosophers say that evil action requires malice, or sadistic pleasure, or defiance, and the philosophers who advocate folk versions of the psychologically thick account can appeal to these folk uses to bolster their respective cases. While these advocates of thick accounts can muster some folk support for their respective views, there remains something troubling about the

96  Assessing Thick and Thin Accounts diversity of conditions that have been proposed as the psychological hallmark of evil action. There is comparatively broad agreement amongst the folk as to the psychological hallmark of cowardly action, or of dishonest action (although, admittedly, philosophers do argue over the fine-grained details in these cases). If evil action really is a psychologically thick concept, then we need to explain why both the folk and the philosophers propose a range of quite different conditions as the psychological hallmark of evil, and why some even reject the idea that there is such a hallmark. One possible explanation for the diversity of folk opinion on this issue is that there are distinct cultural traditions that influence how people think about evil, and how people use the word “evil”. Speculation about these influences and their possible outcomes is easy enough. It could be the case that people whose moral outlook is shaped by Christianity, in which the archetypal figure of Satan acts maliciously, sadistically and defiantly, are more likely to judge that some or all of these psychological conditions are required for evil action (e.g. Rosenbaum 1998, 291). It is also possible that Jews and Germans, for whom the Holocaust looms large, are more likely to use “evil” in a way that fits with psychologically thin accounts, according to which evils can flow from indifference or a misguided sense of duty, as well as from malicious defiance. Perhaps members of East Asian cultures, in which characterological explanations of actions are comparatively less prevalent, would be less likely to judge that there is a psychological hallmark of evil action (cf. Doris 2002, 105). On a more fine-grained level, it could be that people who watch horror movies or read fantasy fiction are more inclined to believe that the word “evil” has supernatural connotations. If hypotheses such as these are correct, then we should be able to discover stable differences in linguistic practice between these groups, and we could expect members of the same cultural group to be more likely to use the word “evil” in similar ways. The discovery of such patterns might lead us to conclude that there is not a single unified concept of evil that is used by the folk, but that there are several similar but distinct folk concepts, all of which go by the name “evil”. Advocates of the psychologically thin account are likely to offer an alternative, less flattering, explanation of the diversity that is apparent in folk thought about evil. It could be the case that there is a single, widely-shared folk intuition that underlies most uses of the term “evil”, but that this intuition is fairly vague; namely, that evils are culpable wrongs that are somehow worse or more wrong than ordinary wrongs. If the nature of the difference between evils and ordinary wrongs is not specified by the folk intuition, individual speakers who are asked to describe the nature of evil may feel pressed to fill in the details in some way. In doing so, it would be natural to point to something that would make wrong actions worse. There are well-known psychological conditions—malice, sadistic pleasure, and

Assessing Thick and Thin Accounts  97 defiance—which are present in many archetypal evil actions, and which commonly are judged to be exacerbators of wrong actions. People who are grasping for the distinctive feature of evil action are likely to latch onto one of these conditions. However, any one of these psychological conditions makes an action morally worse. The danger is that we will simply latch onto the first of these conditions that we notice as being present in the examples that are ready to hand, and declare that this condition is what tips an action out of the category of the merely wrong and into the category of evil. Someone who first notices the anti-Semitic malice that drove Hitler’s actions, for instance, might then conclude that malice is a necessary condition for evil action, whereas someone who first notices Dennis Rader’s sadistic pleasure could then conclude that sadistic pleasure is necessary for evil. If this is the case, then we would end up with a number of fairly precise thick accounts of evil action, none of which conform closely to the comparatively loose folk thought about evil. Rather than matching well-established patterns of folk usage of the term “evil”, each of the various thick accounts constitutes a stab in the dark that goes beyond the vague intuition that evils are culpable wrongs that are extreme in some way. The more precise the stab, the less likely the resulting account will fit with the wide range of folk thought about evil. Until empirical research is conducted into folk intuitions about evil, we are not in a position to judge whether the diversity in folk thought about evil is due to the influence of several distinct cultural traditions, or is the result of uncoordinated attempts to locate an exacerbator, or has some other explanation. In any case, the diversity of folk claims about evil action suggests that we will not be able to decide between the folk versions of the psychologically thick and the psychologically thin account of evil action merely by appealing to what ordinary people think and say about evil. Since both kinds of account can be made to fit well enough with the intuitions of at least some of the folk, we will have to go beyond folk intuitions in our search for the best theory of evil action.

5.2  Non-folk Thick Accounts and Folk Intuitions Before we move beyond the realm of folk intuition, though, we ought to consider more closely the status of what I  have called non-folk versions of the psychologically thick account of evil action, such as those put forward by Garrard and Morton. It is particularly difficult to fit this kind of philosophical theory with what ordinary people say about evil. According to Morton, A person’s act is evil when it results from a strategy or learned procedure which allows that person’s deliberations over the choice of actions not to be inhibited by barriers against

98  Assessing Thick and Thin Accounts considering harming or humiliating others that ought to have been in place. (Morton 2004, 57)

Garrard claims that an act is evil if it is wrongful, and if the agent silences the reasons against doing the act, which reasons are themselves metaphysical silencers, and where the agent’s reasons for doing the act are members of the class of considerations which are in this case metaphysically silenced. (Garrard 1998, 55)

It might seem that these non-folk thick accounts are more attractive than their folk rivals in so far as they allow us to endorse both intuition (6), that evil actions can be banal, and intuition (7), that there is a psychological hallmark of evil action. Yet if the psychological hallmark of evil is a complex property concerning internal barriers and processes of deliberation, as Morton and Garrard propose, ordinary claims about evil action will turn out to be less straightforward than other philosophers have supposed. For example, let us consider a case in which a mother has murdered her child, and several journalists have described her action as evil. Garrard’s account implies that the action is evil only if the deliberating mother silenced the reasons that count conclusively against even considering doing that act. Yet it seems implausible that the journalists who say that this murder was an evil action are thereby affirming that the murderer silenced rather than outweighed the reasons which count against performing actions of that type. The journalists who make this judgement might not have considered the murderer’s deliberations in any detail, let alone undertaken the more complex task of dividing considerations into those that are metaphysical silencers and those that are not. Similarly, it is possible for some folk who have no knowledge about psychological violence-inhibiting mechanisms to judge that an action is evil. If part of what makes an action evil is that it meets these complex psychological conditions, how can the non-philosophical folk, who know nothing about these conditions, successfully wield the concept of evil? In their own defence, Morton and Garrard could say that they are offering an account of the true nature of evil action, not an account of what non-philosophers who use the phrase “evil action” believe evil to be. They might claim to be offering an account of the property of being an evil action, rather than an analysis of the concept of evil action. In doing so, Garrard and Morton would be favouring what I am calling the extensional method over the inferential method. This kind of approach is permissible and often fruitful in many other cases in which we define a term that has an existing folk usage. As Kripke and Putnam claim, the essential nature of gold is that of having the atomic number 79, even though many of the folk who make judgements about which things

Assessing Thick and Thin Accounts  99 are gold are not aware that this is the case, and certainly need not judge that a substance has the atomic number of 79 when they judge that it is gold (Putnam 1975, 241; Kripke 1980, 116–19; Devitt and Sterelny 1987, 70). The folk recognize that there is a type of metal that is yellow, heavy, and that does not tarnish or corrode, and are able to sort particular objects that are made of this metal from those that are not. Arguably, the folk believe that gold is that stuff, whatever that stuff turns out to be (Braddon-Mitchell and Jackson 1996, 67). It could be that evil action, like gold, has a hidden essence. In this case, Morton and Garrard could claim that the psychological hallmark of evil action is a complex property concerning deliberation, but that many of the folk are unaware of this property, and that folk judgements that an action is evil need not include the judgement that the action has this property. It is possible that the folk are able to sort the evil from the non-evil actions by latching onto a set of manifest properties: they recognize that there is a class of wrong actions that are worse than ordinary wrongs, that are particularly deplorable or horrific. In judging these actions to be evil, the folk judge them to be especially deplorable and horrific wrongs, perhaps, but what really makes these wrongs evil is the fact that they possess the hidden psychological property identified by Morton or Garrard. In order to justify taking this kind of approach, Morton and Garrard would still have to show that their respective accounts fit well enough with folk judgements about evil (cf. Kingsbury and McKewon-Green 2009, 172–4). People are happy to accept the scientists’ claim that gold has a hidden essence because prior to the development of the periodic table and atomic theory there was already broad and stable agreement on which things in the world count as gold and which things do not, and the scientists’ account of the nature of gold did not require significant revision of the folk extension of the concept of gold. Virtually everything that the pre-scientific folk thought was gold turns out to be a substance that has the atomic number 79, and everything that is a substance with the atomic number 79 would have been judged by the pre-scientific folk to be gold (Putnam 1975, 235). If it had turned out that a significant portion of the stuff that the folk judged to be gold had an atomic number of 77, then it would be much harder to argue that having the atomic number 79 really was the essence of gold. As we move away from the domain of the natural sciences it becomes increasingly difficult to argue that the things that the folk talk about have hidden essences of this kind. As Robinson points out, when we speak of abstractions—knowledge, virtue, causality, probability—we cannot anchor our conversation by planting the relevant thing on the table or decanting the relevant stuff into a test tube, saying “This is what we are talking about”. (Robinson 2009, 335)

100  Assessing Thick and Thin Accounts We cannot put a sample of evil into a test tube. Worse than that, there is significant and recalcitrant disagreement over the extension of the concept of evil action. We cannot argue that evil is the hidden property that these things have in common if we cannot settle on which specific things qualify as being amongst these things (cf. Jackson 1998, 131; Finlay 2008, 363). This creates real problems for non-folk accounts of evil action. For instance, Garrard’s account might imply that one act of torture is evil because the torturer silenced the considerations which count against performing that action, but that another act of torture is not evil because the torturer considered the rights of the victim but mistakenly outweighed those rights in his deliberation. When we do not have clear agreement amongst the folk over whether one, both, or neither of these instances of torture are evil actions, it is hard to see why Garrard’s hidden property deserves to be described as the essence of evil action. Even if it is true that Garrard’s hidden property marks out an interesting and important class of actions, we cannot simply stipulate that this is the class of evil actions. A further difficulty for Morton and Garrard concerns the epistemic accessibility of what they claim is the psychological hallmark of evil action. It would be surprising if an account of evil action implied that we are very rarely able to tell whether any particular action is evil. According to the thin account of evil action, an action is evil if and only if it is an extreme culpable wrong. There are some cases in which we cannot discern whether an action is an extreme culpable wrong. When a harmful action was the result of provocation, we might be unsure whether the action was morally wrong, and when we are ignorant of the rational capacities of the agent, we are unable to tell whether the agent is culpable for the action in question. Yet in very many cases we do know enough about an action to judge with confidence that it is an extreme culpable wrong. The folk versions of the thick account of evil action suggest that we need to know more about the agent’s psychological states before we can ascertain whether an action is evil. Usually, though, this is knowledge that we can acquire. In many cases there is clearly observable evidence that indicates that the agent bears malice towards the victim, or takes sadistic pleasure in the action.1 Evidence of defiance in an extreme wrongdoer is harder to come by. Defiance can be revealed in the agent’s description of his own motives, as in the case of Augustine’s confession of delinquent theft (Augustine 1951, 24 (Confessions, Bk II, 4)). In the absence

1   There are also cases in which perpetrators conceal malice and sadistic pleasure, and pretend to be “merely doing what is necessary”. Sometimes, on the basis of a broader pattern of behaviour, or the lack of plausible alternative motives, we infer that malice or sadistic pleasure were present despite not being observable to witnesses of the act.

Assessing Thick and Thin Accounts  101 of such a confession, defiance is often inferred from the fact that the action in question is obviously immoral and the agent is obviously sufficiently mentally competent to know right from wrong. This inference is risky, and often is made without proper consideration of the alternative possibility; namely, that the agent is generally rationally competent, but is deluded about what is morally right in this particular situation (e.g. Rosenbaum 1998, 69–70). In any case, both the thin account and the folk versions of the thick account allow that we often have clear evidence that an action is evil. The question at hand, though, is whether Morton’s and Garrard’s non-folk versions of the thick account, which posit comparatively complex properties as the psychological hallmark of evil, would allow us to judge with confidence that particular actions are evil. I will argue that Morton and Garrard respond to this pressure in quite different ways.

5.3  Assessing Non-folk Thick Accounts Morton believes that the label “evil” does more than dramatize a sense of disapproval, and that calling an action “evil” is meant to be psychologically informative (Morton 2004, 13, 64). According to Morton’s barrier theory of evil, the psychological hallmark of evil action is that it results from a strategy or learned procedure which allows that person’s deliberations over the choice of actions not to be inhibited by barriers against considering harming or humiliating others that ought to have been in place. (Morton 2004, 57)

In order to discern whether a particular wrong action is evil, then, we must be able to tell whether it results from this kind of strategy or learned procedure, and in order to do this, we must know which barriers against considering harming or humiliating others ought to have been in place, how these inhibiting barriers work, and how they can fail to function. Initially it seems that this would require us to be familiar with the psychological research into violence-inhibiting mechanisms that Morton references (Morton 2004, 42). It also seems that we must know about the perpetrator’s past actions, so that we can distinguish wrongs that are the result of a strategy or learned procedure from those that are mere uncharacteristic aberrations. According to Morton, “an isolated act of anger . . . will not be evil, because it does not result from a process that the person uses regularly” (Morton 2004, 57). This suggests that evil acts can flow only from a stable character trait or process of thought that reliably and non-accidentally produces such acts. Some actions that have the superficial appearance of being evil will turn out not to be evil at all, on Morton’s view, because they did not possess the complex psychological property that he specifies. At least, this is how things appear at first glance.

102  Assessing Thick and Thin Accounts Yet as Morton applies his account to a wider range of examples, it becomes apparent that his posited psychological hallmark is broader than it seemed. Morton goes on to say that, while evil people must be disposed regularly to perform evil acts, it is “important to keep the acts and the people separate, since anyone can do an evil act”, including a person who is not evil (Morton 2004, 65). If Morton believes that someone who usually has good motives can uncharacteristically do something evil, then he thinks that the so-called strategy that produces an evil act need not be stable or regular after all. Moreover, the so-called strategy need not be intended by the evildoer as a means by which to overcome inhibitions against harming others, nor need the evildoer be aware of the barrier. Such a strategy, on Morton’s view, is present even in psychopaths and sociopaths (Morton 2004, 66). The extreme wrongs committed by psychopaths and sociopaths count as evil, according to Morton, because their actions are not blocked by deliberative barriers that ought to have been in place, regardless of whether psychopaths and sociopaths actually have any violence-inhibiting psychological barriers. Morton also believes that the so-called strategy or learned procedure for overcoming barriers that ought to have been in place could be acquired in virtually any way. It could be the product of “violentization”, or of cultural indoctrination, or of self-deception, or, presumably, of any biological processes that produce psychopathy or sociopathy (Morton 2004, 66). Since there ought to be psychological barriers against considering performing any extreme culpable wrong, and since seemingly any process which allows someone to perform an extreme culpable wrong counts as a strategy or learned procedure, in Morton’s sense, I think that Morton ultimately fails to identify a distinctive psychological hallmark of evil action. What looked like a substantive psychological condition that would divide evils from other extreme culpable wrongs turns out to be something so broad that it is present in all cases of extreme culpable wrongdoing. In contrast, Garrard’s proposed psychological hallmark of evil action really is a substantive condition that will divide extreme culpable wrongs into two classes. Garrard claims that in evil actions a metaphysical silencer—a consideration for action that morally should rule out consideration of countervailing reasons— is itself psychologically silenced in the agent’s deliberation. According to this theory, the evildoer gives no deliberative weight to that which should settle the issue, to that which should lead the agent to not even consider performing the action. Garrard draws the notion of silencing from John McDowell, who uses it in a neo-Aristotelian account of virtue (McDowell 1978, 31–42). The difference between the virtuous and the continent agent, Aristotle tells us, is that the continent agent must overcome bad desires in order to act rightly, whereas the virtuous agent has no bad desires to overcome (Aristotle 1985, NE, 1146a10). Those potential

Assessing Thick and Thin Accounts  103 bad desires are psychologically silenced in the virtuous agent, in that they do not even feature as reasons to be outweighed in that agent’s process of deliberation. We ought to accept Garrard’s claim that some wrong actions are the product of the deliberative outweighing of the reasons that actually count conclusively against those actions, whereas other wrong actions are the product of deliberation in which those reasons are not given any weight at all. If the fact that an agent silences the considerations that count against performing his wrong action is an exacerbator of the wrong, then Garrard can argue that she has identified a psychological hallmark of evil action, and that this hallmark explains why evils are worse than non-evil wrongs.2 While the complex psychological property tagged by Garrard does mark out an interesting category of wrong actions, she also needs to show why this category deserves to be called “evil”. Garrard’s distinction between evils and non-evil wrongs seems to be very close to Aristotle’s distinction between vice and incontinence (although it is not quite equivalent to that distinction).3 In light of this similarity, we have reason to doubt whether Garrard’s psychological hallmark captures the extremity of evil action. Just as there is room for disagreement over whether, as a rule, incontinence is worse than vice (Aristotle 1985 NE, 1146b, 1150b29), it is unclear whether psychological silencing generally is an exacerbator of wrongs. It might seem particularly bad to harm a victim without a thought for her moral standing, but there is also something strikingly deplorable about harming a victim despite having taken her moral standing into account. In any case, the distinction between vice and incontinence seems unlike the distinction between evils and non-evil wrongs because there can be actions that are vicious, in an Aristotelian sense, but of merely trivial importance. Let us consider two minor but gratuitous acts of cruelty—say, of mocking someone in a public setting—that are identical in their expression and their effects, the first of which is a vicious act, performed by a person who never even considers the reasons against acting in this way, and the second of which is closer to being incontinent, in that

  It is possible to believe that there is a psychological hallmark of evil action, but that this hallmark does not explain why evils are extreme, or why evils are worse than non-evil wrongs. Advocates of such an account should include a separate extremity condition, like the one that I set out in Chapter Three. However, several advocates of thick accounts of evil action, including Steiner, McGinn, and Garrard, appear to believe that the presence of the psychological hallmark is what makes evil actions extreme. 3   Aristotle’s incontinent man acts against his rational judgement as to what is best, whereas Garrard’s analogous non-evil wrongdoer outweighs considerations against the wrong action. Garrard’s non-evil wrongdoer need not have judged that the action in question was wrong. Incontinence requires a very strong and philosophically puzzling kind of psychological conflict within the deliberating agent, whereas Garrard’s analogous category of non-evil wrongdoing simply requires that the agent recognize that there are some reasons for and some reasons against the wrong action. 2

104  Assessing Thick and Thin Accounts the agent considers but outweighs the reasons that count conclusively against the action. It is implausible that the first of these actions is evil, while the second is merely wrong. Rather, both are merely minor wrongs. One challenge for Garrard is to show that her account does not imply that the first of these actions is evil, even though the agent silenced the consideration that should have ruled out considering performing that action. Perhaps Garrard achieves this end by arguing that only extreme wrongs are ruled out by metaphysical silencers, whereas pro tanto reasons for performing minor wrongs are only ever metaphysically outweighed, rather than silenced, by competing reasons. This would make Garrard’s account of evil even more complicated. Rather than take this step, it may be preferable for Garrard to simply add to her account an independent condition, such as the one I set out in Chapter Three, which ensures that all evils are extreme wrongs. Even if Garrard can rule out the possibility of trivial evils merely via her silencing condition, it is not clear that the psychological property that Garrard identifies is a necessary condition for evil action. It is hard to see how we could tell whether an agent completely silenced a consideration during deliberation, or merely gave that consideration insufficient weight. Deliberation need not be a temporally drawn out process, and often it is not transparent to the deliberating agent herself. The behaviour of an agent who silences considerations of a certain kind would frequently be indistinguishable from that of an agent who severely undervalues those considerations without silencing them. Since we rarely have epistemic access to these facts about deliberation, it seems unlikely that our judgements about which actions are evil will map onto these facts. Arguably, just this kind of mismatch is evident in Garrard’s own discussion of specific evil actions. In the following passage, Garrard offers a number of examples of evil actions that are not extremely harmful: The most obvious example of this is the case of the Iraqis during the occupation of Kuwait who, having shot a young boy, demanded money from the boy’s family to pay for the bullet. Here the principal suffering was caused by the killing; but it is the charging for the bullet which strikes most people as evil. Another such example is that of the sadistic voyeur, who chooses to observe, with intense relish, the sadistic acts of another. Here nothing is added to the sum of suffering by his voyeuristic behaviour; nonetheless we might wish to call it evil. Finally consider the person who, with gloating enjoyment, slowly tortures a cat. In the scale of the world’s sufferings it barely registers, nonetheless it can plausibly be called evil. (Garrard 2002, 327)

If we were employing Garrard’s account of evil, then we would not know whether these actions are evil unless we know that the agents completely silenced the moral considerations that count conclusively against the actions that they actually performed. Garrard does not explicitly discuss whether this is the case, nor is it clear

Assessing Thick and Thin Accounts  105 that she even considers the question. In many cases of, say, sadistic cat torture, it is plausible that the torturer is well aware of the moral considerations against his actions, gives them some deliberative weight, but fails to realize that these considerations should have completely silenced any competing considerations in favour of the torture. If Garrard’s own account is correct, such cat torturers do not do evil at all. In these cases, at least, it seems that Garrard’s intuitions about evil are being guided by something much closer to Steiner’s or McGinn’s folk versions of the thick account of evil action, according to which the psychological hallmark of evil action is not deliberative silencing, but sadistic pleasure. Garrard’s account implies that evil actions are never wrong actions in which the agent considered and mistakenly outweighed the moral considerations which counted conclusively against those actions. No matter how extreme the wrong, or how deliberate and malicious the action, she believes that it does not count as evil unless the agent gave no weight at all to the conclusive reason which counted against it. Garrard admits that this approach seems deeply counterintuitive to many people (Garrard 1998, 56). In defence of her view she points out that several Nazi war criminals, who clearly performed evil actions, describe themselves as giving no moral weight whatsoever to the suffering of the people whom they murdered (Garrard 1998, 59). Yet this is not the kind of evidence that Garrard needs. Even advocates of the psychologically thin account of evil action, who think that evil action is merely extreme culpable wrongdoing, will agree that these Nazis performed evil actions. In order to support her theory Garrard needs to focus on cases in which, for instance, Nazi war criminals performed horribly wrong actions, the performance of which required them to overcome some degree of internal moral conflict in the process of deliberation. If her account is correct, the actions of these Nazis are not evil, but the actions of their colleagues who suffered no such inner conflict are evil. It is possible that Himmler suffered this kind of psychological conflict. Bennett says of Himmler, Although his policies ran against the human grain to a horrible degree, he did not sandpaper down his emotional surfaces so that there was no grain there, allowing his actions to slide along smoothly and easily. He did, after all, bear his hideous burden . . . , [suffering] a variety of nervous and physical disabilities, including nausea and stomach convulsions. (Bennett 1994, 300)

Perhaps this internal conflict was not a moral conflict that registered in deliberation, and Himmler merely felt squeamish about performing actions that he believed had no moral costs. Yet it is not implausible that Himmler was aware to some degree of the moral status of his victims, but severely undervalued the

106  Assessing Thick and Thin Accounts importance of their status, overriding these considerations for what he thought was the greater good of Germany. If we discovered that this were the truth about Himmler’s deliberations, it is far from clear that we would then judge that, while some Nazis might have performed evil actions, Himmler did not. Advocates of non-folk versions of the thick account of evil action face a particular difficulty in explaining why a particular complex and hard-to-access psychological property counts as the feature that distinguishes evils from other wrongs. The more complex and more hidden the property, the less likely it is that any of the folk who use the concept of evil are thinking about that property, or would even be able to recognize the presence of that property. This would not be an insurmountable problem if there were widespread agreement on the extension of the concept of evil action, and if we discovered that the hidden psychological property was possessed by all and only those actions. As we have seen, though, this is not the case. There is significant disagreement over which particular actions are evil, and Garrard’s account clashes with some widely held intuitions about specific cases. Thus we have no reason to suppose that the property identified by Garrard is the psychological hallmark of evil action. In contrast, Morton’s suggested psychological hallmark is so broad that it fails to carve off evils from other extreme culpable wrongs. It is hard to see how his theory differs from the thin account of evil action, according to which evil actions are extreme culpable wrongs that come from any motives.

5.4  The Original, Natural, or Most Useful Concept I have argued that widely shared intuitions about evil action do not strongly favour the psychologically thin account over the folk versions of the thick account, but that common intuitions do clash with Garrard’s non-folk version of the thick account. As we have seen, though, appeal to folk intuitions is not the only method by which we can evaluate competing philosophical theories. Perhaps by utilizing other criteria we will be able to find grounds for favouring the thin conception of evil over its thick rivals, or vice versa. For instance, we might argue that one account of evil deserves our acceptance because that account gets at the original meaning of the word “evil”, or because it helps us to carve the world at its natural joints, or because it picks out a more useful concept than its rivals. Let us proceed by considering whether an appeal to the original meaning of “evil” will be of any use as an evaluative criterion. In some cases in which there is disagreement over the meaning of a word or phrase it is not uncommon to hear people say that the original meaning is the real meaning, and that new usages of the word or phrase are corruptions of that genuine meaning. The true meaning of

Assessing Thick and Thin Accounts  107 “begging the question”, for instance, is “including the conclusion of an argument in its premises”. People who use “begs the question” to mean “prompts the question” are accused of misusing and corrupting the phrase, even though the so-called misuse is very widespread. Firmly policing the current use of a word or phrase to ensure that it conforms with its original meaning could be justified on several grounds beyond mere deference to tradition. Maintaining the original meaning of “begging the question” would help current speakers understand older texts in which that phrase occurs, and might also help to preserve and make salient an important concept that we should not allow to fall into disuse. It could be argued that the original meaning of “evil” is captured by those thick accounts which imply that evildoers act defiantly and maliciously. For example, it might be thought that talk of evil originated in a traditional religious worldview, in which extreme wrongdoers were supposed, like Satan, to be sadistic, malicious, and defiant. If this were the case, then Arendt could be accused of misusing and corrupting the term “evil” when she claimed that evil can flow from banal motives. Arendt’s claim that evil is banal did seem shocking to many of her readers (Neiman 2003, 271–2). This could be taken as evidence that she was attempting to modify or subvert the existing meaning of that word. Even if it were true that, prior to the publication of Eichmann in Jerusalem, the word “evil” more uniformly picked out a thick concept of evil, it is not the case that an original meaning of a word must be the real or the best meaning. As we all know, the meanings of a great number of words have changed dramatically over time, and in many cases there are no good pragmatic reasons for resisting these changes. While it is important that a reader of nineteenth-century novels knows the original meaning of the word “gay”, someone who maintains that “gay” truly means happy rather than homosexual is linguistically naïve. Regardless, it is not clear that the original shared meaning of “evil” really did pick out a psychologically thick version of the concept, because it is not clear how far back we have to go in order to locate the original meaning. Why stop at a time in which “evil” was more uniformly used to express something captured by the thick accounts, if, indeed, there was such a time? Why not go back even further to a time in which “evil” was also applied to things that are non-morally bad and trivially bad, such as evil diets and evil accidents? Perhaps a more persuasive argument in favour of the thick accounts of evil action can be constructed via an appeal to natural kinds. Even if a term is used in a plurality of ways by the folk, we might favour one of those uses on the grounds that the other uses fail to carve the world at its natural joints (Dupré 1993, 22; Sider 2001, 192). The concept of biological species, for example, has been used in a range of different ways in different communities and in different eras. Some people have

108  Assessing Thick and Thin Accounts sorted animals into species and genera by outwardly visible morphological similarity, in which case whales and dolphins are likely to be grouped with fish, and marsupial mice grouped with placental mice. Other people have sorted animals into species and genera by similarities in inner morphology, or by line of descent, or by membership in sexually reproductive groups. Given that different groups of people mean different things by the term “species”, and given that we have various classificatory aims, John Dupré argues, we should accept that many of these species concepts are valid (Dupré 1993, 57). Most biologists and philosophers disagree with Dupré ’s conceptual pluralism, arguing instead that some conceptions of species are superficial, explanatorily useless, and fail to pick out the fundamental pieces of the biological world, and ought to be rejected on those grounds (Slote 1966; Sober 1984; Hull 1987; Jackson 1998, 34–5). Similarly, a conceptual scheme which categorized metals by their outward appearance rather than by their molecular structure would be accused of failing to pick out genuine natural kinds. It is plausible that the variety in folk talk about evil reflects several different folk conceptions of evil. Could it be that one of these conceptions is preferable to the others in virtue of the fact that it latches onto an important natural kind? It is not clear that there are natural kinds in the realm of morality in exactly the same sense in which there are natural kinds that can be picked out in the domains of physics, chemistry, or biology (Finlay 2008, 363). It is uncontroversial to claim that the facts about atoms, elements, compounds, and genes are independent of human thought, but it is more contentious to claim that moral facts exist independent of our practices, our emotional responses, and our desires. While some concepts in the domain of science are thought to pick out natural kinds because those kinds feature in laws of nature or in causal explanations, it is not clear that moral properties feature in laws of nature or in extra-human causal explanations at all (Jackson 1998, 131). Nevertheless, there are weaker senses in which some moral categories are more natural than others, and hence some meaning can be given to the claim that there are moral natural kinds. If some conceptions of evil fail to pick out even these weaker moral natural kinds, we might have a reason to reject those conceptions. There are three fairly obvious grounds on which we could argue that a moral concept fails to pick out a moral natural kind. The first such ground is that the concept in question is derivative or complex rather than fundamental. Some concepts within ethics, such as good, bad, right, and wrong, do seem to play comparatively basic roles in that they cannot be decomposed into simpler concepts, and that other moral concepts incorporate or modify them. The more fundamental the concept, perhaps, the greater its claim to picking out a moral natural kind. It is implausible that either the thick or the thin accounts of evil pick out completely

Assessing Thick and Thin Accounts  109 fundamental moral concepts that cannot be broken down into more basic constituents (cf. Card 2002, 12). As we have seen, these accounts include as building blocks the prior concepts of wrong action and culpability. The thin account picks out a concept that clearly is more fundamental than those identified by the various thick accounts, all of which mark out subsets of the extension of the thin conception via the addition of psychological conditions. The fact that the thin account possesses this kind of comparative naturalness does not imply that it gets at the best concept of evil. Moral thought involves not only basic concepts like right and wrong, but many complex concepts, such as gratitude, disrespect, and retribution. The ability to use complex moral concepts to draw fine-grained distinctions between cases is a sign of intellectual sophistication and acuity, not of unnatural or superficial thought. We have no reason to suppose that the concept of evil ought to pick out one of the comparatively basic properties rather than one of the complex properties. Therefore, this criterion of naturalness does not count against the thick accounts of evil action. A second possible ground on which to argue that a moral concept fails to pick out a natural kind is that it is unsuitably arbitrary or unstable. Even amongst more fine-grained and complex moral concepts, there is a distinction to be drawn between those that pick out a seemingly natural group and those whose extension seems random, heterogeneous, or unstable in comparison. For instance, a society could employ in its moral system a concept that picks out and prohibits the actions that are offensive to the leader of that society. Given the capricious and poorly-informed attitudes of the leader, the extension of this concept would change dramatically over time, and the various members of the extension might have very little in common with respect to their intrinsic properties. It would not be implausible to argue that this kind of concept, despite its utility in that society, fails to pick out a moral natural kind, and hence that it is inferior to other moral concepts that seem more stable or less arbitrary. However, none of the thick or thin accounts of evil action that we have considered pick out concepts that are particularly prone to this kind of arbitrariness or instability. The more fine-grained thick accounts of evil action latch onto classes of actions that are comparatively more homogenous, given that all actions within such a class are similar with respect to the motives and affect of the evildoers. Yet the class of actions picked out by the thin account of evil action is no more heterogeneous or unstable than the classes picked out by other important moral concepts, such as the concepts of wrong, harmful, or permissible. This criterion of naturalness also gives us no reason to favour either the thick or the thin accounts of evil action. A third possible ground on which to argue that a moral concept fails to pick out a moral natural kind is that the extension of that concept is

110  Assessing Thick and Thin Accounts non-accidentally empty. If there never are any actions, motives, or persons who fall under that concept, then, no matter how simple and stable the concept happens to be, we might argue that it does not allow us to carve reality at all, let alone to carve reality at its joints. Marcus Singer appears to endorse this argument, claiming that “If there are no clear cases of evil, there can be no concept of evil and no theory of evil” (Singer 2004, 191). The general version of this argument from non-instantiation is not persuasive. For instance, the concept of a vacuum appears to pick out a significant natural kind even if there is no perfect vacuum in the universe, and the concepts of moral sainthood and full Aristotelian virtue seem to play an important role in moral theories even if those ideals are never actually instantiated (Wolf 1982, 419; Kamtekar 2004, 485). In any case, this argument from non-instantiation will not be convincing when applied to the specific cases of thick or thin accounts of evil action because none of these accounts point to a concept that has an empty extension. It is true that there are a great many more thin evil actions than thick evil actions, but even the most detailed of the folk thick accounts, which specifies that evils must be malicious, sadistic, and defiant, straightforwardly applies to some real actions performed by serial killers and war criminals (Stone 2009). Even if we admit that each of the folk thick accounts and the thin accounts of evil action that I had tabled picks out a respectable moral natural kind, as I think we should, we could argue in favour of one account on pragmatic grounds. The fact that one account is more useful than the others might convince us that it identifies the best concept of evil, and that we ought to steer people away from using the other conceptions of evil that are in play in contemporary moral discourse. Before we could construct such an argument, though, we would need to identify the ways in which a moral concept could be more or less useful. One criterion to which we might appeal is the size of the extension of the concept, or the frequency with which the concept could properly be applied. Since there are many more thin evil actions than thick evil actions, it might seem that the thin account of evil is thereby more useful than a thick account. Yet, as Daniel Haybron points out while defending his account of evil personhood, the “utility of a notion is not simply a function of how often we may be warranted in applying it” (Haybron 2002a, 278) Another factor that influences the utility of a concept is the amount of information that it carries. On this criterion, it is the psychologically rich thick conceptions of evil action that appear to be more useful than their thin rivals, in that these thick conceptions give us comparatively more information about the actions to which they are applied. Laurence Thomas, in the course

Assessing Thick and Thin Accounts  111 of distinguishing harmful action from evil action, suggests that a thick concept of evil is preferable because it is informationally rich: Nothing is gained by treating all harmful acts as evil acts. For, philosophically, we would still have to distinguish, among the class of evil acts, between intentional and nonintentional evil acts. And among intentional evil acts we would then need to distinguish between those that issue from delight in causing the harm of another and those that do not. And so on. By the time we finished our litany of distinctions, we would be mirroring the very distinction, between harmful and evil acts, that we had obliterated in the first place. This does not make for philosophical progress. (Thomas 1993, 78)

When I first wrote on the topic of evil, I also suggested that a thick concept of evil action is preferable on the grounds that it carried more information (Russell 2006, 103). While it is true that both applicability and informational richness are factors that affect the utility of a moral concept, it is clear that thin moral concepts and thick moral concepts can be very useful. The best kind of moral theory will include both thick and thin concepts. Our moral theories would be less systematic and more cumbersome if they included only the most detailed and psychologically rich concepts, but they would be blunt and crude if they contained only the barest concepts such as right, wrong, good, and bad. Moreover, we cannot expect any particular moral concept to be both maximally informative and maximally applicable. As we pack more information into a concept, we typically reduce the number of cases to which the concept applies. There is a trade-off between breadth of applicability and richness of informational content. Contrary to Thomas’ argument, the fact that thin accounts of evil action fail to include some important moral distinctions does not show that they pick out a moral concept that is less useful than one of the thick concepts of evil. I no longer think that a thick concept of evil is preferable to a thin concept on grounds of usefulness. Advocates of the thin account could simply claim that the concept of evil action is widely applicable, and that the more fine-grained distinctions between types of motive and affect can be made using other concepts. In this chapter, I have argued that we have good reason to reject a non-folk version of the thick account of evil action, according to which there is a hidden psychological hallmark of evil action. However, appeals to folk usage, to natural kinds, and to usefulness do not count decisively in favour of either the thin account or the folk versions of the thick account. In the next chapter, I will move on to consider some more complex philosophical objections to the thin and the thick accounts, in the hope that these might help us to identify the best contender amongst the theories of evil action.

6 Conceptual Pluralism about Evil Action In this chapter, I will continue assessing the competing thick and thin accounts of evil action by focusing on two sets of more sophisticated philosophical arguments. The first set concern the question of whether there is a qualitative difference between evil actions and non-evil wrongs. Several philosophers have claimed that there is such a difference, and that this difference gives us a reason to favour a thick account of evil action. I will argue that there are stronger and weaker versions of the qualitative difference thesis, and that even advocates of thick accounts of evil action do not have good reason to think that evil actions are qualitatively distinct from ordinary wrongs in the stronger sense. The second set of arguments can be developed from Neiman’s claim that people who use a psychologically thick concept of evil action rank wrongs incorrectly and fail to identify the real motives of evildoers. I will show how thick accounts of evil action can survive this objection unscathed. In the absence of any convincing arguments for favouring the folk thick accounts over the thin account of evil action, or vice versa, I will conclude that we ought to adopt a restricted conceptual pluralism about evil action.

6.1  Qualitatively Distinct Evil According to the psychologically thin account, an action is evil if and only if it is an extreme culpable wrong. Arendt, Card, Neiman, and Formosa think that it is the extremity condition that sets evils apart from many other ordinary culpable wrongs. Some philosophers say instead that evil actions are distinct from non-evil culpable wrongs not merely in virtue of being more extreme. Steiner, Garrard, and de Wijze maintain that evil actions are not simply more wrong or more harmful than ordinary wrong actions, but that evil constitutes a qualitatively different kind of wrongdoing. I have listed as intuition (8) the claim that evil action is qualitatively distinct from ordinary wrongdoing. It would be a stretch to describe this as a folk intuition. Typically

Conceptual Pluralism about Evil Action  113 it is expressed by philosophers rather than by non-philosophers, and it is a technical metaphysical claim about kinds of difference. Nonetheless, it could turn out that the belief that evil is qualitatively distinct from ordinary wrong is somehow implicit in folk thought about evil. It is the kind of claim that sounds true and profound, even before we are sure what it means. Regardless of whether the folk already believe that evils are qualitatively distinct from non-evil wrongs, there could be other reasons to endorse intuition (8). For example, we might think that only by positing a qualitative difference between evils and non-evil wrongs could we rule out the possibility of actions falling into a grey area between evil and non-evil, or that only by positing a qualitative difference could we show why the concept of evil ought to be preserved. My first task in this chapter is to consider whether evil actions are qualitatively distinct from ordinary wrongs, and, if so, whether this gives us a reason to favour a thick account of evil action. In order to answer this question we must get a firmer grasp on the vague notion of a qualitative difference between two classes of action. On one possible reading of intuition (8), every account of evil action is guaranteed to imply that there is a qualitative difference between evils and non-evil wrongs, because evil (whatever it turns out to be) is a quality that is possessed by evil actions and not by non-evil actions. I will call this the cheap version of the qualitative difference thesis. No one has any reason to deny the cheap version of the qualitative difference thesis, but nor will the cheap version be of assistance to anyone who is looking for reasons to favour one of the competing accounts of evil action over its rivals. A more substantial, but still fairly weak reading of the qualitative difference thesis is offered by Todd Calder. Calder claims that the concept of evil action is qualitatively distinct from the concept of wrong action in the sense that there is at least one essential property of evil action that is not an essential property of wrong action. Initially Calder suggests that in an evil action extreme harm “must be intended”, but that, in contrast, intending to inflict harm is not a necessary condition for wrong action (Calder 2013, 183). This is a puzzling claim. In an earlier paper Calder accepts that it is implausible that evildoers must intend to inflict extreme harm, if by “intend” we mean to imply that the infliction of harm was desired as an end by the agent (Calder 2002, 53). It could also be argued that some culpably reckless wrongs that are appropriately connected to extreme harms can count as evil. In such a case, extreme harm was an actually unforeseen but rationally foreseeable consequence of the action, so the harm clearly was not intended by the agent.1 Regardless of whether he agrees that culpably reckless wrongs could be evil, Calder should accept that intending extreme harm is not an essential 1   If malice, sadistic pleasure or defiance are necessary conditions for evil action, then culpably reckless wrongs fall short of being evil. Calder, though, denies that these are necessary conditions for evil action (Calder 2002, 54, 57).

114  Conceptual Pluralism about Evil Action condition for evil action, given he believes that some acts of sadistic voyeurism are evil despite the fact that the voyeur does not intend to inflict harm (Calder 2002, 56; Calder 2013, 188). Calder goes on to claim that an action can be evil only if it is produced by an “e-motivation”, by which he means an “inexcusable intention to bring about, allow, or witness . . . significant harm . . . for an unworthy goal” (Calder 2013, 188). Being produced by such a motive is not a necessary condition for wrong action, Calder believes, because some actions that are performed for worthy goals or with excusable intentions still count as wrong, according to some viable and popular consequentialist accounts of wrong action (Calder 2013, 182–4). I agree with Calder that we should allow for the possibility of wrong actions in which the wrongdoer is not properly held responsible for the harm that results, and that, in contrast, if an action is evil, then the evildoer is always culpable for the appropriately connected harm. An evil action is a moral discredit to the agent, but not every wrong action, on consequentialist accounts of wrongdoing, at least, is a moral discredit to the agent. Culpability is an essential condition for evil action, but not an essential condition for wrong action, broadly conceived.2 The concept of evil action is qualitatively different from the concept of wrong action, as Calder interprets qualitative difference. Calder’s version of the qualitative difference thesis is more demanding than the cheap version, but is still comparatively weak. I will go on to examine a stronger version in a moment. For now, though, we should note that Calder’s claim about qualitative difference does not give us any reason to favour a psychologically thick account of evil action. This much is apparent from Calder’s own claims about so-called e-motivation. Calder explicitly allows that e-motivation can include malicious desires to inflict extreme harm or merely instrumental desires to do so, sadistic pleasure or lack of sadistic pleasure, and defiance or an absence of defiance (Calder 2002, 54–7). Even though Calder says that evil action is qualitatively distinct from ordinary wrongdoing, his account of evil action should be classified as psychologically thin. Calder’s account suggests that if we are told that an action is evil, we are being told that the agent had no excuse for doing it, but we are not being told why the agent did it, what the agent was feeling when he did it, or what the agent believed about the rightness or wrongness of the action in question. Calder does not aim to defend a thick account of evil action. Instead he aims to show that evil is not reducible to the very wrong, and hence that the concept of evil should not be eliminated from our moral thinking (Calder 2013, 178). I agree with Calder’s conclusion concerning the importance of the concept of evil, but   I specify what I think is the relevant kind of culpability in Chapter Two.

2

Conceptual Pluralism about Evil Action  115 I doubt that his argument concerning qualitative difference does much to advance this goal.3 Nor does Calder’s account solve any worries that we might have about vagueness at the boundary between evils and non-evil wrongs. Calder’s account implies that there are many cases in which a non-evil culpable wrong differs from a particular evil action only in that the harm to which it is connected is less extreme. Given that Calder offers a vague criterion for the requisite extremity of the connected harm—“a harm that a normal rational human being would take considerable pains to avoid” (Calder 2013, 188)—his account suggests that there will be many actions that fall into a grey area between the clearly evil and the merely culpably wrong. Stephen de Wijze’s account of evil action is similar to that offered by Calder. De Wijze says that evil actions stand out not in virtue of the magnitude of their effects, but rather are qualitatively different from merely bad or very bad acts (de Wijze 2002, 211–12). The qualitative difference between evil actions and non-evil actions, he suggests, is that the evil actions meet at least one of three disparate conditions: A. There is a deliberate violation of persons with the intention to dehumanise (that is, deny basic respect and dignity to) those powerless to retaliate. B. The action or project will gratuitously inflict, or bring about, one or more of ‘the Great Harms’ to sentient beings with the relevant moral standing. C. The action or project (or professed morality) seeks to annihilate the ‘moral landscape’ (de Wijze 2002, 218).

Like Calder, de Wijze does not assume that all evil actions have similar motives. Condition A makes reference to distinctive motives in the evildoer and powerlessness in the victim, and Condition B makes reference to gratuitous infliction of extreme harm, but Condition C makes no reference to motives, intentions or the status of the victims. It is plausible that all extreme culpable wrongs, including non-malicious, non-sadistic and non-defiant wrongs, are actions that seek to annihilate the moral landscape, in de Wijze’s sense (Russell 2007b, 668). Advocates of the psychologically thin account would applaud de Wijze’s suggestion that there is no unified psychological hallmark of evil. Someone who thinks   Calder says that “if evil is just very wrong, we can safely abandon the concept of evil” (Calder 2013, 178). I think that Calder’s own account implies that evil actions are simply extreme culpable wrongs. But if this is the case, then the concept of evil action is distinct from the concept of culpable wrongdoing merely by virtue of being more extreme (that is, evils are simply those culpable wrongs that are connected to worse harms). If he is consistent, Calder should still conclude that we can safely abandon the concept of evil. I think that the concept of evil could be useful and legitimate even if evils turn out to be non-qualitatively distinct from ordinary wrongs, and even if the concept of evil can be reduced to a combination of more basic moral concepts. I discuss these issues in more depth in Chapter Eleven. 3

116  Conceptual Pluralism about Evil Action that evils are clearly and sharply distinct from ordinary wrongs would not, on these grounds, have reason to favour de Wijze’s disjunctive account over the psychologically thin account. Like Calder, de Wijze could say that the concept of evil is qualitatively distinct from the concept of wrong in virtue of the fact that culpability is a necessary condition for evil action but not for wrong action. However, this weak kind of qualitative difference is possessed by all of the accounts of evil action defended in the recent philosophical literature, as far as I can tell, so appealing to this kind of qualitative difference gives us no reason to favour one of these accounts over its rivals. Several of the philosophers who have claimed that evil is qualitatively distinct from ordinary wrong do take this to be grounds for favouring some accounts of evil action over their competitors. In particular, Steiner, Garrard, and Haybron appear not to be endorsing merely the cheap version of the qualitative difference thesis, nor merely Calder’s version. The challenge at hand is to figure out what these philosophers mean when they say that there is a qualitative difference between evils and ordinary wrongs, and then go on to consider whether this would give us a reason to favour psychologically thick accounts of evil action.

6.2  Strongly Qualitatively Distinct One possible alternative is to read claims about qualitative difference as pointing to a phenomenological difference; that is, a felt or experienced difference. A sensation of pain is qualitatively different from a sensation of hunger, in that the feel of pain differs from the feel of hunger. In many cases it feels like something to perform an action, to contemplate or observe an action, and to be acted on. One kind of action could be described as being qualitatively distinct from another kind in that performing, contemplating, observing, or receiving actions of the first kind feels different to performing, contemplating, observing, or receiving actions of the second kind. If this is what they mean by a qualitative difference, Steiner, Garrard, and Haybron would be claiming that evil action feels different to ordinary wrongdoing. Some philosophers have claimed that evil actions are phenomenologically different from ordinary wrongs. For instance, de Wijze writes: Evil, in a different way from merely wrongful actions, leaves behind a moral residue which, if it is possible to remove, requires a special ritual of purification. The horror, the disgust and incomprehension evoked by evil suggests a qualitative difference, something that distinguishes it from wrongful or even very wrongful acts. (de Wijze 2002, 213)

Conceptual Pluralism about Evil Action  117 Morton makes a similar claim: We have a visceral revulsion from extremely evil acts. The revulsion is most vivid when the acts involve physical violence, but it extends to other acts produced by similar patterns of motivation, even if they do not have the same emotional immediacy. Evil acts have a quality that in ancient times would have made us fear that the Gods might send a plague in reprisal, rather than simply making us despair at the incapacities of mere mortals to manage their lives together. (Morton 2004, 13)

The thought that evil actions typically are phenomenologically different from ordinary wrongs is relatively uncontentious. It is a claim that could be endorsed by advocates of the psychologically thin account as well as by advocates of the various thick accounts. Almost all of us care about morality and are emotionally affected by violations of morality, so we should expect that contemplating or suffering an extreme wrong usually would feel different to contemplating or suffering a less extreme wrong. As I have argued in Chapter Three, though, it would be a mistake to suppose that a certain kind of psychological reaction on the part of victims or observers—be it horror, disgust, or incomprehension—is a necessary condition for an action’s counting as evil. Some victims and some witnesses do not have normal or well-attuned reactions, so individual reactions to wrongdoing are an unstable and unreliable gauge of moral significance. In any case, when Steiner, Garrard, and Haybron say that there is a qualitative difference between evil and non-evil, they believe that this goes beyond a mere phenomenological difference. Instead they draw a certain kind of distinction between qualitative differences and quantitative differences. For example, Steiner notes that some people use evil as a “wrong intensifier”, where the intensification consists of a mere increase in the severity of the properties that are possessed by ordinary wrongs. The correct view, according to Steiner, is that evil acts are distinguished from ordinary wrongs through the addition of an extra quality that is completely absent in the performance of ordinary wrongs, just as aggravated assault is distinguished from ordinary assault via the addition of an intent to murder, or via the use of a dangerous weapon (Steiner 2002, 184). Garrard seems to agree, claiming that there is a qualitative, and not merely quantitative, difference between evil acts and other wrongful ones; evil acts are not just very bad or wrongful acts, but rather ones possessing some specially horrific quality. (Garrard 2002, 321)

Daniel Haybron focuses on evil persons rather than evil actions, but he too thinks that “the difference between bad and evil is not merely one of degree. It is a qualitative difference” (Haybron 2002a, 262). Haybron criticizes accounts of evil which imply that the line between evil and non-evil is “rather vague, with a number of cases being neither

118  Conceptual Pluralism about Evil Action clearly evil nor clearly not” (Haybron 2002a, 262). A theory of evil, Haybron claims, “ought to accommodate the idea that there is a significant moral boundary between the evil and the merely bad” (Haybron 2002a, 262). It would be even better, Haybron believes, if we could give an account of evil that vindicated the intuition that evil does not come in degrees; that “evil is an all-or-nothing affair” (Haybron 2002a, 262–3). While the connection between evil action and evil personhood is complex, Haybron suggests that an evil action is one that “manifests . . . the sensibilities characteristic of an evil person” (Haybron 2002a, 280), and it is fair to assume that he thinks that evil actions will be qualitatively rather than quantitatively distinct from non-evil wrongs. When Steiner, Garrard, and Haybron make claims about qualitative difference, they go beyond the cheap version of the qualitative difference thesis, and beyond Calder’s version as well. Arguably, they believe that evil actions differ from ordinary wrongs in virtue of possessing some particular property that is a necessary condition for evil action, and that is not possessed to any degree by any non-evil wrongs (Russell 2007b, 661–2; cf. Calder 2013, 179). I will call this the strong version of the qualitative difference thesis. The strong version of the qualitative difference thesis is far more contentious than the cheap version, and than Calder’s version. Nonetheless, Steiner, Garrard, and Haybron do offer accounts of evil that seem designed to fit with the strong version, and some of other recent philosophical accounts of evil action are incompatible with this thesis. The strong version of the qualitative difference thesis might help us to assess competing accounts of evil action. Moreover, an account that included the strong qualitative difference thesis would allow us to draw a sharp dividing line between evil actions and non-evil culpable wrongs; an outcome that is desired by Haybron, and probably by others. I have tried to clarify the strong version of the qualitative difference thesis. Those who wish to defend it would also be required to address more detailed questions about the nature of qualitative difference.4 It seems that both Garrard and Steiner do believe that there is an essential property of evil actions that is not possessed to any degree by any non-evil wrongs.5 They propose that there is a psychological hallmark   Consider an action that is motivated by an excessive desire for money and another that is motivated by an excessive desire for food. If we make fine-grained distinctions between motives, these two actions are qualitatively distinct, but if we classify motives more broadly we would say that both actions are motivated by greed, and that, with respect to their motives, the actions are not qualitatively distinct. Another problem concerns complex qualities of actions. Imagine that one person mows her front lawn mainly out of civic pride but partly out of a desire to annoy her neighbour, and that a second person mows his front lawn mainly out of a desire to annoy his neighbour, but also partly out of civic pride. These two actions might seem to be merely quantitatively distinct, in that they are the product of more or less of the same kinds of motive. Yet it could be said that they are qualitatively distinct, in that the second action possesses the complex quality of being motivated primarily by a desire to annoy, and the first action lacks this quality altogether. 5   Haybron also posits this kind of difference between evil persons and non-evil persons, and I will assess Haybron’s account in Chapter Eight. 4

Conceptual Pluralism about Evil Action  119 of evil action, but they do not include a separate extremity condition, so they think that the only difference between evils and non-evil culpable wrongs is the presence or absence of the psychological hallmark. In contrast, thick accounts that do include a separate extremity condition do not imply that there is an essential property possessed by evil actions that is not possessed at all by any non-evil wrong actions, and nor does the psychologically thin account. If we were committed to the strong version of the qualitative difference thesis, then we might use this as grounds for favouring either Garrard’s or Steiner’s thick account of evil action. While I agree that is would be advantageous to draw a sharp line between evils and non-evil wrongs, ceteris paribus, I think that we should not endorse the strong version of the qualitative difference thesis, because doing so would commit us to implausible claims about the extension of the concept of evil action. This is evident in Garrard’s and Steiner’s respective accounts. As we have seen in the previous chapter, Garrard believes that evils alone have a complex property involving deliberative silencing of a certain class of reasons for action. Since this kind of silencing is an all-or-nothing property, Garrard’s account allows us to draw a sharp dividing line between evils and non-evil wrongs. Yet, as I argued in the previous chapter, Garrard’s account clashes with widely shared folk judgements about which actions are evil. No matter how bad a wrong action is, Garrard would have us say, it cannot be evil unless the decisive considerations against performing that action were totally silenced in the agent’s deliberation. The serial killer who considers the moral status of his victims for a second, but then outweighs this with thoughts of the sadistic pleasure that he gets from the crime does not do evil, according to Garrard’s account. Rather than accept this result, we are better off admitting that many evils are only quantitatively distinct from non-evil culpable wrongs. Steiner also rejects the need for a separate extremity condition, and believes that the distinctive property that marks evil out from non-evil wrongs is sadistic pleasure: “[e]‌vil acts are wrong acts that are pleasurable for their doers” (Steiner 2002, 189).6 Being pleasurable really is a property that is possessed by some wrong actions and not at all by other wrong actions, and for this reason Steiner’s account may seem attractive to those who want to draw a sharp distinction that leaves no grey area between evils and ordinary wrongs. The weakness of Steiner’s account is that it clashes with common intuitions regarding the extension of the concept of evil action. It implies that any wholly unpleasant actions could not be evil, when in fact very many people would judge that some wholly unpleasant actions performed   Presumably Steiner means that evils are wrong actions in which the agent takes pleasure in either the wrongness of the act or a wrong-making feature of the act, not merely pleasure in any feature of the action. See Russell (2007b, 669). 6

120  Conceptual Pluralism about Evil Action out of a misguided conscience could be evil nonetheless (Bennett 1994, 300). Steiner’s account also pushes trivial wrongs into the category of evil action. The cost of claiming that a kind of motive provides a sharp non-quantitative distinction between evils and non-evil wrongs is that any wrong action which possesses that motive must subsequently be judged as evil. Steiner would have us believe that an act of shoplifting which is pleasurable to the shoplifter is evil (Steiner 2002, 190). Yet pleasurable shoplifting is merely a minor wrong, and intuitively falls far short of being evil. I think that Steiner’s account actually picks out the concept of sadistic wrong action, not the concept of evil action. Every sadistic wrong action is strongly qualitatively different from every non-sadistic wrong, but it is not plausible that every evil action is strongly qualitatively different from every non-evil culpable wrong. Arguably, objections of this kind will undermine any theory which proposes that a particular psychological condition not only is necessary for evil action, but also provides a sharp, non-quantitative distinction between evil actions and non-evil wrongs. There are examples of merely minor wrongs that are pleasurable to the wrongdoer, and are malicious, and are defiant. There are examples of merely minor culpable wrongs that are the product of the psychological silencing of good reasons against performing actions of that type. I contend that the difference between these minor wrongs and evil actions is the extremity of the harms to which the actions are appropriately connected. Admittedly, some differences between types of harm—between physical assault and murder, for instance—are qualitative rather than quantitative differences. Death is not simply more damage. Yet the difference between mild pain and overwhelming pain, such as that administered in torture, plausibly is a quantitative difference. Overwhelming pain is what you get when you take mild pain and keep adding more pain. There will be cases that fall into the grey area on this linear scale between mild and overwhelming. There is no clear and sharp distinction to be had here. Malicious, sadistic, and defiant infliction of mild pain is quantitatively distinct from malicious, sadistic, and defiant infliction of overwhelming pain, but the first clearly is not evil, and the second plausibly is. Thus, even advocates of thick accounts of evil action should not say that evils are strongly qualitatively distinct from ordinary wrongs rather than quantitatively distinct from ordinary wrongs. Of course, no one has any reason to object to the cheap version of the qualitative difference thesis. Evil actions are evil, and non-evil actions are not. Nor should we reject Calder’s version of the qualitative difference thesis. Evil actions are essentially culpable wrongs, but it is plausible that not all wrong actions are culpable wrongs. This is a fact that is respected by both thick and thin accounts of evil action. We should not, however, go so far as to endorse the strong version

Conceptual Pluralism about Evil Action  121 of the qualitative difference thesis. Advocates of psychologically thick accounts of evil ought to admit that sometimes the difference between an evil action and a non-evil action is simply that the evil action inflicted much more pain. This means that we should reconcile ourselves to the fact that there will be actions that fall into a grey area between the clearly evil and the non-evil.

6.3  Neiman’s Arguments for the Thin Account I have argued that intuition (8), concerning qualitative difference, does not give us any reason to favour thick accounts of evil action over their thin rivals. Let us move on to consider a set of arguments that could be taken to favour the kind of thin account of evil action that is advocated by Arendt, Neiman, Card, and Formosa. Neiman has argued that employing a thick conception of evil action, according to which malice, sadistic pleasure or knowing defiance are necessary conditions for evil action, is likely to lead to certain kinds of mistaken moral judgements (Neiman 2003, 271–2). In order to avoid falling into these moral errors, she suggests, we ought to use a psychologically thin conception of evil action instead. This is an argument that goes beyond mere appeals to folk intuitions about the nature of evil action, and instead evaluates competing accounts in terms of the usefulness of the concepts that they pick out. If most of the folk do not currently think that evil actions could come from almost any set of motives, Neiman could say that they ought to change the way that they think about evil in order to avoid making significant moral errors. I will claim that Neiman’s arguments depend on an implicit assumption about the ranking of evil actions compared to non-evil wrongs; namely, that any evil action is morally worse than any non-evil wrong. This assumption is interesting, contentious, and worth examining in its own right. Neiman herself does not explicitly endorse it, but some other philosophers do. For instance, Claudia Card maintains that “evils, the worst wrongs people do, deserve to be taken seriously and to receive priority of attention over lesser wrongs” (Card 2002, 28). Paul Formosa offers a more detailed elaboration: Evils are an important subclass of moral wrongs because they deserve, over other wrongs, our very strongest moral condemnation; they deserve, over other wrongs, priority when it comes to preventing, minimizing and combating them; and they need to be responded to in different ways to other wrongs. (Formosa 2008, 231)

Before we can properly evaluate Card’s and Formosa’s claims we ought to tease apart two different theses that both seem to endorse. The first is what I will call the deplorability thesis: that any evil action is morally worse than any non-evil

122  Conceptual Pluralism about Evil Action wrong, in the sense of being morally more deplorable and more worthy of strong condemnation than any non-evil wrong. The second is what I will call the priority of prevention thesis: that the prevention of any evil action ought to take priority over the prevention of any non-evil wrong. Acceptance of the priority of prevention thesis presumably comes via prior commitment to the deplorability thesis. We could argue that we have a more pressing reason to prevent any evil than to prevent any mere wrong on the grounds that an evil is always morally worse than a mere wrong. Both the deplorability thesis and the priority of prevention thesis seem plausible at first glance. Upon examination, though, the priority of prevention thesis turns out to be false. Our moral aim in preventing certain actions is not merely to reduce the amount of wrong that is done or to minimize the severity of wrong actions, but also to reduce the amount of undeserved harm that is suffered. Some actions that are less morally wrong than evil actions, in the sense of being less deplorable, are foreseeably much more harmful than evil actions. For instance, some evil actions foreseeably will not have very bad consequences at all, because the evildoer is very unlikely to succeed in inflicting the extreme harm that he intends to inflict. A terrorist who adds poison to the water supply in a malicious attempt to kill thousands of innocent people thereby performs an evil action, even if, unbeknownst to the terrorist, there are secret safety mechanisms in place that will detect and neutralize the poison. His action is evil, and should be prevented if possible, but if we have limited resources at our disposal, we ought to give priority to the prevention of other more harmful but non-evil wrongs.7 Similarly, some wrong actions have foreseeably catastrophic consequences but fall short of being evil actions because the agent was non-culpably unaware of the magnitude of the likely consequences. Let us consider the case of a thief who steals some drugs from a hospital without knowing that those drugs are irreplaceable and are needed in order to save the lives of five hundred people who have been struck down in an epidemic. This theft is wrong, and the hospital staff can foresee its catastrophic consequences. The prevention of this non-evil wrong ought to take priority over the prevention of many comparatively less harmful but evil actions (cf. Morton 2004, 10).

7   This kind of counterexample depends on the assumption that there can be harmless evil actions (or, at least, evil actions that are not extremely harmful). Some philosophers, including Card, reject this assumption (Card 2002, 20). Card would claim that the terrorist who added poison to the water tried to perform an evil deed, but failed. I have argued for the claim that there can be harmless evil actions in Chapter Two. Even if we accepted Card’s claim that an action can be evil only if it is harmful, the priority of prevention thesis remains vulnerable to the second kind of counterexample discussed above: the case of catastrophically harmful but non-evil wrongs.

Conceptual Pluralism about Evil Action  123 We ought to reject the priority of prevention thesis, but the deplorability thesis seems more defensible. The word “evil”, like “atrocity”, denotes an extremity. While there can be trivial wrongs, there could not be a trivial evil or a trivial atrocity. Thus we might suppose that any evil action is more morally grave and more deserving of condemnation than any non-evil wrong. One of the difficulties we face when comparing the wrongness of actions is that actions can be individuated in various ways. When an angry pedestrian threatens to harm a beggar, he might also succeed in scaring the beggar and in alerting passers-by to the risk of violence. Depending on our purposes, we could count this as one action or as three. Even more difficult problems arise when comparing actions. Let us suppose that we want to compare a case in which someone puts poison in the water and kills one hundred people with a case in which someone shoots one hundred people, one after the other. Is the first a case of a single action and the second a case of one hundred separate actions? Alternatively, should we bundle the individual murders in the second case into a single complex action: a killing spree, for instance? The plurality of possible individuations of actions threatens to undermine the process of comparing and ranking wrongs. Despite this problem, the idea that evils are always worse than non-evil wrongs remains appealing to some philosophers. As we have seen, it is explicitly endorsed by Card and Formosa. One reason why we should care about the deplorability thesis is that it appears to play an important background role in Neiman’s arguments against using a psychologically thick conception of evil action. The first of Neiman’s arguments runs as follows: those who assume that all evil actions are malicious, sadistic, or defiant are doomed to rank wrongs incorrectly, and thus to overlook some wrongs that deserve greater attention. In order to rank wrongs correctly, we should take evil to be as described in the thin account. Adam Morton offers an inchoate version of this argument: Much misery happens not because a small number of people act out of hatred or sadism but because a large number of people act with limited care or imagination. And many real horrors are the result of the intelligent choices of people we could not easily call evil. . . . Thinking in terms of evil can make it harder to see this: we are pushed to thinking that awful things result from awful motives, to placing the blame for horrors on those whose motives we can see as awful and alien. (Morton 2004, 5–6)

The risk identified by Morton threatens to arise not from thinking in terms of evil per se, but from thinking that all evil actions are motivated by hatred or sadism. In tracking the hatred and sadism, we risk overlooking the serious wrongs that were not produced by hatred and sadism.

124  Conceptual Pluralism about Evil Action While discussing the Holocaust, Neiman also seems to endorse this argument. She claims that “Auschwitz embodied evil that confuted two centuries of modern assumptions about intention.” Neiman continues: Those assumptions identify evil and evil intention so thoroughly that denying the latter is normally viewed as a way of denying the former. Where evil intention is absent, we may hold agents liable for the wrongs they inflict, but we view them as matters of criminal negligence. Alternatively, anyone who denies that criminal intention is present in a particular action is thought to exonerate the criminal. This is the source of the furor that still surrounds Arendt’s Eichmann in Jerusalem. . . . The conviction that guilt requires malice and forethought led most readers to conclude that Arendt denied guilt because she denied malice and forethought. . . . (Neiman 2003, 271–2)

But, as Neiman points out, Arendt did not exonerate Eichmann. Rather, Arendt suggested that evil can be banal, and that Eichmann was an evildoer even if he lacked malice towards his victims. Only by employing a psychologically thin conception of evil action, it seems, is Arendt able to give non-malicious but extreme wrongs the strong condemnation that they deserve. Neiman goes on to say that “simple, demonic conceptions of evil often deliberately obscure more insidious forms of it”, such as the banal, institutionalized form of evil we see in Arendt’s Eichmann (Neiman 2003, 285). Neiman’s claims about the risk of employing a psychologically thick conception of evil can be extrapolated into a more formally structured argument. The argument relies on the deplorability thesis, which Neiman herself does not explicitly endorse, but which is required if the argument is to be valid. For the sake of simplicity I will assume in this argument that a thick account of evil action implies that an action could be evil only if it malicious, sadistic, and defiant, although the argument could be run against any of the folk versions of the thick account of evil action. In order to make the argument clear we require an example of an extreme culpable wrong that is not malicious, sadistic, or defiant, and an example of an extreme culpable wrong that is malicious, sadistic, and defiant. Let our first example be Arendt’s Eichmann ordering the murder of millions of Jews, but doing so without malice, sadistic pleasure, or defiance.8 Our second example will be the malicious, sadistic, and defiant murder of a twenty innocent people in a racist hate-crime bombing. The argument runs as follows: 1. If an action is evil then it is morally worse than any non-evil wrong. 2. If a thick account of evil action is correct, then Eichmann’s order is not an evil action, but the hate-crime bombing is an evil action.

8   In referring to this action as that performed by “Arendt’s Eichmann”, we should keep in mind the fact that the actual Eichmann did act out of malicious anti-Semitism.

Conceptual Pluralism about Evil Action  125 Therefore, 3. If a thick account of evil is correct, then the hate-crime bombing is morally worse than Eichmann’s order. 4. Eichmann’s order is morally worse than the hate-crime bombing. Therefore, 5. A thick account of evil action is incorrect. The first premise of this argument is the deplorability thesis, the second premise follows from the definitions of evil action offered by the thick and thin accounts, and the third premise can be derived from the preceeding two premises. The fourth premise is a comparative evaluation of specific cases. Although the fourth premise is open to challenge, it too seems plausible, given the magnitude of the harm wrongly inflicted by Eichmann and the clear culpability he bears for his action. Yet if we accept all of these premises, then we ought to reject the various accounts of evil action which suggest that evil action must be malicious, sadistic, or defiant. By thinking of evil in accordance with the thick account, we would be led to misorder wrongs. In contrast, no such misordering of wrongs is produced by a psychologically thin conception of evil action. According to the thin account, Eichmann’s order does count as an evil action despite its not being malicious, sadistic, or defiant, and hence this account does not compel us to rank it as less grave than other evils. There is a second, related argument against a psychologically thick account of evil action. It is that by employing a thick conception of evil action we are more likely to ascribe malicious, sadistic, or defiant motives to extreme wrongdoers even in cases in which there is clear evidence that no such motives lay behind their wrongdoing. Neiman maps out this chain of reasoning as a warning: The apparent absence of malice or forethought [in many people who committed atrocities during the Holocaust] has proved so disturbing that many observers prefer to argue they were present in subterranean form. Writers like Goldhagen argued that behind a mask of relative tolerance, German culture contained particularly virulent forms of antisemitism. The appeal of such claims derives less from historical accuracy than from philosophical naïveté. An old-fashioned picture of evil as inevitably connected to evil intention is more soothing than alternatives. (Neiman 2003, 271)

After the terrorist attacks of 11 September 2001, Neiman points out, many people who were familiar with Arendt’s examples of non-malicious, non-sadistic, and non-defiant evil actions were unsure whether they ought to call the terrorists’ actions evil, for fear of implying that all evildoers were like the terrorists (Neiman 2003, 285). Yet the terrorists’ actions were evil, Neiman maintains, just as much

126  Conceptual Pluralism about Evil Action as the non-malicious, thoughtless actions performed by Arendt’s Eichmann were evil. The solution to the problem, it seems, is to continue to use the concept of evil but to make it clear that evil actions need not be malicious, sadistic, or defiant. “Surrendering the word evil to those who perceive only its simplest forms”, Neiman cautions, “leaves us fewer resources with which to approach the complex ones” (Neiman 2003, 286). Once again, we can press this argument into a clearer form using the examples of Arendt’s Eichmann ordering of the murders of many Jews, and a malicious, sadistic, and defiant murder of twenty innocent people in a hate-crime bombing: 1. If an action ø is morally worse than an evil action, then ø is evil as well. 2. The hate-crime bombing is an evil action. 3. Eichmann’s order is morally worse than the hate-crime bombing. Therefore, 4. Eichmann’s order is an evil action. 5. If a thick account of evil action is correct, then every evil action is malicious, sadistic, and defiant. Therefore, 6. If the thick account of evil action is correct, Eichmann’s order was malicious, sadistic, and defiant. 7. Eichmann’s order was not malicious, sadistic, and defiant. Therefore, 8. The thick account of evil is incorrect. The first premise of this argument follows from the deplorability thesis. The second premise is an evaluative claim that would be accepted by almost all of the philosophers who are engaged in this debate, as well as by many people from outside of the philosophical community. The third premise is an evaluative claim that is made plausible via a comparison of the amounts of harm for which the two agents are properly held responsible in light of these specific actions. The fourth premise follows from the first three. The fifth follows from the definition of evil action contained in the most demanding of the thick accounts, but the argument, suitably modified, could run against any of the thick accounts of evil action. If we accept all of these premises, this argument would undermine a thick account of evil action. If we employed a thin conception of evil action instead, then we could accept the first four premises without thereby being led to ascribe malicious, sadistic, or defiant motives to Arendt’s Eichmann.

Conceptual Pluralism about Evil Action  127

6.4  Responding to Neiman’s Arguments Neiman’s arguments suggest that, even though both the thick and the thin accounts of evil action are supported to some degree by folk intuition, and even though both accounts pick out important moral natural kinds, we have a decisive reason to favour a thin account. The reason, it seems, is that only a thin account will lead us to rank wrongs correctly and to ascribe motives correctly. How ought advocates of the thick accounts of evil action respond to these arguments? One way to block the arguments would be simply to deny that Eichmann’s order is morally worse than the hate-crime bombing that kills twenty innocent people. For instance, we could claim that the malice and sadistic pleasure that are manifested in the bombing make it morally more deplorable than Eichmann’s order, even though the scale of harm inflicted by Eichmann’s action far exceeds that inflicted by the hate-crime murder. I think that such a claim would be implausible. Particular kinds of vicious motives can increase the magnitude of wrongness of an action, but so too can the magnitude of the harm that the wrongdoer intends to inflict, and Arendt’s Eichmann intentionally inflicted a staggeringly large amount of harm, regardless of whether he did so maliciously, sadistically, or defiantly. The most contentious premise in Neiman’s arguments is, in fact, the deplorability thesis: if an action is evil then it is morally worse than any non-evil wrong action. It is not surprising that people who think that evils are merely extreme culpable wrongs either implicitly assume the deplorability thesis, or merely state it without arguing for it. According to the thin account, evil actions are distinct from non-evil culpable wrongs simply in that they are more extreme, which makes it seem that any evil action would have to be morally worse than any non-evil wrong. Yet people who think that the concept of evil is psychologically thick would take this assumption to be highly contentious. Advocates of thick accounts, including Steiner, Perrett, McGinn, Singer, and Kekes, think that the concept of evil action is analogous in some respects to what might be called a vicious action concept, such as cowardly action, greedy action, or disloyalty.9 Just as a wrong action can be a disloyal betrayal only if the agent meets quite detailed and specific psychological conditions, advocates of thick accounts say that an action can be evil only if the agent acts out of malice, with sadistic pleasure, or in defiance of morality. But no one would suggest that any action that is a disloyal betrayal is morally worse than any action that is not a   Advocates of the thick account do not think that the concept of evil action is analogous in all respects to these vicious action concepts. Saying that an action is a dishonest betrayal does not imply that it is morally wrong, whereas saying that an action is evil does imply that it is morally wrong. 9

128  Conceptual Pluralism about Evil Action disloyal betrayal, or that any action that is greedy is morally worse than any action that is not greedy. Why, then, should advocates of a thick account of evil action agree that any evil action is morally worse than any non-evil action? They ought to respond by saying that some non-evil wrongs are morally worse than some evil actions. Maliciousness makes an action morally worse, ceteris paribus. Sadistic pleasure makes an action morally worse, ceteris paribus. Knowing defiance of morality makes an action morally worse, ceteris paribus. But other things often are not equal, and an evil action can be trumped in moral magnitude by a non-evil but extremely harmful culpable wrong. By rejecting the deplorability thesis, advocates of the thick accounts can block both of Neiman’s arguments against their view. For example, they can admit that Eichmann’s order, as described by Arendt, was morally worse than the hate-crime bombing, and admitting this need not lead them to misorder wrongs or falsely to ascribe malicious or sadistic motives to all extreme wrongdoers. They should maintain that the order given by Arendt’s Eichmann was an atrocious action that was morally worse than many evil actions, but that it was not itself an evil action. The fact that Neiman’s arguments against the thick accounts of evil action are not decisive does not imply that she had no reason to be worried about the misordering of wrongs, or about false ascription of malice and defiance to extreme wrongdoers. These errors do occur, and they are extremely likely to occur in people who accept the deplorability thesis and also take evil action to be psychologically thick. Moreover, recent advocates of the thick accounts of evil action have not stated clearly that they reject the deplorability thesis, and this failure makes them seem vulnerable to Neiman’s arguments. What Neiman fails to note, though, is that there are two quite different ways in which we could avoid falling into the errors of misordering wrongs and falsely ascribing malicious motives. The first is the route that she favours: combining the deplorability thesis with a thin conception of evil action. The second, which she overlooks, is rejecting the deplorability thesis and employing a thick conception of evil. Simplistic thinking about evil ought to be avoided, as Neiman points out. Yet people who suppose that evil requires malice, sadistic pleasure, or defiance are not doomed to think about evil in a simplistic way. Ultimately, Neiman’s arguments do not give us a reason to favour the thin account of evil action over its thick rivals.

6.5  Defending Conceptual Pluralism When we encounter a range of competing philosophical theories it is reasonable to hope that we will be able to identify grounds for favouring one of them. As we have seen, there are many conflicting accounts of evil action on offer. I have argued that

Conceptual Pluralism about Evil Action  129 we have decisive grounds to reject some of these accounts. For instance, I argued that Steiner and McGinn are mistaken in supposing that an action is evil if and only if it is a sadistically pleasurable wrong, that Singer is mistaken in claiming that evils can be identified via a connection with incomprehensibility, and that Garrard is mistaken in thinking that evils are marked out by a complex property involving deliberative silencing. In addition to arguing against these particular theories I have attempted to clarify and flesh out the theoretical options in this field. What all of us should agree on is that an action is evil only if it is an extreme culpable wrong, and I have argued for a particular view of the nature of that extremity. Beyond that common core there is significant disagreement between people who think that an action is evil only if it also possesses a psychological hallmark, such as being malicious, sadistically pleasurable, or defiant, and people who believe that every extreme culpable wrong is an evil action. In this and the previous chapter, I have considered a range of arguments that might lead us to prefer a thin account of evil over its thick rivals, or vice versa. I have argued that we have good reason to reject the non-folk version of the thick account offered by Garrard, but I have failed to identify clear grounds for favouring the thin account over its folk thick rivals, or vice versa. It is surprising and somewhat dispiriting to reach this conclusion. Both the thin account and the folk versions of the thick account fit well enough, though far from perfectly, with the scattered array of folk judgements about evil actions. Both the thin account and its folk thick rivals pick out important natural kinds within the moral sphere, and both kinds of account are applicable and useful. The more sophisticated philosophical arguments considered in this chapter also fail to indicate which approach should win out. I have argued that advocates of thick and thin accounts alike ought to reject the claim that evil actions are strongly qualitatively distinct from non-evil wrongs, so intuitions about qualitative difference do not favour thick accounts. Moreover, utilizing a thick concept of evil will not necessarily lead to errors in ranking wrongs or in ascribing motives, so Neiman has not identified grounds for preferring a thin account. If I am right in suggesting that there are several viable accounts of evil action but no single best account, then we ought to accept conceptual pluralism about evil action (cf. Braddon Mitchell and Nola 2009, 8). The conceptual pluralism that I advocate is not the unrestricted view that any possible account of evil is as good as any other, or that anything goes when it comes to thinking about evil. Rather, it is the view that there are a number of folk conceptions of evil action, as reflected in competing traditions of use of the word “evil”, that some of these folk conceptions can be shaped into coherent and viable philosophical accounts of evil action, and there are no good reasons for thinking that any one of these viable accounts

130  Conceptual Pluralism about Evil Action picks out what really are the evil actions (cf. Dupré 1993, 57; Sider 2001, 189). I contend that philosophers who claim that one of these accounts is the single correct account of evil action are either making an unjustified claim, or are merely stipulating that they will use the term “evil” to express one particular conception, and encouraging us to follow suit. One consequence of adopting a restricted conceptual pluralism about evil action is allowing that some disputes over whether a particular action is evil are merely linguistic disagreements. Roughly speaking, a merely linguistic disagreement occurs when interlocutors agree in their beliefs and value judgements but disagree over whether a particular sentence is true (cf. Robinson 2009, 325). For instance, let us imagine that René claims that there is a chandelier hanging in the local restaurant, The Little Snail, and Alain disagrees. As René and Alain proceed to argue about this, they might gradually discover that they agree entirely about the shape, structure, and material constitution of the light fitting that hangs in The Little Snail, but that Alain believes that chandeliers must feature cut glass or crystal, whereas René believes that chandeliers can be made out of pieces of opaque plastic. In this case, René and Alain are employing two different conceptions of chandelier, but they agree about all of the non-linguistic facts regarding the light fitting that hangs in The Little Snail. Moreover, there is no deeper fact about what really is a chandelier, and no good pragmatic reasons for using one conception rather than the other. What Alain ought to say is, “Well, if that is what you mean by ‘chandelier’, then I agree that there is a chandelier hanging in that restaurant” (cf. Jackson 1998, 32). Similarly, in some cases of disagreement over whether an action is evil, it will turn out that the interlocutors are simply employing two different conceptions of evil action, and that they agree on every detail of the agent’s motives and intentions, and every detail of the action’s relation to moral duties and the magnitude of its harmful effects, and on the appropriate practical response to the action. Moreover, if each conception of evil fits well enough with folk usage and is philosophically respectable, then there is no deeper fact about which conception picks out the real evil actions. In such a case there is no significant moral disagreement between the interlocutors after all. If René and Alain agree that Himmler’s actions were extreme culpable wrongs, and were malicious but not defiant or sadistically pleasurable, and deserve our strongest moral condemnation, but they merely disagree as to whether an evil action must be sadistically pleasurable, then it appears that their disagreement is merely linguistic. “If that is what you mean by evil”, Alain might say, “then we can agree that Himmler’s actions were not evil.” When we discover that a dispute is merely linguistic, we expect that the disagreement will dissolve, and that neither party will continue to argue about the original issue.

Conceptual Pluralism about Evil Action  131 In suggesting that some disputes over whether particular actions are evil are merely linguistic disagreements, I am not claiming that all such disputes will dissolve in this way (cf. Sider 2001, 198–200). Some conflicting claims about evil really do bite rather than slide past one another. For example, let us imagine that the relatives of children who have been killed as so-called collateral damage in military bombings claim that the military actions were evil, and that a politician maintains that the actions were not evil, but were a necessary consequence of waging a just war. It is unlikely that this disagreement will simply dissolve when the victims’ relatives point out that they are employing a thin conception of evil according to which malice, sadistic pleasure, and defiance are not necessary for evil action. The politician is likely to say that, even if we agree to use a thin conception of evil action, this bombing was not evil. The reason that this particular dispute over evil does not dissolve into a merely linguistic disagreement is that both sides in the dispute agree that all evils are extreme culpable wrongs, but the victims’ relatives see this bombing as a culpable wrong, whereas the politician sees it as a justifiable right action that has some terrible consequences. Similarly, non-dissolving disagreements over whether an action is evil might occur when interlocutors have different views about the degree of moral extremity of an action that they agree is morally wrong, and when they disagree as to whether the agent was motivated by malice, or was knowingly defiant of morality. If we accept a restricted conceptual pluralism about evil actions, then we will diagnose some but not all disputes over whether a particular action is evil as being merely linguistic disagreements in which neither party is making a mistake, and in which there is no actual disagreement in substantive moral judgements. Given that several conceptions of evil are reasonable and no single conception compulsory, sometimes the most apt question is not “Is this action really evil?”, but “Which conception of evil is in play?” (cf. Robinson 2009, 328). This refusal to say unambiguously what evil actions really are may seem pragmatically costly. It could be argued that conceptual pluralists allow an avoidable ambiguity to persist, and thus perpetuate confusion in discussions involving evil. When two people disagree over whether an act of terrorism was evil, it might well be that they agree on all of the fine-grained moral facts and merely are employing different but equally respectable conceptions of evil. Yet the disputants in this case might remain unaware that they are employing different conceptions of evil, and hence mistakenly believe that they are having a substantive disagreement. If philosophers came down in favour of one and only one of the viable conceptions of evil, then perhaps we could stamp out this kind of misunderstanding. Even if we have no grounds for favouring one conception as the right conception, it could be argued that we ought to choose one

132  Conceptual Pluralism about Evil Action arbitrarily and hope that our philosophical advocacy would eventually lead to a widely accepted univocal meaning of “evil action”. It is far from obvious that even the most unified and vociferous philosophical advocacy could reform linguistic practice and resolve the current ambiguity of “evil action”. However, choosing one conception and trying to stamp out the others is not our only available means of fostering clear thinking and effective communication regarding evil. Another is to raise awareness of the fact that multiple conceptions are in play, some but not all of which are viable, and to encourage people who condemn certain actions as evil to specify which of the viable conceptions of evil action they wish to employ. By way of analogy, let us consider the word “parent”. It is not the case that just anyone can correctly be described as the parent of a particular child, yet there are two well-established but divergent uses of the word. “Parent” can be used to express the concept of a biological parent or the concept of the socio-legal parent of a child. Of course, it is possible for someone who is not the biological parent of a child to be the socio-legal parent of that child, and vice versa. This ambiguity can give rise to confusion, and some disagreements about who are the parents of a particular child will turn out to be merely linguistic disagreements. The solution to this problem, though, is not for philosophers to mandate that the real parents are biological parents, for instance, and to maintain that anyone who claims that someone is a parent merely in virtue of being a legal guardian and caregiver is making some kind of mistake (cf. Jackson 1998, 32). Rather, it is to note that in some particular cases it is important to ask whether “parent” means biological parent or socio-legal parent. I think that the concept of evil action is like the concept of parenthood in this respect. There are divergent but defensible traditions of use of the term “evil action”, and it is not the case that one and only one of these picks out the real evil actions. When someone claims that an action is evil, it is important to ask whether the claimant takes this to mean that it is an extreme culpable wrong, or that it is also malicious, sadistic, or defiant. In the remainder of this book, I will usually employ the psychologically thin conception of evil action, but I will draw attention to those cases in which things might go differently were we to adopt one of the viable thick conceptions instead.

7 Aggregative Accounts of Evil Personhood Over the past few chapters, I have evaluated competing accounts of evil action, rejected some of them, and recommended that we carefully specify which of several viable conceptions of evil action is in use on any particular occasion. Evil action is only part of the story. There are many occasions on which the word “evil” is applied not to an action but to a person. Hitler was evil, we are often told, and so were some other war criminals and serial killers such as Idi Amin and Freddie West. As is the case with the concept of evil action, the concept of evil personhood clearly is supposed to denote an extreme category, but it is not immediately obvious what it would take for a person to count as evil. Some philosophers maintain that there are no evil persons, and that we would be better off if we excised the concept of evil personhood from our moral thinking. I suspect much of the current scepticism regarding evil flows from concerns about evil personhood rather than evil action, and that we have good reason to be more cautious in judging that a person is evil than in judging that an action is evil. In the next three chapters, I will develop and evaluate two competing answers to the question of what it would take for a person to count as evil. Roughly speaking, the first answer is that an evil person is someone who has done more than a specified amount of evil, and the second answer is that an evil person is someone who would do evil if given the chance. I will call these the aggregative account and the dispositional account, respectively. In this chapter, I aim to develop the strongest possible version of the comparatively neglected aggregative account, according to which being an evil person is equivalent to actually having done enough terribly wrong things. In Chapter Eight, I will argue that a suitably modified dispositional account is preferable to its aggregative rival. While I concluded that there is not a single best account of evil action, I do think that there is a best account of evil personhood, and hence I conclude that

134  Aggregative Accounts of Evil Personhood evil personhood is a dispositional property. Both the aggregative and dispositional accounts suggest that the concept of evil personhood can be understood in terms of a connection to the concept of evil action, but, in Chapter Nine, I will go on to consider whether these accounts need to be supplemented by a condition concerning evil feelings.

7.1  Intuitions about Evil Persons We have seen that Marcus Singer and Daniel Haybron suggest that the concept of evil action can be derived from a prior concept of evil personhood (Haybron 2002a, 280; Singer 2004, 190). However, neither have explained this derivation in sufficient detail, and I have argued that we do not have a good reason to believe that such a derivation is possible. I have joined most of the contemporary philosophers who have written about evil by first giving an account of evil action, and then hoping to build a theory of evil personhood that depends, at least in part, on that prior account (e.g. Card 2002, 21; Calder 2003, 367; Morton 2004, 66; Kekes 2005, 2; Formosa 2008, 233; Russell 2010b, 233). Given that I have advocated a restricted conceptual pluralism about evil action, I intend to construct an account of evil personhood that is compatible with at least the thin account of evil action and the folk versions of the thick account of evil action. It might also turn out that my theory of evil personhood is amenable to some people who reject my claims about the nature of evil action, but who agree that we ought to build an account of evil personhood on a prior account of evil action. I began constructing an account of evil action by identifying and assessing eight widely held intuitions. Now that we are moving on to think about evil persons, I will compile a similar list of intuitions about evil persons. In doing so, I will call on common folk judgements about the extension of the concept, although here, too, there is significant disagreement over who deserves to be called evil. In trying to get at the meaning of “evil person” we must also delineate the inferential role that is played by the judgement that a person is evil. Our method is not only to figure out who the folk commonly judge to be evil, but also to discover what, in judging them to be evil, the folk take them to be. Perhaps we will discover that folk intuitions about evil personhood are mutually inconsistent. Alternatively, we could discover that a definition of evil personhood that respects the folk inferential role of the concept fails to refer to any actual persons. On either of these grounds we may decide to reject some of the folk intuitions about the nature of evil personhood, or to revise common judgements about the extension of the concept of evil personhood.

Aggregative Accounts of Evil Personhood  135 The following seven intuitions about evil persons are particularly deserving of our attention. Some are more contentious than others, but all have some prima facie plausibility, and have been endorsed in recent philosophical literature. (9)  There are some actual evil persons. (10)  Evil persons are rare. (11)  Evil persons deserve our strongest moral condemnation. (12) In some cases the fact that S is an evil person helps to explain why S performed an evil action. (13)  Not every evildoer is an evil person. (14)  It is possible to become an evil person by performing evil actions. (15) It is possible, though very difficult, for an evil person to become a good person. Intuition (9), that there are some actual evil persons, is denied by many people, but fervently asserted by many others. It is endorsed by journalists who declare that terrorists are evil (Morrow 2003, 18), by psychiatrists who say that serial killers are evil (Stone 2009, 243), by historians who say that war criminals are evil (Rosenbaum 1998, xxii), and by philosophers who catalogue a range of cases from psychopaths and religiously inspired murderers to concentration camp commandants and military despots (Card 2002, 21–2; Singer 2004, 190; Kekes 2005, 128; Formosa 2008, 235; Calder 2013, 189). It is possible that all of these people are confused or mistaken, and that evil persons exist only in fiction or in our imaginations (Cole 2006, 173; cf. Rosenbaum 1998, xxi). I will return to consider scepticism about the existence of evil persons in Chapter Ten. For now, though, I will try to develop an account of evil that could vindicate the common claim that there are actual evil persons. Intuition (10), that evil persons are rare, is also relatively uncontentious, and is endorsed by many of the contemporary philosophers who have written about evil. For instance, Garrard assumes that it is implausible that very large numbers of people are evil (Garrard 2002, 321), and Haybron agrees that “[e]‌vil persons are probably neither common nor easily identified, but they do comprise a genuine and interesting class of moral characters” (Haybron 2002a, 278; cf. Kekes 1990, 5). Intuition (11) is also widely shared. Evil denotes a moral extremity when applied to persons, as when applied to actions. Singer is right to declare that “ ‘evil’ is the worst term of opprobrium that can be applied to a human being” (Singer 2004, 190; cf. Kekes 2005, 58). The remaining intuitions are, perhaps, more controversial. Let us consider intuition (12): that evil personhood might help us explain evil actions. While Cole has

136  Aggregative Accounts of Evil Personhood argued that it is never informative to say that “He did it because he was evil”, Cole himself notes that many of the folk offer precisely this explanation of the extreme wrongs performed by child murderers or terrorists (Cole 2006, 141, 212). These particular explanations certainly are contentious, but perhaps there are other cases in which similar claims are both true and informative (Kekes 1990, 7; McGinn 1997, 77–87). For instance, we might say that Melville’s character Billy Budd killed Claggart because he was provoked into an uncharacteristic rage, but that, in contrast, Dennis Rader committed a particular murder because he was an evil person who characteristically sought opportunities to torture and kill innocent victims. I will explore the question of the explanatory role of evil personhood in much more detail in Chapter Ten. For now, though, I claim that we have a defeasible reason to reject any accounts of evil personhood which imply that the claim that “He did it because he was evil” could never be true and informative. Intuition (13) is that not every evildoer is an evil person. An evildoer is simply someone who has performed at least one evil action. A broad range of cases could be offered in support of the claim that not every evildoer is an evil person. In some cases an evil action seems to be anomalous and out of character for the evildoer. Such actions might deserve our strongest condemnation even though the person who performed them is far from the worst kind of person, so it seems plausible that these evildoers are not evil persons. Similarly, some terrible deeds might be the product of the kind of coercion or extreme stress that is suffered by soldiers in wartime. If these deeds are still within the domain of culpable wrongs, then we might call them evil actions despite maintaining that the evildoers did evil because of their difficult circumstances, and not because they were evil themselves. Some evildoers might also fail to be evil persons because they condemn their past evildoing and have undergone some kind of moral reform. If any of these claims sound convincing, then we have a defeasible reason to favour accounts of evil personhood that allow that not every evildoer is an evil person (cf. Card 2002, 22; Garrard 2002, 321; Haybron 2002a, 279; Calder 2003, 367; Morton 2004, 65; Singer 2004, 190; Formosa 2008, 234; Russell 2010b, 232). While explicit advocacy of intuition (13) is plentiful, support for intuition (14)— that it is possible to become an evil person by performing evil actions—is comparatively anecdotal. It is plausible that some political leaders, such as Mao Zedong, are at first well-motivated, but in the struggle to achieve and maintain power are led to perform horribly wrong actions, and, with the acquisition of absolute power, are corrupted absolutely. Such leaders may originally have been strongly opposed to the suggestion that they ought to kill their opponents, but end up ordering murders on a whim. Some soldiers who have worked as torturers come to lose all resistance to the horrendous tasks that they routinely perform (Wolfendale 2007,

Aggregative Accounts of Evil Personhood  137 175), and some even acquire a taste for torture, and seek out opportunities to inflict extreme suffering on innocent victims (Simpson and Bennett 1985, 109). It seems that the repeated performance of evil actions could lead to desensitization and fetishization, and thus could turn someone into evil person (Morton 2004, 38). Finally, let us consider intuition (15): that it is possible, though very difficult, for an evil person to become a good person. A striking feature of the paradigmatic examples of evil persons is their recalcitrance. When someone says that the serial killer Ted Bundy, for instance, was an evil person, they imply that we could not have turned him into a good person simply by sitting him down and explaining to him the error of his ways, or even by threatening him with extreme punishment. Bundy knew exactly what he was doing, and when he escaped from prison for the second time, he quickly returned to murdering women. An evil person cannot easily become good, and this is why we might be tempted to say that an evil person is a write-off who is unfit for any society (Haybron 2002a, 278; cf. Vargas 2010, 76), or that we should not expect to be able to reason with evil terrorists. Philip Cole goes further than this, suggesting that, if any person really was evil, then that person could not be redeemed by any means (Cole 2006, 236). Yet it is not clear that an evil person could not be made good by any means whatsoever. For instance, it does not seem incoherent to claim that even though Ted Bundy was an evil person, a skilled team of neurosurgeons could have rewired his brain such that he ceased to be evil, or that God could have reached down and miraculously converted him into a virtuous man. Some philosophers clearly believe that even without these fantastic or miraculous interventions, an evil person could become good through a slow and difficult process of moral reform and redemption (Card 2002, 21; Thomas 2003, 201). Psychiatrist Michael Stone points to the spree killer Archie McCafferty as an example of an evil man who, Stone believes, eventually was rehabilitated and redeemed (Stone 2009, 144–6). We want an account of evil personhood that allows for at least the logical possibility that an evil person could become a good person, but that explains why it is very difficult for such a transformation to take place in the actual world.

7.2  The Regularity Account of Evil Personhood An adequate account of evil personhood, in addition to being internally coherent, should fit with as many of these seven intuitions as possible. Most of the recent accounts of evil personhood are dispositional. According to such theories, an evil person is a person who is disposed to do evil things. This dispositional model might appear to be the only real option for a theory of person evaluation. It could be argued that the evaluation of a person is simply equivalent to the evaluation of

138  Aggregative Accounts of Evil Personhood that person’s character, and that character is an essentially dispositional notion, as is reflected in dispositional theories of virtue and vice. Yet not all person evaluations are evaluations of dispositional character. For instance, a person is a criminal in virtue of having actually committed a crime, and a person is a hero in virtue of having actually performed at least one heroic deed. Even if it is true that only someone with a certain kind of character would perform such a deed, having that character alone is insufficient for being a hero or for being a criminal. It makes no sense to say that Sven was a great hero even though he fell off the longboat and drowned before he got the chance to do anything heroic. Some philosophers have claimed that the only way in which a person could be evil is for that person to have performed an evil deed or deeds. Others suggest that the only way a person could be evil is for that person to have performed, or attempted to perform, or intended to perform an evil deed or deeds. Advocates of this non-dispositional approach see evil persons as more closely analogous to moral criminals than to vicious persons. This non-dispositional approach has been considered and dismissed by Morton (2004, 65), and by Haybron, who maintains that we ought not “confuse the evil person with the moral criminal” (Haybron 2002a, 261). Yet Morton and Haybron have considered only the most simplistic of the non-dispositional accounts. We should explore the structure of such accounts in more detail and develop the strongest possible version before evaluating it in comparison with its dispositional rivals. In what follows I will propose a series of possible non-dispositional accounts of evil personhood, moving step by step towards what I think is the most appealing version of this kind of theory. The first non-dispositional theory of evil personhood that I will consider is the regularity account. According to the regularity account, S is an evil person if and only if S frequently or regularly performs evil actions. Stone seems to endorse this view when he claims that, when we call a person evil, “we imply that the person can be counted on to commit such acts habitually and often” (Stone 2009, 23). John Kekes maintains that “[p]‌eople can be said to be evildoers if they habitually perform evil actions” (Kekes 2005, 2), and that, if “agents are regular sources of evil . . . we can identify them, and not merely their actions and character traits, as evil” (Kekes 1990, 48), It is not clear whether Kekes intends regular evildoing to be a necessary as well as a sufficient condition for evil personhood, but if he does then he is committed to the regularity account. The regularity account of evil personhood fits neatly with intuitions (9)–(11) and (13)–(15). Since some but not many people regularly perform evil actions, the regularity account implies that evil persons exist but are comparatively rare. If evil persons regularly perform the worst kind of culpable wrongs, then it is plausible that evil persons deserve our strongest condemnation. Since many evildoers are

Aggregative Accounts of Evil Personhood  139 not frequent evildoers, the regularity model implies that not every evildoer is an evil person. If counting as an evil person is a matter of performing evil actions sufficiently regularly, then it does seem that a person can become evil by performing evil actions. Moreover, since someone can frequently perform evil actions for a period of his life but then cease to perform evil actions frequently, the regularity account implies that an evil person could become a good person. Despite these strengths, it appears that the regularity account of evil personhood does not fit with intuition (12); that is, it does not allow that evil personhood could help to explain the performance of evil actions. If being an evil person were equivalent to performing evil actions regularly, then it would be circular to say that S performed that evil action because S is an evil person. At best, the claim that S is an evil person could alert us to a pattern in S’s actions, and the fact that S is an evil person might be taken to explain some of S’s actions in so far as it implies that those actions are not anomalous. But if the regularity account were true then the fact that S is an evil person could not causally explain why S performed certain actions, and we might prefer an account of evil personhood that can support genuine causal explanations of actions in terms of evil. The more significant objection to the regularity account, though, is that it does not fit well with common intuitions about which actual persons count as evil. Let us consider a brooding spree killer such as Martin Bryant, who lived a troubled but comparatively quiet life until 28 April 1996, when he went on a shooting spree in Port Arthur, Tasmania, in which he murdered thirty-five people (Wainwright and Totaro 2009). A brooding spree killer does seem to be a good candidate for evil personhood even if, during the years before the spree, he merely fantasized about killing and planned the attack, and did not inflict or attempt to inflict any serious harm. Advocates of the regularity account might claim that such a spree killer became an evil person only at the moment when he began regularly to perform evil actions. This seems implausible, though, especially since a possible explanation of the fact that the would-be killer planned his attack is that, during those years of planning, he was an evil person. An evil person could perform his evil actions not habitually and regularly but sporadically, infrequently, or all in rush. The irregularity of his evildoing does not disqualify him from being an evil person.

7.3  Aggregative Accounts of Evil Personhood By adjusting the regularity account slightly we could bring our theory of evil personhood more closely into alignment with common intuitions concerning brooding spree killers. Rather than setting a threshold for evil personhood in terms of the frequency or regularity of evil actions, we might set that threshold in terms of

140  Aggregative Accounts of Evil Personhood an aggregate. According to the simplest version of the aggregative account, S is an evil person if and only if S performs more than a specified number of evil actions. As we shall see, though, there are a range of alternative aggregative accounts according to which we ought to add up not the number of S’s evil actions, but some other things that are related to S’s actions. Thus, speaking generally, aggregative accounts of evil personhood suggest that S is an evil person if and only if the aggregate of something closely related to S’s actions exceeds a specified threshold.1 All aggregative accounts have a significant advantage over the regularity account in that the aggregative accounts preserve the distinction between evildoer and evil person while allowing for the existence of evil persons who perform evil actions very infrequently, sporadically, or all in short period of time. Let us consider in more detail the basic aggregative account according to which S is an evil person if and only if S performs more than a specified number of evil actions. Something like this view seems to be advocated by Marcus Singer, who defines an evil person as one who knowingly performs, wills or orders [evil actions], or remains indifferent to them when performed by another in a situation where one could do something to stop or prevent them. (Singer 2004, 196)

Singer does not specify a precise numerical threshold, but he does go on to claim that someone “who engages in a pattern of [evil action] is evil” (Singer 2004, 197).2 Such an account fits neatly with intuitions (9)–(11). We could set the threshold at a sufficiently large number of evil actions such that some but not many persons count as evil, and it seems obvious that such repeat evildoers deserve our strongest condemnation. The basic aggregative account also accords with intuitions (13) and (14). So long as not every evildoer performs more than the specified number of evil actions, not every evildoer is an evil person, but evildoers can become evil persons by repeatedly performing evil actions. Like the regularity account, the basic 1   There is also room for variation on the issue of whether a person whose aggregate exceeds the threshold becomes an evil person at the moment his aggregate exceeds the threshold, or whether a person who exceeds the threshold at some point in his life counts as an evil person retrospectively, and thus was an evil person before he exceeded the threshold. 2   It is hard to square Singer’s derivative definition of evil person with his earlier claim that evil applies primarily to persons and secondarily to acts. Moreover, there is room for dispute over whether Singer’s definition of evil person actually commits him to the basic aggregative account. Arguably, ordering an evil action is itself an evil action, in that the person who issues the order is morally responsible for the harm that wrongfully will be inflicted if the action is successfully carried out. Similarly, failing to prevent extreme suffering when it would be easy to do so seems to be an evil action. It is not clear what Singer means by “willing” evil actions. If “willing” means striving to perform evil actions, then Singer’s definition ties evil personhood very closely to evildoing. If “willing” means intending to perform evil actions, then Singer’s account is better understood as a disjunctive action-intention account, similar to that proposed by Formosa (2008). Such accounts are discussed later in this chapter.

Aggregative Accounts of Evil Personhood  141 aggregative account does not allow that evil personhood can causally explain evil action, but this defect might be worth accepting if it is outweighed by the theory’s other strengths. Unfortunately, though, the basic aggregative account also fails to accord with common intuitions about which specific persons count as evil. It is plausible that a person could be evil even though he manages to perform only one evil action, if that single action is extreme even within the category of evils. If we accept the basic aggregative account but also want to allow that a misanthropic suicide bomber, who kills himself and fifty victims in a single action, is an evil person, then we would have to set the threshold for the number of evil actions required for evil personhood at one. Yet, by collapsing the distinction between evildoers and evil persons, such a threshold would fail to respect intuition (13). Advocates of the aggregative account could avoid this problem by aggregating a range of things other than the number of evil acts performed by S. For instance, they might claim that S is an evil person if and only if the aggregate of the undeserved harm caused by S’s wrong actions exceeds a specified threshold. According to this basic aggregate-of-harms account, not every evildoer is an evil person, but a one-off evildoer could have wrongly inflicted enough harm to count as an evil person. The basic aggregate-of-harms account suffers from two obvious defects, though. The first defect is that it mistakenly includes the harms caused by S’s wrong actions for which S is not culpable, such as accidental, unforeseeable, or unpreventable harms. S’s unforeseeably calamitous wrongdoing does not seem to be relevant to the question of whether S is an evil person (Card 2002, 20). The second defect of the basic aggregate-of-harms account is that it fails to include the harms that S attempts but unluckily fails to inflict, which seem to be relevant to the question of whether S is an evil person. We can avoid both of these problems by modifying the account such that S is an evil person if and only if the aggregate of the undeserved harm for which S can be held morally responsible, plus the undeserved harms for which S would have been morally responsible had S’s actions succeeded as intended, exceeds a certain threshold. This more sophisticated version of the aggregate-of-harms account of evil personhood is attractive in many respects, but it remains vulnerable to a significant objection concerning the aggregation of minor harms. Let us consider Methuselah, who we may suppose is a comparatively ordinary person who possesses some virtues and some minor vices, but who lives for a very long time. Since Methuselah lives for so long he ends up committing a great many culpable wrongs, most of which cause a small amount of undeserved harm for which Methuselah is morally responsible. If Methuselah lives for long enough, the aggregate of these harms will end up exceeding any plausible threshold set by an aggregative model of

142  Aggregative Accounts of Evil Personhood evil personhood. According to the sophisticated aggregative account, Methuselah would count as an evil person even though he never performed a single action that would have been sufficiently harmful to count as evil. The extremity in the amount of harm caused by Methuselah is due to the immense number of Methuselah’s harmful acts, even though none of Methuselah’s particular acts was very harmful on its own. In light of these examples it seems that evil persons are more closely connected to evil action than is entailed by the sophisticated aggregative account of evil personhood.3 Perhaps we can avoid this problem by aggregating the undeserved harms that are connected only to S’s evil actions, in which case S could be an evil person only if S has at least attempted to inflict an extreme harm. According to this account, S is an evil person if and only if the aggregate of the undeserved harms that are or would have been caused by S’s evil actions, had all of those actions succeeded, and for which S is or would have been morally responsible, exceeds a certain threshold. This aggregative account of evil personhood appears to be superior to its predecessors. It suggests that unforeseeably calamitous wrongdoers and Methuselah will fall short of evil personhood so long as they do not perform any evil acts, but that brooding spree killers and unlucky attempters could count as evil. Moreover, it allows that evil persons exist but are rare, that evil persons deserve our strongest moral condemnation, that not every evildoer is an evil person, and that a person could become evil by performing evil actions.

7.4  Aggregating from Evil to Good There remains a significant problem for all of the aggregative accounts considered so far. We have seen that the regularity account of evil personhood fits with intuition (15); that is, it allows that an evil person could become a good person by no longer performing evil actions with sufficient regularity. Moral transformations of this kind might well be uncommon, but they do seem to be possible (e.g. Stone

  Advocates of psychologically thick accounts of evil action, including Kekes, McGinn, Steiner, and Perrett, will be especially likely to reject an account of evil personhood that merely aggregates the harms for which S is morally responsible. According to psychologically thick accounts, evil actions can flow only from specified kinds of motives, whereas harmful wrong actions can flow from almost any set of motives. Thus, it would be odd for an advocate of the psychologically thick account of evil action to suggest that S could be an evil person despite the fact that S totally lacks the distinctive motives required for evil action. In contrast, advocates of the psychologically thin account of evil action, including Arendt, Card, Neiman, and Formosa, do not think that there are any distinctive motives required for evil action, so they should be comparatively more willing to allow that Methuselah could turn out to be evil. Yet even advocates of the thin account might maintain that evil persons cannot count as evil merely in virtue of racking up an enormous number of minor harms. 3

Aggregative Accounts of Evil Personhood  143 2009, 144–6). Laurence Thomas imagines a man who performed evil actions as a Nazi during the Second World War, escaped to the USA and became rich, and remorsefully donated his money to Jewish charities while living on modest means for fifty years (Thomas 2003, 201). Thomas suggests that in 1945 this man was an evil person, but that in 1995 he was not. Yet according to the aggregative accounts considered so far, once S has crossed the threshold for evil personhood it is logically impossible that S could cease to be an evil person by becoming good. This would imply that even God or a skilled team of neurosurgeons could not turn S back into a person who is not evil. We ought to consider whether there is any way to modify the aggregative account of evil personhood so as to allow that someone like Thomas’ ex-Nazi do-gooder could turn from an evil person into a good person. We could achieve this result by claiming that it is possible for an evil person to lose his evil status by doing so much good that he counterbalances the harm he has inflicted.4 According to this balanced aggregative account we might claim that S is an evil person if and only if the aggregate of the undeserved harms that are or would have been caused by S’s evil actions, had all of those actions succeeded, and for which S is or would have been morally responsible, minus the benefits for which S is morally responsible and the benefits for which S would have been morally responsible had S’s actions succeeded, exceeds a certain threshold. Kekes suggests that a similar kind of balanced aggregate is centrally important in moral evaluations of persons: Moral achievement depends on the balance of good and evil one has caused. From the moral point of view, this is the most important question about moral agency and about the moral judgment of character. (Kekes 1990, 105)

If we evaluate the ex-Nazi do-gooder against this standard, he might well have transformed from an evil person into a good person. This result counts in favour of the balanced aggregative account of evil personhood. Unfortunately there are other examples of possible moral transformation that are not covered by the balanced aggregative account. Let us imagine another ex-Nazi who, like the do-gooder, was strongly committed to murdering Jews, but who eventually is filled with remorse and never commits evil actions again   In order to ensure that the benefits produced by S’s actions are a moral credit to S, we should stipulate that the benefits must accrue to people other than S himself, and that the people to which the benefits accrue are not people who deserve not to receive benefits. An ex-Nazi who secures many benefits merely for himself or for his ex-Nazi cronies, who deserve not to receive benefits, would not thereby counterbalance the harm done by his evil actions. In contrast, an ex-Nazi who acts so as to produce great benefits for other people who are not undeserving might thereby come to lose his status as an evil person. 4

144  Aggregative Accounts of Evil Personhood after 1945. This ex-Nazi, unlike the do-gooder, does not even attempt to perform enough beneficial acts to counterbalance the undeserved harm for which he is responsible. Instead he lives a long and quiet life of modest virtue. Some people would accept that in 1995 this modestly virtuous ex-Nazi is no longer an evil person, even though it remains true that he is a moral criminal and is properly held responsible for his past wrongdoing. If this person no longer holds false moral views, and is no longer any kind of moral danger, it might seem misleading to call him an evil person. There is room for disagreement as to whether the relevant moral transformation could have taken place in this case. It could be argued that an evil person can cease to be evil only by making amends for his evil deeds, which the ex-Nazi do-gooder has done, but the so-called modestly virtuous ex-Nazi has not. In many cases it is true that we can make amends for wrongdoing by acting so as to benefit those whom we have harmed, or so as to benefit other people generally. However, this is not the only possible reaction to our own wrongdoing that can help to transform our moral status. Perhaps the modestly virtuous ex-Nazi believes that the harm he inflicted was so great that he could never outweigh it with beneficial deeds, but he apologises, renounces his past actions and attitudes, and either accepts punishment or punishes himself severely for his own wrongdoing. If this change in the ex-Nazi appears to be deep-rooted and long-lasting, we might judge that in 1995 he no longer is an evil person despite his not having outweighed his evil deeds with beneficial actions. The fact that the balanced aggregative account cannot make room for this kind of moral conversion renders it less plausible. A different modification of the aggregative account could allow for a moral transformation from evil to good even in the case of the modestly virtuous ex-Nazi. Instead of introducing a balanced aggregate of undeserved harms minus benefits, we could take the aggregate of harms from a segment of S’s life, rather than from S’s life as a whole. According to the segmented aggregative account, S is an evil person if and only if, in a specified period of time, the aggregate of the harms that are caused by S’s evil actions, and for which S is morally responsible, plus the harms for which S would have been morally responsible had S’s actions succeeded, exceeds a certain threshold. When we consider the segment of the Nazi’s life that stretches from 1940 to 1945, we might judge that he is an evil person due to the extreme amount of undeserved suffering for which he is morally responsible. When we consider the segment of that same person’s life that stretches from 1990 to 1995, we might judge that he is no longer an evil person because during those later years he has not attempted to perform actions that, if successful, would have caused much undeserved harm at all.

Aggregative Accounts of Evil Personhood  145 The segmented aggregative account is attractive in many respects. It fits with intuitions (9)–(11) and (13)–(15), and seems to accord with widely held intuitions regarding brooding spree killers, unlucky attempters, unforeseeably calamitous wrongdoers, Methuselah, ex-Nazi do-gooders, and moderately virtuous ex-Nazis. Unfortunately, though, it looks like the segmented aggregative account is internally inconsistent. Let us imagine that S lives from T1 to T6 (that is, from Time 1 to Time 6), and that between T1 and T2 S performs several extremely harmful evil deeds, the aggregate of which exceeds any plausible threshold for evil personhood. Further, let us suppose that between T3 and T6 S performs no very harmful actions and some good deeds. Is S at T3 an evil person? According to the segmented aggregative model the answer to this question will depend on which segment of S’s life we consider. If we consider the segment of S’s life from T1 to T3, then S is an evil person at T3, but if we consider the segment of S’s life that stretches from T3 to T6, S is not an evil person at T3. These segments are equal stretches of time, so it is hard to see how either could be ruled out as not relevant to S’s moral status at T3. Therefore, according to the segmented aggregative account, S at T3 both is an evil person and is not an evil person. This apparent contradiction seems to give us a good reason to reject the segmented aggregative account of evil personhood. There are a range of ways in which advocates of the segmented aggregative account could try to avoid this contradiction. One option is to maintain that the aggregative criterion is merely a fallible indicator of evil personhood, just as a cricket or baseball player’s batting average in the current season can be used as a fallible indicator of that player’s current batting ability or prowess. If we adopt this model, we could admit that S at T3 meets the segmented aggregative criterion but maintain that he is not actually an evil person at T3, as is revealed by his subsequent failure to perform evil actions. Yet if we treated the aggregative criterion as merely an epistemic criterion, we would need to provide an additional metaphysical account of the necessary and sufficient conditions for evil personhood. If this metaphysical account pointed to S’s current dispositions to perform evil actions, then we would have shifted to the more popular dispositional model of evil personhood. Alternatively, if the metaphysical account was also aggregative, it seems that it too would be vulnerable to the objections that I have already raised. Another way in which we might try to prevent the segmented aggregative account from generating contradictions is to claim that lives fall into natural segments, and hence that S at any particular moment ought to be morally assessed with reference to only one segment of his life. Given that we are often not in a position to tell where the natural divisions fall, and hence which segment of S’s life is relevant to his status as an evil person, it is possible that what seems to be clear evidence for S’s being an evil person at T3 is actually misleading. If this account is to

146  Aggregative Accounts of Evil Personhood be anything more than an ad hoc solution to this problem, though, it must include an explanation of why the natural divisions fall where they do. Perhaps such an explanation could be given in non-dispositional terms. For instance, we could claim that instances of evildoing fall into patterns, and that the boundaries of the segments of a life that underlie the facts about evil personhood are the boundaries between stretches of regular evildoing and stretches of very infrequent evildoing. As we have seen, though, it is not plausible that evil persons must be regular evildoers. There are a variety of reasons, including prudent planning or a dearth of opportunity, for which an evil person might not perform evil actions for a long stretch of time (e.g. Stone 2009, 238). The most plausible way to divide a life into natural segments of this kind is to locate points at which that person’s dispositions change: moments of religious conversion, or moments at which temptation breaks through and destroys a hitherto effective barrier against wrongdoing. Yet if we identify the boundaries of natural segments in these dispositional terms, it is hard to see how we could avoid conceding that it is the dispositional qualities of S that determine S’s status as an evil person, and that aggregates of evildoing are mere fallible indicators of this underlying property. There are several other ways in which we could prevent the segmented aggregative account from generating contradictions without collapsing it into a dispositional account. For the ease of formulation of these views, let us call a segment of a specified length of S’s life in which S has exceeded the aggregative threshold for evil personhood an e-segment, and a segment of that length in which S has not exceeded the aggregative threshold for evil personhood a non-e-segment. We could avoid contradictions by making e-segments trump non-e-segments, or vice versa. For instance, we could claim that S is an evil person at T if and only if S falls into an e-segment at T, even if S also falls into a non-e-segment at T. On this account, the fact that S does fall into an e-segment at T3 trumps the significance of the fact that S also falls into a non-e-segment at T3, and S straightforwardly comes out as an evil person at T3. Alternatively, we could claim that S is an evil person at T if and only if S falls into an e-segment at T and does not fall into a non-e-segment at T, in which case S straightforwardly comes out as not an evil person at T3. Finally, we could avoid contradictions by building truth-value gaps into the segmented aggregative account. On such an account, if S at T falls into both an e-segment and a non-e-segment, then it is neither true nor false that S is an evil person at T. It is not clear which, if any, of these complex versions of the segmented aggregative account of evil personhood best match our intuitions about evil persons. Given the complexity of these accounts, some advocates of an aggregative approach might prefer to stick with the comparatively simple balanced aggregative account, although in doing so they would have to deny that the so-called modestly

Aggregative Accounts of Evil Personhood  147 virtuous ex-Nazi could have ceased to be an evil person. In any case, the balanced aggregative account and the segmented aggregative account are the strongest of the aggregative theories of evil personhood. Both versions fit with many of our intuitions about evil persons, including the intuitions that evil persons exist but are rare, that evil persons deserve our strongest moral condemnation, that not every evildoer is an evil person, that it is possible to become an evil person by performing evil actions, and that it is difficult but possible for an evil person to cease being evil and become a good person. If we are deciding whether we should adopt an aggregative account of evil personhood, it is these sophisticated theories that should be tested against their dispositional rivals. It is to this task that I now turn.

7.5  Evil Personhood and Luck As we have seen, none of the aggregative accounts allow that evil personhood could count as a non-circular explanation of evil action. They force us to reject as incoherent the claim that “He did it because he was evil”. In addition to this problem, all of the aggregative accounts of evil personhood have difficulty in respecting some widely shared intuitions about evil personhood and luck. Aggregative accounts suggest that being an evil person is a matter of doing or attempting to do certain things. Yet what a person does is partly explained by what that person is like and partly explained by the environment in which that person is located. There are many cases in which we choose to move from one environment to another, and in which we shape the features of our environment, but even in such cases there remains much about the environment that is beyond our control. In this sense, many features of the environment that influence our actions are a matter of luck. There has been much disagreement amongst philosophers over the possibility of various kinds of moral luck (Nagel 1979, 28; Williams 1981, 21; Rescher 1993, 154). It is not plausible that the matter of whether S is an evil person must be immune to all kinds of luck. The conditions in which S was raised and the quirks of S’s genome are both a matter of luck, but they greatly influence the kind of person that S becomes. It seems that whether S turns out to be an evil person depends to some degree on what Nagel calls “constitutive luck” (Nagel 1979, 32). Nonetheless, there are some cases in which two persons are exactly alike but attempt to perform very different actions because, by chance, they happen to be in different environments. These cases of what Nagel calls “circumstantial luck” are problematic for aggregative accounts of evil personhood (Nagel 1979, 34). Let us consider two misanthropes, S1 and S2, who maliciously desire to cause suffering and loss of life. Both S1 and S2 carefully plan to carry out separate acts of

148  Aggregative Accounts of Evil Personhood suicide bombing, each of which, if successful, will kill one innocent person in addition to themselves. For the sake of argument, let us assume that the amount of harm caused by each one of these actions, were it successful, would not be sufficient on the aggregative accounts to make either S1 or S2 an evil person. Suppose that both S1 and S2 attempt the bombing, that S1 is successful, but that, by chance, S2’s equipment malfunctions and he is unsuccessful. At this point, according to the aggregative accounts, neither S1 nor S2 is an evil person. Yet suppose that, undeterred by his failure, S2 goes on to make several other attempts on different targets which also fail due to chance. So long as the murders of these various targets are taken to be distinct harms, the undeserved harm that S2 wrongfully attempts to inflict far outweighs that inflicted by S1, and carries S2 over the threshold for evil personhood. Some people might simply agree that S2 is a morally worse person than S1, given that he made more evil plans than S1, and that he was persistent and hence deeply committed to inflicting harm. But let us suppose that, had S1’s attempt failed, he too would have continued trying to inflict carnage. In this situation it is hard to maintain that S2 is an evil person but that S1 is not.5 It does not seem intuitively plausible that unlucky but persistent suicide bombers are more likely than first-time-lucky suicide bombers to be evil, but this is exactly the result that appears to be secured by aggregative accounts. Let us consider another case featuring two more misanthropes, S3 and S4, who maliciously desire to cause suffering and loss of life, but who, by chance, live in very different societies. S3 lives in a corrupt society in which wealthy and well-connected people who commit atrocities can pay off the authorities and escape punishment. In contrast, S4 lives in a society in which a very efficient police force reliably detect and severely punish all extreme wrongdoing. S3 happens to be wealthy and well-connected, and goes ahead and satisfies his desire to cause extreme suffering and loss of life. S4 has the same level of wealth and the same desires as S3. Had he lived in the same society as S3, he too would have committed many evil deeds. Because S4 knows that he runs a very high risk of punishment, though, he never even attempts to commit an extremely harmful wrong. S3 is an evildoer and S4 is not an evildoer, and this fact might lead us to judge that S3 is deserving of punishment whereas S4 is not. But are we willing to say that S3 is an evil person whereas S4 is not, even though S4 is perpetually frustrated, and wishes that he lived in a society in which he could carry out his malicious fantasies without the risk of punishment?

5   We could try to avoid this problem by reverting to an aggregative account that focuses on the actual undeserved harms inflicted by S, and ignores the undeserved harm that would have been inflicted by S were S’s actions successful (cf. Card 2002, 20). This would have the implausible consequence that unsuccessful attempts to inflict harm are never sufficient for evil personhood.

Aggregative Accounts of Evil Personhood  149 It is not clear how we could alter aggregative accounts of evil personhood so as to allow that first-time-lucky suicide bombers are no less likely to count as evil than their unlucky but persistent comrades, nor so as to allow for the possibility that a perpetually frustrated misanthrope is an evil person. Advocates of aggregative accounts could simply maintain that perpetually frustrated misanthropes are frightening, horrible, and deeply immoral, but not evil. Unless they can provide a more detailed account of this distinction, and point to intuitively plausible cases in which it should be made, this response seems to be ad hoc. When we ask whether a person is evil it seems we are making a broad assessment that takes into account what that person would do if, by chance, he was placed in a different situation: if his attempts were unsuccessful, or if threats of punishment were removed. This is not equivalent to asking what the person would do if he was a different kind of person. Surely that question is not relevant in our evaluation of the person as he is now. But if we want to get at the moral worth of a person as he is now—if we want to know whether he now is morally admirable or despicable—then we need to know what he is disposed to do in a range of situations. Since it is plausible that an evil person might, through luck, fail to perform any evil actions, it seems that when we ask whether a person is evil, we are hoping to evaluate the dispositional character of that person, not merely that person’s actions.

7.6  Adding Intentions Some philosophers, including Card and Formosa, claim that a person could be evil in virtue of his evil actions or his evil intentions. It is worth considering whether this kind of disjunctive account can make room for our intuitions about evil persons and luck within a broadly aggregative framework. Card’s view appears to be that, in order to be sufficient for evil personhood, these intentions must effectively issue in attempts to inflict extreme harm (Card 2002, 21–2). Thus, Card’s addition of effective intentions will not allow for the possibility of a perpetually frustrated evil person who, due to fear of punishment, never even tries to inflict extreme harm. In contrast, Formosa seems to allow that having completely ineffective intentions to perform evil actions could be sufficient for being an evil person. He claims that an “evil person is an unreformed person who repeatedly perpetrates, or at least intends to perpetrate, evil acts” (Formosa 2008, 233).6

6   Note that Card and Formosa do not suppose that the intention to perform an evil action must be the intention to perform an action that is evil under that description. According to Card and Formosa, someone can intend to perform an evil action while mistakenly thinking that the action is morally right.

150  Aggregative Accounts of Evil Personhood Formosa’s account is attractive in that it allows that not all evil persons are evildoers, but it is not clear that Formosa goes far enough. We can imagine that the society in which the perpetually frustrated misanthrope lives is so efficient in detecting and punishing extreme wrongdoing that he never even forms the intention to perform a specific evil action, even though he is strongly disposed to perform such actions when in a more lenient environment, and wishes for that reason that he were placed in such an environment. This perpetually frustrated misanthrope seems to be an evil person despite his lack of evil intentions. I contend that neither the formation of evil intentions nor the performance of evil actions are necessary for evil personhood. We should add this to our list of intuitions concerning evil persons: (16) Not every evil person performs, attempts, or intends to perform evil actions. Formosa might simply reject intuition (16). He could maintain that if the perpetually frustrated misanthrope is an evil person, he must have some kind of intention to perform evil actions. We could say that he has a general intention to perform evil action, but never forms an intention to perform any specific evil action. It is not clear, however, that this mental state is best characterized as an intention to act rather than a goal or a disposition to form particular intentions. Perhaps instead the perpetually frustrated misanthrope has a conditional intention to perform an evil action if placed in a lenient environment, or if given the opportunity. Again, it is not clear that this really is an intention rather than a disposition to form particular intentions. Moreover, it would be difficult to individuate and aggregate such general and conditional intentions. This difficulty is apparent in Formosa’s definition of “evil person”. When Formosa claims that an “evil person is an unreformed person who repeatedly perpetrates, or at least intends to perpetrate, evil acts” (Formosa 2008, 233), it is not clear whether he means that repeated intentions to perpetrate evil actions are sufficient for evil personhood, or simply that the intention to perpetrate evil acts is sufficient for evil personhood. If Formosa means the first, then the perpetually frustrated misanthrope would fail to count as evil on his account, because he does not repeatedly form specific intentions directed at specific evil actions. If Formosa means the second, then the formation of a single intention to perform more than one evil act would be sufficient for evil personhood, even if that intention is short-lived and uncharacteristic. Thus it is not clear whether Formosa’s account could be stretched to include the perpetually frustrated misanthrope as an evil person. Regardless of whether general goals or dispositions to form intentions to perform evil actions can be included within Formosa’s account, it appears that the

Aggregative Accounts of Evil Personhood  151 having of these goals or dispositions could be sufficient for evil personhood. If we accept that these general goals or dispositions are sufficient for evil personhood, then we have abandoned an aggregative account of evil persons. Rather than adding up the acts, attempts, or undeserved harms inflicted by S, or adding up S’s intentions to perform specific harmful acts, we are instead evaluating S in light of S’s dispositions. I have argued that luck can prevent an evil person from performing evil actions and from forming specific evil intentions. It is also plausible that luck repeatedly can cause a person who is not evil to perform evil actions. For instance, a soldier who is unlucky enough to be placed in a stressful and dangerous situation for a sustained period in wartime might commit a large number of extremely harmful culpable wrongs, even though he would not have committed any extreme wrongs had, by chance, he been posted elsewhere. In some sense of the word “characteristic”, these actions are characteristic of the unlucky soldier. He is disposed to perform such actions when placed in that stressful and dangerous environment, and this disposition is part of his character. However, in such cases we might judge that the evildoer does not have an unusually bad moral character, and that his repeated performance of evil actions should be explained by his being placed in that environment rather than by his being a very bad kind of person. Thus, in another important sense, an evildoer’s repeated evil actions can be described as uncharacteristic (cf. Kekes 2005, 209). Aggregative accounts of evil personhood would count the unlucky soldier as an evil person. In so far as we share the intuition that the unlucky soldier is not an evil person we have a further reason to reject the aggregative accounts of evil personhood.

7.7  The Appeal of Aggregative Accounts I have argued that aggregates of evil acts, intentions, or harms are not necessary or sufficient for evil personhood. Rather, it seems that a certain kind of disposition to perform evil actions is sufficient for evil personhood, regardless of whether the disposition is manifested in evil actions or specific intentions.7 A dispositional model of evil personhood holds a significant advantage over aggregative models in that it respects intuition (12). For instance, a dispositional account allows that it could be true and informative to claim that, while Billy Budd killed Claggart because he was

7   It is not clear whether such a disposition is both sufficient and necessary for evil personhood. It is possible that a person who has deeply morally corrupt feelings, such as a sadistic voyeur, could count as evil even though he lacks a disposition to perform evil actions. This possibility will be discussed in detail in Chapter Nine.

152  Aggregative Accounts of Evil Personhood in an uncharacteristic rage, Freddie West killed his many victims because he was an evil person. It is also easy to see how a dispositional model will neatly respect intuitions about evil persons who are not evildoers, and about evildoers who are not evil persons, including the morally reformed evildoers. An essential part of the moral reform of a vicious person is the alteration of that person’s character. Given these advantages of the dispositional model, it is worth asking why the aggregative approach deserves our attention, and why it has attracted some support. As we have seen, Haybron suggests that the popularity of aggregative accounts is based on a confusion between two kinds of moral extreme: the moral criminal, who has performed the worst kind of actions, and the evil person, who has the worst kind of character (Haybron 2002a, 272). Since people with the worst character are disposed to perform the worst kind of actions, there will be substantial overlap in the extension of these two concepts. Moreover, since our evidence regarding S’s character is S’s observable actions, the clearest evidence that S is an evil person will be S’s repeated performance of evil actions (Haybron 2002a, 272). Confusion between epistemically accessible evidence and metaphysical criteria could lead someone mistakenly to judge that an aggregate of actions, intentions, or harms is not merely a fallible indicator of evil personhood, but is necessary and sufficient for evil personhood. Haybron is right to suggest that such a confusion could explain why some people have defended aggregative accounts. Yet if this were a complete explanation, we should expect advocates of the aggregative account to fall back to a dispositional model once the so-called confusion is pointed out. Not all of them do fall back. For instance, Formosa suggests that someone who would do evil given the chance “is not an evil person until she turns her evil fantasies into evil intentions and evil acts”, because there is an “important moral difference” between merely fantasizing about evil and actually doing or intending to do it (Formosa 2008, 235). Advocates of a dispositional account, including Haybron, can admit that this is an important moral difference while denying that it marks the boundary between merely bad persons and evil persons. Formosa would need to argue that the aggregative approach is superior to the dispositional approach on other grounds. There are two points at which the dispositional account appears to be vulnerable to this kind of attack, which I shall sketch here and then go on to explore in more detail in the next chapter. First, it could be argued that the dispositional account fails to respect intuition (11): that evil persons deserve our strongest moral condemnation. The dispositional account seems to allow for the possibility of a blameless evil person, such as the perpetually frustrated misanthrope, who is disposed to do extreme wrong but has not yet done anything extremely wrong. How could he deserve the strongest kind of moral condemnation, and, if he does not deserve the

Aggregative Accounts of Evil Personhood  153 strongest condemnation, how could he count as an evil person? Second, it could be argued that almost all of us are disposed to perform evil actions under some circumstances, such as those present in Milgram’s obedience experiment. Thus, if we were to accept the dispositional model, perhaps we would have to reject intuition (10) concerning the rarity of evil persons. In contrast, aggregative models fit neatly with the intuition that evil persons are rare, and with the intuition that evil persons deserve our strongest moral condemnation. We have seen that more sophisticated versions of the aggregative account are available, and that these theories can make room in the class of evil persons for brooding spree killers, unlucky attempters, and suicide bombers, while excluding calamitous wrongdoers, Methuselah, and morally reformed evildoers. The aggregative approach is attractive in important respects, and appears to have some strengths that the dispositional approach lacks. Contrary to Haybron’s suggestion, the appeal of the aggregative accounts is not best explained as resting on a mere confusion over criteria. Nonetheless, we have seen that even the strongest versions of the aggregative accounts of evil personhood are vulnerable to many objections. It is not clear that aggregates can capture widely-shared intuitions about the explanatory power of evil personhood, nor common intuitions about evil persons and luck. In contrast, dispositional accounts allow evil personhood to play an explanatory role, and appear to fit more neatly with common intuitions concerning first-time-lucky suicide bombers, unlucky soldiers, and perpetually frustrated misanthropes. In the next chapter, I will attempt to develop a version of the dispositional account that is not vulnerable to the objections concerning blameless evil persons and the rarity of evil persons, and thus fits more neatly with intuitions (10) and (11).

8 Dispositional Accounts of Evil Personhood Aggregative accounts of evil personhood do not fit well with common intuitions about the explanatory power of evil personhood, the possibility of moral reform, and the relationship between evil and luck. In contrast, a dispositional account of evil personhood can allow that evil is explanatory, that an evil person can become good, and that luck can prevent evil persons from doing evil, and lead non-evil persons to do evil. Yet the dispositional account of evil personhood implies that some evil persons have done no wrong, and hence are blameless, which seems to clash with the intuition that evil persons deserve our strongest moral condemnation. Moreover, since it is likely that a large proportion of us are disposed to perform evil actions in some environments, the dispositional account threatens to clash with the intuition that evil persons are rare. In this chapter, I will consider a range of possible modifications to the dispositional account that could bring it more closely into alignment with our intuitions about moral condemnation and the rarity of evil persons. I will conclude that a more sophisticated version of the dispositional account is the single best account of evil personhood.

8.1  The Basic Dispositional Account of Evil Let us remind ourselves of the common intuitions that an account of the concept of evil personhood should aim to respect:

(9)  There are some actual evil persons. (10)  Evil persons are rare. (11)  Evil persons deserve our strongest moral condemnation. (12) In some cases the fact that S is an evil person helps to explain why S performed an evil action.

Dispositional Accounts of Evil Personhood  155

(13)  Not every evildoer is an evil person. (14)  It is possible to become an evil person by performing evil actions. (15) It is possible, though very difficult, for an evil person to become a good person. (16)  Not every evil person performs, attempts, or intends to perform evil actions. We have seen that not even the most sophisticated aggregative accounts of evil personhood fit well with intuitions (12) and (16). People sometimes try to explain the performance of an evil action by saying “He did it because he was evil”, but such an explanation would be incoherent if evil personhood consists in an aggregate of evil acts or an aggregate of harms. Aggregative accounts also cannot allow that, due to luck, an evil person might never perform or even intend to perform any evil actions. These problems that face the aggregative accounts of evil personhood appear to be avoided by a basic dispositional account, according to which S is an evil person if and only if S is disposed to perform evil actions. If we adopt an account of dispositions that allows for dispositions to be causes of their manifestation (Armstrong 1968; Mackie 1977), or, at least causally relevant to their manifestation (McKitrick 2005), the basic dispositional model of evil personhood will make sense of the claim that “He did it because he was evil”. No doubt this explanatory claim strikes some readers as contentious, and I will assess it in more detail in Chapter Ten. Another attractive feature of the basic dispositional account is that it makes sense of the fact that there can be evil persons who, by chance, never intend or attempt to perform any evil actions, just as there can be fragile objects that, by chance, never break. The dispositional account is particularly appealing in relation to the possibility of moral conversions. While a balanced aggregative account allows that an evil Nazi war criminal could become a good person by counterbalancing his evil actions with good deeds, the dispositional account suggests that an evil Nazi war criminal might become a good person simply by losing his disposition to perform evil actions and living a life of moderate virtue. If the moderately virtuous ex-Nazi is no longer the kind of person who would perform those horrible deeds, then it is quite possible that he no longer is an evil person. As we shall see in Chapter Nine, there might be other conditions, including remorse, that are required for this kind of change in an evildoers’ moral status. The basic dispositional account of evil personhood fits well with intuitions (9), (12), (15), and (16). Since some persons are disposed to perform evil actions, this account implies that some persons are evil. If being an evil person is equivalent to possessing a disposition to perform evil actions, it is possible that S’s evil action is

156  Dispositional Accounts of Evil Personhood explained by the fact that S is an evil person. Since S’s dispositions can change over time, the basic dispositional account implies that S can change from being an evil person to being a good person. Finally, since dispositions can fail to be manifested in behaviour due to luck, this account implies that some evil persons might fail to perform, attempt, or intend evil actions due to luck. It might appear that the basic dispositional account clashes with intuition (13): that not every evildoer is an evil person. It could be argued that if S performs an action ø then S must have been disposed to some degree to ø, even if øing is not the kind of thing that S does often, or the kind of thing that S is likely to do. Thus, it seems, every evildoer must have been disposed to some degree to perform evil actions, and every evildoer would count as an evil person on the basic dispositional model. Yet it is commonly claimed that S did ø even though S is not disposed to ø. We might say that Dave cried when he heard that his team won the premiership even though Dave is a hard man who is not disposed to cry, or that Trev dropped the ball even though he is skilful player who is not disposed to drop the ball. In such cases the claim that S is disposed to ø is taken to imply that S is strongly disposed to ø, or that S is markedly disposed to ø. Arguably, philosophers who claim that a specific evildoer is not an evil person are thinking of a case in which the evildoer was not strongly disposed to perform evil actions, but did so on this occasion. In order to allow that not every evildoer is an evil person, the basic dispositional account of evil personhood should state that S is an evil person if and only if S is strongly disposed to perform evil actions.1 The basic dispositional model also fits neatly with intuition (14):  that it is possible to become an evil person by performing evil actions. There are various processes by which S’s performance of evil actions might change S’s disposition to perform such actions. If S’s performance of extremely violent actions has some pay-off for S, such as the alleviation of a threat or the achievement of social power, then S’s violent behaviour might be reinforced, and S might come to view himself as the kind of person who is feared and respected because he will fly off the handle at the slightest provocation (Morton 2004, 35–8). If S possesses a normal aversion to inflicting extreme harm but repeatedly inflicts such harm— say, when placed in a stressful and dangerous situation—then S might come to   There is room for disagreement over how strongly S must be disposed to ø in order for us to say without qualification that S is disposed to ø. Moreover, the relevant strength varies depending on which kind of action we are considering. For instance, S could count as forgetful in spite of being disposed to remember the vast majority of what she knows, but S could count as honest only if S is disposed to tell the truth in almost all situations. Perhaps the rough rule is that S counts as being disposed to ø if and only if S is markedly likely to ø, compared to some base level of likelihood that varies from one action type to the next. 1

Dispositional Accounts of Evil Personhood  157 be desensitized to the suffering of others, and hence more strongly disposed to inflict harm in future (Morton 2004, 42–3; Wolfendale 2007, 175). Some people might develop a taste for the adrenaline rush that comes with violating moral norms, or for the sense of power than comes from breaking the will of another person (McGinn 1997, 77). The serial killer John Wayne Gacy, for instance, reported killing his first victim in what he believed was self-defence, but, he said “That’s when I realized that death was the ultimate thrill” (Cahill 1986, 349). After this, Gacy’s behavioural dispositions changed; he began killing regularly. Finally, if S has performed evil actions uncharacteristically, subsequent reflection on the wrongness of those actions might threaten S’s positive self-image. One way in which S could respond to this threat would be to explain the actions as having been chosen by someone else, rather than by S himself (Wolfendale 2007, 145), but another possible response is for S to abandon his judgement that the actions were wrong and instead see them as being justified (Zimbardo 2007, 220). If this post-hoc justification of S’s specific wrong actions ushers in a change in S’s general attitude to actions of that type, then S’s disposition to perform such actions in future might be greatly strengthened, and S would have become an evil person via performing evil actions. We have seen that the basic dispositional account of evil personhood accords well with intuitions (9), (12), (13), (14), (15), and (16), and in many respects is more appealing than the aggregative models. It is no surprise that some philosophers have leaned towards a dispositional account of evil personhood. Morton points out that the “simplest slide from acts to people is to call someone evil when they produce a lot of evil acts”, but he concludes that “it makes more sense to think of an evil person as one whose personality has as a central element some way of negotiating the barriers [against inflicting extreme harm]” (Morton 2004, 65). Calder also gives a dispositional account of what it takes for a person to have evil character (Calder 2003, 367). Garrard assumes that many evildoers fail to count as evil persons because they do not possess “settled dispositions” to do evil (Garrard 2002, 321), so it seems she holds that a disposition to perform evil actions is at least necessary for evil personhood. Haybron endorses an “affective-motivational” account of evil personhood which is dispositional rather than aggregative, but which includes extra conditions that are not present in the basic dispositional model (Haybron 2002a, 269).

8.2  Blamelessness and Rarity Despite its many strengths, the basic dispositional account appears to be vulnerable to two significant objections. The first objection concerns intuition (11): that

158  Dispositional Accounts of Evil Personhood evil persons deserve our strongest moral condemnation. Aggregative accounts fit comparatively neatly with this intuition because they imply that S is evil only if S has wrongfully attempted to inflict extreme harm. Granted, such evildoers may have grown up in violent homes, and may have done evil only when placed in difficult circumstances. In such cases, coming to know the evildoers’ circumstances tends to undermine or weaken our moral condemnation of them (Watson 1987, 272–4). Yet even in such cases these people have performed extreme culpable wrongs, and hence our moral indignation and our blame seems warranted. If we judge that the agent’s background and circumstances are sufficiently difficult to be exculpatory, then we ought not judge that the agent’s actions are evil. In contrast, the basic dispositional account implies that S might be an evil person even though S has performed no wrongs. Yet how can someone who has performed no wrongs be blameworthy for anything? And if such a person is not blameworthy, how can he be deserving of the strong moral condemnation that comes with being labelled evil? Advocates of the basic dispositional account could respond to this challenge by accepting that moral condemnation is tied to blameworthiness, but claiming that someone who is strongly disposed to perform evil actions is blameworthy for having acquired that disposition, regardless of whether the disposition manifests itself in action. If S’s choices foreseeably led S to become evil, then we might blame S for being evil regardless of whether S actually performs an evil action. This response will not solve the problem, though, because it is possible for us to acquire behavioural dispositions through no choices of our own. That such unchosen acquisition is logically possible is demonstrated by the example of a Frankensteinian creature who is designed to be evil and then is built from scratch. No actual people are built in this way, though, so we might think that every actual person who acquires a strong disposition to do evil does so through his or her own choices, and hence can be held responsible for his or her resultant character. Yet there are many actual cases in which a morally decent person accidentally suffers a head injury that radically changes his behavioural dispositions. Some victims of injury to the ventromedial prefrontal cortex report loss of affect related to moral situations, and become markedly more likely to inflict harm on others as a result of this accidentally acquired mental deficit (Goldberg 2001, 146–7). A certain kind of brain injury can turn ordinary, decent people into emotionally volatile murderers (Stone 2009, 308–9). It is not clear whether such brain injuries result in rational as well as emotional impairment (Fine and Kennett 2004; Maibom 2005, 242). If the injury produced a sufficient rational impairment we might judge that the injured person no longer has the rational capacity required for moral responsibility, and henceforth would

Dispositional Accounts of Evil Personhood  159 be able to perform extremely harmful actions but not evil actions. It could also be argued that emotional impairment alone could remove the injured person from the domain of moral responsibility. Nonetheless, it seems possible that via an accidental injury a person blamelessly could acquire a strong disposition to perform evil actions while retaining some moral sensibility and the capacity for moral reasoning. Stone suggests that the serial killer and torturer Richard Starrett fits this description (Stone 2009, 235–6). In addition to such cases, it is plausible that a person blamelessly could acquire a strong disposition to perform evil actions by inheriting an unusual genetic mutation (Stone 2009, 320–1), or by being born into and raised in an extremely abusive family (Stone 2009, 223–32). Every moral theory has some difficulty explaining how we can be aware of the extra-agential contributing causes of behaviour, yet still hold people morally responsible for their actions. The relevant issue at this point is not merely that dispositional accounts of evil personhood also face this problem. Rather, it is that dispositional accounts, but not aggregative accounts, imply that S can be an evil person before S acts badly, while S remains conspicuously blameless. Advocates of the dispositional approach could simply deny intuition (11), and claim that until evil persons actually perform evil actions, they deserve non-moral resentment or disapproval but do not deserve moral condemnation. Yet it is plausible that by calling a person evil we are not merely expressing disapproval, but are mor­ ally condemning that person in the strongest terms. If evil persons deserve our strongest moral condemnation, then the dispositional account implies that some blameless persons deserve our strongest moral condemnation. This result seems counterintuitive. The second objection to the basic dispositional account of evil personhood concerns intuition (10): that evil persons are rare. This intuition is widely shared, but it needs to be explored in more detail. The claim that evil persons are rare could be read either as a modal claim expressing some kind of necessity, or as a merely material claim that applies to the actual world as it is. For example, the claim that the intellectual elite are comparatively rare is a modal claim that remains true even if the general intelligence of the population increases dramatically, because the concept of the elite marks the far end of the statistical distribution, and its reference shifts as the distribution of intelligence in the population shifts. In contrast, the claim that anorexics are comparatively rare is a merely material claim that might be true now, but need not remain true. When we say that evil persons are rare I think that we are making a merely material claim analogous to the claim that anorexics are comparatively rare. We can imagine a world in which no one is evil, just as we can imagine a world in which the vast majority of people are evil. The widely shared intuition is that only a small proportion of actual persons are evil.

160  Dispositional Accounts of Evil Personhood A theory that implied that a very large proportion of actual persons are evil would not thereby be incoherent, but such a result would be surprising. Aggregative theories of evil personhood can easily account for the fact that the vast majority of actual persons are not evil, since the vast majority of actual persons have not wrongfully attempted to inflict extreme amounts of harm. At first glance it might appear that the basic dispositional account of evil personhood secures the same result. We might take the fact that not many persons perform evil actions as evidence in support of the claim that not many persons are strongly disposed to perform evil actions. Yet we should not forget that the vast majority of us have not been placed in the kind of environments that are most likely to elicit extreme wrongdoing. Some such environments are highly stressful, dangerous, and demoralizing. For instance, let us consider the Sonderkommando: prisoners in Nazi concentration camps who collaborated with the guards. Primo Levi notes that the Sonderkommando performed many extremely wrong actions, but, he claims: I believe that no one is authorised to judge them. . . . I would invite anyone who dares pass judgment to carry out upon himself, with sincerity, a conceptual experiment:  Let him imagine, if he can, that he has lived for months or years in a ghetto, tormented by chronic hunger, fatigue, promiscuity and humiliation; that he has seen die around him, one by one, his beloved; that he is cut off from the world, unable to receive or transmit news; that, finally, he is loaded onto a train, eighty or a hundred persons to a boxcar; that he travels into the unknown, blindly, for sleepless days and nights; and that he is at last flung inside the walls of an indecipherable inferno. (Levi 1989, 59)

It seems likely that quite a large proportion of us would wrongly inflict extreme harm when placed in such environments. However, it is not clear that such wrongdoers are culpable for their wrong actions. If the extremity of the conditions exculpates the Sonderkommando, then we should not judge that their actions are evil. The fact that a large proportion of us are disposed to perform extremely harmful actions when placed in similarly desperate circumstances would not imply that a large proportion of us are evil. There are other environments that do not seem sufficiently desperate or threatening to be exculpatory in which a surprisingly large proportion of people have performed extreme wrongs. A large number of Germans willingly participated in the persecution and execution of Jews in the Holocaust, but many of the perpetrators of those atrocities were not themselves in extremely dangerous and stressful conditions, and many were not compelled by orders and threats to act as they did (Goldhagen 1997, 379). Similarly, a surprisingly large number of people participated in the Rwandan genocide despite a lack of exculpatory conditions (Prunier 1995, 243–8). In light of the widespread participation in these atrocities, it seems

Dispositional Accounts of Evil Personhood  161 we must conclude that a surprisingly large proportion of Germans in the 1940s and Rwandans in the 1990s were disposed to perform evil actions. It is possible that the people of these nations were atypical, and that, while many of them were evil, it remains the case that evil persons are very rare in the human population as a whole. The worrying alternative possibility is that these populations were not atypical, and that a great many of us would have committed evil actions had we been unlucky enough to have been placed in those societies. This possibility gains credence from Stanley Milgram’s obedience experiment, in which the experimental subjects were asked by an authority figure to give an increasingly painful and dangerous series of electric shocks to a so-called learner when that learner made mistakes in answering questions. Deliberately giving apparently painful and life-threatening electric shocks to an innocent person is a culpably wrong action that is sufficiently extreme to count as evil, at least according to psychologically thin conceptions of evil action.2 The surprising result of the Milgram experiment was that 65 per cent of experimental subjects were willing to administer shocks right up to the final voltage, at which point the subjects had been given reason to believe that they were not only hurting but clearly endangering the life of the learner. Milgram himself claimed: If a system of death camps were set up in the United States of the sort we had seen in Nazi Germany, one would be able to find sufficient personnel for those camps in any medium-sized American town. (Blass 2000, 35–6)

The Milgram experiment has been widely replicated at other times and in other cultures. While rates of compliance vary, the clear finding is that a surprisingly large portion of experimental subjects will inflict extreme undeserved suffering on another human being when instructed to do so by a lone authority figure (Blass 2004, 301–5). The best explanation for this result is that a surprisingly large portion of humans are disposed to perform actions of this type. If these actions count as evil, and if we accept the basic dispositional account of evil personhood, then we must say that a surprisingly large portion of the human population are evil persons. It seems that the basic dispositional account of evil personhood would lead us to conclude that there can be blameless evil persons who have done nothing wrong,

2   It is unlikely that advocates of some psychologically thick accounts of evil action would claim that the subjects in the Milgram experiment performed evil actions, because the subjects in the Milgram experiments did not act out of malice, nor did they take pleasure in their wrongdoing or inflict harm for its own sake. However, it is plausible that most of those subjects performed wrong actions that they themselves knew to be wrong. Thus, if we stipulated that an action is evil if and only if it is a defiant extreme culpable wrong, then the subjects of the Milgram experiment did perform evil actions.

162  Dispositional Accounts of Evil Personhood and that a very large portion of humans are evil. In contrast, if we employ an aggregative account of evil personhood then we would be able to say that only some of the comparatively rare persons who actually perform evil actions count as evil, and all evil persons are blameworthy in that they have performed evil actions. In so far as we share intuitions (10) and (11), then, it seems that we have a reason to prefer aggregative to dispositional accounts of evil personhood. There are various ways in which we could respond to this problem. We could simply switch back to an aggregative theory. As we have seen, though, even the more sophisticated aggregative accounts clash with other deeply held intuitions concerning the explanatory power of evil personhood, the possibility of moral conversions, and the relation between evil and luck. In order to avoid these costs we might take a second option: a hybrid account according to which an aggregate and a disposition each are necessary and jointly are sufficient for evil personhood. By setting a double hurdle for evil personhood, a hybrid account would secure the result that evil persons are rare, and since one of the necessary conditions for evil personhood is a disposition, this model allows that an evil person could become a good person by losing that disposition. A hybrid account also comes closer to making sense of the claim “He did it because he was evil”. There is much that is attractive about a hybrid theory, but it fails to allow for the possibility of reliably deterred evil persons who, due to fear of punishment, never attempt to inflict extreme harm. If luck can prevent an evil person from ever doing evil, then we have a good reason to reject any hybrid account that includes an aggregate as a necessary condition for evil personhood. A third possible response to the problem is for us to stick with the basic dispositional account and bite the bullet. In this case we would have to accept both that a blameless person can be evil and that a large proportion of us are evil, even though, for most of us, our evil characters are never manifested in evil actions. Granted, it is counterintuitive to claim that a great many seemingly ordinary people who perform no extreme wrongs are evil nonetheless. Yet such a claim is not incoherent, and the historical and experimental facts that prompt that claim are themselves genuinely surprising. Some philosophers have argued that social psychologists have discovered that, contrary to expectation, no one is morally virtuous (Harman 2000, 177; Doris 2002, 24–5). Has the twentieth century also revealed that, contrary to expectation, a large proportion of human beings are evil? Many of us will baulk at this conclusion, partly because it requires that we take seriously the possibility that we ourselves are evil. It is tempting to dismiss our reaction as mere moral vanity, yet it is not clear that vanity could wholly account for our reaction. Many of the subjects in the Milgram experiment do not strike us as evil persons even after we have seen them administer the shocks. While these

Dispositional Accounts of Evil Personhood  163 people clearly are disposed to perform extreme culpable wrongs and have done so, it still seems that they fall short of being morally the worst sort of people. It is worth asking whether there are ways in which we could modify the basic dispositional account so as to bring it more closely into alignment with our intuitions concerning the rarity of evil persons, without sacrificing its advantages over its aggregative rivals. If we cannot do so, perhaps we should endorse a restricted conceptual pluralism in relation to evil personhood as well as evil action. That is, perhaps we should conclude that neither the aggregative account nor the dispositional account identifies the clear best contender for the property of evil personhood, and hence there is no fact of the matter as to which of these two properties really counts as evil personhood (cf. Sider 2001, 189).

8.3  Thoroughness of the Disposition One interesting possible modification of the dispositional account can be extracted from Daniel Haybron’s affective-motivational account of evil personhood, according to which evil persons not only are disposed to perform evil actions, but are “not aligned with the good to a morally significant extent” (Haybron 2002a, 269). In other words, Haybron takes evil persons necessarily to be so thoroughly disposed to wrongdoing that they “have no good side, but are consistently vicious” (Haybron 2002a, 269).3 This view is echoed in comments made by Phillip Cole, who suggests that extreme wrongdoers such as Hitler and Mengele were not evil because even they had a morally good side (Cole 2006, 196–8). Haybron and Cole are suggesting that evil persons must be thoroughly bad; that is, morally bad in every respect. When Haybron and Cole add this condition to their respective accounts of evil personhood, they do not do so with an eye to defending dispositional accounts against aggregative rivals, yet it is obvious that such a thoroughness condition could be put to this use. Nonetheless, it is not easy to see exactly which modification of the basic dispositional account will produce the desired result. Given that evil actions must be extreme rather than minor wrongs, it is not plausible that S is an evil person if and only if S is disposed never to do anything morally virtuous or admirable, because such a person might also be disposed never to do anything extremely vicious or deplorable. It would also be too demanding to claim that S is evil only if S is strongly disposed to perform evil actions in every situation

3   Haybron also claims that an evil person must be thoroughly bad not just in their dispositions to act, but also in their moral affect; that is, that an evil person must not feel the pull of morality and must not feel appropriate guilt or shame about his own moral failings (Haybron 1999, 134–6). For current purposes, I will focus merely on thoroughness of dispositions to act.

164  Dispositional Accounts of Evil Personhood and at every opportunity. An evil person who has no good side might well take a rest from evildoing and watch some television without ceasing to be evil. It seems that the most plausible version of a thorough dispositional account of evil personhood must include a connection to extreme wrongdoing and include a thoroughness condition that is just strong enough to exclude the possibility of the evil person having a good side. Let us say, then, that the thorough dispositional account of evil personhood is the view that S is an evil person if and only if S is strongly disposed to perform evil actions and is strongly disposed not to perform morally virtuous actions, where virtuous actions are morally right actions performed from morally admirable motives. The thorough dispositional account of evil personhood appears to hold all of the advantages of the basic dispositional model, with the additional benefit of securing the result that evil persons are rare. According to this account the vast majority of the subjects in the Milgram experiment are not evil because, while they are disposed to perform evil actions (in the psychologically thin sense of “evil action”), they have a good side nonetheless. For instance, they live normal family and social lives in which they perform many admirable and morally upright actions. The problem with the thorough dispositional account, though, is that it suggests that evil persons are too rare. Haybron is aware that the addition of a thoroughness condition strikes many people as too demanding, “because most of those whom we tend to call evil—war criminals, torturers, mass murderers, etc.—would surely be excluded by the account” (Haybron 2002a, 270). According to Haybron, the fictional mafioso Tony Soprano is not an evil person, because, despite his many acts of extreme violence, Soprano is a devoted and in some ways admirable family man (Haybron 2002b, 65–6). While Haybron leaves open the possibility that actual evil persons exist, he is unsure whether even Hitler was an evil person, because it is quite possible that Hitler had a good side (Haybron 2002a, 270). Haybron finds this thought “distressing” (Haybron 2002b, 74), but Cole seems comparatively happy to accept that, since no actual person is thoroughly bad, no actual person is evil (Cole 2006, 173). For many people the implication that even Hitler was possibly not an evil person would count as sufficient grounds for rejecting the thorough dispositional account of evil personhood.4 It simply is not plausible that we revise our assessment that Hitler was evil when we see that he had respectful relationships with some of his staff, or that he was kind towards dogs.5 Nor do we revise our judgement that the 4   Advocates of a psychologically thick account of evil action that includes defiance of morality as a necessary condition for evil action might claim that, since Hitler mistakenly believed that what he was doing was morally right, he did not perform evil actions. Most accounts of evil personhood built on this foundation would also imply that Hitler was not an evil person. 5   In order to support the claim that Hitler was evil and hence thoroughly bad, Haybron suggests that all of Hitler’s superficially admirable actions were motivated by selfishness or habit rather than a

Dispositional Accounts of Evil Personhood  165 BTK killer Dennis Rader was evil when we discover that he was a church and cub scout leader who took a sustained break from killing while he was raising his children (Stone 2009, 237–8). It is worth asking why the thorough dispositional account of evil personhood has attracted advocates. I will set out and examine three assumptions that might lead people to accept the thorough dispositional account:  the mirror thesis, the end point thesis, and the qualitative difference thesis. We might be led to the thorough dispositional account via the assumption of the mirror thesis; namely, that an evil person must be a mirror-image of the morally best sort of person (McNaughton 1988, 135; Steiner 2002, 185). Peter Brian Barry, who supports the mirror thesis, suggests that the morally best kind of person is someone like Martin Luther King, Oskar Schindler, or Gandhi, who has done much good, and who has been widely admired, and whom the folk might call a moral saint (Barry 2011, 202). I think that this kind of person is more accurately called a moral hero, and in many cases it is clear to all concerned that a moral hero, like Schindler, is not the morally best kind of person (cf. Haybron 2002a, 274). Barry agrees that many actual moral heroes are morally flawed in significant respects, and hence he concludes that an evil person who has a good side really is the mirror image of the morally best kind of person; that is, the mirror image of a flawed moral hero. Clearly, the idea that an evil person is the mirror image of a flawed moral hero does not support the thorough dispositional account of evil personhood, as Barry admits (Barry 2011, 202). There is a stronger version of the mirror thesis that would support Haybron’s thorough dispositional account. This is the claim that an evil person is the mirror image of the morally best kind of person; namely, a person who possesses Aristotelian virtue, or a person who is a moral saint in Susan Wolf ’s sense of that term (Haybron 2002a, 274). The morally best kind of person, in this more demanding sense, is morally admirable in all respects. Wolf ’s moral saint is someone who at every opportunity performs the morally best of the available actions, and who is “as morally worthy as can be” (Wolf 1982, 419). The mirror image of this is a person who always does what is morally worst out of the available options. Aristotle’s virtuous man has moral knowledge and is committed to morality (Aristotle 1985, NE, 1113a32–3). He has well-trained desires and moral emotions, and possesses practical wisdom which allows him to determine which action is right even in complex situations (Aristotle 1985, NE, 1144b20 n33). The virtuous man has the strength to genuine valuing of other people (Haybron 2002b, 74–5). This is possible, but highly speculative. The problem for Haybron is that it also seems quite possible that Hitler did value some people, and that he would still count as an evil person even if he had a good side.

166  Dispositional Accounts of Evil Personhood do what is virtuous even when it is costly and difficult (Aristotle 1985, NE, 1152a20). A mirror image of that person would be bad in all respects. He would be committed to bad moral values, and could be counted on to act viciously in all situations. It is plausible that the mirror image of Aristotle’s virtuous man would be an evil person, as would the mirror image of Wolf ’s moral saint. Even so, we have no reason to believe that every evil person is a mirror image of the best kind of person, nor that we gain insight into evil personhood by looking for perverse reflections of the virtuous. By way of analogy, let us consider what it takes to be a good or bad tennis player. An extremely good tennis player must possess a mastery of the many different activities that are part of the game—serving, returning, and running around the court hitting forehand and backhand ground strokes, volleys, lobs, smashes—and must consistently excel at all of these activities. Yet it is not the case that someone can be a really bad tennis player only if she is bad at every one of these activities. Since each activity is required regularly in every game, a player who is very good at some but very bad at others could count as a very bad tennis player who is easily beaten. The skills and activities that are required by moral virtue are more varied and complex than those required when playing tennis, and range over a far broader domain. For this reason, full virtue is an exacting ideal that requires an integrated unity of good character traits (Aristotle 1985, NE, 1144b32–1145a1; Hursthouse 1999, 156). But, as with tennis, someone can be very bad at morality without being bad at every activity that is required in the game of moral virtue. An evil person might have correct moral beliefs, but simply not care about being morally good himself. Conversely, an evil person might have great commitment to doing what he judges to be morally right, but horribly misguided beliefs about which actions are morally right. Some evil persons pursue extremely immoral goals doggedly and consistently, but others are erratic and opportunistic in their evildoing. These differences indicate that not every evil person is a mirror image of the best kind of person. The analogy between tennis and moral virtue, while imperfect, is borne out by actual examples of evil persons. Some brutal and sadistic war criminals were comparatively virtuous in their dealings with people within their own culture but violently hostile towards members of some other cultures. Indeed, it is plausible that some of the morally worst kind of people in the actual world are better able to achieve their immoral goals because they possess some moral knowledge and some morally admirable character traits. For instance, people who play important organisational roles in carefully coordinated atrocities are better able to spread death and destruction if they are courageous, committed, and loyal. Similarly, people with psychopathic or sociopathic tendencies

Dispositional Accounts of Evil Personhood  167 who are better able to blend in with society by virtue of having good sides are thereby more likely to inflict great suffering on the people who, having come to trust them, have let down their guard (e.g. Stone 2009, 230). If we assumed that evil persons must be mirror images of fully virtuous persons or moral saints, then these chillingly successful and reliable real-life evildoers would not count as evil persons simply because they are not bad in every respect. The implausibility of this result suggests instead that we should reject the strong version of the mirror thesis. Another route to the thorough dispositional account is via the assumption that S is an evil person only if there could not be a morally worse person than S. This is the assumption that evil marks an end point on a moral spectrum, rather than an extended zone at the bad end of the spectrum. Yet even if we grant that for every vicious person who has a morally good side there is a worse possible person who is thoroughly bad, it is not plausible that the concept of evil applies only to the worst possible person. Acceptance of the end point thesis would imply that necessarily no actual person is evil, because for any actual bad person we could always imagine a slightly worse person who took one more innocent life, or who felt slightly more pleasure while inflicting torture. Moral concepts that are guaranteed to lack any actual referent might be useful in allowing us to mark out positive ideals at which we can aim, but it is not clear that the negative counterpart of such concepts could serve any such practical purpose. Even if there were such a possible purpose, there is no evidence that our everyday uses of the concepts of evil and evil person fit with the end point hypothesis. For instance, people who believe that Ted Bundy was evil are not at all moved to revise that judgement if they encounter another serial killer who was just like Bundy, but murdered several extra victims. A final route to the thorough dispositional account of evil personhood is via the assumption that evil persons are qualitatively rather than quantitatively distinct from ordinary bad persons. As we have seen in Chapter Six, several philosophers have made a similar claim with respect to the distinction between evil action and ordinary wrong action (de Wijze 2002, 211; Garrard 2002, 321; Steiner 2002, 184). I have argued that we should not conclude that this difference is qualitative in the strong sense (Russell 2007b). The failure of these attempts should make us wary of similar attempts to show that an evil person must be qualitatively rather than quantitatively distinct from a merely bad person. Haybron makes one such attempt, claiming that evil does not come in degrees, and that there must be a sharp, qualitative difference between evil and non-evil persons, because otherwise the distinction between evil and bad would not matter (Haybron 2002a, 262). A person who is strongly disposed to perform evil actions is merely quantitatively distinct from a person who is less strongly disposed to

168  Dispositional Accounts of Evil Personhood perform evil actions, so Haybron would reject the basic dispositional model of evil personhood. Arguably, it is Haybron’s desire to locate a qualitative difference between bad and evil persons that leads him to claim that evil persons must be bad in every respect. A merely bad or vicious person does not have that quality of “thorough badness” to any degree, and thus could be said to be strongly qualitatively distinct from any person who has a good side. Haybron’s account might have been persuasive if it fitted fairly well with ordinary usage of the concept of evil person. Yet it is clear, even to Haybron himself, that by adding a thoroughness condition to our account of evil we would depart “from ordinary ascriptive practice” (Haybron 2002a, 271). We might have had a reason to revise popular usage if Haybron’s account offered significant benefits that could not be provided by rival accounts, but this is not the case. We ought to reject Haybron’s suggestion that the distinction between bad and evil persons could matter only if it marked a sharp, qualitative difference between the two. The distinction between heavy and light is practically important, but it is merely quantitative and admits of vague cases on the boundary. Similarly, the distinction between being highly resistant to temptation and being easily tempted is merely quantitative and admits of vague cases on the boundary, but it is a deeply significant moral distinction nonetheless. There is no obstacle to claiming that the distinction between evil and bad persons matters a great deal, even if it marks a merely quantitative difference. We have seen that there are three assumptions—the mirror thesis, the end point thesis, and the qualitative difference thesis—that might be used to support the thorough dispositional account of evil personhood, but that none of those assumptions are warranted. Moreover, the thorough dispositional account appears to clash with intuition (9); namely, that there are some actual evil persons. Even though the thorough dispositional account would fit well with intuition (10), that evil persons are rare, we should consider alternative methods of strengthening the basic dispositional account.

8.4  Fixity of the Disposition Another way in which we could modify the basic dispositional account so as to accord with the intuition that evil persons are rare is to maintain that all and only evil persons have an especially highly fixed and strong disposition to perform evil actions. The strength of S’s disposition to ø in conditions C is, roughly speaking, a measure of the likelihood that S will ø when in C, or the frequency with which S will ø when in C. In contrast, the flexibility or fixity of a disposition is a measure of how easy it is, over time, to change the strength of that disposition. For instance, S

Dispositional Accounts of Evil Personhood  169 might be strongly disposed to mispronounce French words, but if we could quite easily teach S correct French pronunciation, then we would say that this strong disposition is highly flexible. In contrast, S might be strongly disposed to get drunk, and if we could not easily change the strength of this disposition via rational persuasion, training, the promise of reward or the threat of punishment, or other available means, then this disposition would be not only strong but highly fixed. Claims about the fixity of character are common in ethics. Virtue ethicists maintain that a person is genuinely virtuous not only if she can be counted on to do the right thing here and now, but is firmly committed to correct moral values and cannot easily be corrupted (Aristotle 1985, NE, 1100b12–15; Hursthouse 1999, 28). It is easy to see why a genuinely virtuous person must have a strong and highly fixed disposition to act appropriately. A firm commitment to correct moral values is not only more admirable than a flexible commitment, but more useful. We count on other people to do the morally right thing in planning and coordinating our lives. If we assumed that the evil person is simply a mirror-image of the extremely virtuous person, then we would conclude that S is evil only if S has a highly fixed disposition to do the morally wrong thing. However, we have seen that this strong version of the mirror thesis is not warranted. A person could be extremely morally bad without being the psychological mirror image of Wolf ’s moral saint, or of Aristotle’s virtuous man. We would need independent grounds to justify the claim that S is evil only if S has a highly fixed disposition to perform evil actions. Such grounds are provided by another common intuition about evil persons: that an evil person is a moral write-off. Although it is widely held that it is possible for an evil person to become a good person, it would be jarring to say that S is an evil person but just needs a good talking to, or that S is an evil person right now but perhaps next week could be trusted with the care of your children. Cole assumes that an evil person would be beyond “communication and negotiation, reform and redemption” (Cole 2006, 236). Haybron similarly suggests that “regarding individuals as evil amounts to treating them as moral write-offs” (Haybron 2002a, 278). This implication underlies the common application of the concept of evil to the enemy in political and military settings. If the enemy are evil, then the only way to remove the threat from that enemy is to destroy them, scare them off, or constrain them; rational persuasion and diplomacy are out of the question. Similarly, when we call a serial killer an evil person we indicate that he should be “expelled, destroyed, or otherwise expunged from our lives” because it is overwhelmingly likely that he will keep doing evil regardless of our attempts at reform (Vargas 2010, 76). In light of this intuition we might propose the fixed dispositional account of evil personhood, according to which S is an evil person if and only if S is strongly and

170  Dispositional Accounts of Evil Personhood highly fixedly disposed to perform evil actions. According to this account it could be the case that a very large proportion of us are strongly disposed to perform evil actions, for instance, in a Milgram scenario, but that only a very small proportion of us are evil because the vast majority of us could be reformed or redeemed and lose our disposition to do evil. If this is the case, then it looks like the fixed dispositional account will hold all of the advantages of the basic dispositional model while securing the result that evil persons are rare. While it is true that a dispositional account of evil personhood requires a fixity condition in order to accord with common intuitions, it is not clear that our dispositions to commit extreme wrongs when placed in Milgram scenarios or in volatile and threatening political situations could easily be changed. There does seem to be strong evidence that people who perform extreme wrongs when in such situations can change their patterns of behaviour. For instance, many people who administered torture in military settings subsequently go on to lead morally normal lives in which they perform no major wrongs (Wolfendale 2007, 162; Zimbardo 2007, 287). Unfortunately, though, this absence of extreme wrongdoing in later life does not show that the ex-torturers have lost the disposition to perform evil actions when placed in those unusual scenarios. Perhaps they retain the disposition and merely are morally lucky enough not to find themselves again in the conditions that elicit its manifestation. We have no good evidence that this is not the case. If we accept the fixed dispositional account of evil personhood, we should conclude that, given what we know, it is quite possible that a large proportion of us are evil.

8.5  Dispositions Indexed to Conditions There is an extra modification of the dispositional model, though, that promises to deliver the result that evil persons are rare. Let us approach this modification by considering the more basic disposition of flammability. Roughly speaking, X is flammable if and only if X will burn by itself, once lit. However, whether or not X will burn by itself is highly dependent on other factors, including the amount of oxygen in the atmosphere around X. A treated woollen fabric will not burn by itself in normal conditions, but will burn by itself when in very highly oxygenated conditions. Would this imply that this fabric is flammable after all? Arguably, we would say that this fabric is not flammable because the disposition of flammability is indexed in some way to certain background environmental conditions. X is flammable if and only if X would burn by itself, once lit, when in conditions C. If a large proportion of us are disposed to do evil in some conditions, but only a small proportion of us are disposed to do evil in a specified set of conditions C, then we might claim that evil persons are those who are strongly and highly fixedly

Dispositional Accounts of Evil Personhood  171 disposed to do evil in conditions C, and hence that only a small proportion of us are evil persons. What are the relevant conditions to which a disposition is indexed? We might be tempted to say that the relevant conditions are the normal or typical conditions. X is flammable if and only if X would burn by itself when in normal conditions, and S is an evil person if and only if S is strongly and highly fixedly disposed to perform evil actions in normal conditions. Let us call this the fixed and normally-indexed dispositional account of evil personhood. Since the Milgram scenarios and extreme political conditions are unusual rather than normal, this account seems to imply that evil persons are comparatively rare. It is important to note that if we index dispositions to normal conditions and there is a change in which conditions are normal, then there will be corresponding changes in the dispositions of objects. This might be an acceptable consequence with respect to some dispositions. If flammability is indexed to normal conditions then we might say that a piece of wood is flammable on Earth but not flammable on the moon, where the normal conditions include no oxygen, even though the intrinsic properties of that piece of wood do not change if it is taken to the moon. It is not clear whether we would be equally happy to say that person S who is strongly and highly fixedly disposed to perform evil actions only in Milgram scenarios could become an evil person simply by being placed in an environment in which Milgram scenarios are common, despite there having been no change in S’s intrinsic properties. It is unsettling to think that we could turn S into an evil person simply by placing him in a society in which pressure from immoral authorities was common. It is certainly true that such a person becomes much more likely to inflict extreme harm when placed in this new oppressive environment. Yet there is something implausible about the claim that, if two persons are identical in their intrinsic and historical properties, one could be an evil person and the other not. It is not clear exactly why this seems implausible. Perhaps it is because we think it would be unfair to judge that a person has become evil simply by being moved into a different environment. As we have seen already, though, dispositional accounts seem committed to the claim that someone could blamelessly become an evil person, so intuitions about the fairness of evil attributions might not be defensible. Perhaps instead we are intuitively committed to the general view that significant changes in a person’s moral status require changes to the intrinsic properties of that person. In any case, we could avoid claiming that S could become evil merely due to a change in S’s normal background conditions by fixedly indexing the relevant disposition to the conditions that were normal or typical at the time and place in which the concept of evil personhood was developed. A similar kind of historically

172  Dispositional Accounts of Evil Personhood indexed disposition might underlie the property of being red, such that X is red if and only if S looks red to normal observers in conditions that were normal in the time and place in which the concept of red was developed. If there were a radical change in normal light conditions on Earth, we might coherently say that my old red shirt is still red even though it now looks blue to ordinary observers in the conditions that now are normal, and that the shirt could cease to be red only if it changed its intrinsic properties. Similarly, we might offer a fixed and historically indexed dispositional account of evil personhood, according to which S is evil if and only if S is strongly and highly fixedly disposed to perform evil actions in the conditions that were normal when the concept of evil was developed. If Milgram scenarios were very rare in these original conditions, and if only a small proportion of us are strongly and highly fixedly disposed to perform evil actions when outside of Milgram scenarios, then evil persons would be rare and would remain rare even if Milgram scenarios became much more common. By fixing the reference on these historical conditions, we guarantee that S could become evil only by changing his intrinsic properties. This fixed and historically indexed dispositional account of evil personhood probably secures the result that evil persons are rare, and seems attractive in many other respects. Nonetheless, this account is vulnerable to two related objections. First, it might seem that advocates of this account are making an ad hoc modification to the dispositional model merely to secure the result that evil persons are rare. Second, it is not clear exactly which features of the original normal conditions are relevant to the question of whether S’s actions will reveal whether S is an evil person or not. It is plausible that now we are still in those original normal conditions, and thus that the strength of S’s dispositions to perform evil actions in our current environment is a clear indicator of whether S is evil. Yet conditions have changed greatly in many respects over the last fifty years with the invention of computers, mobile phones, the internet, and so on. Why should we say that, despite these changes, we are still in the original conditions with respect to the disposition that is necessary for evil personhood? Arguably it is because those changes in computing technology strike us as irrelevant to the questions of whether an evil person is more or less likely to perform evil actions, and whether a non-evil person is more or less likely to perform evil actions. In contrast, a dramatic change in the level of political coercion in our society might lead us to say that the strength of S’s disposition to do evil in these new conditions does not indicate whether S is an evil person, because these new conditions are not relevantly similar to the original conditions to which the dispositional concept of evil person is indexed. In order to differentiate

Dispositional Accounts of Evil Personhood  173 between relevant and irrelevant changes to background conditions, we must be relying on some deeper conception of which background conditions are relevant to the disposition in question. How are we to specify the conditions that are relevant to the disposition that is sufficient for evil personhood? Perhaps we should note that there are some conditions, including Milgram scenarios and coercive political environments, in which people who perform extremely wrong actions tend to describe themselves as being alienated from those actions in some way. They might claim to have been forced to perform the actions, to have had no choice, to have been tricked into acting in that way, to have not been able to do what they really wanted to do. This suggests that there are an identifiable set of conditions in which our autonomy is limited, and in which our actions might not reflect our true selves.6 Of course, what S is disposed to do under such difficult conditions is part of S’s character, but we might claim that whether S is an evil person depends on what S is disposed to do when S is not under the kinds of pressure that typically would alienate him from his actions; that is, when S is in what I will call autonomy-favouring conditions. I will call this the fixed and autonomy-favouring dispositional account of evil personhood, according to which S is an evil person if and only if S is strongly and highly fixedly disposed to perform evil actions when in autonomy-favouring conditions. Even if it is true that a very large proportion of people are disposed to perform evil actions under some conditions, it is likely that only a small proportion of people are disposed to perform evil actions when they are able to do what they really want to do. Thus, the fixed and autonomy-favouring account of evil personhood seems to imply that, given what we know, evil persons are comparatively rare. Moreover, indexing the disposition to autonomy-favouring conditions does not seem to be an ad hoc move designed simply to secure the result that evil persons are rare. It is plausible that there is a morally significant difference between what S is disposed to do when she is not deceived, threatened, coerced, or pressed, and what S is disposed to do when her autonomy is limited. It is not the case that all of S’s dispositions and actions under these limiting conditions are morally irrelevant. These dispositions could count against S in some kinds of moral evaluation. So long as they are culpable wrongs, actions performed under these conditions can count as evil, and it is proper to hold a person responsible for the evil actions that he performs whether he is an evil person or not. But S is morally the worst kind of   To say that S’s autonomy is limited in such circumstances is not to say that S cannot perform any free actions when in such circumstances, nor that S is not morally responsible for the actions that she performs in these circumstances. 6

174  Dispositional Accounts of Evil Personhood person only if S is strongly disposed to do evil even when he is not pushed in that direction by circumstance. Another reason for believing that it is not ad hoc to index the disposition to autonomy-favouring conditions is that the resulting account allows for the possibility that, if we made certain empirical discoveries, we should conclude that evil persons are actually far more common than we had supposed. If the threat of punishment for extreme wrongdoing in our society were removed and there were a sudden and dramatic increase in the proportion of people who perform evil actions, then we would have reason to believe that a large proportion of those people had been evil all along. Once they were placed in an autonomy-favouring environment in which they were not being restricted by threats of punishment they were able to do what they really wanted to do, which was to perform evil actions. Conversely, if the threat of punishment for extreme wrongdoing were suddenly imposed on a comparatively corrupt society and the proportion of the population who do evil decreased dramatically, we would not necessarily have reason to believe that we had decreased the proportion of evil persons in that society. In such a case we would merely have prevented some would-be evildoers from doing what they really wanted to do. As we have seen, it is possible for there to be perpetually frustrated evil persons who are not evildoers. I contend that the question of whether S is an evil person is settled by whether S is strongly and highly fixedly disposed to perform evil actions when in those conditions in which S is able to do what he really wants to do. The fixed and autonomy-favouring dispositional account of evil persons fits well with intuition (9):  that there are some evil persons. It allows that most serial killers and many war criminals are evil because they are disposed to perform extreme wrongs even when they are not being instructed to do so by authority figures, and even when they are not living in intense and difficult social conditions. The fixed autonomy-favouring dispositional account also fits with intuition (10): that evil persons are rare. Given what we know, the vast majority of people are not strongly disposed to perform extreme wrongs when allowed to do what they really want to do. By indexing the disposition to autonomy-favouring conditions we also go some way towards accommodating intuition (11): that evil persons deserve our strongest moral condemnation. Although this account still allows that a person blamelessly can become evil, it suggests that many people who are disposed to perform evil actions only when in autonomy-limiting circumstances are not as deplorable as evil people, who are disposed to do evil even when not being threatened or pushed. As with all of the other dispositional accounts of evil personhood, the fixed autonomy-favouring account fits with intuitions (12)–(16). It makes sense of

Dispositional Accounts of Evil Personhood  175 the claim that “He did it because he was evil”, it allows that not every evildoer is an evil person, that it is possible to become an evil person by performing evil actions, that it is possible, though difficult, for an evil person to become a good person, and that not every evil person intends or attempts to perform evil actions. As with the other dispositional accounts, the fixed autonomy-favouring account allows that an evil person could be blameless, but this is the price of accommodating the intuition that there can be perpetually frustrated evil persons. It might appear that the fixed autonomy-favouring account is too narrow in that it removes military torturers and war criminals from the category of evil persons if they are not disposed to perform extreme wrongs when in autonomy-favouring conditions. We should remember, though, that even if these are not evil persons, they remain evildoers, and are blameworthy and deserving of condemnation for their actions. The fixed autonomy-favouring dispositional account of evil personhood is more appealing than any of its aggregative or dispositional competitors. In Chapter Six, I  argued that there are several viable philosophical accounts of evil action, and that none of these identify a uniquely most deserving candidate for what really is evil action. For this reason I advocate a restricted conceptual pluralism about evil action. Given that there is a single best account of evil personhood, I do not think that we ought to add an extra layer of pluralism at the level of personhood. However, my account of evil personhood is parasitic on a prior account of evil action. An evil person is someone who has a certain kind of disposition to perform evil actions in certain circumstances. Thus, there is a sense in which I am a conceptual pluralist when it comes to evil personhood, too. Someone is an evil person if and only if she has the right kind of disposition to perform evil actions, but in any particular dispute about whether S is an evil person we need to specify which of the several viable conceptions of evil action we wish to employ in our fixed and autonomy-favouring account of evil personhood.

9 Evil Feelings The evaluation of a person is more complex than the evaluation of an action. There are many facets of a person that might not be revealed in an action, nor contribute to that action. Moreover, persons typically exist for much longer than actions. The extra breadth and extra temporal depth that is available in evaluations of persons suggests that the question of whether a person is evil could be much broader and temporally deeper than the question of whether an action is evil. In the previous chapter, I argued that S is an evil person if and only if S has a strong and highly fixed disposition to perform evil actions when in autonomy-favouring conditions. This account includes extra breadth in so far as it points to dispositions rather than to actions, and extra temporal depth in so far as it suggests that evil persons have highly fixed dispositions that are very difficult to alter over time. Yet it might be thought that by focusing merely on dispositions to act in certain ways this account of evil personhood leaves out a crucial part of our evaluation of a person: the evaluation of that person’s feelings. If there are such things as evil feelings, and if evil feelings are separable from a disposition to perform evil actions, then we ought to explore two further possibilities: that having evil feelings is necessary for evil personhood, and that having evil feelings is sufficient for evil personhood. In this chapter, I will explore these questions and defend the conclusion that possessing a certain kind of disposition to have a certain kind of evil feelings is enough to make a person evil.

9.1  What are Evil Feelings? Our feelings, in the everyday sense of that term, include a broad range of psychological states and events: emotions, sensations, and desires. Many of what we would call feelings contribute towards our dispositions to act in various ways, and some questions couched in terms of feelings are best understood as questions about what we are likely to do in certain situations. When we ask someone how she feels about kayaking, or how she feels about the conservative

Evil Feelings  177 candidate, what we really want to know is whether she is likely to go kayaking when given the opportunity, or whether she is likely to vote for the conservative candidate. Similarly, some questions about a person’s disposition towards extreme wrongdoing could be couched in terms of how that person feels about performing those extreme wrongs herself. Yet questions about feelings are not exclusively questions about what people would choose or would not choose to do when placed in various situations. Emotions are richer than mere behavioural dispositions in that they include affective feels, and one and the same behavioural disposition could be accompanied by a range of distinct sets of emotions, or, indeed, by affective flatness. Feelings can be the object of moral evaluation. A feeling of anger in response to a perceived insult, for instance, might strike us as unwarranted and shameful, or as petty and self-righteous but excusable, or as justified and admirable (D’Arms and Jacobson 2000, 65; Hurka 2001, 45). Thus we can ask which specific feelings or emotions should be judged to be morally the worst kind of feelings. Which, if any, might warrant being called evil feelings? The most likely candidates come from within the class of antisympathetic feelings (McGinn 1997, 62; Haybron 1999, 135– 6): feelings of pleasure at another’s pain, harm, or suffering, and feelings of pain at another’s happiness or success. These are not merely pleasures that are caused by the suffering of others or pains that are caused by the pleasures of others. Rather, antisympathetic feelings are intentionally directed at the pleasure or suffering of others, in the same way that your fear of the guard dog, for instance, is an emotion that is directed at the guard dog; the dog is what you fear. There are some atypical cases in which it is fitting or even admirable to feel pleasure at another’s pain or pain at another’s pleasure. It is fitting that I am overjoyed by the fact that my paralysed friend has started feeling pain in his legs, because his pain is a sign that the medical treatment is working. Similarly, it is fitting that I am dismayed by the fact that my friend is experiencing heroin-induced ecstasy, because her pleasure indicates that her time in rehab was ineffective. What does appear to be morally bad is taking pleasure in another’s pain in and of itself, and being pained by another’s happiness in and of itself (Hurka 2001, 19). These are antisympathetic feelings, properly construed. Antisympathetic feelings need not be simple pleasures and pains. I could be pained by another’s happiness because I am envious of that person’s success and want it for myself, or simply because I  feel malicious hatred towards that person. Similarly, pleasure in witnessing the suffering of another person could be grounded in a sadistic sexual perversion or in a belief that the victim deserves to suffer. Antisympathetic feelings include these fine-grained and cognitively rich emotions.

178  Evil Feelings Being disposed to have antisympathetic feelings is a moral vice. It is one form of hating what is good and loving what is bad (Hurka 2001, 11–16). The fact that such feelings are vicious does not entail that they are evil feelings. Vices can be trivial or mild, but evil is a term that we reserve for those things that deserve very strong moral condemnation. If any feelings deserve to be called evil, they must be morally extreme in some respect. There are a range of ways in which we could try to capture the requisite gravity that would mark out some antisympathetic feelings as evil feelings. We could claim that all antisympathetic feelings are vicious, but that only extremely intense antisympathetic feelings are evil. This claim would be implausible, though. Taking intense pleasure in the minor misfortune of a colleague is morally bad, but, despite its intensity, this feeling is too morally trivial to count as evil. Alternatively, we could claim that it is the extremity of the harm or the suffering in which the pleasure is taken that could make a feeling evil. Taking pleasure in the slow and painful death of the victim of a car accident does seem to be a morally deplorable feeling, and might have the kind of gravity that is necessary for evil. Perhaps it is also true that the duration and intensity of S1’s own feeling of pleasure at S2’s extreme suffering must exceed a certain threshold in order to qualify as an evil feeling, and that as the intensity of this pleasure is increased the feeling becomes more morally deplorable. But unless these pleasures are directed at extreme suffering, harm, or wrongs, I contend, they would not be sufficiently grave to count as evil feelings at all. It seems plausible that some instances of schadenfreude could count as evil feelings. What should we say about the converse; that is, feelings of envy or displeasure that are directed at extreme feelings of happiness or at extreme good fortune? If our evaluations of antisympathetic feelings were symmetrical, then envy of extreme happiness would be significantly morally worse than envy of mild happiness. Yet envy which is directed at small pleasures and minor triumphs arguably is more morally deplorable than envy of extreme happiness, which can be comparatively easier to understand and to excuse. In any case, it seems that the feelings of schadenfreude and envy are not equally morally bad, even if the unhappiness or happiness at which they are directed happen to be equally extreme. It is not clear how we could explain or justify this evaluative asymmetry. Taking pleasure in another’s extreme pain just strikes us as being more morally corrupt and deplorable than being pained by another’s extreme pleasure. It is certainly true that the desire, or, worse, the intention, to destroy another person’s happiness is deeply immoral, but merely feeling peeved that another person is ecstatically happy falls short of the worst kind of depravity. This kind of envy might seem perverse and petulant, but not evil. In contrast, imagine how shocked you would be to discover that one of your friends enjoys watching footage of torture and political executions.

Evil Feelings  179 Perhaps this kind of schadenfreude is so morally deplorable that it warrants being called an evil feeling. Some instances of schadenfreude strike us as morally much worse than others, though. It is not clear that taking pleasure in contemplating the fictional suffering of non-existent people is morally bad at all. Giggling at cartoon violence is not a sign of moral corruption. Even when pleasure is directed at the suffering of real people, it is not always deeply deplorable. A student who, while reading Herodotus, enjoys contemplating the destruction of the Persian fleet could be described as taking pleasure in contemplating the suffering and destruction of real people, but perhaps both the temporal distance between ourselves and the victims and our ignorance of which particular people were suffering render such pleasure merely mildly vicious, if, indeed, it is vicious at all. In contrast, it seems that someone who takes pleasure in directly witnessing the extreme suffering of a particular person here and now does have a morally deplorable feeling. There are some special cases in which, arguably, there is nothing morally wrong with feeling happy upon directly witnessing another actual person’s intense suffering, because that suffering is deserved. Thomas Aquinas claims that the blessed in heaven will “see perfectly the suffering of the damned”, because by doing so their own happiness will be rendered “more delightful to them” (Aquinas 1942, 107, Summa Theologica III, Q. 94, Art. 1). In cases of witnessing justly deserved punishment it is hard to separate pleasure that is taken in witnessing the dispensation of justice—a pleasure that seems not to be morally corrupt—from the more questionable pleasure that is taken in witnessing the suffering itself (cf. Aquinas 1942, 109–10, Summa Theologica III, Q. 94, Art. 3). After all, the suffering is a necessary constituent of the punishment, which, arguably, is necessary for the restoration of justice. Yet anticipated bliss upon witnessing the suffering of all sinners can seem uncomfortably close to a sadistic form of gloating (cf. Bennett 1994, 301; Calder 2013, 183–4). When we are observing well-deserved suffering, perhaps the most we ought to feel is a grim satisfaction at justice being done. However, in cases in which the witnesses of justly inflicted suffering were themselves the victims of the wrongdoing that is being punished, even joyful pleasure seems not to be a sign of moral corruption. For example, many Italians took great pleasure in witnessing the murder of Mussolini, and many Iraqis cheered at the hanging of Saddam Hussein. Given that these witnesses were victims of these oppressive dictators, their joyful schadenfreude strikes many of us as excusable or even warranted. There is a fine line between dishing out justice and gratuitously torturing extreme wrongdoers, as becomes clear when viewing Quentin Tarantino’s fantastical anti-Nazi revenge film Inglorius Basterds. We want the wrongdoers to suffer, but should we enjoy watching them being mutilated? Similar difficulties

180  Evil Feelings confront us when we note that members of a violently oppressed minority group often take pleasure in witnessing the extreme suffering of innocent members of the oppressive majority, including young children. Such feelings are vicious and unjust, but some people might argue that they are a primal response that has been provoked by sustained wrongdoing, and hence are excusable in the circumstances. Beyond these murky waters of justified or excusable schadenfreude there are cases in which inexcusable morally depraved pleasure is taken in witnessing or contemplating the extreme suffering of particular innocent victims. Out of all of the possible candidates, these pleasures seem most deserving of being labelled evil feelings.1 It is obvious that such feelings often accompany or give rise to desires and intentions to perform extremely harmful wrongs. Nevertheless, these feelings are distinct from these desires and intentions. It is possible for there to be an ideologically misguided war criminal who commits mass murder but who has no evil feelings, because he takes no pleasure in performing or contemplating his actions. As we shall see, it is also possible for there to be a person who has evil feelings but never intends or desires to perform any evil actions. I have defended the view that S is an evil person if and only if S has a strong and highly fixed disposition to perform evil actions when in autonomy-favouring conditions. Until now I have referred to this as a dispositional account of evil personhood, as opposed to an aggregative account. Now it is clear that this dispositional account could be described more precisely as a disposition-to-act account of evil personhood. Some people might dismiss this view of evil personhood as being too narrow, on the grounds that it makes no mention of evil feelings. For instance, if having evil feelings is necessary for being an evil person, then we would expect evil feelings to feature explicitly in our account of evil personhood. It would be a mistake to suppose that a disposition-to-act account of evil personhood could not possibly include evil feelings as a necessary condition for evil personhood. I have argued that the most plausible accounts of evil personhood are parasitic on accounts of evil action, and some viable psychologically thick accounts of evil action include pleasure in the infliction of harm as a necessary condition for evil action (McGinn 1997, 69; Perrett 2002, 304; Steiner 2002, 189). If a disposition to perform evil actions is necessary for evil personhood, and if evil actions must be accompanied by sadistic pleasure, then a disposition to take pleasure in the extreme suffering of others will be a feature possessed by all evil persons.   Failure to feel anything at all in response to witnessing extreme suffering might also strike some people as being sufficiently bad to count as evil. Yet feeling pleasure in such circumstances seems to be even worse than feeling nothing. If evil feelings are to pose a challenge to the existing dispositional account of evil personhood, then the challenge will be strongest in the case of inexcusable schadenfreude. For this reason, I will focus on this kind of case. 1

Evil Feelings  181 Philosophers who take all evil persons to be affectively similar to Satan and sadistic serial killers need not augment the existing dispositional account of evil personhood with an extra condition that refers to evil feelings, so long as they specify that evil actions necessarily involve sadistic pleasure. However, the claim that evil action requires sadistic pleasure is contentious. We have seen that many philosophers, including Arendt, Code, Neiman, Morton, and Formosa, favour psychologically thin conceptions of evil action that do not include such sadistic pleasure as a necessary condition. If we build a dispositional account of evil personhood on these thin foundations then it appears that someone could be an evil person despite not being at all disposed to have evil feelings. It is plausible that this is exactly the result desired by most advocates of thin accounts of evil action, in that it allows them to label as evil persons the kind of ideologically committed extreme wrongdoers who take no pleasure in inflicting carnage, but do so in the misguided conscientious pursuit of some deeply immoral end, be it the elimination of homosexuals or securing the racial purity of the Fatherland. Thus, the question of whether evil feelings are necessary for evil personhood is most naturally absorbed into the question of whether evil feelings are necessary for evil actions.2

9.2  Quadriplegics and Voyeurs A more pressing challenge to the existing dispositional account of evil personhood arises from the suggestion that having evil feelings could be sufficient for evil personhood, independent of a disposition to perform evil actions. This suggestion conflicts with the accounts offered by some philosophers, who maintain that every evil person is disposed to perform evil actions (Card 2002, 21; Morton 2004, 66). While there might be good reasons to defend such a position, it is also possible that these philosophers have simply overlooked a number of cases in which a person might have a strong disposition to have evil feelings, but no disposition to perform evil actions. In order to assess the plausibility of the disposition-to-act account that I defended in the previous chapter, we must consider these cases in more detail. The first such case is that of the malevolent quadriplegic who appreciates evildoing and strongly desires to perform evil actions herself, but is not disposed to do so because she is incapable of inflicting extreme harm on anyone. Daniel Haybron   It would be possible for advocates of a psychologically thin account of evil action to stipulate that evil personhood also requires evil feelings, although it is hard to see what is attractive about such a view. Advocates of this view should add a condition to the dispositional account, thus claiming that S is an evil person if and only if S is strongly and highly fixedly disposed when in autonomy-favouring conditions to perform evil actions and to have evil feelings. 2

182  Evil Feelings judges that the handicap of such a person “does not make her anything better than evil” (Haybron 2002a, 264), and thus he appears to accept that having evil feelings is sufficient for being an evil person. In contrast, Paul Formosa claims: there is an important moral difference between merely wishing and fantasizing about evil, as the quadriplegic does, and actually intending and inflicting evil, which the quadriplegic does not and cannot do. . . . [W]‌hile the quadriplegic remains a vile human being, she is not an evil person until she turns her evil fantasies into evil intentions and evil acts. (Formosa 2008, 235)

Formosa is right to point out that there is a moral difference between the person who merely fantasizes about evil and the person who does evil, but this does not imply that only one of the two could be an evil person. The malevolent quadriplegic described by Haybron is not blameworthy or punishable for performing any evil actions, but the fact that she falls short of doing evil is of no moral credit to her. While we ought to agree with Haybron’s claim that the malevolent quadriplegic is evil, this admission does not undermine the dispositional account of evil personhood that I advocated in the previous chapter. According to that account S is evil if and only if S is strongly and highly fixedly disposed to perform evil actions when in autonomy-favouring conditions. Autonomy-favouring conditions are those conditions in which a person is able to do what she really wants to do, as opposed to conditions in which she never gets the chance to do what she wants to do, or is manipulated into acting against her better judgement, or is pressed by difficult circumstances to perform actions from which she would otherwise refrain. It is plausible that the malevolent quadriplegic is disposed to do evil when in autonomy-favouring conditions. A quadriplegic is never in autonomy-favouring conditions in relation to many kinds of action that require bodily motion. In the case of the malevolent quadriplegic, her quadriplegia is the only thing that prevents her from doing the evil things that she wants to do. In this respect the malevolent quadriplegic is in the same moral category as the perpetually frustrated misanthrope who never performs an evil action merely due to a well-justified fear of punishment. Both are classed as evil persons by the autonomy-favouring dispositional account. A second example does raise a significant problem for my dispositional account of evil personhood. Let us imagine a harmless sadistic voyeur who takes great pleasure in observing and contemplating the extreme suffering of others, but is not disposed to inflict any extreme harms, nor to act in such a way that he is morally responsible for increasing the likelihood that such harms are inflicted (cf. McGinn 1997, 66; Calder 2002, 56; Haybron 2002a, 265). Of course, it is likely that the vast majority of actual sadistic voyeurs are not harmless. They do act in ways

Evil Feelings  183 that increase the likelihood of such harms being suffered, and are blameworthy and culpable for doing so. Many sadistic voyeurs inflict the harms themselves, or pay to view footage of the infliction of extreme harm, thereby providing a market and encouraging others to abuse, torture, or even murder innocent victims. Other sadistic voyeurs culpably fail to step in to prevent extreme suffering when they have the opportunity to do so. In our imagined case, though, the harmless sadistic voyeur does none of these things. He merely enjoys observing and contemplating the extreme suffering of others in cases in which he could not intervene. Even if he were in completely autonomy-favouring conditions, in which he could do what he wants to do, our harmless sadistic voyeur would not cause others to suffer. The harmless sadistic voyeur appears to fall short of being an evil person on the current version of my dispositional account. Yet such a person clearly is deeply depraved. When we imagine him rubbing his hands together with glee while watching victims hopelessly trapped in the wreckage of a car accident, it is hard not to see him as an evil person. A third example exacerbates the problem. Let us consider a person who has performed many evil actions in the past, who has subsequently lost the disposition to perform such actions, but who does not regret or renounce his wrongdoing. Indeed, he takes pleasure in reminiscing about the evils he has perpetrated. If this case of the unrepentant nostalgic former evildoer is to be interestingly distinct from that of the malevolent quadriplegic, we must imagine that he has lost the disposition to perform extreme wrongs not because some limit has been placed on his powers, nor because he now lacks the opportunity to do evil, but for some other reason. Perhaps he has simply lost the taste for doing evil, just as someone who was a keen dancer in her youth might think back fondly on her dancing days, but have no desire to dance any more. We should be relieved that this unrepentant former evildoer is no longer a risk to others, but his nostalgia indicates that he has not been made good or redeemed. Upon discovering that he feels no remorse, and that he smiles when reminiscing about the malicious murders he committed, I think that he might strike us as an evil person despite the fact that he is no longer disposed to perform evil actions. Thus, the autonomy-favouring disposition-toact account of evil personhood appears to be incomplete. Advocates of the dispositional account could respond to this objection by claiming that, so long as we accept a sufficiently broad notion of evil actions, both the harmless sadistic voyeur and the unrepentant nostalgic former evildoer are disposed to perform evil actions. It is not plausible that every evil action succeeds in causing extreme harm. For example, wholehearted attempts to murder that, by chance, fail to inflict any harm can be evil actions nonetheless. In Chapter Three, I claimed that an action can be evil only if it is appropriately connected to

184  Evil Feelings extreme harm, and unsuccessfully attempting to inflict extreme harm is one such connection. Some of the things that are done by the harmless sadistic voyeur and the unrepentant former evildoer similarly seem to be actions that are connected to extreme unjustified harms in such a way that they would qualify as evil. Rather than acts in which the agent intends and attempts to produce extreme unjustified harm, they are acts in which the agent appreciates extreme unjustified harm. These acts of appreciation include deliberately seeking out cases of extreme suffering, voluntarily directing one’s attention to that suffering, choosing to linger at the scene, and imagining or recalling such suffering, all of which are done for the purpose of enjoying that suffering (cf. Calder 2002, 56). Many of these acts of appreciation involve voluntary bodily movements, but some could be merely mental acts. So long as these are voluntary actions, though, it seems that the existing dispositional account of evil personhood easily can make room for the harmless sadistic voyeur and the unrepentant nostalgic former evildoer. Both of them are disposed to perform evil actions, broadly construed, so both of them could count as evil persons.

9.3  Involuntary and Conflicted Evil Feelings While many instances of taking pleasure in an event are properly called acts of appreciation, there are also cases in which a person passively rather than actively takes pleasure in something. Perceptual experiences can be thrust upon us without our consent, and these experiences can be pleasurable nonetheless. Memories or imaginings can come to us unbidden, even compulsively and against our will. Moreover, many of our emotional responses to our experiences are automatic, cognitively impenetrable, and involuntary (Griffiths 1997, 92–7; Foucher and Tappolet 2006, xiii). The existence of such passive appreciatings suggest the possibility of involuntary evil feelings; that is, involuntary appreciatings of undeserved extreme suffering. It is not plausible that everyone who has this kind of feeling is an evil person, but it might seem that someone who is strongly and highly fixedly disposed to have such feelings is evil, whether or not he is disposed to perform any evil actions. The delighted passive spectator who revels in the extreme suffering that he happens to witness is clearly not as morally deplorable as the person who voluntarily chooses to do extreme wrong, but the delighted passive spectator has a terrible character flaw nonetheless. Yet the contrary position also has some appeal. It could be argued that these feelings, insofar as they are involuntary, cannot be a significant moral discredit to the person who has them. Calder claims that “it would not be evil to just happen upon a mugging and witness (and perhaps even take pleasure in) the victim’s harm” (Calder 2002, 56). If the delighted

Evil Feelings  185 passive spectator cannot help but feel that way, then why should we hold his feelings against him? (cf. Hurka 2001, 46). Let us delay our answer to the question of whether the delighted passive spectator is an evil person in order to consider another complication in our evaluation of persons in light of their feelings. Feelings within a single person can be conflicted. Not only can a person’s feelings drag her in incompatible directions, but a person can feel both positive and negative emotions about the one state of affairs. We also have many emotional responses that are directed at our own feelings, and a pleasant first-order feeling can be overlaid with a unpleasant second-order feeling directed at the first feeling, or with a negative evaluative judgement directed at the first feeling (Frankfurt 1988, 21). For instance, an animal lover might enjoy the gustatory sensations of eating veal but feel ashamed about doing so, and might also judge that there are compelling moral arguments against eating veal. Some of the people who have evil feelings will suffer from no such internal conflict, but others will be deeply conflicted, and will feel bad about either actively or passively taking pleasure in the extreme suffering of others. This kind of internal conflict will be particularly common amongst those who have evil feelings but lack a disposition to perform evil actions. Someone who enjoys witnessing suffering but who judges that his sadistic enjoyment is immoral would be comparatively less likely to go out and inflict extreme suffering precisely because his feelings are conflicted. We have seen that actions differ from feelings in two important respects. Generally speaking, actions are under the direct voluntary control of the agent, whereas feelings more often are not. There is also something conclusive about an action, in that an action is the outcome of all of an agent’s competing desires and drives in the moments before the performance of the action. What S freely chooses to do at time T is a good, albeit imperfect, indicator of S’s overall beliefs, attitudes, and evaluative judgements at T. Some feelings are harmonious, settled, and conclusive, but many feelings are fleeting and conflicted, and do not even come close to representing the conclusive judgement or the overall attitude of the person who has them. A person who is strongly disposed to perform evil actions is disposed to have some voluntary and conclusive connection to the infliction of extreme harm, but a person who is strongly disposed to have evil feelings might be related to extreme harms in nothing more than an involuntary and inconclusive manner. It is perhaps too harsh to say that someone who has fleeting and conflicted evil feelings is an evil person. However, there are other cases in which a person’s wholehearted evil feelings conceivably could constitute good grounds, in and of themselves, for saying that the person is evil. In order to make sense of this as yet unclear connection between evil feelings and evil personhood, we need to consider several more elaborate accounts of

186  Evil Feelings the conditions in which having evil feelings might be sufficient for being an evil person. In the remainder of this chapter, I will examine three possible reasons for which someone who regularly has involuntary evil feelings might nonetheless strike us as falling short of being an evil person. The first such case is that in which a person’s evil feelings are not indicative of evil values or inverted values. The second such case is that in which the evil feelings are not really the person’s own feelings, in virtue of not being narratively integrated into the person’s life. The third such case is that in which the evil feelings are repudiated by the person who has them.

9.4  Evil Values and Narratively Integrated Evil Feelings Expanding the dispositional account of evil personhood so as to include both actions and feelings might seem to be unnecessarily messy, or even ad hoc. It could be argued that there is a more fundamental feature of a person, namely, that person’s set of values, which determines whether that person is evil. Perhaps we should say that a person is evil if and only if that person has inverted values or evil values, and that such values could manifest themselves in evil actions as well as in evil feelings. In this case, having evil feelings would be sufficient for being an evil person only when those feelings are an indication of evil values. There are several different ways in which we could understand the notion of a person’s values. We could think that S’s values correspond to the moral principles that are compatible with S’s typical or characteristic actions. Alternatively, we might think that S’s values are the moral principles that S herself judges to be correct, whether she acts in accordance with them or not. Perhaps S’s values are the moral principles that are both endorsed by S and that guide S’s planning. Another approach would lead us to think instead that S’s values are those of S’s desires and attitudes that S herself desires that she have, and hence with which S identifies (Bratman 2007, 47–67; Frankfurt 1988, 21–2). According to all of these accounts, S’s values are supposed to reflect S’s deep commitments, and on any of these accounts it is possible that some of S’s feelings and actions conflict with S’s own values. To say that S has evil values, then, is to say that S endorses moral principles that in fact are compatible with evil action, or that S not only has evil feelings or desires but that she desires that she has those feelings or desires, or that S identifies with her evil desires and attitudes, or something else along similar lines. Someone who has fleeting, conflicted evil feelings and who disapproves of those feelings so strongly that she never acts on them seems to fall short of having evil values.

Evil Feelings  187 Let us consider the proposition that evil feelings are sufficient for evil personhood only when those evil feelings are part of that person’s evil values. It is plausible that someone who has evil values thereby counts as an evil person. In contrast, it is not clear that someone who lacks evil values thereby falls short of being an evil person, or even that someone who does not have evil values must be a significantly morally better person than someone who does have evil values. As I claimed in Chapter Eight, not every evil person is the mirror image of the ideally morally virtuous person. Some extreme wrongdoers might be so erratic and inconsistent in their behaviour and their moral judgements that we are unable to ascribe to them any stable and coherent set of values. Even some predictable and consistent extreme wrongdoers could be judged as falling short of having any moral values on the grounds that they have not engaged in the kind of higher level, intellectualized moral reflection that is required to generate a set of values. Yet it is not clear that these erratic and unreflective extreme wrongdoers are morally superior to people who have reflected on and endorsed immoral principles. Unreflective sociopaths who never endorse general principles or reflect on the propriety of their first-order desires, and thus who lack values, might be evil persons nonetheless. Perhaps some bureaucratic evildoers, such as Arendt’s Eichmann, are so unreflective and thoughtless that they also lack moral values altogether. We should grant that some of the people who lack values altogether also lack the rational capacities that are required for responsible moral agency. Such a person falls outside of the category of evil. But others might lack values despite being rationally capable of detecting right and wrong, of reasoning about consequences, and of forming intentions and sticking to them. If such people, in addition to lacking values, are strongly and highly fixedly disposed to perform evil actions when in autonomy-favouring conditions, then they are evil. It is too demanding to claim that having evil values is necessary for being an evil person. Given that we cannot unify our account of evil personhood by appealing to underlying evil values, perhaps we should include both dispositions to perform evil actions and dispositions to have evil feelings as sufficient conditions for being an evil person. As we have seen, though, not all persons who have evil feelings seem bad enough to be described as evil. It is worth considering whether we can identify the cases in which evil feelings are sufficient for evil personhood by appealing to the ownership of feelings. If a person has evil feelings that are not really her own, in some sense, then we might conclude that her evil feelings are not enough to make her an evil person, whereas a person who has evil feelings that are fully her own might thereby count as evil. What is it, though, for one’s feelings to really be one’s own? Frankfurtian accounts of a person’s values give us one model of ownership (Frankfurt 1988,

188  Evil Feelings 21–2). As we have seen, though, this model is too demanding for our purposes. Someone can be evil despite not having reflected on and taken ownership of the evil parts of his character. In a related discussion of situationism and moral responsibility, John Doris sketches a less demanding account of psychological ownership. Doris’ aim is to distinguish between the unconscious motives for which we hold the agent responsible from those for which we do not. He claims that we are properly held responsible for some motives that we have never consciously endorsed, and even for some motives that we would not consciously endorse were they brought to our attention (Doris 2002, 140–2). Doris also thinks that there are some motives, such as those flowing from obsessive compulsive disorder, for which we ought not hold a person responsible. The motives for which we are responsible, according to Doris, are those that can be narratively integrated into the stories of our lives. Motives that cannot be narratively integrated strike us as mere tics and glitches, rather than significant parts of ourselves (Doris 2002, 143). Evil feelings differ from the subconscious motives that Doris is considering, not least in that they involve conscious pleasure. Nonetheless, we could try applying Doris’ criterion of ownership in the hope of identifying the cases in which evil feelings are sufficient for evil personhood. According to this proposal, S’s evil feelings are sufficient for S being an evil person if and only if those evil feelings can be narratively integrated into the life story of S. This approach appears to be attractive in that it allows that an erratic and unreflective person who lacks values all together could count as an evil person nonetheless, so long as his evil feelings are an integral part of his life story, while also allowing that in other cases evil feelings are mere tics or glitches that do not indicate evil personhood. This approach will be useful only if we have a clear grasp of what counts as grounds for narrative integration, and it is not obvious that we do. We should not endorse the general claim that regular and typical feelings are narratively integrated in a person’s life story, and that unusual or atypical feelings are not. For some people compulsive feelings are all too regular and typical, even if they are experienced as being an extra-personal intrusion. Conversely, atypical feelings can be deeply significant at pivotal moments in a person’s life, and thus an essential part of a personal narrative. When a characteristically hostile and exacting person forgives her father for having abandoned her, for instance, it is natural to see her feelings of forgiveness on this occasion as being central to her life story, despite their being atypical and uncharacteristic. Similarly, it would be churlish to claim that a cowardly person who uncharacteristically performs a courageous action does not deserve credit for doing so, on the grounds that he acted from a motive that, supposedly, was not really his own (Hurka 2006, 72; cf. Doris 2002, 131).

Evil Feelings  189 We should allow that typical feelings can be narratively excluded from a life story, and that atypical feelings can be narratively integrated. In order to do this we must appeal to some richer notion of the story of a life. We tell stories about our lives for a range of reasons: to entertain or instruct others, to justify our actions and make ourselves seem attractive to others, to regulate our own behaviour, to predict our own actions and to make plans, and to make sense of our joys and sorrows. Whether a motive or feeling is deemed to be owned or disowned in a person’s life story will depend on which story is being told, and for what purpose. For example, when someone who has obsessive compulsive disorder is trying to justify or excuse her actions to others she might offer a narrative in which her compulsive motives are an externally imposed interruption to her life. When she is trying to make sense of her struggle she might instead describe those motives as a crucial, if unwanted, part of her identity. Perhaps this story cannot be told without including the fact that she herself feels driven to perform cleaning rituals, that she herself needs things to be arranged symmetrically, and so on. Unless we have grounds for privileging one of these stories over its competitors, it is not clear how an appeal to ownership of feelings could assist us with our project regarding evil personhood. The evil feelings of some passive delighted spectators could be narratively integrated in so far as these feelings are accepted and enjoyed wholeheartedly by those spectators whenever they occur, and arguably such people would count as evil. Yet we can also imagine a self-loathing delighted passive spectator who enjoys observing innocent people being harshly treated by others, but despises himself for doing so. If such a person lives in a slave-owning society he might regularly be confronted with the unjust infliction of extreme suffering, and he might see his struggle to reject his own feelings of sadistic pleasure as central to the story of his life. This person does seem to have narratively integrated evil feelings that are far less peripheral than any mere tics or glitches, but it is not at all clear that he himself is evil. The fact that he has accepted that these feelings are his own, instead of evasively dismissing them as nothing more than an alien intrusion, can render him more rather than less morally admirable. It is not always the case that a person who has narratively integrated evil feelings will thereby be morally superior to a person who excludes his evil feelings from his life story.

9.5  Unrepudiated Evil Feelings I have been exploring the possibility that in some but not all cases passive or involuntary evil feelings are sufficient for evil personhood. I have suggested that a person with evil values would count as evil, but that a person who is too erratic or unreflective to have any values could count as evil as well. Moreover, I have argued

190  Evil Feelings that someone who has narratively integrated evil feelings might or might not be an evil person, and that accepting ownership of evil feelings can render a person more admirable rather than more deplorable. Let us consider a possible third way to identify the conditions in which passive evil feelings are sufficient for evil personhood. It seems that the cases in which someone who has evil feelings falls short of being an evil person are those in which that person condemns and rejects those feelings. It is the repudiation of his own evil feelings that would lead us to judge that a delighted passive spectator is far from being the worst kind of person. It is the fact that the unrepentant nostalgic former evildoer does not feel remorseful about his extreme wrongdoing that leads us to judge that he is still an evil person, even though he has lost his disposition to do evil. Thus, it is plausible that S is an evil person if S has unrepudiated evil feelings. According to this view, repudiation of bad feelings is a moral credit to the repudiator. His repudiation of these feelings renders him a morally better person than he otherwise would be, ceteris paribus. If this is the case, though, we will have to specify more carefully what we mean by repudiation. Having mixed or conflicted feelings is not sufficient for having repudiated feelings. For example, a person can feel guilty about having lied to his mother about where he is going on his birthday, feel annoyed that he is feeling guilty on his birthday, yet still judge that it is fitting that he should feel guilty, given his actions. In this case he has conflicted feelings, but he has not repudiated his guilt. Repudiation of a feeling requires a judgement that the feeling is inappropriate. If the repudiation of a bad feeling is to be a moral credit to the repudiator, then the feeling must be repudiated because it is judged to be morally bad or because it is judged to have some morally bad-making qualities, and not merely because it is judged to be imprudent or embarrassing.3 But a person deserves little, if any, credit for making a correct moral judgement unless he is also moved, to some degree, to act in accordance with the judgement. If the repudiation of a bad feeling is to be a moral credit to the repudiator then the repudiator must be moved by her moral judgement to try to stifle that feeling, and to prevent that feeling from issuing in wrong action. An amoralist who correctly judges that his sadistic evil feelings are morally bad but simply does not care about morality has not really repudiated his evil feelings. Therefore, S’s feelings are repudiated by S, in the relevant sense, when S judges them to be morally bad or to have a bad-making

  Morally creditable repudiation need not be motivated by the judgement that the feeling in question is bad and a de dicto desire to not have feelings that are bad. For instance, it is morally creditable to repudiate a feeling on the grounds that the feeling is callous, or that it is invidiously envious. We might call these qualities “bad-making features” of a feeling, analogous with the wrong-making features of actions. 3

Evil Feelings  191 feature, and when S for this reason tries to stifle those feelings and prevent them from issuing in wrong action. If a person fails to repudiate his involuntary evil feelings, he must have a deeply depraved moral character. This depravity could consist in a kind of moral stupidity or idiocy which leads the person to judge incorrectly that his feelings are not extremely morally bad, or it could consist in the person’s failure to care sufficiently about his having feelings that he himself correctly judges to be evil. A person who does repudiate his involuntary evil feelings still possesses a significant moral weakness, but at least he recognizes and struggles to correct the flaws in the affective part of his character. Even if his evil feelings stubbornly persist, in repudiating them he has responded to his own involuntary moral flaws in a morally admirable way. We might say that such repudiation is a sign that, despite his evil feelings, he has good values. It is this moral repudiation, rather than any judgement that the feelings are owned by the agent or are a crucial part of the agent’s life story, that ought to sway our moral evaluation of a person who has evil feelings. It is a terrible thing to have evil feelings, but it is worse to have such feelings and not to reject them as immoral. The fact that S has unrepudiated evil feelings seems to be sufficient for S’s being an evil person. However, this condition concerning evil feelings needs to be refined in a number ways that mirror the refinements in my dispositional account of evil personhood set out in the previous chapter. Given common intuitions about the explanatory power of evil personhood and the logical possibility of the creation of a fully-formed evil person, we should say that S is an evil person if S is strongly disposed to have evil feelings, whether that disposition has been manifested in occurrent feelings or not. Identifying S’s evil personhood with a disposition allows us to make sense of the claim that S enjoyed observing the extreme suffering because S is an evil person, or that S would enjoy doing so in future because S is an evil person. A disposition to perform evil actions must exceed a certain strength if it is to be a sufficient condition for evil personhood, and the same seems to be true of a disposition to have evil feelings. Someone who is disposed to have evil feelings only once every few years is, ceteris paribus, not as bad as someone who is disposed to have evil feelings regularly. Since an evil person is a moral write-off who is beyond our ability to reform, we should claim that S is an evil person if S has a strong and highly fixed disposition to have unrepudiated evil feelings. The dispositional account of evil personhood that I set out in the previous chapter indexes the relevant behavioural disposition to autonomy-favouring conditions. According to this account, someone is an evil person if he is strongly and highly fixedly disposed to do evil when he is able to do what he really wants to do, but he might not be evil if he is disposed to perform evil actions only when pushed

192  Evil Feelings to do so by manipulative or difficult circumstances. It is hard to say whether a similar limitation ought to be placed on the extra condition concerning evil feelings. On the one hand, repudiation of our own feelings is an option that is almost always available to us, regardless of the environmental conditions in which we are situated. Even those who have many evil feelings while in difficult, oppressive circumstances can easily escape evil personhood simply by repudiating those feelings. Perhaps a person who would be provoked by difficult circumstances into evil feelings and who would endorse and wallow in those feelings, regardless of our attempts to persuade him that his feelings are immoral, deserves to be described as evil. On the other hand, there are environmental conditions that make it much more likely not only that people will perform evil actions, but that people will have evil feelings, and in which such feelings are closer to being excusable. Members of an ethnic group who have been systematically persecuted are more likely to enjoy witnessing the undeserved extreme suffering of innocent members of the oppressive group, and might enjoy hearing of the murder of their oppressor’s children. Arguably, this could be an evil feeling, but it seems uncharitable to say that someone who is disposed to have such feelings only when under these oppressive conditions is an evil person. It seems fair to say that S is an evil person if S has a strong and highly fixed disposition to have unrepudiated evil feelings when in autonomy-favouring conditions. In any case, it appears that a certain kind of disposition to have evil feelings, independent of a disposition to perform evil actions, is sufficient for evil personhood. It should be clear that I do not intend that this sufficient condition replace the autonomy-favouring dispositional account of evil personhood, but to be added to it. Someone can be an evil person despite not being disposed to have evil feelings at all, so long as he has the requisite disposition to perform evil actions (and so long as evil actions are taken not to require sadistic pleasure on the part of the wrongdoer). My full account of evil personhood is as follows: S is an evil person if and only if S is strongly and highly fixedly disposed to perform evil actions when in autonomy-favouring conditions, or S is strongly and highly fixedly disposed to have unrepudiated evil feelings when in autonomy-favouring conditions. While this addition to the dispositional account of evil personhood does make the account significantly more complex, it seems to be warranted by the fact that dispositions to feel are a morally significant part of a person, and, arguably, are not reducible to dispositions to act. Yet it could be argued that this apparent revision concerning evil feelings is actually not a revision at all, but is better understood as an elucidation of an important and often overlooked kind of evil action: namely, the action of failing to repudiate one’s own evil feelings. If we are willing to count

Evil Feelings  193 this failure to repudiate as an action, then it remains the case that S is an evil person if and only if S is strongly and highly fixedly disposed to perform evil actions when in autonomy-favouring conditions. The reason that the self-accepting passive delighted spectator and the nostalgic unrepentant former evildoer count as evil persons is that they are strongly and highly fixedly disposed to do something evil, namely, to leave their own evil feelings unrepudiated. In contrast, the self-loathing passive delighted spectator does not remain passive in response to the feelings that arise in him unbidden. He does what he ought to do: he repudiates those feelings and struggles to prevent them from issuing in wrong action. The repentant former evildoer also does not remain passive with regards to his past evil actions, but repudiates them and feels guilt and remorse. People who have no disposition to perform evil actions and who repudiate their evil feelings are morally far superior to people who have unrepudiated evil feelings. Arguably, the latter but not the former ought to be judged as evil.

9.6  Repudiated Evil Actions If we agree that repudiating evil feelings can save someone from being an evil person, what should we say about those cases in which someone is strongly and highly fixedly disposed to perform evil actions but repudiates those actions, feeling remorse, condemning them as wrong and struggling not to perform them? Roy Perrett appears to accept the claim that remorse or repudiation is not felt by evil persons. He offers as supporting evidence the self-reported opinions of prisoners who are serving time for murder, who describe pleasure in wrongdoing and lack of remorse as necessary for evil (Perrett 2002, 306). Haybron points out that the serial killer Jeffery Dahmer reported feeling appalled by his own behaviour, and he suggests that Dahmer’s remorse, if genuine, is a mitigating factor in our assessment of Dahmer as a person (Haybron 1999, 135). Haybron seems unsure whether to class Dahmer as evil in light of this good side to his character, and ultimately settles on an account of evil personhood that implies that anyone who has a good side is not evil (Haybron 2002a, 269). While Haybron is right to suggest that an active conscience is a good thing, it is not the case that everyone who feels remorse thereby escapes evil personhood. In the previous chapter, I argued that a conflicted person who has morally mixed motives can be evil nonetheless, and the good parts of his character can allow him to be even more successful in wreaking havoc. The genuine episodes of remorse that are felt by a fixedly paedophilic priest who has molested and abused many children, for instance, could give us the false impression that he has changed his ways, and thus lead to his being forgiven, trusted, and given more opportunities

194  Evil Feelings to commit the extreme wrongs to which he remains disposed.4 When a remorseful person remains strongly disposed to do extreme wrong in autonomy-favouring conditions his remorse is not irrelevant to our evaluation of his character, but it falls well short of indicating his moral reform and redemption, and its moral significance is swamped by that of his persistent and awful disposition to perform evil actions. It is tempting to see contrary claims by murderers as nothing more than hollow attempts at self-justification, unless that remorse is matched by a change in the wrongdoers’ behavioural dispositions. However, genuine remorse does seem to be necessary in ordinary cases of moral reform. In those rare cases in which we would be willing to judge that an evil person has become a good person, he must have repudiated all of the extreme wrongs that he knows he has performed (cf. Thomas 2003, 201; Formosa 2008, 234).5 There remains one further possible objection to the claim that a strong and highly fixed disposition to have unrepudiated evil feelings is sufficient for evil personhood. It seems that being disposed to perform evil actions that inflict extreme harm on innocent victims is morally much worse than being disposed to enjoy witnessing such harm (Calder 2002, 56). Performing extremely wrong actions is more deplorable than having very bad feelings. If we supposed that an evil person was the worst possible person, than whom none worse could be conceived, then it would be implausible to claim that a disposition to have unrepudiated evil feelings was sufficient for evil personhood. Yet there is a significant cost to that supposition: it guarantees that there will be no actual evil person, because for every actual extremely depraved person we can imagine one who is slightly worse, who takes even more pleasure in wrongfully inflicting even more harm. Rather than see evil personhood as an end point of a spectrum, we should see it as a end zone of a spectrum. Only someone who is in that zone could be an evil person, but some of them are significantly morally worse that others. If evil personhood is an extended end zone rather than an end point, the significant question is where we locate the threshold. Given that harmful evil actions are much worse than harmless evil feelings, it is likely that some philosophers will maintain that a disposition to have unrepudiated evil feelings is not sufficient for evil personhood, but that a disposition to perform evil actions is both necessary 4   For a chilling example of this, see the 2006 documentary film Deliver Us From Evil, directed by Amy J. Berg, which features interviews with the paedophile Oliver O’Grady. 5   Both Thomas and Formosa suggest that a reformed evil person will feel remorse for his evil actions. However, Thomas and Formosa fail to note the possibility of unusual cases of moral reform in which an evildoer who has lost the disposition to do evil has also lost all memory of his past evildoing, and hence does not feel remorse for his evil actions. So long as this loss of memory is not the product of self-deception, and so long as the reformed evildoer would feel remorse if he was aware of his past actions, there is no reason to think that his lack of remorse counts against him.

Evil Feelings  195 and sufficient. In order to test this claim, the best we can do is to consider our intuitions about particular cases that fall across this section of the spectrum. I must confess that my own intuitions in relation to evil feelings are less firm than my intuitions concerning evil actions. That said, unrepentant nostalgic former evildoers who have lost the taste for inflicting harm but reminisce fondly about their days working in the Nazi death camps or the killing fields of Cambodia do seem to me to be evil. Passive delighted spectators who enjoy being confronted with extreme suffering and see nothing bad in their having those feelings also strike me as evil persons, although as less morally deplorable than those in the previous category. This is not to say that there are no persons who are morally worse than these. Rather, the unrepudiated evil feelings of these persons place them in the category of those who deserve our extreme moral condemnation, who are not merely vicious but evil. In condemning these people, we should remember that we are not judging them merely on their involuntary feelings, but on their failure to respond to those feelings as they should.

10 Evil and Explanation The concept of evil has been attacked on the grounds that it is primitive, simplistic, and morally dangerous. It certainly is true that a great many claims about evil are vague or ambiguous, and that many public declarations that certain actions or persons are evil are unjustified, and have had terrible effects. Over the course of the past nine chapters, I have tried to develop a clear and philosophically defensible account of the nature of evil actions, evil persons, and evil feelings. I contend that all evil actions are extreme culpable wrongs, and that it is reasonable but not obligatory for a person who is making judgements involving evil to stipulate that she will assume that evil actions must also be accompanied by malice, sadistic pleasure, or defiance. An evil person is someone who is strongly and highly fixedly disposed to perform evil actions when in autonomy-favouring conditions, or who is strongly and highly fixedly disposed to have unrepudiated evil feelings when in autonomy-favouring conditions. My account is not intended to match exactly what every particular person who makes claims involving evil thinks that the word “evil” means. Given the level of disagreement amongst the folk, no account could do that. Nonetheless, I do think that my account fits well enough with folk thought about evil, and that many of the folk who say that certain actions or persons are evil mean something close to what I have set out. No doubt my more elaborate and more precise philosophical account of evil will still strike some critics as misguided and dangerous. In this chapter, I will respond to arguments put forward by Philip Cole, who claims that the concept of evil is explanatorily vacuous, and that there are no actual evil persons. In the following chapter, I will move on to assess further pragmatic arguments to the effect that we would be better off if we ceased using the concept of evil in our moral thinking.

10.1  Is Evil an Explanatory Concept? In his book The Myth of Evil Philip Cole argues that there is no such thing as evil, and that we would be better off if we dropped the concept (Cole 2006, 236). One

Evil and Explanation  197 strand of Cole’s argument concerns supernaturally loaded conceptions of evil. In Chapter One, I noted that some people think of evil in supernatural terms, but that many others have a secular conception of evil, according to which evil actions need not have any connection to supernatural beings or forces, and evil persons need not be literal demons or monsters. Cole agrees that there are conceptions of evil in play in contemporary moral thought that are not supernaturally loaded, and I contend that these ought to be the primary target of our philosophical evaluation (Cole 2006, 13–16).1 Cole thinks that even these secular conceptions of evil are flawed. He assumes that we ought to employ the concept of evil only if doing so is explanatorily useful. More specifically, Cole suggests that a legitimate concept of evil would have to contribute to explanations of why extremely wrong actions were performed, rather than explanations of why certain actions were condemned, or explanations of why the people who performed those actions were ostracized. Cole writes: The fundamental problem is one of explanation, and whether the concept of evil can play any constructive or useful role in explaining human action. Is ‘because she was evil’ ever an explanation, even a partial one? (Cole 2006, 18)

Cole believes that an explanatorily useful concept must have the requisite form to contribute to explanations of actions. Moreover, Cole assumes that an explanatorily useful concept must feature not merely in explanations of possible or fictional actions, but also in some true explanations of actual actions. It is far from obvious that the concept of evil could be a legitimate and useful moral concept only if it is explanatorily useful in this sense. Many moral concepts, such as the concepts of good, right, bad, and wrong, appear to be purely evaluative or prescriptive, and hence appear not to serve the explanatory function identified by Cole. Nonetheless, we could not build an ethical theory without such basic, non-explanatory concepts. Other important moral concepts, such as the concepts of harmful action and disrespectful action, give us information about the effects of actions rather than explain why those actions were performed. As Cole points out, Raymond Gaita accepts that “evil has no explanatory power” but maintains that we need the concept in order to describe and understand some cases of extreme wrongdoing (Gaita 2000, 46; Cole 2006, 64). However, Cole claims that people who use the concept of evil mean it to be “part of an explanation” of why the purported evildoer performed the purportedly evil action (Cole 2006, 7). For instance, some people say that terrorists or child murderers did what they 1   In Cole’s terminology, I will be setting aside the monstrous conception of evil, and focusing on the pure and impure conceptions of evil.

198  Evil and Explanation did because they were evil. Since many people who use the concept of evil believe that it is explanatory in Cole’s sense, we ought to consider whether it can be put to explanatory work, or whether, as Cole suggests, the concept turns out to be explanatorily vacuous. There are many things that we might want to know when we ask for an explanation of an action. Often we want to know which environmental conditions prompted the action. In many cases we also want to know about the psychological states of the agent. Sometimes we want to know whether it was this set of motives or that set of motives that moved the agent to act. Sometimes we want to understand why an agent chose this action rather than another of the available alternatives. Sometimes we want to understand why it was this agent, rather than another agent, who performed the action. Sometimes we desire a synchronic explanation that specifies the agent’s motives just prior to the action itself. Sometimes we want a diachronic explanation that sets out the processes that, over time, led the agent to have those motives. A concept will be explanatorily useful, in Cole’s sense, if it plays a role in explanations of these kinds. My account of evil suggests that there is not a single, unified concept of evil. We must distinguish the concepts of evil action, evil person, and evil feeling. Furthermore, I have argued that there are several viable conceptions of evil action, none of which picks out the real or the best concept of evil action. In asking whether the concept of evil is explanatorily useful, we ought to consider all of the more precise concepts within this broader group. First let us think about the concept of evil action. Imagine that an ordinary person describes an action—say, the torture and murder of a political prisoner—as evil. Is it plausible that, in doing so, she is offering some kind of explanation of why the action was performed? In comparison, let us consider the claim that an action is compassionate. This claim carries information about the motive from which the action was performed, and hence could help to explain the action. Of course, the fact that an action is compassionate does not completely explain why it was performed, and does not provide any information about the diachronic process by which the agent became the kind of person who is moved to act compassionately. Yet by claiming that the action was compassionate we are ruling out the possible motivational sets for the action that do not include the agent’s being aware of and being moved by the suffering of another. The concept of compassionate action is psychologically informative and explanatorily useful, even though the fact that the action is compassionate is both a merely synchronic and a merely partial explanation of the action. Whether the concept of evil action is similarly explanatorily useful depends, in part, on which specific conception of evil action is under consideration. As we have seen, the psychologically thin account of evil action implies that evil actions

Evil and Explanation  199 are culpable wrongs that are appropriately connected to extreme harms. The fact that an action is evil in the thin sense does not tell us much at all about the motives that lie behind that action.2 Philosophers who endorse the psychologically thin account of evil actions stress that such actions could come from a very diverse range of motives: from deliberate and defiant malice, misguided moral conviction, or banal and thoughtless pursuit of career advancement (Card 2002, 3; Neiman 2003, 252; Kekes 2005, 174; Arendt 2006, 252). The notion of thin evil action, unlike that of compassionate action, has barely any psychological content, and thus the claim that an action was thinly evil tells us almost nothing about why the action was performed. However, Cole would admit that some people employ psychologically thick conceptions of evil action. According to Cole’s notion of “pure evil” an action is evil only if it flows from a desire to inflict suffering for its own sake (Cole 2006, 15). In my terminology this corresponds to the view that an action is evil only if it is performed out of malice. The concept of pure evil action is analogous to the concept of compassionate action in that both of these concepts point to a desired end—either the infliction or the alleviation of suffering—that is pursued for its own sake by the agent through the action. Other philosophers who employ thick conceptions of evil action clearly take them to contribute towards explanations of those actions (e.g. McGinn 1997, 91; Garrard 2002, 332; Steiner 2002, 190). The burden of proof thus lies with Cole to show that a psychologically thick conception of evil action is not explanatorily useful. Regardless of whether the concept of evil action is explanatory, it is prima facie plausible that the concept of evil personhood has the requisite form to fit into explanations of extremely wrong actions. In some cases in which we encounter extreme wrongdoing we are tempted to say “He did it because he was evil”. As Cole points out, many journalists and commentators attempted to explain the murder of James Bulger by claiming that the child’s killers, Venables and Thompson, were evil (Cole 2006, 124–7). The psychiatrist Michael Stone also points to evil personality as a cause of evil action (Stone 2009, 313–18). Several philosophers share these folk intuitions. Colin McGinn claims that the evil person takes pleasure in inflicting suffering on others, hence removing the victims’ will to live, and McGinn believes that this character trait explains why the evil person performs evil actions 2   It is plausible that all culpably wrong actions must meet some basic psychological conditions regarding the rational capacity of the agent and the availability to that agent of information that is relevant to the action. Since the psychologically thin account implies that all evil actions are culpable wrongs, to call an action evil in the thin sense is to rule out at least some possible motivational sets. This might allow the thin conception of evil action to scrape over the threshold for explanatorily useful concepts.

200  Evil and Explanation (McGinn 1997, 77–87). John Kekes maintains that many evil actions are inexplicable unless we posit evil characters (Kekes 1990, 7). Some folk intuitions are misguided. In order to figure out whether evil actions really can be explained by evil personhood, we must remind ourselves of what it would take for someone to be an evil person. I have argued that S is an evil person if and only if S has an evil character, that is, S is strongly and highly fixedly disposed to perform evil actions (or to possess unrepudiated evil feelings) when in autonomy favouring conditions. On such an account evil personhood and character depend on a prior account of evil actions and attitudes. I have argued that we ought to accept a restricted conceptual pluralism about evil action, which admits several viable psychologically thick and psychologically thin conceptions. Given that the concept of evil personhood depends on a prior concept of evil action, this conceptual pluralism carries through to evil persons, although no extra layer of pluralism is added with this step. Thin evil persons are persons who are strongly and fixedly disposed to perform culpable extreme wrongs from any set of motives, be they malevolent, malicious, self-interested, or banal. Corresponding to the various folk versions of the psychologically thick conceptions of evil action are various thick evil persons; for instance, persons who are disposed to perform culpable extreme wrongs and take pleasure in doing so, or persons who are disposed to perform culpable extreme wrongs in knowing defiance of morality, and so on. The question of whether evil personhood is explanatorily useful, in Cole’s sense, boils down to the following: Can the fact that S is an evil person, on some or all of these conceptions of evil personhood, help to explain the fact that S performs terribly wrong actions? It is not surprising that the various distinct notions of evil personhood appear to have differing degrees of explanatory power. The fact that a particular soldier is an evil person on McGinn’s psychologically thick dispositional account, for instance, seems to explain why it was he, rather than other soldiers, who tortured and murdered a prisoner. A person who is evil in this sense is strongly and highly fixedly disposed to take pleasure in inflicting extreme suffering on others. For those who adopt a thin dispositional account of evil personhood, the concept of evil personhood is comparatively less informationally rich. The motives lying behind psychologically thin evil actions are diverse, and this implies that the character traits of people who are strongly and fixedly disposed to perform such actions could be similarly diverse. On this model, evil persons might be malicious and defiant wrongdoers, or moral idiots who think that the harm that they cause is justified in so far as it advances their self-interest, or sociopaths who simply do not register the fact that they are harming others

Evil and Explanation  201 (Morton 2004, 66). Since the thin dispositional account of evil personhood does not specify which of these motives are built into the evil person’s character, that conception gives us comparatively less information about why the wrong action was performed. We might worry that explanations invoking psychologically thin evil personhood are circular and hence uninformative. Garrard writes: If the question is: “Why did he commit an evil act?” and the answer is: “Because he’s evil”, then indeed little has been explained. Here we are far too close to that paradigmatic case of non-explanatoriness, the appeal to opium’s dormative powers to account for its ability to send people to sleep. (Garrard 2002, 332)

Yet even advocates of psychologically thin accounts of evil can use the concept of evil personhood in non-circular explanations of evil actions. The key to avoiding circular explanations is identifying the right question. If our question were “Why did taking the opium put him to sleep?”, then it would be uninformative to say that it was because opium has dormative virtue. Similarly, if our question were “Why did his character cause him to perform an evil action?”, then it would be uninformative to say that it was because his character is evil. Yet in many cases that is not our question. Often we want to know why he performed an evil action rather than an alternative action that was not evil, and why it was he rather than another person who performed the evil action. These are analogous to the questions “Why did he fall asleep now rather than stay awake?”, and “Why did he rather than other people fall asleep now?”. In response to these questions the claim that “He fell asleep because he took opium, which has dormative virtue” is clearly an informative alternative to other possible explanations, such as that he fell asleep because he did not get enough sleep last night and ate a heavy lunch. As we have seen, it is widely accepted that not every evildoer is an evil person; that is, that persons who are not evil sometimes perform actions that are evil (Card 2002, 22; Haybron 2002a, 279; Morton 2004, 65). The claim that he performed an evil action because he is an evil person evil is an informative alternative to other possible explanations, such as that he performed that action because he was overcome by a sudden and uncharacteristic desire for revenge, or that he performed that action because he uncharacteristically failed to consider the harm that his action was likely to cause. It is obvious that there are many other questions that need to be answered when we are trying to explain an evil action. However, the concept of evil personhood appears to help us to perform some explanatory tasks.

202  Evil and Explanation

10.2  Is Evil a Complete Explanation? I have argued that a range of explanatory roles can be played by the various viable concepts of evil action and the concept of evil personhood that I outlined in the preceding chapters. Some people remain sceptical about the explanatory power of the concept of evil. For instance, the psychologist Simon Baron-Cohen urges us “not to be satisfied with the word ‘evil’ as an explanatory tool” (Baron-Cohen 2011, 100), and the historian Inga Clendinnen claims that the word “evil” is “of no use whatsoever when it comes to teasing out why people act as they do” (Clendinnen 1998, 104). Philip Cole, who agrees with Baron-Cohen and Clendinnen, goes on to give a more elaborate set of arguments designed to show that any apparent explanatory power possessed by the concept of evil is merely illusory. Cole’s arguments are worth considering in more detail, not only because they challenge the legitimacy of the concept of evil, but because they will help us to identify the proper uses and the possible misuses of that concept. There are two grounds on which Cole argues that the concept of evil has no explanatory power. The first is that the concept has the wrong form to function in explanations. The second is that, even if it had the right form to function in explanations, the concept does not apply to any actual cases in the real world, and thus does not feature in any explanations of actual events. Let us consider these objections in turn. The first of Cole’s arguments is that the concept of evil purports to provide a complete explanation for wrongdoing, when in fact it cannot, and hence that the concept is fundamentally flawed. Cole rightly points out that some people appeal to evil when they are unable to explain why some children who have abusive upbringings go on to perform terribly wrong actions, whereas other children with very similar upbringings become morally decent human beings. Richard Bernstein claims that social sciences and psychology help us understand why people come to have certain desires and behavioural tendencies, but this “never adds up to a complete explanation of why individuals make the choices that they do. There is always a gap, a ‘black hole’, in our accounts” (Bernstein 2002, 235). Cole suggests that people try to fill this explanatory gap with the concept of evil personhood. The explanation for the murder of the toddler James Bulger is that the ten-year-old perpetrators, Jon Venables and Robert Thompson, are evil, or so we are told. However, Cole claims, “evil itself is a ‘black-hole’ concept; it does not so much close the gap in our account as make it more apparent. What we have here is the illusion of closure” (Cole 2006, 142). Even in what might seem to be the clearest cases in which evil is explanatory, Cole maintains, no real explanation has been offered.

Evil and Explanation  203 If we submit the Devil to any detailed examination to try to understand why he afflicts humanity, “because he is evil”, paradoxically, fails to explain anything about him and his motivations. (Cole 2006, 19)

Cole thinks that people mean “she was evil” to be a “complete description with nothing missing” that can “play a full role in an explanation of a human action” (Cole 2006, 7). Yet, Cole argues, the fact that S is evil could not be a complete explanation for any of S’s actions. To fully understand why the murder of James Bulger occurred, for instance, we would have to know about the history of the perpetrators; their horrendous home lives, their difficulties in school, and their own abuse at the hands of others (Cole 2006, 132–41). The claim that the perpetrators, Venables and Thompson, are evil does not convey any of this information, and hence it could not be a complete explanation of their action. Thus, Cole concludes, evil is not explanatory. Contrary to Cole’s claims, it is deeply implausible that evil character is supposed to provide a complete explanation for a particular action, and that evil is explanatorily useful only if it can provide such a complete explanation. To say that S is an evil person is to say that S is strongly and highly fixedly disposed to perform evil actions in certain conditions or that S possesses highly fixed unrepudiated evil attitudes. The primary difficulty in claiming that the concept of evil personhood is supposed to provide a complete explanation of actions is that attitudes and behavioural dispositions are never complete explanations of the performance of specific actions. A complete explanation of the action might well include reference to the attitudes and behavioural dispositions of the agent, but it would have to be supplemented at the synchronic level by extra information about the environmental conditions or the beliefs of the agent that triggered the exercise of the disposition, and supplemented at the diachronic level by extra information about how the agent came to have those dispositions in the first place. For instance, let us imagine that Sarah sees a spider and immediately runs out of the room. The fact that Sarah is arachnophobic does not, on its own, completely explain why Sarah suddenly ran out of the room. A complete synchronic explanation of that action would also include the fact that Sarah saw a spider in the room, that Sarah believed that there were no spiders immediately outside of the room, that Sarah has functioning legs, and so on. A complete diachronic explanation of the action would include an account of how Sarah came to be arachnophobic; an account that might include Sarah’s traumatic childhood encounters with spiders as well as natural selection on Sarah’s ancestors that produced an instinctive fear of spiders. None of this shows that arachnophobia is not an explanatorily useful concept. The fact that Sarah is arachnophobic helps to explain why Sarah ran out of

204  Evil and Explanation the room. Is there any reason to suppose that evil character differs from arachnophobia in this respect? When trying to explain an action, it is common for people to believe that they have answered all relevant explanatory questions when in fact they have provided a partial explanation, or no explanation at all. For example, imagine that we are wondering why Tony took offence at a seemingly innocuous remark, and that Christopher says that Tony took offence because Tony is Italian. Christopher might suppose that no further explanation of Tony’s action is needed. Yet the fact that Tony is Italian clearly is not a complete explanation of his having taken offence, and when Christopher offers it as if it were a complete explanation he has made a mistake. This does not show that there is something wrong with the concept of being Italian, nor that the fact that Tony is Italian can never play a role in an explanation of Tony’s actions. What we have in this case is merely the misuse of that concept. This instance of misuse is not a misapplication, because Tony really is Italian. Rather, Christopher fails to recognize or to respect the proper consequences of the application of this concept (cf. Robinson 2009, 337). Similarly, it is plausible that some people mistakenly judge that the fact that S is an evil person answers all of the open questions about why S performed an extreme wrong. Cole identifies the torture in Abu Ghraib as one case in which some misguided people reason in this way: When we are asking why the American guards in the Iraqi jail inflicted gross suffering upon their victims while others in the identical situation chose not to, then “because they were evil” completes our accounts of them, and tells us everything we need to know. (Cole 2006, 18)

Even if we do judge that the actions performed in Abu Ghraib were evil, we would need much more evidence concerning the strength and stability of the evildoers’ dispositions before we would be convinced that they were evil persons. Moreover, Cole is right to point out that the fact that the soldiers were evil persons would not tell us everything that we need to know. It would not tell us why those specific soldiers came to have those behavioural dispositions in the first place, nor what environmental conditions triggered the exercise of the dispositions, nor how widespread those dispositions are. Yet we have no reason to accept that the concept of evil personhood could be explanatorily useful only if it somehow carried all of that extra information. Someone who supposes that we can completely explain the actions in question by saying that the guards were evil has misused the concept of evil personhood, but that does not suggest that the concept itself is flawed or explanatorily vacuous.

Evil and Explanation  205 Cole might be tempted to argue that explanations which invoke the concept of evil personhood, unlike explanations which invoke other characterological concepts, are always intended to be stand-alone explanations, and could not be used merely as part of a richer synchronic and diachronic explanation of actions. Yet this is not the case. For instance, it is perfectly coherent for people to agree that Stalin was an evil person but to disagree about how Stalin came to be evil. It could have been because he was mistreated as a child, or because he had a genetic defect, or because he suffered some kind of brain injury, or because he was corrupted by absolute power, or because of some combination of these conditions (e.g. Stone 2009, 337). These varied attempts to identify the causes of a particular case of evil personhood show that evil personhood itself is not always intended to provide a complete explanation of actions. It is true that people who use the concept of evil personhood as if it provided a complete explanation of actions are making a mistake, but it is also true that many people use that concept without falling into this error. Another way to assess the explanatory usefulness of the concept of evil is to ask whether we could explain just as much and just as well without it. Cole encourages us to do this, rejecting the use of the concept of evil in favour of what he calls the psychological conception of extreme wrongdoing (Cole 2006, 17). It is worth considering whether Cole’s suggested alternative approach allows him to explain extremely wrong actions. According to Cole’s so-called psychological conception, the proper explanation of an act of extreme wrongdoing, such as the murder of James Bulger, is not that the wrongdoers were evil, but that they were violently mistreated themselves as children, that they lacked positive role models, and so on (Cole 2006, 140–1). An immediate problem for this argument is the fact that at least some of the worst serial killers, including Clifford Olsen and Dennis Rader, had “ordinary non-traumatic home environments” and lacked other “risk factors” such as head injury or mentally ill parents (Stone 2009, 285, 336). Not all extreme wrongdoing can be explained by abuse or trauma that the perpetrator suffered as a child. Another problem for Cole’s argument is that many children who grew up under such conditions did not go on to commit unprovoked murder (e.g. Stone 2009, 222, 289). Let us call those people the virtuous underprivileged. We need to explain why Venables and Thompson committed murder but the virtuous underprivileged, who lived under the same conditions, did not. Cole admits that in order to explain this difference in behaviour we need to posit dispositional differences between Venables and Thompson, on the one hand, and the virtuous underprivileged on the other (Cole 2006, 143). Yet this behavioural disposition towards extreme wrongdoing is not equivalent to evil character, Cole claims, because evil

206  Evil and Explanation is supposed to be an active cause of wrongdoing, whereas this disposition is only a passive causal factor that contributed to wrongdoing. The real causal explanation of Venables’ and Thompson’s wrongdoing remains the fact that they were mistreated as children. The explanation of the fact that Venables and Thompson, but not the virtuous underprivileged, committed murder is not that Venables and Thompson had something—evil character—that the others lacked, but rather that Venables and Thompson lacked something that the others had; they lacked the “ability to resist responding violently to their situation” (Cole 2006, 143). Cole’s distinction between active and passive causes is obscure, but he offers two analogies to help bring out what he has in mind: Suppose there is a wall built to resist flood water, but suppose certain areas of the wall have internal structural weaknesses, such that when the flood water comes parts of the wall stay in place and resist, but other parts collapse. There is only one active cause here, the flood water, and we do not have to suppose there was some other active cause at work to make the walls collapse, such as evil saboteurs—the only other causal factor we need for a complete explanation is that there were structural weaknesses within the structure of parts of the wall. (Cole 2006, 143)

Cole also asks us to consider a precariously placed bucket collecting drops of water that gradually fills and then collapses catastrophically. Although the change in behaviour [of the bucket] is sudden and dramatic, its causes are not, and just because the drops of water on previous days did not lead to this catastrophic outburst of destructive violence, this does not show that they were not part of the causal process leading up to it. . . . By analogy, a child could be on the receiving end of abuse and other kinds of negative experience until he or she reaches a saturation point which triggers a radical change of behaviour, perhaps extreme and catastrophic violence. . . . [I]‌t is extremely implausible to suppose that all children have the same saturation point. . . . The difference in where their saturation points lie plays a role in the complete explanation of the violent behaviour [of the children] but it is not an active causal feature, and it is in the supposed need for an extra active causal factor to make the difference between them that evil gains a foothold in the explanation. (Cole 2006, 144–5)

We have already examined some of the mistaken assumptions that underlie Cole’s argument for the so-called psychological conception of extreme wrongdoing. It simply is not true that social and psychological diachronic explanations of wrong actions must be in competition with synchronic explanations that appeal to evil character. We might claim that the physical and psychological abuse suffered by Venables and Thompson gradually turned them into evil persons, and their evil character helps to explain why they performed evil actions. Of course, we might claim instead that Venables and Thompson were not evil persons at all, because they were young enough not to have stable characters, or because their extremely

Evil and Explanation  207 wrong actions were anomalies that did not reflect their characters. Yet this would not imply that the concept of evil personhood is not explanatorily useful; merely that it is not useful in explaining the killing of James Bulger. The repeated murders committed by adults such as Ted Bundy, in contrast, clearly reflect a strong and settled disposition in the wrongdoer to perform extremely harmful culpable wrongs. The fact that Bundy is an evil person does help to explain why he committed those murders. The analogies that Cole draws between extreme wrongdoing and collapsing walls and toppling buckets are not apt, and the failure of these analogies reveals a significant weakness of his own explanation of extreme wrongdoing. Cole’s point is that small contributing causes can accumulate and, once they exceed a saturation point, can trigger sudden and harmful events. This is certainly true in cases of collapsing walls and also in some cases of human action. However, the kind of human actions that we might try to explain by appealing to evil personhood are not undirected collapses or psychological failures, which may or may not be harmful. Rather, they are deliberate attempts to inflict extreme harm on others, sometimes for its own sake. Often the actions in question are not sudden responses to immediate threats, but are carefully planned, and carried out many times over. Some of the actions in question form part of a socially coordinated and explicit effort to inflict suffering on others. Cole suggests that we should explain any such action by pointing to an accumulation of contributing harms inflicted on the agent and a saturation point within the agent, and see the action as a collapse; an inability to resist the pressure any longer. Yet this leaves a conspicuous gap—a black-hole, we might say—in the explanation. Why does the so-called collapse result in the agent deliberately and repeatedly seeking to harm others? Why is the resulting action directed and purposeful in this way? Why, for instance, does the collapse not consist in a mental breakdown, a refusal to interact with other people, or a fit of tears? This is left utterly mysterious and unexplained by Cole. In contrast, those of us who posit the existence of evil persons are able to go some way to explaining these patterns of repeated extreme wrongdoing.

10.3  Is Any Actual Person Evil? I have argued that the concept of evil personhood has the right form to contribute to explanations of actions. However, this concept will do explanatory work in the real world only if it applies to some actual persons. Cole argues that, even if the concept of evil personhood had the right kind of form to be explanatory, it still would be unexplanatory in virtue of the fact that no actual person is evil. According to Cole, the pure conception of evil suggests that ordinary

208  Evil and Explanation human beings are capable of inflicting suffering for its own sake (Cole 2006, 15). The impure conception of evil suggests that humans are capable of causing suffering not for its own sake, but “for some other human end, such as power, wealth, security, or the greater collective good” (Cole 2006, 16). It is clear from his other claims, though, that Cole believes that an evil person is someone who is not merely capable of inflicting suffering either for its own sake or for some other end, but someone who is very strongly and fixedly disposed to do so. For instance, Cole maintains that an evil person would be beyond “communication and negotiation, reform and redemption” (Cole 2006, 236). We have seen that evil actions, on any plausible account, are connected to actual or possible extreme harms, not merely to minor suffering, and hence that an evil person cannot be merely trivially bad, but must be extremely bad. Again, it seems that Cole would agree to this condition. He claims that by thinking of people as evil we suppose that they deserve “complete condemnation” (Cole 2006, 236). Cole argues that no one deserves complete condemnation, and no one is sufficiently strongly and fixedly disposed towards extreme wrongdoing so as to count as purely or impurely evil. Let us turn our attention to the question of whether anyone deserves complete condemnation. There are two distinct ways in which condemnation of a person might be complete. First, it might be the strongest kind of condemnation of a person. Let us call this the extreme condemnation of a person. Second, it might be condemnation of the person in every respect, such that every action, attitude, and disposition of the person is condemned. Let us call this the thorough condemnation of a person. Cole suggests that no actual person deserves complete condemnation. Yet Cole explicitly aims to provide an account of wrongdoing that allows that “moral judgment—sometimes of the most severe kind—is possible” (Cole 2006, 148). He writes: we are rejecting evil as a category that can meaningfully be applied to human beings. We can still condemn such people to the fullest extent of our judgment. . . . The rejection of the discourse of evil does not rule out moral condemnation and legal punishment of the most extreme kind . . . (Cole 2006, 173)

In this passage, at least, Cole appears to endorse extreme moral condemnation of some actual persons. When Cole claims that no one deserves complete condemnation, it seems that he means that no one deserves thorough condemnation. For instance, Cole suggests that even such extreme wrongdoers as Hitler and Mengele were not evil persons because each had a good side, and hence did not deserve thorough condemnation (Cole 2006, 196–8), a view that seems to be shared by Haybron (2002a, 270).

Evil and Explanation  209 In Chapter Eight, I argued that an evil person need not be bad in every respect and on every occasion. Indeed, some evil persons are more effective in their evildoing because they possess some admirable qualities such as courage and loyalty. Cole’s own definitions of pure and impure evil do not imply that evil persons are thoroughly bad. Moreover, Cole and Haybron offer no evidence that the everyday uses of the concept of evil conform with the assumption that only thoroughly bad people could be evil. Haybron even admits that his account “departs from ordinary ascriptive practice” (Haybron 2002a, 271). If we want to determine whether Dennis Rader, say, is evil, it would be irrelevant to ask whether he reliably was kind to his parents, whether he was honest and diligent in the workplace, or whether he occasionally made a generous donation to charity. What matters is whether he is strongly and fixedly disposed to inflict extreme harm on some innocent people. Cole is right to suggest that by claiming Rader is evil we imply that he deserves complete moral condemnation, in some sense. Yet it is a certain kind of extremity of vice, not thoroughness of vice, that provides the gravity required for evil personhood. Cole’s claim that everyone has a good side, even if true, does not show that no one is evil. Cole’s suggestion that an evil person would be beyond “communication and negotiation, reform and redemption” provides a stronger challenge to the view that some actual persons are evil (Cole 2006, 236). According to Cole, no one is sufficiently strongly and fixedly disposed towards extreme wrongdoing so as to be irredeemable, hence no actual person is evil, and the concept of evil personhood can play no role in explanations of actions. Cole does not state clearly what it means for an extreme wrongdoer to be redeemed. It might be suggested that an extreme wrongdoer has been redeemed if and only if he has been forgiven by those wronged, or by the community, or by God. Alternatively, perhaps a wrongdoer has been redeemed when someone else pays his moral debt for him. However, it seems that S could be redeemable in either of these senses even if S was an evil person, so long as S’s redeemers were sufficiently forgiving or were capable of paying S’s debt. The fact that S is redeemable, in these senses, tells us more about the redeemer than the redeemed. A more relevant case in which an extreme wrongdoer might be described as having been redeemed is that in which he has lost the disposition to perform extreme wrongs or to have evil feelings. This change might well be necessary for redemption, but it is not sufficient. For example, let us imagine that a regular extreme wrongdoer suffers a terrible head injury and loses virtually all of his behavioural dispositions. He no longer is strongly disposed to do wrong, but that is because he no longer is disposed to do anything morally significant at all. It is implausible that such a wrongdoer has been redeemed via his head injury. The sense of redemption

210  Evil and Explanation that seems to be relevant for Cole’s argument is that in which an extreme wrongdoer becomes a good person who is strongly disposed to do right rather than wrong.3 Cole rejects the view that any actual person is incapable of “reform and redemption” (Cole 2006, 236), so it is plausible that he intends his claim that every person can be redeemed to mean that every person can be made good. In Chapter Eight, I argued that an evil person would have a strong and highly fixed disposition to perform evil actions when in autonomy-favouring conditions. The strength or weakness of S’s disposition to ø in conditions C is a measure of how likely it is that S will ø when placed in C. The fixedness or flexibility of S’s disposition to ø in conditions C is a measure of the ease with which, over time, the strength of that disposition can be changed. If Cole is to argue convincingly that no actual person is evil, he will have to show that evil persons would have dispositions towards extreme wrongdoing that exceed a certain degree of strength and a certain degree of fixity, rendering those persons unpersuadable and irredeemable, and that no actual person has such dispositions. Not everyone who performs extremely wrong actions is strongly disposed to do so. An extremely wrong action might be uncharacteristic; due to chance or circumstance rather than the agent’s character. Even if we agree that there are some actual evil actions, we should believe that not every evildoer is an evil person. It is plausible that some extreme wrongdoers could have been talked out of their wrong actions had they been engaged in rational argument before the act. Yet there are many examples of people, including dictators, war criminals, and serial killers, who repeatedly have performed extremely wrong actions. Some of these wrongdoers performed these actions in difficult or coercive circumstances, in which they were threatened, ordered, bribed, or misled, but many others acted in circumstances that could not be described as difficult or coercive. They actively sought out and created opportunities in which they could inflict extreme harm. We have reason to believe that many of these extreme wrongdoers could not have been talked out of their actions. Some of these wrongdoers were acting in the name of a political or religious cause to which they were deeply committed, and carried on with their wrong actions despite explicit and repeated attempts at negotiation. Some other extreme wrongdoers, including some serial killers, appear to have been acting compulsively, in which case negotiation and persuasion probably would have failed. It is hard to see how we could deny that such people are very strongly disposed to perform extreme wrongs.

3   “Good person” should be taken to mean someone who is good enough, rather than someone who is morally good in every respect, or morally good to an extreme degree.

Evil and Explanation  211

10.4  The Fixity of Our Dispositions If Cole is to show that no actual person is evil, his remaining option is to argue that no actual people are fixedly disposed to perform such actions, but that evil people would be so disposed. It is difficult to assess exactly what degree of fixity is implied in the concept of evil personhood. If it were true that the concept implied absolute fixity of character, then it would be contradictory to say that S once was an evil person but now is a good person, or even that S is an evil person but might one day become a good person. As I argued in Chapter Eight, it is not clear that such claims are contradictory. For instance, we might claim without contradicting ourselves that the thirty-year-old Ted Bundy was an evil person, but that by the age of thirty-one he could have been made into a good person by radical brain surgery, or by a miraculous act of God. In contrast, it would be very odd—arguably incoherent—to claim that a thirty-year-old Ted Bundy was an evil person but that if we treated him nicely and taught him some moral philosophy, by the time of his thirty-first birthday he might not be evil anymore (cf. Haybron 2002a, 277; Vargas 2010, 74). Cole seems to be right in thinking that an evil person, in virtue of being evil, could not easily be taught or trained to be a good person. An evil person would have a strong disposition towards extreme wrongdoing, and that disposition would have a high degree of fixity, but would not necessarily be fixed absolutely.4 According to some ethicists, similar conditions hold true for ascriptions of virtue. If S really is an honest person, then S would be strongly disposed to tell the truth, and that disposition would have a high but not absolute degree of fixity. It is important to note that the implicit ascription of fixity of character that follows from the ascription of virtue is a forward-looking fixity only. Aristotle claimed that a young person is not virtuous to begin with, but can acquire highly fixed virtue through training (Aristotle 1985, NE, 1103b21–25; 1100b12–15). The forward-looking fixity of character implied by ascriptions of virtue to S is compatible with the claims that S’s virtuous character was not possessed by S at some point in the past, and that S’s virtuous character is the contingent product of training,   It is not clear what we would say if there were a dramatic change in our technological abilities such that we were able cheaply and easily to mould the behavioural dispositions of everybody via neurosurgery or drugs. In such circumstances, we might say that henceforth no one has a highly fixed character because now every person easily can be made good or made bad, and thus that now no one is virtuous and no one is evil. Yet it would be odd to claim that every virtuous person could suddenly cease to be virtuous due to the creation of a new technology, whether that technology has been applied to them or not. Alternatively, we might fix upon what were the ordinary means for manipulating character when our concepts of virtue and evil were developed, and maintain that a person is virtuous or evil only if their character cannot easily be changed by those means. 4

212  Evil and Explanation and is far from being innate or inevitable. Arguably, the same holds true for ascriptions of evil personhood. There is no contradiction in claiming that S used not to be an evil person and need not have turned out that way, but came in fact to be evil through being mistreated, or through selfishness and self-deception, or through indoctrination. An evil person, then, is strongly disposed to commit extreme wrongs, and that disposition has a high degree of fixity but is not necessarily absolutely fixed, and that disposition was not necessarily innate or inevitable, but necessarily is fixed in a forward-looking direction. As we have seen, Cole claims that the child murderers Venables and Thompson were not evil persons. His argument for this conclusion rests on the assumption that an evil person must have an innate or inevitable fixed disposition towards extreme wrongdoing, rather than a disposition that arose contingently and is fixed merely in a forward-looking direction. For instance, Cole claims that “Venables and Thompson were damaged children, not evil monsters—they were already broken” (Cole 2006, 145). Similarly, he writes: And this is perhaps the crucial point that the “moral monsters” thesis overlooks, or prefers to ignore. Thompson and Venables, as with other children who kill, were not born killers. They became killers. (Cole 2006, 137)

Cole is probably right to claim that Venables and Thompson were not doomed by their genes to commit murder (cf. Stone 2009, 321–2). Yet the question of whether Venables and Thompson are evil is not settled by whether their disposition towards extreme wrongdoing was flexible at some point in the past, such that, had they been raised in decent homes, they might have become virtuous citizens. Rather, the question turns on whether they have a sufficiently strong disposition towards performing extreme wrongs that is highly fixed in a forward-looking direction. The best argument for the conclusion that Venables and Thompson were not evil is that they were still children at the time of the murder, so, even if they were culpable for their wrong actions, they were still flexible enough to be moulded by ordinary means into comparatively virtuous persons. Admittedly, the empirical claim about flexibility of character in this argument is speculative. After all, Venables and Thompson might never have come to understand sentences written in English had they been raised differently, but once they were ten years old they had acquired a strong and highly fixed disposition to do so. Given our relative lack of information about Venables and Thompson as they have grown into adulthood, it is hard to tell whether their respective dispositions towards extreme wrongdoing were more or less fixed than their disposition to understand English. Phil Roberts, a police officer who interviewed one of Bulger’s killers, said, “They were evil. I think they would have killed again” (Morrison 1997, 230–1; Cole 2006,

Evil and Explanation  213 126). Whether or not we agree with Roberts’ prediction, at least he follows up his ascription of evil by pointing us in the right direction: not towards the flexibility of the boys’ dispositions in the past, but towards the forward-looking fixity of their disposition to extreme culpable wrongdoing. The murder of James Bulger is far from the ideal test case for the claim that some actual persons are evil. The fact that Venables and Thompson were only ten years old at the time of the killing raises questions about their culpability. Many people believed that the young offenders did not have firmly fixed characters, and that they had been rehabilitated by the time they were released back into the community. The authorities gave new identities to Venables and Thompson to protect them from reprisals, and this means that we now have very patchy information about their long-term behaviour, although we do know that Venables was sent back to prison on child pornography charges in 2010. We have limited access to the evidence that would confirm or disconfirm the hypothesis that Venables and Thompson are strongly and highly fixedly disposed towards extreme wrongdoing when in autonomy-favouring conditions. The challenge for Cole is to provide reason for believing that no actual person has such a disposition. Correspondingly, the challenge for those of us who do claim that some persons are evil is to provide reason for believing that some actual human beings not only have performed extreme culpable wrongs, but are strongly and highly fixedly disposed to do so. Both challenges are difficult to meet because it is hard to ascertain the fixity of behavioural dispositions. Even if we agree that S is strongly disposed to ø when in autonomy-favouring conditions, how can we determine whether S’s disposition is highly fixed except by subjecting S to a very broad range of training regimes or other processes that seem to be likely candidates for changing S’s disposition? Many people think that, rather than run this test on extreme wrongdoers, we are morally obliged to prevent them from living in autonomy-favouring conditions where the relevant disposition might be manifested. Instead, extreme wrongdoers should either be killed or locked away in a place in which severe limits are imposed on their freedom to act as they wish. The evidence that would tell us whether anyone is evil has been and often will be difficult to gather. Unfazed by this difficulty, Cole confidently declares that every person is redeemable, and hence that no one is evil (Cole 2006, 173). The strongest support Cole can offer for this view is a seemingly a priori argument concerning free action. He claims that, as free agents, we are capable of choosing to perform terribly wrong actions. Yet, Cole writes: because we know that under similar situations we are capable of choosing to do such dreadful things to others, we also know that we are capable of choosing not to do them. And if

214  Evil and Explanation those who actually have chosen them are people like us, then they were equally capable of not choosing them. They are, therefore, redeemable. By bringing “evil” into the realm of the human, by humanizing evil, we find that it can be opposed and defeated. (Cole 2006, 167)

A redeemed person is not merely capable of choosing to do what is right, but is fairly reliably disposed to choose to do what is right, and is reliably disposed not to choose the worst wrongs. A redeemable person is someone who can come to have this good disposition. As we have seen, the fact that S is an evil person would imply that S has a disposition to perform extreme wrongs that is fixed to a high degree, but that is not necessarily absolutely fixed. The fact that S is an evil person would imply that S is not easily redeemable by ordinary means, such as rational persuasion, education, training, or punishment. Thus, Cole seems to be arguing that the fact that every extreme wrongdoer could choose to act otherwise implies that every extreme wrongdoer could, by ordinary means, come to have a reliable disposition to choose to do right and not wrong. Clearly, though, the ease or difficulty with which we can change our behavioural dispositions cannot be established by pointing to the bare fact of free agency. The fact that we could choose not to communicate with anyone for the next five minutes does not show that, through ordinary means, we could be stripped of our disposition to communicate with each other. In this case, as in all cases, we must determine the flexibility or fixity of a particular behavioural disposition by looking at real situations in which that disposition varies over time, and at real attempts to change that disposition. It is hard to see how Cole’s claim about free choice assists at all in our efforts to oppose and defeat extreme wrongdoers such as Hitler, Himmler, Goering, and Eichmann. Our primary aim in such cases is to prevent these people from wreaking havoc and inflicting extreme harm. The fact that these war criminals freely could have chosen to act differently did not make the Allied war effort any easier, nor would awareness of that fact have changed the Allied strategy. Part of what made it hard for the Allies and the occupied nations to defeat Nazi Germany was the strength and fixity of the leading Nazi’s respective dispositions to extreme wrongdoing, which rendered them immune to the pre-war attempts at negotiation or persuasion, and which led them resolutely to carry out genocide for a sustained period of time despite great cost and despite the great risk of catastrophic reprisal.5 Nonetheless, the leaders of the Nazi party acted freely, and could have chosen 5   I claim that these Nazi war criminals were evil persons, and that this fact helps to explain why attempts at negotiation and rational persuasion failed. I am not suggesting that we can explain the occurrence of the Second World War merely by pointing to the fact that these men were evil. Obviously an adequate explanation would include much more historical and psychological information.

Evil and Explanation  215 otherwise. Since free action is compatible with firmly fixed dispositions to extreme wrongdoing, recognizing that wrong actions are performed freely does not, in and of itself, help us to prevent those wrong actions. It is despots and war criminals like these, along with adult serial killers, who constitute the best test cases for whether any actual person is evil. None of these people are supernatural or inhuman monsters, literally speaking, but literal monstrosity is not implied by the claim that they are evil. It is quite possible that every one of these people has a good side, and hence that none of them deserve thorough condemnation, but we have seen that evil persons would deserve extreme rather than thorough condemnation. It is quite possible that all of these people might not have developed strong dispositions towards extreme wrongdoing had they been raised differently, but evil persons necessarily have fixed forward-looking, rather than innate or inevitable, dispositions to extreme wrongdoing. Cole could maintain that even these war criminals and serial killers did not possess a disposition to commit terrible wrongs that was fixed to a sufficiently high degree in a forward-looking direction, and thus they could have been redeemed and made good by rational persuasion or some other ordinary means. We would only know for sure that their respective dispositions were sufficiently strongly fixed if we had tried every available means of persuasion. Since we did not try every means, the possibility of redemption has not definitively been ruled out. Nonetheless, given the many examples of sustained extreme wrongdoing in the face of rational persuasion and the threat of violent reprisal, believing that no actual person is evil seems to require a wild leap of faith. Jack Unterweger, an Austrian murderer who served 15 years of a life term, convinced the authorities, prison psychiatrists, and many members of the public that he was remorseful and had reformed. Upon being released from prison, Unterweger promptly went back to strangling prostitutes (Stone 2009, 241). Consider also the case of Clifford Olsen, the Vancouver serial killer who murdered eleven children. Asked by a journalist what he would do if he were ever released from prison, Olsen replied “I’d take up where I left off.” When asked if he felt remorse in light of the suffering experienced by the families of his victims, Olsen said “If I gave a shit about the parents, I wouldn’t have killed the kids” (Stone 2009, 350). In light of the evidence available, we ought to conclude that Unterweger and Olsen are not redeemable. They are evil. Cole has failed to provide convincing arguments for the view that the concept of evil is explanatorily vacuous. We have seen that the concept of evil personhood has the right kind of form to fit into explanations of evil actions. In many cases the evidence of sustained extreme wrongdoing strongly indicates that there are some actual evil persons. In these cases it makes sense to say “He did it because he was

216  Evil and Explanation evil”. It would be a mistake to suppose that the fact that the person was evil provides anything like a complete explanation for the action in question, but it would also be a mistake to suppose that this fact can contribute nothing to explanations of evil actions. We should not forget that not every evildoer is an evil person. Even if we have doubts as to whether any actual person has a sufficiently fixed character to count as an evil person, we should judge that many actual actions are evil. As we have seen, the various viable conceptions of evil action, especially the psychologically thick conceptions of evil action, can play an important role in explanations of action.

11 Doing Away With Evil? In this final chapter, I will address some remaining pragmatic and moral objections to my defence of the concept of evil. Even if there are some evil actions and some evil persons, as defined in my account, critics might argue that the ambiguity and vagueness inherent in the concept of evil is an obstacle to clear communication and precise moral judgement. Some critics also claim that we should not employ the word “evil” in public moral discourse, nor the concept of evil in our private moral judgements, because thinking in terms of evil has morally bad consequences. Cole suggests that the concept of evil introduces a dangerous Manichean worldview, and that are many cases in which public pronouncements that certain people are evil have morally disastrous effects. By demonizing the enemy, we destroy the chance of negotiation and rational engagement, and encourage atrocities. In response to these moral arguments I concede that the concept of evil—especially the concept of evil personhood—is often misused, and should be used sparingly and handled with great care. Yet this is true of some other fundamental moral concepts as well. Ultimately, we have good reason to judge that some actions and some persons are evil, and we should be willing to make these judgements public.

11.1  Pragmatic Objections Throughout this book I have repeatedly divided the folk concept of evil. I argued that the word “evil” is sometimes used as a label for a malevolent supernatural force, but that this usage does not reveal the core of the secular moral concept of evil. I divided the concept of evil action from those of evil feeling and evil person, and argued that there are several viable conceptions of evil action. The variety and complexity in my account might prompt a pragmatic argument against the use of the concept of evil. Folk moral discourse undeniably is messy, and one of our aims in doing moral philosophy is to reduce confusion and facilitate clear

218  Doing Away With Evil? communication regarding moral issues. If the word “evil” is so vague and ambiguous that it requires clarification whenever it is used, perhaps we should stop thinking in terms of evil, and instead use more precise and univocal substitutes. In evaluating this argument we must keep an eye on the difference between the word “evil” and the concept of evil. Someone who wishes to purge the concept of evil from her thinking might aim to do so by dropping the word “evil” from her moral utterances; by no longer silently saying to herself that terrorist bombings are evil, or that the United States Government is evil, and so on. Yet it is implausible that concepts are tied so tightly to specific words that someone who does not publicly or privately utter a specific word is guaranteed not to employ the concept that typically is picked out by that word (Jackson 1998, 33; Kingsbury and McKewon-Green 2009, 179). For example, the concept that is usually picked out by the word “father” can also be picked out by the words “male parent”. Someone who regularly judges and says that there are male parents is using the concept of father, even if she never uses the word “father”. Questions about how best to foster clear communication are focused on the use of the word “evil”, rather than the concept of evil. Nonetheless, philosophers who think that the concept of evil plays an important role in moral thought should respond to this pragmatic challenge. One way to defend the use of the word “evil” would be to argue that other words or phrases that we might use in its place are incapable of capturing the semantic content or the expressive force that is conveyed by “evil”. It certainly is true that some attempts to substitute alternative terms for “evil” are hopelessly inadequate. Baron-Cohen suggests that we “substitute the term ‘evil’ with the term ‘empathy erosion’ ” (Baron-Cohen 2011, 4). Clearly this will not work as a substitute for “evil” as applied to actions. Nor will it serve the purpose that Baron-Cohen appears to intend; namely, that it replace “evil” as applied to persons. Not every person who lacks empathy should be morally condemned as evil, as some might fall below the bounds of moral responsibility, and others might be emotionally flat but not strongly or fixedly inclined to extreme wrongdoing. Baron-Cohen’s proposed substitution contentiously would rule out the possibility that some extremely harmful actions, such as those performed in a Milgram scenario, are evil even though the perpetrators feel empathy for the victim. Another simple but implausible option would be to substitute the word “atrocity” for “evil action” (cf. Card 2002, 9). Every atrocity is extremely harmful, yet not every evil action inflicts harm, so some evil actions are not atrocities.1   Card does not believe that we can substitute “atrocity” for “evil” because “[m]‌any evils lack the scale of an atrocity” (Card 2002, 9). However, it is possible that Card’s use of the “atrocity paradigm” leads her to think that every evil deed is extremely harmful (Card 2002, 20). 1

Doing Away With Evil?  219 While there is no other single English word that is a plausible substitute for “evil”, I have argued that at least the semantic content of “evil action”, “evil person”, and “evil feeling” can be captured by more complex definitions setting out necessary and sufficient conditions. Thus I am not inclined to say that we must use the word “evil” if we want to express the judgement that an action or a person is evil. In very many cases I think that we would aid clear communication by expressing our moral judgements in terms of the more complex definitions. When people are disagreeing over whether Osama Bin Laden was an evil person, it would be helpful to break this question down into more basic components: Was Bin Laden strongly disposed to perform the worst kind of wrong actions? Was this disposition highly fixed? Did he act out of malice towards his victims? Did he knowingly defy morality? These more fine-grained questions are important, and by asking these questions we will be better able to sort substantive moral disagreements from mere linguistic disagreements. Even though we can break down the comparatively vague question of whether an action or a person is evil into more fundamental and more precise questions using other terminology, I think we should not simply stop using the word “evil”. There are at least two reasons for this. The first is that, independent of its semantic content, “evil” has an expressive force that is not possessed by other semantically equivalent phrases that might be used as substitutes. As I argued in Chapter Two, the expressive force of the word “evil” depends on the moral commitment of the person who utters the word, but for ordinary people who are committed to morality, the claim that an action is evil functions as a clear and powerful expression of disapproval (e.g. Morrow 2003, 7). Saying that the action is an extreme culpable wrong may well convey the same factual information and the same negative evaluative judgement, but it lacks the arresting emotional punch of the claim that the action is evil. In Garrard’s words, [t]‌he terrible massacres of the twentieth century, the hideous and endless ingenuity of its tortures, seem to require description in terms of evil . . . because other kinds of moral condemnation do not capture their nightmarish horror. (Garrard 2002, 321)

It is understandable that people whose lives are impacted by extreme wrongdoing want to denounce it in the strongest possible terms. Proclaiming that certain actions are evil is also a particularly useful method of drawing attention to those actions, and calling others to help to prevent such wrongs in future. For instance, let us consider a survivor of Auschwitz who says that Dr Mengele was an evil man guilty of evil deeds. It is one thing to encourage the survivor to elaborate and explain her judgement using more precise language, but it would be wrong to ask her to soften the emotional force of her condemnation by retracting the claim that Mengele was evil.

220  Doing Away With Evil? A second reason for maintaining the use of the word “evil”, despite its vagueness and ambiguity, is that “evil” plays a significant role in folk moral discourse. Not every word that plays a role in folk discourse ought to be preserved and defended by philosophers. The fact that many ordinary people talk about ghosts, for instance, does not imply that philosophers ought to engage in ghost talk as well. Yet when a particular domain of folk discourse is coherent and does not rest on any kind of error, philosophers who are theorizing about that domain have a strong pro tanto reason to explain and employ the key terms in that folk discourse (Kingsbury and McKewon-Green 2009, 172). As moral philosophers our aims ought to include those of engaging with non-philosophers, making sense of their moral claims, and helping them understand the relationship between moral concepts and the nature of moral judgements. In this case I suspect that the achievement of these goals would be hampered by the removal of the word “evil” and the replacement of it with more precise and elaborate paraphrases. By way of analogy, let us consider the concept of freely chosen action. Folk discourse about free will is firmly entrenched, and is connected to many evaluative judgements concerning issues such as responsibility, punishment, happiness, and political rights. When we engage in philosophical reflection about free will we discover that there are several distinct concepts that are candidates for counting as the concept of freely chosen action: action caused by one’s own desires absent coercion, action caused by the desires that one desires to desire, action that accords with one’s desires (whether or not it is caused by those desires), action caused by a non-physical self, and so on. Arguably some of these, including the libertarian conception of free action, ought to be rejected on metaphysical grounds, but there remain several metaphysically respectable and useful concepts that are candidates for being described as the concept of freely chosen action, none of which is obviously the thing that most deserves the label “free action”. Philosophers who hold this view might go so far as to endorse a restricted conceptual pluralism with regards to free action, but even if they do not, they should conclude that folk claims about free action are vague or ambiguous, and require clarification (cf. Sider 2001). In the face of this superfluity of reasonable options it might be suggested that we would aid clear communication by simply refusing to say that any action is freely chosen, and trying to replace the folk vocabulary involving “free choice” with more precise and univocal claims, such as that an action was caused by the agent’s desires, or that it was caused by the desires that agent desires to desire, and so on. Yet refusing on these grounds to say that any action is freely chosen is actually more likely to undermine clear communication. For example, my refusal to say that Sam’s voting for the Greens was a freely chosen action would lead some people to conclude that I must believe that Sam does not deserve to be praised or

Doing Away With Evil?  221 blamed for his action, and that Sam had no control over his action. Even if I am a conceptual pluralist about free action, it would be pragmatically better for me to keep using the word “free”, and to specify answers to the more detailed questions as to whether his action meets the criteria for the various possible philosophical conceptions of free action. Similarly, a philosopher who recognizes that there is vagueness and ambiguity in folk talk about evil, and who wants to communicate clearly with the folk on questions regarding extreme wrongdoing and extreme vice, would invite more confusion if she refused to use the word “evil”, and denied that any actions are evil. What we ought to do instead is explain and clarify our claims that certain actions are evil. The fact that it is difficult to communicate clearly when we use the word “evil” is not unusual. The same kind of vagueness and ambiguity that is inherent in the folk term “evil” can be found, to varying degrees, in most of our moral vocabulary. Even the claim that an action is morally right is ambiguous: it could mean that the action is amongst the morally permissible options, or that it is the single morally best option out of those available, or that it is a permissible action that was performed from admirable motives. The appropriate philosophical response to this ambiguity is not to say that there is no such thing as a morally right action, and that we ought to replace this vague talk of rightness with more precise language. Rather, we should agree with the folk that there are right actions, and, when necessary, specify which sense of rightness is in play in any particular discussion. The fact that we often need to do the same thing with relation to claims about evil does not show that we could communicate more clearly if we dropped the word “evil” from our moral vocabulary.

11.2  Moral and Political Costs Some philosophers have suggested that thinking in terms of evil has morally and politically bad consequences. Claudia Card points out that some critics believe that the discourse of evil “calls forth disturbing emotions, such as hatred, and leads to such disturbing projects as revenge” (Card 2002, 28). If this were the case, then we might believe that even if there are good epistemic reasons for judging that some actions and some persons are evil, we ought to do away with the concept of evil, and should drop the word “evil” from our moral discourse. Phillip Cole is one philosopher who believes that evil-talk is a “highly dangerous and inhumane discourse and we are better off without it” (Cole 2006, 21). This also could be described as a pragmatic objection to the concept of evil, but it is focused on moral costs rather than purported difficulties in clear communication.

222  Doing Away With Evil? Cole thinks that people who use the concept of evil create a unrealistic opposition between two kinds of person; they demonize their opponents rather than seeing them as ordinary and rational human beings. For example, Cole claims that media characterization of murderers as evil creates “an imagined binary between the victim and the victimiser” (Cole 2006, 127). While we might be tempted to describe the Holocaust as evil, Cole suggests that we ought to resist this temptation. The discourse of evil is so dangerous that we must try to do without it. In the face of events like the Holocaust, we have to remember that one major factor in bringing it about was a particular discourse of evil, the anti-Semitism that drove Hitler and his followers, the belief that the Jews represented a cosmic evil enemy bent on the destruction of the German people and civilisation in general. (Cole 2006, 236)

Cole argues that the discourse of evil is used by the powerful for political purposes: The historical record shows that [the discourse of evil] has been directed against weak and marginalised groups who present very little danger to those in power. The real threat to the powerful comes from the loss of popular support by the masses, and the way to keep their loyalty is to direct their hostility—or ressentiment—towards an “evil enemy”, and convince them that the only way to be safe from this terrible enemy is to remain loyal to the ruling caste. . . . The group picked out as the “evil enemy” can never be that genuinely dangerous, because then, of course, the ruling caste would face a genuine threat to their power. (Cole 2006, 74)

This political power play lies behind the historical cases of witch trials and vampire scares (Cole 2006, 77–89), and more recent characterization of migrants as evil (Cole 2006, 215). Like Cole, I believe that we ought not divide people into two and only two categories: the good and the evil. Like Cole, I believe that the Nazis were mistaken when they characterized the Jews as an evil enemy, and I agree that this kind of sweeping characterization can be an effective tool in facilitating unjust acts of war and genocide. Like Cole, I believe that the politically powerful can and often do seize on minor threats and paint them as evil enemies, and sometimes try to justify acts of war and destruction by falsely describing their opponents as evil. Like Cole, I believe that innocent people were murdered because they were mistakenly judged to be evil witches and vampires. Like Cole, I think we ought not characterize people as evil simply because they are migrants. Where Cole and I disagree is on the question of how best to explain and respond to these moral errors. Cole thinks that these moral errors occur because people use the concept of evil. I believe that these moral errors occur when people misuse the concept of evil. Many people can and do employ the concept of evil judiciously in their moral thinking without falling into the kind of moral mistakes that rightly are condemned by Cole.

Doing Away With Evil?  223 First let us consider Cole’s suggestion that use of the concept of evil introduces an unrealistic dualism, a binary opposition between good and evil people that does not conform with the moral facts. The false dualism in question is not a metaphysical dualism between matter and spirit or between physical beings and supernatural beings, but is a moral dualism. In order to assess Cole’s argument we need to ask what would make a binary opposition between moral concepts false or misleading. All moral systems require concepts that are opposed, and hence all moral systems incorporate binary thinking or dualism in some sense of those terms. Every moral system must incorporate the claim that some things are good and worthy of admiration or pursuit, while other things are not, and that some actions ought to be performed and others not. Cole’s objection to the concept of evil cannot be merely that it is opposed to the concept of good. Perhaps the best way to interpret Cole’s argument is to read him as suggesting that people who use the concept of evil are thereby forced to sort people into two and only two categories—good and evil—and that those two categories alone are inadequate. A moral system that forced us to class every person as either good or evil and precluded any other moral evaluations certainly would be unrealistically dualistic. Such a system would leave an evaluative gap between the good and the extremely bad. Yet Cole has given us no reason to believe that the use of the concept of evil somehow prevents us from deploying a whole range of other moral concepts that allow us to evaluate persons, motives, and actions across a spectrum stretching between the extremities of heroic moral virtue and evil. The philosophers who have defended the concept of evil agree that evil actions are morally extreme and appalling, and that evil persons are deserving of our strongest condemnation, so it would be utterly baffling if they then declared that every wrong action is evil and every bad person is evil. We coherently can judge that Stalin was evil but that Nixon was merely deceitful and corrupt, or that Hitler was evil but that Oskar Schindler was certainly less than evil, and in many ways was morally admirable. Cole might argue that, while it is coherent and possible to use the concept of evil and still make fine-grained moral evaluations right across the spectrum, in practice people who use the concept of evil become dualistic moral thinkers. Yet there are plenty of thinkers, including Card, Neiman, Kekes, and McGinn, who judge that some people are evil without thereby exhibiting a Manichean tendency and assuming that all people are either extremely good or extremely bad. When critics of the concept of evil assume that everyone who uses that concept employs a simplistic binary moral system, those critics are guilty of exactly the kind of cartoonish oversimplification that they attribute to their opponents. It is worth asking how Cole and some others came to believe that using the concept of evil would lead to an unrealistic binary moral outlook which leaves

224  Doing Away With Evil? an evaluative gap between two extremes. Perhaps the source of this error is the apparent shift in use of the word “evil” over time. As we have seen, originally “evil” was used as a synonym for “bad”, but later was restricted to the moral domain and took on connotations of extremity. Under the original use of “evil”, the dichotomy between good and evil was close to being exhaustive. Every value would have been said to be an instance of either good (that is, positive value) or evil (that is, negative value), leaving only a neutral space occupied by things that are neither valuable nor disvaluable. Even trivial and non-morally bad things, according to this use of the word, could properly be described as evil. But if we hold onto the idea that all values can be put into these two classes while shifting the meaning of “evil” so that it is restricted to the moral sphere and comes to have connotations of extremity, we would produce a false dichotomy: the view that everything which has moral value is either good or extremely bad. The way to avoid the false dichotomy, though, is not to drop the word “evil” from our vocabulary, but simply to recognize that the category of evil is extreme, and that the vast majority of wrong actions fall short of being evil, and the vast majority of vicious persons fall short of being evil. We have seen that using the concept of evil need not lead to the acceptance of an unrealistic binary moral system. Let us move on to consider Cole’s claim that by using the concept of evil we falsely demonize our enemies and produce terrible political consequences. Here too Cole points out a genuine mistake in moral thinking, but fails to recognize that we can avoid making this mistake without abandoning the concept of evil. Perhaps the most important fact that Cole overlooks here is that there is a distinction between evil actions and evil persons, and that there is virtual unanimity amongst philosophers who use the concept of evil that not every evildoer is an evil person. In judging that an action is evil we do not commit ourselves to the view that the agent in question is irredeemably morally corrupt, much less that the agent is a supernaturally demonic being. We might judge that the acts of torture committed in Abu Ghraib were evil, but that at least some of the perpetrators were not evil persons. For instance, Private Lynndie England is culpable for having performed extremely wrong actions, but if she is not strongly and highly fixedly disposed to do that sort of thing when in autonomy favouring conditions, then she is not an evil person (cf. Zimbardo 2007, 372–3). Evildoing is something that we often try to prevent by engaging the evildoer in moral argument and other forms of rational persuasion, or by designing social environments that will make it less likely to occur (cf. Russell 2009b, 461–2). When the evildoer is not an evil person, it is reasonable to hope that these methods will succeed. By using the concept of evil we do not necessarily demonize our enemies and shut down the possibility of diplomacy, dialogue, and moral reform.

Doing Away With Evil?  225 If it is to have any force, the warning against demonizing our enemies must be directed at the use of the concept of evil personhood rather than the concept of evil action. I have argued that an evil person is someone who is strongly and highly fixedly disposed to perform evil actions when in autonomy-favouring conditions, and hence that, for practical purposes, an evil person is a moral write-off, whom we cannot expect to listen to our moral arguments or to be a suitable candidate for other attempts at moral reform. If this is what it means to demonize a person, then in judging that our enemy is an evil person, we are demonizing the enemy. Cole is right to point out that there are many cases in which a person is judged too readily to be evil, on the basis of scant evidence. Some such judgements are motivated by an irrational fear of foreigners and outsiders, and some are the result of cynical propaganda or media hysteria. As Cole points out, sometimes the so-called evil enemy is not a genuine danger at all (Cole 2006, 74). Yet these are misapplications of the concept of evil personhood. There are other cases in which we have strong evidence that a particular person is evil, and in which it is rational to treat that person as a write-off. Many war criminals and serial killers exhibit a pattern of recalcitrant evildoing which indicates that they are beyond rational persuasion and cannot be made good. When we condemn Stalin and the Liberian warlord Charles Taylor as evil persons, it would be implausible to suggest that our targets of condemnation are not genuinely dangerous men, or that in judging them to be evil we are inventing an enemy in order to distract or manipulate the populace. The fact that the concept of evil personhood often is misapplied does not imply that it has no proper application. Cole’s argument against the concept of evil can be undermined further via a comparison with other moral concepts, many of which are misused with disastrous effect. For example, the concept of moral heroism is regularly misapplied. Terrorist bombers and war criminals are mistakenly praised as moral heroes by their deluded supporters. Young and impressionable minds can be pressed into bloody service via a desire to emulate these so-called heroes. Yet should we respond to this danger by claiming that we are better off without the concept of moral heroism? Should we deny that William Wilberforce, Susan B. Anthony, and Martin Luther King are moral heroes, whose achievements we should find admirable and inspiring? The reasonable response in this case is to argue that the concept of moral heroism is often misapplied and can be dangerous, but that there really are moral heroes whom we ought to recognize, admire, and whose successes we should strive to emulate. The often misused concept of moral heroism is defensible, and ought to be used, albeit with care. It is also true that the concept of moral duty or moral obligation has regularly been misapplied. Many perpetrators of atrocities during the Holocaust, including Himmler and

226  Doing Away With Evil? Eichmann, told themselves that, despite feeling troubled by their actions, they were meeting their obligations (Bennett 1994, 300; Pauer Studer and Velleman 2011, 352). The fact that the concept of duty can be put to such terrible use does not show that we ought to deny that there are any moral duties, nor that we should expunge the notion of duty from our moral thinking. Arguably, we must use the concept of moral duty itself in our evaluation of this massive moral failure. The concept of evil, like the concept of moral heroism and the concept of duty, can be misused. In light of this danger, we ought to urge caution when people rush to judge that certain actions or persons are evil. Our difficulty in ascertaining the fixity of a person’s character should give us an extra reason to exercise caution in applying the concept of evil personhood. If we say that an action or person is evil, we must also be ready to explain in more detail the nature of that judgement, and to engage with the more fine-grained questions concerning motive and character that lie beneath its surface. But these warnings do not indicate that there is a special problem with the concept of evil. Rather, they are particular instances of general warnings that apply to all of our moral thinking. In this book I have offered a secular philosophical account of the moral concepts of evil action, evil person, and evil feeling. In doing so I have tried to avoid the temptation to latch onto an interesting moral concept and merely stipulate that this is the concept of evil. I have also tried to avoid the opposing error of assuming that evil is whatever the folk believe it to be. Given the messy nature of folk moral discourse, no coherent account of evil could do that. Moreover, there are several complex philosophical puzzles—including questions about extremity, incomprehensibility, qualitative differences, and fixity of character—about which the folk are unlikely have strong or well-informed intuitions. My aim has been to walk a path between these two extremes, to construct an account of evil which delineates a coherent and useful moral concept that fits closely enough with folk discourse to warrant being called “the concept of evil”. My story about evil is incomplete in many respects. It is parasitic on a prior account of wrong action, and hopefully is compatible with several competing theories of wrong action. It also depends on a prior account of culpability, and I have not engaged in any detail with important connected questions concerning the boundary between illness and depravity, and the possibility of group agency. Nonetheless, I hope that I have shown that the concept of evil has an important role to play in contemporary secular moral thought.

Bibliography Aquinas, Thomas (1942), The Summa Theologica, Part III (Supplement), London: Burns Oates and Washbourne. Arendt, Hannah ([1951] 1967), The Origins of Totalitarianism, London: Allen and Unwin. Arendt, Hannah (1978), The Jew as Pariah, New York: Grove. Arendt, Hannah ([1963] 2006), Eichmann in Jerusalem, New York: Penguin. Aristotle (1985), Nicomachean Ethics, translated by Terence Irwin, Indianapolis: Hackett. Armstrong, David. M. (1968), A Materialist Theory of the Mind, London: Routledge. Augustine (1951), The Confessions of St. Augustine, translated by F. J. Sheed, London and New York: Sheed and Ward. Austin, J. L. (1979), Philosophical Papers, Oxford: Oxford University Press. Baron-Cohen, Simon (2011), Zero Degrees of Empathy, London: Penguin. Barry, Peter Brian (2011), “In Defense of the Mirror Thesis”, Philosophical Studies 155: 199–205. Bennett, Jonathan (1994), “The Conscience of Huckleberry Finn”, in Peter Singer (ed.), Ethics, pp. 294–305, Oxford: Oxford University Press. Bentham, Jeremy (1948), An Introduction to the Principles of Morals and Legislation, New York: Hafner. Bernstein, Richard J. (2002), Radical Evil:  A  Philosophical Interrogation, Cambridge: Polity Press. Blass, T. (2000), Obedience to Authority: Current Perspectives on the Milgram Paradigm, Mahwah, NJ: Lawrence Erlbaum. Blass, T. (2004), The Man Who Shocked the World, New York: Basic Books. Braddon-Mitchell, David and Jackson, Frank (1996), Philosophy of Mind and Cognition. Oxford: Blackwell. Braddon-Mitchell, David and Nola, Robert (2009), “Introducing the Canberra Plan”, in D. Braddon-Mitchell and R. Nola (eds), Conceptual Analysis and Philosophical Naturalism, pp. 1–20, Cambridge, MA: MIT Press. Bratman, Michael (2007), Structures of Agency, Oxford: Oxford University Press. Brink, David (1989), Moral Realism and the Foundations of Ethics, Cambridge: Cambridge University Press. Browning, Christopher (1998), Ordinary Men: Reserve Police Battalion 101 and the Final Solution in Poland, New York: HarperCollins. Cahill, Tim (1986), Buried Dreams:  Inside the Mind of a Serial Killer, New York: Bantam Books. Calder, Todd (2002), “Towards a Theory of Evil Acts: A Critique of Laurence Thomas’s Theory of Evil Acts”, in Daniel Haybron (ed.), Earth’s Abominations: Philosophical Studies of Evil, pp. 51–61, New York: Rodopi. Calder, Todd (2003), “The Apparent Banality of Evil: The Relationship Between Evil Acts and Evil Character”, Journal of Social Philosophy 34, 3: 364–76.

228  Bibliography Calder, Todd (2013), “Is Evil Very Wrong?”, Philosophical Studies 163: 177–96. Capote, Truman (2000), In Cold Blood, London: Penguin. Card, Claudia (2002), The Atrocity Paradigm, New York: Oxford University Press. Cesarani, David (2007), Becoming Eichmann: Rethinking the Life, Crimes, and Trial of a “Desk Murderer”, London: De Capo Press. Clendinnen, Inga (1998), Reading the Holocaust, Melbourne: Text Publishing. Cole, Phillip (2006), The Myth of Evil, Edinburgh: University of Edinburgh Press. D’Arms, J. and Jacobson, D. (2000), “The Moralistic Fallacy” On the ‘Appropriateness’ of the Emotions”, Philosophy and Phenomenological Research 61, 1: 65–90. de Wijze, Stephen (2002), “Defining Evil: Insights from the Problem of ‘Dirty Hands’”, The Monist, 85, 2: 210–38. Devitt, Michael and Sterelny, Kim (1987), Language and Reality. Blackwell: Oxford. Devitt, Michael (2012), “Wither Experimental Semantics?” Theoria 27, 73: 5–36. Dews, Peter (2008), The Idea of Evil, Oxford: Blackwell. Doris, John (2002), Lack of Character:  Personality and Moral Behaviour, Cambridge: Cambridge University Press. Duff, Antony (1977), “Psychopathy and Moral Understanding”, American Philosophical Quarterly 14, 3: 189–200. Dupré, John (1993), The Disorder of Things: Metaphysical Foundations of the Disunity of Science, Cambridge, MA: Harvard University Press. Eagleton, Terry (2010), On Evil, New Haven, CT: Yale University Press. Fine, Cordelia and Kennett, Jeanette (2004), “Mental Impairment, Moral Understanding and Mitigation:  Psychopathy and the Purposes of Punishment”, The International Journal of Law and Psychiatry 27, 5: 425–43. Finlay, Stephen (2008) “The Error in the Error Theory”, Australasian Journal of Philosophy 86, 3, pp. 347–69. Formosa, Paul (2008), “A Conception of Evil”, The Journal of Value Inquiry 42: 217–39. Foucher, Luc and Tappolet, Christine (2006), “Introduction: Modularity and the Nature of Emotions”, Canadian Journal of Philosophy, Supplementary Volume 32, vii–xxxi. Frankfurt, Harry (1988), The Importance of What We Care About. Cambridge: Cambridge University Press. Gaita, Raymond (2000), A Common Humanity: Thinking About Love and Truth and Justice, New York: Routledge. Garrard, Eve (1998), “The Nature of Evil”, Philosophical Explorations, 1: 43–60. Garrard, Eve (2002), “Evil as an Explanatory Concept”, The Monist 85, 2: 320–36. Goldberg, Elkhonon (2001), The Executive Brain: The Frontal Lobes and the Civilized Mind, New York: Oxford University Press. Goldhagen, Daniel (1997), Hitler’s Willing Executioners:  Ordinary Germans and the Holocaust, New York: Random House. Griffiths, Paul (1997) What Emotions Really Are, Chicago: University of Chicago Press. Hare, R. M. (1952), The Language of Morals, Oxford: Clarendon Press. Harman, Gilbert (2000), Explaining Value, New York: Oxford University Press. Haybron, Daniel (1999), “Evil Characters”, American Philosophical Quarterly 36, 2: 131–48. Haybron, Daniel (2002a), “Moral Monsters and Saints”, The Monist, 85, 2: 260–84.

Bibliography  229 Haybron, Daniel (2002b), “Consistency of Character and the Character of Evil”, in Daniel Haybron (ed.), Earth’s Abominations:  Philosophical Studies of Evil, pp. 63–78, New York: Rodopi. Hitchens, Christopher (2002), “Evil”, 31 December 2002, Slate . Hoess, Rudolph (2000) Commandant of Auschwitz, London: Phoenix. Hull, D. (1987), “Genealogical Actors in Ecological Roles”, Biology and Philosophy, 2: 168–83. Hurka, Thomas (2001), Virtue, Vice and Value, New York: Oxford University Press. Hurka, Thomas (2006), “Virtuous Act, Virtuous Dispositions”, Analysis 66, 1: 69–76. Hursthouse, Rosalind (1999), On Virtue Ethics, Oxford: Oxford University Press. Jackson, Frank (1998), From Metaphysics to Ethics:  A  Defence of Conceptual Analysis, Oxford: Oxford University Press. Jackson, Frank (2001), “Precis of From Metaphysics to Ethics”, Philosophy and Phenom­ enological Research 62, 3: 617–24. Johnston, Mark (2011), Saving God: Religion after Idolatry, Princeton: Princeton University Press. Joyce, Richard (2001), The Myth of Morality, Cambridge: Cambridge University Press. Joyce, Richard (2011), “The Error in ‘The Error in the Error Theory’”, Australasian Journal of Philosophy 89, 3: 519–34. Kamtekar, Rachana (2004), “Situationism and Virtue Ethics on the Content of Our Character”, Ethics 114: 458–91. Kant, Immanuel (1994), On a Supposed Right to Lie from Altruistic Motives, in Peter Singer (ed.), Ethics, pp. 280–1, Oxford: Oxford University Press. Kant, Immanuel (1996), “Religion Within the Boundaries of Mere Reason”, in Immanuel Kant, Religion and Rational Theology, translated and edited by Allen W.  Wood and George di Giovanni, pp. 39–216, Cambridge: Cambridge University Press. Kant, Immanuel (1997), Critique of Practical Reason, translated and edited by Mary Gregor, Cambridge: Cambridge University Press. Kekes, John (1990), Facing Evil, Princeton, NJ: Princeton University Press. Kekes, John (1998), “The Reflexivity of Evil”, Social Philosophy and Policy, 15, 1: 216–32. Kekes, John (2005), The Roots of Evil, Ithaca, NY: Cornell University Press. Kemp, K. W. and T. Sullivan (1993), ‘Speaking Falsely and Telling Lies’, in Proceedings of the American Catholic Philosophical Association 67: 151–70. Kingsbury, J and J. McKewon-Green (2009), “Jackson’s Armchair:  The Only Chair in Town?”, in D. Braddon-Mitchell and R. Nola (eds), Conceptual Analysis and Philosophical Naturalism, pp. 159–82, Cambridge, MA: MIT. Kripke, Saul (1980), Naming and Necessity, Cambridge, MA: Harvard University Press. Leiter, Brian (2002), Nietzsche on Morality, London: Routledge. Levi, Primo (1989), The Drowned and the Saved, translated by Raymond Rosenthal, New York: Vintage. Levi, Primo (2000), “Introduction”, in Rudolph Hoess, Commandant of Auschwitz, London: Phoenix. Locke, John (1987), An Essay Concenring Human Understanding, edited by Peter Nidditch, Oxford: Clarendon.

230  Bibliography Lycan, B. (1988), Judgment and Justification, Cambridge: Cambridge University Press. McDowell, John (1978), “Are Moral Requirements Hypothetical Imperatives?”, Proceedings of the Aristotelian Society, supp. vol. 52: 31–42. McGinn, Colin (1997), Evil, Ethics and Fiction, Oxford: Clarendon Press. Mackie, J. L. (1985), Ethics: Inventing Right and Wrong, Harmondsworth: Penguin. Mackie, J. L. (1977), “Dispositions, Grounds and Causes”, Synthese 34: 361–70. McKitrick, J. (2005), “Are Dispositions Causally Relevant?”, Synthese 144: 357–71. McNaughton, David (1988), Moral Vision, Oxford: Basil Blackwell. Maibom, Heidi (2005), “Moral Unreason: The Case of Psychopathy”, Mind and Language 20, 2: 237–57. Midgley, Mary (2001), Wickedness, London: Routledge. Milgram, Stanley (1974), Obedience to Authority, New York: Harper and Row. Milton, John (1993), Paradise Lost, edited by Roy Flannagan, New York, Macmillan. Morrison, Blake (1997), As If, London: Granta. Morrow, Lance (2003), Evil: An Investigation, New York: Basic. Morton, Adam (2004), On Evil, New York: Routledge. Nagel, Thomas (1979), Mortal Questions, Cambridge: Cambridge University Press. Nichols, Shaun (2004), Sentimental Rules: On the Natural Foundations of Moral Judgment, Oxford: Oxford Universtity Press. Neiman, Susan (2003), Evil in Modern Thought, Melbourne: Scribe. Nietzsche, Friedrich (1996), On the Genealogy of Morality, translated by M. Clark and A. J. Swensen, Indianapolis: Hackett. Nolan, Daniel (2009), “Platitudes and Metaphysics”, in D. Braddon-Mitchell and R. Nola (eds), Conceptual Analysis and Philosophical Naturalism, pp. 267–300, Cambridge, MA: MIT Press. Pauer-Studer, Herlinde and Velleman, David (2011), “Distortions of Normativity”, Ethical Theory and Moral Practice 14, 3: 329–56. Paulson, S. (2006), “The Flying Spaghetti Monster”, Salon, 13 October 2006. . Plantinga, Alvin (1990), “God, Evil, and the Metaphysics of Freedom”, in Marilyn McCord Adams and Robert Merrihew Adams (eds), The Problem of Evil, pp. 81–109, New York: Oxford University Press. Perrett, Roy (2002), “Evil and Human Nature”, The Monist 85, 2: 304–19. Prunier, Gérard (1995), The Rwanda Crisis:  History of Genocide, New  York:  Columbia University Press. Putnam, Hilary (1975), Mind, Language, and Reality (Philosophical Papers, vol. 2), Cambridge: Cambridge University Press. Rawls, John (1972), A Theory of Justice, Oxford: Clarendon. Rescher, Nicholas (1993), “Moral Luck”, in Daniel Statman (ed.) Moral Luck, pp. 141–66, Albany, NY: State University of New York Press. Robinson, Denis (2009), “Moral Functionalism, Ethical Quasi-Relativism, and the Canberra Plan”, in D. Braddon-Mitchell and R. Nola (eds), Conceptual Analysis and Philosophical Naturalism, pp. 315–48, Cambridge, MA: MIT. Rosenbaum, Ron (1998), Explaining Hitler, New York: HarperCollins.

Bibliography  231 Rosenbaum, Ron (2009), “The Evil of Banality”, Slate, 30 October . Russell, Luke (2006), “Evil-Revivalism Versus Evil-Skepticism”, The Journal of Value Inquiry 40: 89–105. Russell, Luke (2007a), “What Even Consequentialists Should Say About the Virtues”, Utilitas 19, 4: 466–86. Russell, Luke (2007b), “Is Evil Action Qualitatively Distinct From Ordinary Wrongdoing?”, Australasian Journal of Philosophy 85: 659–77. Russell, Luke (2009a), “He Did It Because He Was Evil”, American Philosophical Quarterly 46, 3: 267–82. Russell, Luke (2009b), “Is Situationism All Bad News?”, Utilitas 21, 4: 443–63. Russell, Luke (2010a), “Evil, Monsters and Dualism”, Ethical Theory and Moral Practice 13, 1: 45–58. Russell, Luke (2010b), “Dispositional Accounts of Evil Personhood”, Philosophical Studies 149: 231–50. Russell, Luke (2012), “Evil and Incomprehensibility”, Midwest Studies in Philosophy 36: 62–73. Searle, John (1983), Intentionality, Cambridge: Cambridge University Press. Shafer-Landau, Russ (2004), Whatever Happened to Good and Evil, New  York:  Oxford University Press. Shakespeare, William (1974), Richard III, London: Macmillan. Shakespeare, William (2008), Titus Andronicus, edited by Eugene M. Waith, Oxford: Oxford University Press. Sider, Theodore (2001), “Criteria of Personal Identity and the Limits of Conceptual Analysis”, Philosophical Perspectives 15, 189–209. Simpson, John and Bennett, Jana (1985), The Disappeared and the Mothers of the Plaza: The Story of the 11000 Argentinians Who Vanished, New York: St Martin’s Press. Singer, Marcus (2004), “The Concept of Evil”, Philosophy 79: 185–214. Slote, M. (1966). “The Theory of Important Criteria”, Journal of Philosophy 63, 211–24. Smith, Michael (1994), The Moral Problem, Oxford: Blackwell. Sober, E. (1984), “Sets, Species, and Natural Kinds: A Reply to Philip Kitcher’s ‘Species’”, Philosophy of Science, 51: 334–41. Steiner, Hillel (2002), “Calibrating Evil”, The Monist, 85: 183–93. Stone, Michael H. (2009), The Anatomy of Evil, New York: Prometheus Books. Thomas, Laurence (1993), Vessels of Evil: Slavery and the Holocaust, Philadelphia: Temple University Press. Thomas, Laurence (2003), “Forgiving the Unforgivable?”, in Eve Garrard and Geoffrey Scarre (eds), Moral Philosophy and the Holocaust, pp. 201–30, Aldershot: Ashgate. Tooley M. (1980), “Alvin Plantinga and the Argument from Evil”, Australasian Journal of Philosophy, 58: 360–76. Van Inwagen, Peter (2008), The Problem of Evil, Oxford: Oxford University Press. Vargas, Manuel (2010), “Are Psychopathic Serial Killers Evil? Are They Blameworthy for What They Do?”, in S. Waller (ed.), Serial Killers—Philosophy for Everyone: Being and Killing, pp. 66–77, Oxford: Wiley-Blackwell.

232  Bibliography Wainwright, Robert and Totaro, Paola (2009), Born or Bred: Martin Bryant, Fairfax: Sydney. Waller, S. and Deal, William E. (2010), “Introduction: Meditations on Murder, or What is So Philosophical About Serial Killers?”, in S. Waller (ed.) Serial Killers—Philosophy for Everyone: Being and Killing, pp. 1–14, Oxford: Wiley-Blackwell. Watson, Gary (1987), “Responsibility and the Limits of Evil: Variations on a Stawsonian Theme”, in Ferdinand Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays in Moral Psychology, pp. 256–86, Cambridge: Cambridge University Press. Williams, Bernard (1981), Moral Luck, Cambridge: Cambridge University Press. Williams, Bernard (1993), Ethics and the Limits of Philosophy, London: Fontana Press. Wittgenstein, Ludwig (1953), Philosophical Investigations, translated by G. E. Anscombe, Oxford: Basil Blackwell. Wolf, Susan (1982), “Moral Saints”, Journal of Philosophy 79, 8: 419–39. Wolfendale, Jessica (2007), Torture and the Military Profession, Basingstoke:  Palgrave Macmillan. Zaibert, Leo (2012), “Beyond Bad:  Punishment Theory Meets the Problem of Evil”, Midwest Studies in Philosophy: The Concept of Evil, 36: 93–111. Zangwill, Nicholas (2003), “Perpetrator Motivation: Some Reflections on the Browning/ Goldhagen Debate”, in Eve Garrard and Geoffrey Scarre (eds), Moral Philosophy and the Holocaust, pp. 89–102, Aldershot: Ashgate. Zimbardo, Philip (2007), The Lucifer Effect: Understanding How Good People Turn Evil, New York: Random House.

Index Abu Ghraib  18, 51, 204, 224 action individuation  123 aggregative accounts of evil personhood  133, 139–47, 151, 153, 160, 162 akrasia  75, 84–5, 104–5 Alford, C. Fred  83, 88 Amin, Idi  133 amoralists  37, 43, 84, 190 antisympathetic feelings  177–9 appropriate connection to extreme harm 61–2, 183–4 Aquinas, Thomas  179 arachnophobia 203–4 Arendt, Hannah  3, 20–1, 35, 69–74, 76–8, 82, 89, 91, 106, 112, 121, 128, 187 Aristotle  84, 102, 165–6, 169, 211 atheism 11, 14, 20 atrocity  38, 40, 83, 166, 218 Augustine  38, 85, 100 Auschwitz  77, 124, 219 autonomy 173 autonomy-favouring conditions  173–5, 182, 192, 213, 224 axis of evil  1

Calder, Todd  35, 41–2, 65, 77, 81, 91, 113–16, 118, 120, 157, 184 Capote, Truman  73 Card, Claudia  18, 35, 42, 45–6, 52, 76–7, 87, 91, 112, 121–3, 149, 221, 223 Cesarani, David  72 character  33, 96, 102–3, 138, 151–2, 156, 158, 188, 203, 205, 211–13 Christianity 11, 96 Clendinnen, Inga  202 coercion 44, 171–2 Cole, Phillip  12, 21, 78, 135–7, 163, 169, 196–9, 202–18, 221–5 compassion 79, 198–9 conceptual analysis  24–30, 93–5, 98–9, 130–1 conceptual pluralism  4, 129–32, 175, 200, 221 concept v. word  14, 16, 19, 41, 93–4, 96, 106, 131–2, 218, 224 consequentialism  39, 41–2, 114 courage v. cowardice  33, 75, 96, 127, 151, 188 criminals  138, 144, 152 culpability  44, 47, 52, 76, 88–9, 114–6, 122, 131, 141, 158, 160, 184, 199, 206, 212 cultural differences regarding evil  96

bad  17–19, 41, 163, 190, 197 banality of evil  35–6, 71–4, 77, 79, 89, 106, 124 Baron-Cohen, Simon  55, 202, 218 Barry, Peter Brian  165 Bauer, Yehuda  20 becoming an evil person  156–9, 205–6, 211–12 Bennett, Jonathan  73, 105 Bentham, Jeremy  17 Bernstein, Richard  202 Bin Laden, Osama  219 binary moral systems  222–4 Blair, Tony  26–7 blameless evil persons  152–3, 158–9, 182 bombing  53, 124, 126, 131, 141, 147–8 böse 19 Brando, Marlon  85 Breivik, Anders  57 Browning, Christopher  78, 81 Bryant, Martin  139 Budd, Billy  136, 151 Bundy, Ted  2, 70, 137, 167, 207, 211 Bulger, James  199, 202–3, 205, 213 Bush, George W.  1, 10, 26–7, 87

Dahmer, Jeffery  6, 193 Dawkins, Richard  20 de dicto desires  85 de Wijze, Stephen  42, 55, 91, 112, 115–16 defiance of morality  36, 59, 70–1, 73–4, 76, 78–9, 80, 83, 84–6, 100, 123, 131 definition  14, 16–17, 22–5, 219 deism 13 deliberative silencing  90–1, 98, 102–3, 119–20 delinquent defiance  59, 85–6 demon/demonic  10, 12, 21, 197, 224–5 deontic judgements  38 deplorability thesis  121–4, 126–8 despair 77 Dews, Peter  10, 82 dictionary definitions of evil  11, 16–17 dispositions  155–6, 168–71, 205 dispositional accounts of evil personhood  133, 137, 149, 151–2, 154–65, 167–75, 180, 182, 184, 191, 210, 213 Doris, John  188 Dr Evil  38, 83 dualism 222–4 Dupré, John  108

234  index Eagleton, Terry  23 Eichmann, Adolf  19, 20–1, 35, 70–4, 76, 124–8, 187, 213, 225 emotion  37, 43, 83, 116–17, 159, 176–7, 184, 218, 221 empathy 218 end point v. end zone  64–5, 167, 194, 223 England, Lynndie  224 envy 178 epistemic accessibility  100–2, 104, 152, 213 error theory about evil  12, 29–30, 64, 115, 164, 196, 205, 221 evaluative judgements  38, 185 evil action  2, 31–2, 34–6, 41, 43, 47, 52, 62–8, 69–74, 76–82, 87–8, 91, 93–5, 101, 118, 132, 134, 175, 184, 192–3, 197, 200, 204, 218, 224 evil, concept v. word  16, 19, 25, 41, 106, 198, 218, 221, 224 evildoer v. evil person  136, 140, 152, 175, 201, 210, 224 evil feelings  5, 134, 176–95 evil person  4, 32–4, 87, 102, 110, 118, 133–75, 183, 186, 189–90, 192–3, 200, 204, 207, 215, 218, 225–6 evil values  186 exacerbators of wrongs  80, 83, 88, 97, 103, 127–8 excuses  6–7, 44–5, 88, 114, 160 extensional method  25–7, 98–100, 134 explanation  5–6, 56, 136, 139, 147, 155, 175, 191, 197–207, 215–16 expressive thickness v. thinness  75–6 expression of disapproval  37–8, 43, 75–6, 159, 219 extremity  18, 46, 49–56, 59–68, 71, 76, 90, 96–7, 103–4, 115, 117–19, 141–2, 178, 209, 223

gold 98–9 Goldhagen, Daniel  81, 125, 160 greed 21–2, 57 grey areas  67, 113, 115, 117–18, 120, 168 group action  7

failed attempts  52, 122, 141, 147–9 fictional portrayals of evil  38, 83, 85, 96, 164, 179 fixity of dispositions/character  168–70, 206, 210–13, 226 flammability 170–1 folk accounts v. non-folk accounts of evil action  79, 88–91, 97–100, 106, 129 folk claims about evil  6, 25, 28, 55, 87, 94–5, 100, 106–7, 112–13, 119, 168, 196, 200, 209, 217–18, 220, 224 forgiveness  2, 11, 188, 209 Formosa, Paul  35, 46, 52, 65, 76–7, 91, 112, 121, 123, 149–50, 152, 182 Frankfurt, Harry  185–8 Fritzl, Josef  51

Kant, Immanuel  20, 40 Kekes, John  35, 41, 44, 46, 65, 77, 82, 88, 91, 127, 138, 143, 200, 223 Kripke, Saul  98–9

gratuitous harm/wrongdoing  80–2, 91–2, 103 Gacy, John Wayne  18, 70, 157 Gaita, Raymond  197 Garrard, Eve  36, 37, 77, 89–91, 97–9, 100, 102–6, 112, 116–19, 129, 135, 157, 201, 219 God  10, 12–16, 83, 209

harm  38–9, 50, 52, 62–7, 75, 80, 122, 197 harmless evil actions  52–4, 55–6, 62–3, 122, 148, 183, 218 Haybron, Daniel  32–3, 64, 76–7, 110, 116–17, 134–5, 138, 152, 157, 163–4, 167–9, 181–2, 193, 208–9 head injury  158, 205, 209 heroism  67, 74, 138, 165, 225 hidden essences  98–100, 106 Himmler, Heinrich  73, 105–6, 130, 213, 225 Hitchens, Christopher  20, 37, 87–8 Hitler, Adolf  22, 73, 78, 80, 95, 97, 133, 163–4, 208, 213, 222–3 Hoess, Rudolph  20, 73 Holocaust  2, 20, 21, 45, 55–6, 70–1, 73, 77–8, 81, 96, 105, 124, 143–4, 160, 219, 222, 225 horror 37, 55 Hurka, Thomas  33, 42, 83, 178, 188 Hussein, Saddam  179 hybrid aggregative/dispositional account of evil personhood 162 incomprehensibility  49, 55–61, 117, 129 inferential method  25, 27, 98, 134 instrumental wrongdoing  80–1, 83, 86, 208 intentions  113, 124, 149–50, 152, 182 intuitions 34, 134 involuntary v. voluntary  184–5, 193

Leiter, Brian  20 lesser of two evils  17, 40 Letterman, David  87 Levi, Primo  20, 45, 73, 160 linguistic disagreement  130–2 Locke, John  17 lying 38–40, 75 malevolent quadriplegic  181–2 malice  36, 41, 70–1, 73–4, 76–9, 80–3, 97, 100, 123–6, 128, 131, 148, 199 Mao Zedong  6, 136 Manicheanism 222–4 Manson, Charles  12 mass shooting  57, 123, 137, 139 McCafferty, Archie  137 McDowell, John  102

index  235 McGinn, Colin  35, 46–7, 63, 80, 84, 88, 91, 94, 127, 129, 199–200, 223 McNamara, Robert  41 Melville, Herman  136, 151 Mengele, Josef  163, 208, 219 mental illness  6–7 metaethics 6 Methuselah 141–2 Milgram experiment  60, 153, 161–4, 170–2, 218 Milton, John  38, 59, 83, 85, 87 minor wrongs  32, 46, 90, 103–4, 120, 123, 141–2, 224 mirror thesis  165–7, 169, 187, 208 monstrosity  70, 79, 197, 212, 215 moral judgement  38, 40, 84, 159, 190, 208 moral luck  122, 147–9, 151 moral saints  110, 165, 169, 187 motives  33, 56, 72, 74–6, 79–82, 113–15, 118, 121, 123–6, 188, 198–200 Morrow, Lance  10, 23, 55 Morton, Adam  35–6, 44, 46–7, 55, 77, 84, 89, 91, 97, 99, 101–2, 106, 117, 123, 138, 157 murder  51, 65–6, 78, 81, 83, 98, 199, 203, 205, 215 Mussolini, Benito  179 narrative integration  188 natural kinds  29, 107 naturalism 9–12, 21 Neiman, Susan  19, 23, 35, 56, 64, 77, 91, 112, 121, 123–8, 223 negligence 44 Nietzsche, Friedrich  20 Nixon, Richard  76, 223 nostalgic evildoers  183, 190, 193, 195 obsessive-compulsive disorder  188–9 Olsen, Clifford  205, 215 O'Rielly, Bill  87 paedophilia 193–4 parent 132, 218 perpetually frustrated misanthropes  148, 150, 152, 162, 164, 182 Perrett, Roy  36, 83, 88, 91, 127, 193 phenomenology of evil  116–17 pragmatic considerations  131, 217, 219–26 priority of prevention thesis  122 pro tanto reasons  38, 40 problem of evil  18–19 psychological barriers  89, 97–8, 101–2 psychological hallmark of evil action  2–3, 77–8, 89–90, 95–7, 103, 115, 129 psychological silencing  90–1, 98, 102–3, 119–20 psychological thickness  39, 76, 79 psychopaths/sociopaths  45–7, 84, 102, 166–7, 187, 200

punishment  53, 81, 148, 162, 174, 179, 213 Putnam, Hilary  98–9 qualitative v. quantitative differences  63, 90, 112–21, 129, 167 qualitative difference thesis  112–21 Rader, Dennis  2, 59, 70, 86, 97, 136, 205, 209 ranking evils/wrongs  50, 62, 64–5, 67–8, 123–5, 127, 129, 167, 194 rarity of evil persons  135, 153, 159–61, 164, 170, 173 rational competence  6–7, 158 Rawls, John  87–8 reform/redemption  137, 142–3, 152, 155, 169, 191, 193–4, 208–9, 213–14, 224 regularity account of evil personhood  138–9 Reid, Richard  53 relativism 6, 58 religion  2, 9, 11, 13–15, 21, 96, 106 remorse  83, 88, 183, 193–4, 215 repudiation 190–1, 193–4 responsibility 7, 45, 75 Rosenbaum, Ron  20, 22, 71–3, 78, 80, 87–8, 95, 101 Rwandan Genocide  2, 57, 160 sadistic pleasure  36, 42, 63, 70–1, 73–4, 76, 78–9, 82–4, 97, 100, 104–5, 119–20, 123–6, 128, 177–84, 189 sadistic voyeurism  54, 82, 104, 114, 182–4, 189 Satan/Devil  10, 12, 21, 38, 58–9, 83, 85, 96, 106, 181, 203 schadenfreude 178–80 Schindler, Oskar  165, 223 secular account of evil  10, 13, 21–2, 197 September 11 terrorist attacks  77–8, 125 Shakespeare 17, 71, 85 sin  14–16, 21, 85, 179 Singer, Marcus  18, 33–4, 37, 44, 46, 52, 55, 59–60, 64, 88, 91, 110, 127, 134–5, 140 Shipman, Harold  57 Smith, Perry  73 Sonderkommando 160 Soprano, Tony  164 species 107–8 spree killing  57, 123, 137, 139 Stalin  205, 223, 225 Starrett, Richard  159 Steiner, Hillel  35, 46–7, 63, 84, 88, 91, 94, 112, 116–20, 127, 129 Stone, Michael  37, 42, 55–6, 74, 83, 137–8, 159, 199 supernatural beings  9, 12–14, 21, 197, 212, 217, 223 supernaturally loaded concepts  10, 13–16, 20, 197 synchronic v. diachronic explanations  198, 203, 205

236  index Tarantino, Quentin  179 Taylor, Charles  225 tennis 166 terrorism  73, 77, 122, 125, 136, 225 thick accounts of evil action  3, 77–90, 95–6, 107, 109–10, 114, 124–9, 199–200 thick v. thin concepts  75, 77, 111, 198 thin account of evil action  3, 70, 74–8, 109–10, 114, 124–9, 181, 198–200 Thomas, Lawrence  35, 46, 52, 82, 84, 88, 91, 110–11, 143 thorough dispositional account  163–4 torture  18, 51, 56, 65–6, 73–4, 83, 100, 104, 136–7, 170, 178, 198, 200, 204, 224 Trevor-Roper, Hugh  73 trivial wrongs  32, 46, 90, 103–4, 120, 123, 141–2, 224 unforeseeable calamities  51, 141 Unterweger, Jack  215 usefulness of a concept  110, 121 utilitarianism 83

vagueness  67, 113, 115, 117–18, 120, 168 Vargas, Manuel  45, 82, 169 Venables, Jon, and Robert Thompson  199, 202–3, 205–6, 212 violence inhibiting mechanisms  89, 97–8, 101–2 virtue and vice  33, 42, 96, 102–3, 110, 127, 138, 144, 162–3, 165–6, 169, 178, 187–8, 198, 209, 211, 223 virtuous underprivileged  205 weakness of will  75, 84–5, 102–3 West, Freddie  133, 152 Williams, Bernard  45–6, 75 witches 29–30, 222 Wittgenstein, Ludwig  24 Wolf, Susan  165, 169 word v. concept  14, 16, 19, 94, 96, 106, 131–2, 218, 224 write-offs  137, 169, 191, 208, 225 wrong  39–40, 42–3, 50, 75, 84–5, 116, 122, 197 Zoroastrianism 11